Author Topic: Statistical arrays  (Read 40143 times)

0 Members and 1 Guest are viewing this topic.

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Statistical arrays
« on: December 09, 2018, 08:49:16 pm »
Dear colleagues,

First of all we heared about interesting properties of TaN resistors statistical arrays.
There are some CERN- related papers describing developments made by great Mr John Pickering.
He seems to like using this technique as it does the job at reasonable price and in relatively simple, elegant way.
Inspired by him I made simple setup to see what can be achieved using off the shelf component.
I used 8x10k NOMCA in SOIC 16 package and wired the circuit as shown on the drawing below.
It is thermally balanced divider by 10.

The setup takes advantage of HP 34401A in ratio mode, Data Precision 8200 as (4-wire) voltage source, thermocouple thermometer.


P.S.  please post examples of statistical arrays applications (schematics) that You know about

« Last Edit: December 30, 2018, 12:30:39 pm by doktor pyta »
 
The following users thanked this post: lukier

Offline Echo88

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: de
Re: Statistical arrays
« Reply #1 on: December 09, 2018, 08:57:21 pm »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #2 on: December 09, 2018, 09:27:18 pm »
I was always concerned about the tracking t.c. of 5ppm/K for Vishay TDP16031002 or NOMCA16031002. I wonder if TaN statistical arrays are any better. But even if they were, they seem to be unobtanium. I once searched for more information, but wasn't able to find the real hints. So any link to related papers are welcome.

https://www.ebay.de/itm/1pcs-TDP16031002BUF-Vishay-Resistor-Networks-Arrays-10Kohms-16/263962507025
https://www.ebay.de/itm/1pcs-NOMCA16031002ATS-Vishay-Resistor-Networks-Arrays-16-pin-10/263957703582

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #3 on: December 09, 2018, 10:25:08 pm »
Now watch this:

1. division ratio error: +12ppm (no adjustment!)

2. T amb=25'C, T heat=50'C
average T.C. of division ratio: 0.32ppm/'C

3. Thermal shock 25'C; -20'C; 25'C
division ratio change: 0ppm

4. Thermal shock 25'C; 100'C; 25'C
division ratio change: 0ppm

Keep in mind 1ppm is a resolution of my simple but very stable setup (further measurements need to be done).


To be precise I used P/N NOMCA16031002ATS bought from Farnell.
« Last Edit: December 11, 2018, 09:04:16 am by doktor pyta »
 
The following users thanked this post: Andreas, EmmanuelFaure, ch_scr

Offline d-smes

  • Regular Contributor
  • *
  • Posts: 101
  • Country: us
Re: Statistical arrays
« Reply #4 on: December 10, 2018, 11:12:09 am »
When you did thermal shock, did you go from 25C to -20C, back to 25C, and then make the measurement?  Or did you go from 25C to -20C and then make the measurement?  I just wondering why the TC didn't show up in the ratio measurement...
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #5 on: December 10, 2018, 11:56:52 am »
2. and 3.
3. and 4. describe situation after thermal shock. So measurements were performed @25'C.
« Last Edit: December 10, 2018, 07:53:15 pm by doktor pyta »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #6 on: December 10, 2018, 04:55:41 pm »
So this means it's not a temperature coefficient [ppm/K] but hysteresis value [ppm] you measured for any case @ 25°C, doesn't it?

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #7 on: December 10, 2018, 06:31:22 pm »
Using more resistors is a known way to improve TC matching and similar in a kind of statistical way. Knowing how the resistors are located on the substrate could go beyond just the statistical improvement. The more normal case is to have a gradient in TC over the chip.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #8 on: December 10, 2018, 07:54:15 pm »
So this means it's not a temperature coefficient [ppm/K] but hysteresis value [ppm] you measured for any case @ 25°C, doesn't it?

-branadic-
3. and 4. is hysteresis in ppm

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #9 on: December 10, 2018, 08:04:27 pm »
Using more resistors is a known way to improve TC matching and similar in a kind of statistical way. Knowing how the resistors are located on the substrate could go beyond just the statistical improvement. The more normal case is to have a gradient in TC over the chip.

True, but here in metrology section only few (2...3?) of us ever tried using this technique.
Besides I haven't found their results.
I needed numbers to have reference point for my further designs.

Kleinstein, could You please point a literature, papers which explains theory behind statistical arrays ?

Offline e61_phil

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Re: Statistical arrays
« Reply #10 on: December 10, 2018, 11:18:27 pm »
Just for my naive understanding:

I wonder how much improvement you're getting from that "mixed" wiring and if that very good behaviour here is only luck?

Such arrays are normally specified with a tracking and an absolute TC. The absolute is worse than the tracking, because the tracking is more or less only the difference in TCs of the resistors. I think with such a layout used here one can improve the thermal homogeneity but I wouldn't expect an order of magnitude improvement only by that.

Which point I'm missing here?


A second point is the increase of the amount of resistors. Averaging out the TC over many resitsors only works if the mean of many resistors is really zero (or nearby). I have no experience with such tan arrays, but I measured a lot of thick film HV resistors (Caddock USF) and resistors with negative TC are very rare.
« Last Edit: December 10, 2018, 11:24:17 pm by e61_phil »
 

Offline pigrew

  • Frequent Contributor
  • **
  • Posts: 680
  • Country: us
Re: Statistical arrays
« Reply #11 on: December 11, 2018, 12:09:18 am »
A second point is the increase of the amount of resistors. Averaging out the TC over many resistors only works if the mean of many resistors is really zero (or nearby). I have no experience with such tan arrays, but I measured a lot of thick film HV resistors (Caddock USF) and resistors with negative TC are very rare.

My naïve understanding:

These are used as voltage dividers. It doesn't matter if the mean TC is zero (or not). It matters that the effective TC of each group is matched. If one group has a +14.5ppm/C and the other is +14.6ppm/C, then the ratio will be very close to constant with temperature, even though the value of the resistors changes significantly with temperature

These resistors are specified as having a tracking coefficient of 5 ppm/C. This means that the ratio of any two resistors will change no more than 5 ppm/C.

With enough resistors, the sample mean TC approaches the population mean TC (assuming random sampling), so the TC of the two samples of resistors should approach equality. In this case, we think that neighboring resistors will have better matched TC and more identical temperatures, so we try to distribute likely-matched resistors into separate groups, so the groups will match each other with fewer resistors required than if we assumed a random sampling of resistor TCs.

 

Offline e61_phil

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Re: Statistical arrays
« Reply #12 on: December 11, 2018, 12:12:08 am »
Ahh  :palm:

Yes, of course you only need to average out the tracking, because your whole divider is in that chip (not like mine, where only the HV part consists of the Caddock resistors) :)

Thanks!
 

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #13 on: December 11, 2018, 12:36:33 am »
Since (on the NOMCA), the resistors are on a ceramic substrate, and the package material is epoxy, one would have to assume that the two materials will behave differently with changes in temperature and humidity (with humidity being the more evil of the two effects).  This would most likely result in a curvature lengthwise-- suggesting that the resistors more near the center of the package will track better than the resistors near the ends of the package.  For better statistical tracking, we can also combine multiple units.  for example, we start with some NOMCA 10K 16-pin (8-resistor) networks.  Now place them side by side on a PCB.  We wire R1 in series on all three packages (resulting in a 30K resistor R1')-- and then do the same with the other 7 resistors in each package.  Now we have 8 resistors (R1' through R8') that have a value of 30K-- with "statistically better" absolute value and TCR tracking.  Let's now make the same 10:1 divider.  We need (3) 30K resistors in series, and (3) 30K resistors in parallel, making 90K and 10K.  It is assumed that resistors next to each other in the network will TCR/humidity track each other better than resistors farther apart from each other-- so, the best arrangement is to "interdigitate" the resistor sets (the 90K and the 10K).  With this in mind, we ignore R1' and R8'.  R2', R4', and R6' are wired in series making 90K, while R3', R5', and R7' are wired in parallel, making the 10K.  Now, wiring the 90K in series with the 10K, we end up with a 10:1 ratio that will track very well-- (at least theoretically, anyway).  My "gut feel" for this arrangement is that it will track within +/-0.1ppm over temperature and humidity.  Not bad actually-- and you can do this over production without selecting parts!
 
The following users thanked this post: TiN

Offline JimmyJo

  • Contributor
  • Posts: 16
  • Country: hk
Re: Statistical arrays
« Reply #14 on: December 11, 2018, 03:27:51 pm »
Dear Doktor Pyta,
I see that you have interleaved the divider resistors of different power dissipation.  Would a current step be useful to determine the power coefficient of ratio?

p.s. Has anyone actually seen the internal construction of of these resistor networks? are they really just 8 resistors side by side on the substrate? 

 

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #15 on: December 11, 2018, 04:43:12 pm »
Dear Doktor Pyta,
I see that you have interleaved the divider resistors of different power dissipation.  Would a current step be useful to determine the power coefficient of ratio?

p.s. Has anyone actually seen the internal construction of of these resistor networks? are they really just 8 resistors side by side on the substrate?

Actually, in his original post, they were NOT described as being "interleaved" (or "interdigitated"), which they should have been.
Edit:  I was looking at the schematic and not the layout.  The layout is done correctly-- (using the inner-most resistors only, and interdigitating the series and parallel connected resistors-- this provides the best chance of tracking).

Yes, the resistors are made on a ceramic substrate one after another.  There is literature on the Vishay website that describes how they are made.  There are also thin-film networks that use a silicon substrate, and these are less desirable for this application (because they flex more).
« Last Edit: December 11, 2018, 05:07:02 pm by Magnificent Bastard »
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #16 on: December 11, 2018, 07:29:51 pm »
Would a current step be useful to determine the power coefficient of ratio?

Other test would be more suitable.
I've just applied 7.07V instead of 10.00V (half of power is dissipated at 7.07V)
34401A in ratio mode changed reading from:
100.0012m @ 10V to
jumping between 100.0011m and 100.0012m @7.07V

so the change is below 1ppm and it is not necessary 100% related  to DUT.
« Last Edit: December 11, 2018, 07:46:54 pm by doktor pyta »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #17 on: December 11, 2018, 08:06:18 pm »
I have no reference for a chance to get more than just the statistical advantage. The point is more from simple reasoning:
The TC of the individual resistors will likely not be purely random but likely will also have some position dependent part, like changing in one direction from one side to the other. In this case the use of interleaved resistors is more efficient in compensating these differences than just averaging over random values.

Also for self heating the interleaving helps to get faster and better temperature matching between the resistors, even with self heating.

I really like the idea of using the array.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #18 on: December 11, 2018, 10:47:08 pm »
Finaly I found a good paper related to our subject.
'20A trapezoidal reference current pulse generator for the evaluation of current transducers'  DOI: 10.1109/I2MTC.2013.6555434


My next step will be testing an array to replace temperature setting divider in LTZ1000 circuit.
From my measurements I expect that one 8-resistor array may do the job.
I have an idea shown on the schematic below.
Divider by 13 could replace 1k/12k divider which might be used only with LTZ1000CH.
Divider by 16 could replace 1k/15k divider which might be used with LTZ1000ACH, unfortunately it will be hot inside. In this case thermal layout seems to be slightly worse.
Off course more possibilities can be obtained using two arrays, but I didn't came up with any universal solution like thermally balanced divider by 14.

If someone has better ideas please share.
« Last Edit: December 11, 2018, 11:47:29 pm by doktor pyta »
 
The following users thanked this post: Magnificent Bastard

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #19 on: December 12, 2018, 01:48:52 am »
Finaly I found a good paper related to our subject.
'20A trapezoidal reference current pulse generator for the evaluation of current transducers'  DOI: 10.1109/I2MTC.2013.6555434


My next step will be testing an array to replace temperature setting divider in LTZ1000 circuit.
From my measurements I expect that one 8-resistor array may do the job.
I have an idea shown on the schematic below.
Divider by 13 could replace 1k/12k divider which might be used only with LTZ1000CH.
Divider by 16 could replace 1k/15k divider which might be used with LTZ1000ACH, unfortunately it will be hot inside. In this case thermal layout seems to be slightly worse.
Off course more possibilities can be obtained using two arrays, but I didn't came up with any universal solution like thermally balanced divider by 14.

If someone has better ideas please share.

This looks good-- no need for the 8th resistor-- you could use a 14-pin device.  3K would be ideal, but not available (ever!) as Vishay NOMCA-- so you can use 2K or 5K.  If you want to special order, you can buy some series SOIC-C at 3K from IRC/TT-Electronics (through Mouser)-- you will have to ask Mouser for a quote, and there will be a minimum buy quantity-- (~100?)  A 14-pin 2K or 5K NOMCA will work just fine, and you might find some in stock (somewhere).

Edit:
Another possiblity is to use 14-pin 1K NOMCA networks (3 of them).  As before, wire all R1's in series to make R1' (a 3K resistor).  Now do the same for R2-R7 making R2' through R7'.  Now wire these 3K resistors (R1' though R7'), as before:  R1', R3', R5', and R7' in series to make 12K; and R2', R4', and R6' in parallel to make 1K.  You now have a 12K:1K divider that will TCR and humidity track within 0.1ppm/K-- and this will be true over production without special parts selection.  Pretty cool!


Aren't Hamon dividers fun and useful?
« Last Edit: December 12, 2018, 02:04:43 am by Magnificent Bastard »
 
The following users thanked this post: doktor pyta

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #20 on: December 12, 2018, 12:05:39 pm »
Has anyone actually seen the internal construction of of these resistor networks? are they really just 8 resistors side by side on the substrate?

zlymex once posted a picture of TDP series array.
https://www.eevblog.com/forum/metrology/any-resistor-series-that-are-stable-wrt-timetemp-but-not-precision-values/msg1056895/#msg1056895

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #21 on: December 12, 2018, 03:09:02 pm »
I wonder how they perform in a 4k : 10k or at least 10k : 25k configuration for the 7V --> 10V boost circuit. Any real world numbers on that?

EDIT: The only used experience seems to be with LM399

-branadic-
« Last Edit: December 12, 2018, 04:04:26 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #22 on: December 12, 2018, 03:24:51 pm »
This will be next step. Will be tested for sure.

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #23 on: December 12, 2018, 07:11:03 pm »
Have ordered a few NOMCA16035001 for experiements as NOMCA14035001 are equal to NOMCA14032001 and NOMCA16032001 simply unobtanium. So the 8th resistor can be used in parallel to some low ohmic resistor for fine adjusting the 10V output voltage.

-branadic-
« Last Edit: December 12, 2018, 07:14:52 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline razvan784

  • Regular Contributor
  • *
  • Posts: 90
  • Country: ro
Re: Statistical arrays
« Reply #24 on: December 12, 2018, 08:57:44 pm »
Played with this (in LTSpice) some time ago as a replacement for 13k-1k.
Tried a layout that is as symmetrical as possible.
Didn't manage to built it yet.
 
The following users thanked this post: doktor pyta, Magnificent Bastard

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #25 on: December 12, 2018, 10:04:08 pm »
The 13:1 divider is a really nice solution. However it would provide only a limited amount of averaging. So I am not so sure it would be really stable and low TC.

For the 15:1 ratio, there is an easier solution: 5 in series and 3 in parallel.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #26 on: December 12, 2018, 10:44:54 pm »
razvan784, thanks for the inspiration!

Below is my version of divider by 14.
1.6k output resistance should not be a problem.
The resistors are grouped in pairs with same power dissipation.

What do You think guys ?


Edit: tempco measured for comparision.
Let's call it layout v.1.

T1=25.2'C
T2=60.4'C
mean T.C. of division ratio -3.9ppm/'C

and

division coefficient measured: 14.00088
« Last Edit: December 31, 2018, 07:11:35 pm by doktor pyta »
 

Offline razvan784

  • Regular Contributor
  • *
  • Posts: 90
  • Country: ro
Re: Statistical arrays
« Reply #27 on: December 13, 2018, 03:38:44 pm »
Thank you doktor pyta. Yes, your design is better, more symmetric.
Kleinstein, I agree that averaging is limited; these designs would work well however if the individual TCs were not just randomly distributed but more like a gradient, as you noted earlier. On the other hand I have done no such measurements; I finally abandoned this design in favor of just using two selected foil resistors.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #28 on: December 13, 2018, 09:34:03 pm »
Update: divider /14 with slightly corrected layout.


Edit: tempco finally measured.

T1=25.2'C
T2=60.0'C
mean T.C. of division ratio 0.41ppm/'C

and

division coefficient measured: 14.00055

« Last Edit: December 30, 2018, 12:39:38 pm by doktor pyta »
 
The following users thanked this post: EmmanuelFaure

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #29 on: December 14, 2018, 09:04:42 am »
For the layout, selecting the resistor positions, the power dissipation should not be that important, if used at a reference, as the power dissipation would be constant. The power dissipation could be important if the divider is used at variable voltage, e.g. to set an amplifier gain.

So It is not sure the 2 nd version is actually better than the 1 st. It is likely more about a gradient in the resistor properties, so to have the resistors well mixed and evenly distributed. Here the 1st version is still not that good as it has the resistors towards ground are in the center - so a difference from center to edges would show up. The 2nd version is more like a little biased left to right.
 
The following users thanked this post: doktor pyta

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #30 on: December 14, 2018, 05:51:09 pm »
The idea is to keep the temperature as uniform as possible across the ceramic substrate.  This will allow for better tracking over time and temperature-- (the Ta2N resistors "age" by heat and oxygen availability-- keeping them close to each other in temperature will aid in uniform aging characteristics).  Also, the farther away from the center of the package, the less the resistors will track; and resistors next to each other will track better than resistors far apart; and this is due to the difference in TCE between the ceramic substrate and the epoxy package; and also-- humidity will have a similar (more pronounced) effect.  The layout should be as symmetrical as you can get it while observing the above constraints.  As oxygen attacks the tantalum nitride compound-- tantalum pentoxide is formed, which is not conductive (and is unaffected by water molecules-- even under power).  Since these are thin-film resistors, that leaves only one drift mechanism-- and they always drift "up" in value.  When a thick enough layer of tantalum pentoxide forms-- the resistors will slow their drift rate to almost zero.  This is what the 90-day burn-in at 125oC is all about-- to artificially age these resistors so that the subsequent drift over time is almost zero-- doing this in an ozone-rich environment will speed up this process.
 
The following users thanked this post: e61_phil, doktor pyta

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #31 on: December 14, 2018, 06:31:28 pm »
I'm pretty sure the resistors are processed in a panel for economic reasons and cut into pieces after PVD process. So one can expect the resistors to be very uniform within one ceramic die. So it's all about heat distribution across the ceramic die.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #32 on: December 14, 2018, 06:36:38 pm »
I wonder if the temperature distribution inside statistical array could be evaluated objectively before practical measurements.
Some simple freware thermal simulation software would be useful but without knowing dimensions, thermal conductivities specifc to the array it may not be good enough.
Any ideas?
« Last Edit: December 14, 2018, 07:15:11 pm by doktor pyta »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #33 on: December 14, 2018, 07:20:28 pm »
Well, it's not that you now nothing about the array... you can estimate the size of the ceramic, you now it's alumina (Al2O3) substrate and know about thermal conductivity and that there are 7 or 8 resistors on it. But I'm sure there is no need to simulate the heat distribution, but to think about how heat will spread.
On the other hand you can use any thermal simulation program available to get an idea of what is thermally going on. If you want to do a real world simulation you need to sacrifice one unit, crack it open and measure the dimensions including the thickness of the TaN layer.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #34 on: December 14, 2018, 08:08:57 pm »
OR, you can just measure it!  Measure R2-R8 while applying power to R1 (vs. no power).  Now do the same by powering R2, and measuring the other resistors.  Same with R3 (and so on) until you have a complete set of measurements.  With some mathematical magic, you should be able to understand how each resistor's dissipation affects the other resistors.  Intuitively, it "feels" like having the higher-dissipation resistors on the ends of the package will help keep the substrate temperature gradients lower (as opposed to having the higher dissipation resistors in the center of the package).  This can be verified with the above experiment.  An additional interesting question is: "Do the temperature gradients of the substrate change when mounted to a PCB vs. a package in "free air"?

Fun stuff!
 

Offline JimmyJo

  • Contributor
  • Posts: 16
  • Country: hk
Re: Statistical arrays
« Reply #35 on: December 15, 2018, 06:53:54 am »
Inspired by Zlymex, I present to you this 1K version of the NOMCA1603.  Upper right corner is Pin1 in both pictures.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #36 on: December 15, 2018, 10:25:08 am »
Nice job JimmyJo!

Below distances measured from the image.

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #37 on: December 15, 2018, 01:09:15 pm »
Since Vishay is not a european company the distance between the resistors is most likely 0,3175mm or 0,0125". There seems to be a small parallax error from the microscope.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #38 on: December 15, 2018, 07:50:01 pm »
Just curious if NOMCA16035001 could perform as temperature stabilized 10k resistance reference just as good as Fluke SL935 from TiN? There are enough 5k resistors inside one package to realize that.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #39 on: December 15, 2018, 08:12:19 pm »
One could use 1 or 2 of the resistors for heating, but even than the resistor is encapsulated in epoxy. So there can be humidity effects.

Extra heating might be useful to compensate for self heating, if the resistors are used for something like a divider with a variable voltage.
 

Offline splin

  • Frequent Contributor
  • **
  • Posts: 999
  • Country: gb
Re: Statistical arrays
« Reply #40 on: December 16, 2018, 03:42:48 am »
One issue with statistical arrays is that they multiply the number of leakage paths and also makes guarding more of a problem. If, for example, you have 10 resistors in series forming the bottom leg of a divider, then to fully guard it you would need 10 guard voltages.

In practice, given the lower voltage drop across each resistor, you would probably compromise by guarding fewer nodes at the expense of potentialy a bit more leakage. The calculations might be interesting if the leakage resistance of the PCB is significantly non-linear with voltage - perhaps due to humidity and/or surface contamination (I've no idea if it is).
 
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #41 on: December 16, 2018, 11:42:54 am »
I was thinking about a pcb with the 5k resistor array connected as a 10k reference on the component side of the board and an oven control circuit on the back or the pcb with a big copper area on the back mounted to some BPR10 resistor forming an oven that is running at 35°C ... 40°C.
The thermistor could be some SMD longterm stability type mounted on the component side close to the resistor array. If permanently powered there should be no problem with humidity after a small run in phase. It's then a question of thermal insulation and oven temperature stability (0.1°C, 0.01°C or 0,001°C). But the oven could be of much smaller size compared to SL935.

-branadic-
« Last Edit: December 16, 2018, 12:55:18 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #42 on: December 30, 2018, 11:09:47 pm »
Little update in first post and Reply #28 and Reply #26.

In short, T.C. :

-from Reply #26: -3.9ppm/'C
-from Reply #28: +0.41ppm/'C

I couldn't believe my eves so I double checked everything and I made multiple measurements.
Both layouts were using the same specimen of NOMCA array.
« Last Edit: December 31, 2018, 08:01:38 am by doktor pyta »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #43 on: December 31, 2018, 12:02:20 am »
Can you rearrange the resistors and repeat the measurement? R7 + R8 as the center resistors, R5 + R6 symmetrical an both sides, followed by R3 + R4 and finally R1 + R2 (see attachement).

-branadic-
« Last Edit: December 31, 2018, 08:49:23 am by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #44 on: December 31, 2018, 12:27:53 am »
@branadic

proposed layout is definitely worth testing.
I will do it within one or two days.

P.S. please remove connections on the layout part of the attached drawing so the readers won't be confused.
I will post new layout drawing soon.

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #45 on: December 31, 2018, 12:23:11 pm »
Measurements of the layout suggested by branadic.
Proper drawings below.


T1=25.3'C
T2=63.6'C
mean T.C. of division ratio -5.8ppm/'C

and

division coefficient measured: 14.00017
 
The following users thanked this post: branadic, Andreas

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #46 on: December 31, 2018, 04:34:47 pm »
There are a few possibilities left, like having R7 and R8 on the outer ends and R5 and R6 in the center.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #47 on: December 31, 2018, 11:31:55 pm »
Hmm,

did you use one and the same sample for all these 3 wirings
or are that different samples for each wiring?

with best regards

Andreas
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #48 on: January 01, 2019, 12:02:12 am »
I used one sample in these 3 layouts. Otherwise it would not make big sense.
Cheers! Happy New Year!

Offline e61_phil

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Re: Statistical arrays
« Reply #49 on: January 01, 2019, 08:47:31 pm »
Did you measure the individual TCs?

I wonder if these measurements are due to the position or due to lucky matching of the resistors for the different sides. Or is that already clear and I overlooked something?
 

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #50 on: January 01, 2019, 09:43:59 pm »
It would be good idea to have individual resistors; T.C. measured.
It may contribute our understanding of the problem.
I have to think how to use my gear to make 8x 4wire measurements at two temperatures easiest way.

At this point my theory of best performance of layout v.2 is that resistors R4, R2 have equal temperatures.
Same with R1, R3.


Offline antintedo

  • Contributor
  • Posts: 45
  • Country: pl
Re: Statistical arrays
« Reply #51 on: January 14, 2019, 12:58:05 am »
I found some time to perform TCR measurements. So far only 1 sample of 10k NOMCA was tested.
1:2 divider built out of this sample measured 0.4ppm/K which seems to agree with the results below.

Code: [Select]
Sample #1
No  T1      R1          T2      R2          dT      TCRavg  TCR/TCRavg
1   26      10.00243    64      9.99943     38      -7.89   77%
2   26      10.00286    64      9.99750     38      -14.11  138%
3   26      10.00363    64      9.99877     38      -12.79  125%
4   26      10.00189    64      9.99913     38      -7.26   71%
5   26.5    10.00284    63      9.99927     36.5    -9.78   95%
6   26      10.00264    63      9.99840     37      -11.46  112%
7   26      10.00179    64      9.99839     38      -8.95   87%
8   26.5    10.00230    64      9.99862     37.5    -9.81   96%

Sample #2
No  T1      R1          T2      R2          dT      TCR     TCR/TCRavg
1   25      10.00220    67      10.00106    42      -2.71   73%
2   25      10.00235    67      10.00032    42      -4.83   129%
3   25      10.00193    67      10.00057    42      -3.24   87%
4   25      10.00160    67      9.99992     42      -4.00   107%
5   24.5    10.00193    67      10.00021    42.5    -4.05   108%
6   25      10.00160    67      9.99982     42      -4.24   113%
7   25      10.00205    67      10.00074    42      -3.12   83%
8   25      10.00214    67.5    10.00056    42.5    -3.72   99%
« Last Edit: July 04, 2019, 06:44:01 pm by antintedo »
 
The following users thanked this post: Andreas, e61_phil, EmmanuelFaure, doktor pyta, alm, razvan784

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #52 on: January 14, 2019, 08:25:20 pm »
Thanks for the measurement,
I never had expected that the stray between neighboured resistors is that large.
I had expected more likely a rising T.C. from one side to the other or a center symmetrical behaviour.

with best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #53 on: January 14, 2019, 08:41:18 pm »
I would have expected more equal values for t.c. between single resistors as the ceramic area is comparable small so are the resistors and they are manufactured in thinfilm technology, so multiple networks are processed in parallel and films are highly uniform over small areas.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #54 on: January 14, 2019, 09:34:41 pm »
I found some time to perform TCR measurements. So far only 1 sample of 10k NOMCA was tested.
1:2 divider built out of this sample measured 0.4ppm/K which seems to agree with the results below.

Code: [Select]
No  T1     R1         T2   R2        dT     TCR
1   26     10.00243   64   9.99943   38     -7.89
2   26     10.00286   64   9.99750   38     -14.11
3   26     10.00363   64   9.99877   38     -12.79
4   26     10.00189   64   9.99913   38     -7.26
5   26.5   10.00284   63   9.99927   36.5   -9.78
6   26     10.00264   63   9.99840   37     -11.46
7   26     10.00179   64   9.99839   38     -8.95
8   26.5   10.00230   64   9.99862   37.5   -9.81

These results seem suspiciously variant (not the absolute values, but the relative TCRs).  Can you describe in detail how you made these measurements?
 

Offline antintedo

  • Contributor
  • Posts: 45
  • Country: pl
Re: Statistical arrays
« Reply #55 on: January 15, 2019, 12:13:41 am »
These results seem suspiciously variant (not the absolute values, but the relative TCRs).  Can you describe in detail how you made these measurements?
The resistor was soldered on a FR-4 SOIC to DIP converter and attached to a stabilized heater with a pretty large heatspreader in between.
The whole setup was put into a lightly insulated aluminum box.

For every measurement wires were soldered to corresponding DIP holes.
I used 34401A in 4-wire mode to perform 3 measurements: first at ~25C then at ~65C then again at ~25C.
Temperature readout was done with a thermocouple placed near the middle of the package. Plenty of time was given to reach thermal equilibrium after heating/cooling cycle and soldering.

After doing all 8 resistors, I repeated the measurements in the same order as first round of measurements.
I didn't write it down, but I think repeatability of resistor values between two runs was well below 10 ppm, otherwise I would have just discarded the results.
Room temperature measurements after the experiment agreed to 7ppm with measurements taken the day before. I also used VHP101 resistor box as a sanity check.

I think the main downside of this setup was mechanical stability and the need to solder to the board between every measurement, potentially leaving flux and skin contaminants on the surface. I will try to eliminate this and re-test, but it has to wait until the weekend.

I had expected more likely a rising T.C. from one side to the other or a center symmetrical behaviour.
Andreas
I was surprised as well. Out of curiosity I tested one more more sample today and got very different results. TCR was much lower, 3.7 ppm average +/- 1 ppm spread with no obvious patterns. It is certainly possible something was wrong about my initial methodology. This time I paid more attention to handling the board when soldering and fixed the board rigidly using bolts instead of springs.
 

Online nfmax

  • Super Contributor
  • ***
  • Posts: 1554
  • Country: gb
Re: Statistical arrays
« Reply #56 on: January 22, 2019, 09:29:08 am »
Maybe the effect of thermally-induced strain changes from the mounting arrangement? All resistors are strain gauges!
 
The following users thanked this post: antintedo

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #57 on: April 10, 2019, 04:06:18 pm »
Since I was interested in the question if the NOMCA arrays are useful for a 7V --> 10V boost circuit and noone has yet answered the question by experiment results I've now ordered the boards with my design shown a few posts before. They will receive soon and I will then perform some measurements such as ratio measurements over temperature. If they perform well I can directly use them for my LTZ1000 references.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: Andreas, e61_phil, 2N3055, ch_scr, Grandchuck

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #58 on: April 13, 2019, 12:03:36 am »
branadic,

I did not have time to test the idea but this is what I came up with (drawing).
Layout for this can be designed symmetrically and I think it could be further improved by paralleling similar dividers.

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #59 on: June 25, 2019, 09:30:32 am »
I was looking at Nomca arrays for the 13:1 divider for a LTZ1000 reference and found similar results as above when using 3x 5K in series with 4x 5K in parallel. Division ratio TC was less than 1 ppm/°C for two units and 2.4 ppm/°C for one unit.

Now, when i look at the measurements presented above by antintedo for the TCs of single resistors in a Nomca array, i'd like to show where one may get by selection. This is only an example, but my conclusion would be that a Nomca array can be used for the 7V to 10V gain stage, if one finds a practical solution for free combinations. As soon as i have time i'll check this with my 5K Nomca arrays.

Regards, Dieter

PS: Sorry for the confusion, but those measurements were not from Magnificent Bastard.
« Last Edit: June 26, 2019, 05:01:13 pm by dietert1 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #60 on: June 25, 2019, 10:02:18 am »
Measuring and selecting TC is a slow process. There is an additional difficulty - soldering and board stress can be part of the TC. So it would need TC measurement after soldering and than adapting the layout to the appropriate connections.  :-//
So it is more like a very limited number of option one could chose from by adding bridges, like leaving out one resistor.

Getting 2.4 ppm/K for the divider is quite a bit. The 3 in series and 4 in parallel version is already averaging over 3 and 4 resistors so that chances are low to get such a poor combination with individual resistors at 5 ppm/K spread.

Just for a test one could try re-soldering the bad divider.

The other point is that often the TC is only the easier part. Long term drift may be of similar importance, with little chance to measure and select upfront.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Statistical arrays
« Reply #61 on: June 25, 2019, 10:15:43 am »
Perhaps a socket could be used for testing, to avoid heat stress to the package?
Like this one?
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #62 on: June 26, 2019, 05:22:34 pm »
Measurement of the TCs of the eight resistors in the array requires a 8x multiplexer and one temperature cycle. Yes, maybe the resistor array can't be soldered after the measurement without affecting it's behaviour. That needs to be decided by measurement.
Anyway, results better than 1 ppm/°C without selection may be good enough to try and control the temperature of the resistor array in a way similar to the heated zener reference.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #63 on: June 26, 2019, 07:53:42 pm »
One normally would not even need 1ppm/K performance in the LTZ1000. Without adjustment (R9) the TC of the LTZ1000 circuit is more like in the 0.2-0.5 ppm/K range which corresponds to some 20-50 ppm/K for the resistors. The individual R9 trimming would also include resistor TC.
However this is only the TC part - it does not help with long term drift.

I don't know how much the soldering effects the TC. Besides TC soldering can effect the short time drift and possibly hysteresis effects. These could be confused with a TC error in an early test. Still a significant drift / error in the TC test is not a good sign, even of not due to actual TC but hysteresis or aging.

I don't think temperature control is helping, as the TC is not the real critical parameter - it's more like aging that is critical and the TC is more used as an indication of stability, with the real thing very difficult to measure. Some burn in after soldering could help - though it would also hide things. Ideally one would measure before burn in and get suspicious if burn in changes too much.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #64 on: June 26, 2019, 08:11:40 pm »
While looking at the 34470A teardown, at 18:40-22:45 for example, they are using rather common resistor parts around the LTZ1000, imho.
« Last Edit: June 26, 2019, 08:28:14 pm by imo »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #65 on: June 26, 2019, 08:27:06 pm »
The demands on the resistors around the LTZ1000 are not that high. It is only of one wants really low long time drift (e.g. < 2 ppm/a) that is really takes really stable resistors (still only some 50 ppm/a). AFAIK the 34470 is not the highest stability - more like having the LTZ1000 for low TC and low noise.

The more critical resistors are in a 7 to 10 V stage. Another point could be a gain stage or divider and possibly a resistor in an ADC, that can effect linearity due to self-heating.  Selft heating could increase the temperature of a resistor by something like 0.1-1 K. So an TC of 10 ppm/K could cause an INL error of some 1-10 ppm.
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #66 on: June 26, 2019, 08:29:43 pm »
While looking at the 34470A , at 37:45 for example, they are using rather common resistor parts around the LTZ1000, imho.

Ok as a volt-nut you have to count the vishay VSMP series (around $15 per pop) as rather common.

the TC of the LTZ1000 circuit is more like in the 0.2-0.5 ppm/K range

from where do you have such large values?
my results with long legs on the non-A version are below 0.06 ppm/K

https://www.eevblog.com/forum/metrology/ultra-precision-reference-ltz1000/msg2301414/#msg2301414

with best regards

Andreas
 
The following users thanked this post: iMo

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #67 on: June 26, 2019, 08:41:38 pm »
Quote
from where do you have such large values?
my results with long legs on the non-A version are below 0.06 ppm/K

This are numbers I can confirm with 400k resistors unpopulated, short legs.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #68 on: June 26, 2019, 09:13:18 pm »
Quote
from where do you have such large values?
my results with long legs on the non-A version are below 0.06 ppm/K

This are numbers I can confirm with 400k resistors unpopulated, short legs.

-branadic-

ok this confirms my opinion that it is not a good idea to shorten the legs on the non-A Version.
(with half shortened legs I had also a larger T.C. than before).

with best regards

Andreas
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #69 on: July 01, 2019, 02:38:18 pm »
In the meantime i have done some multiplexer measurements on three more Nomca 1603 8x 5K arrays. They got mounted on SMD adapters with gold plated pins and cleaned in ultrasonic bath. Then i setup 4 wire R measurements using a HP3456A at 0,1 mA measurement current with averaging to get below ppm resolution. Temperature span was 18 °C. Error on a single resistor TC measurement is about 0,03 ppm/°C.

Results on TCs for each part (average( 8 ), stdev( 8 ), span( 8 )):

D     4,34 +/- 0,62 ppm/°C, max-min = 1,85 ppm/°C
E   11,23 +/- 0,63 ppm/°C, max-min = 1,74 ppm/°C
F   16,20 +/- 1,55 ppm/°C, max-min = 4,51 ppm/°C

According to my measurements those are pretty serious parts, well suited to produce temperature compensated voltage dividers. And it makes complete sense to measure TCs of individual resistors in order to determine a good combination - better than statistics! Combining  resistors properly each of my three parts should give me a 13:1 divider with less than 0,03 ppm/°C on the division ratio. At least that's the result of a simulation using the measured curves of the individual resistors.

Regards, Dieter

« Last Edit: July 01, 2019, 02:41:29 pm by dietert1 »
 
The following users thanked this post: branadic, ch_scr, antintedo

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #70 on: July 01, 2019, 02:50:52 pm »
Thanks Dieter,

since I'm still waiting for my boards to arrive, because of the inability of DHL or customs the boards were sent back to Elecrow, I ordered a zero force adapter

eBay auction: #282277360455

that received today to perform similar measurements. Need to perpare a cable for multiplexer of my Prema 5017SC.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #71 on: July 02, 2019, 10:19:48 pm »
Today i got more evidence for the concept of combining selected resistors within a Nomca 1603 array.

The part B that i mentioned above giving 2.4 ppm/°C showed less than 0.5 ppm change during a 18 °C temperature cycle now after rewiring. It's an upper limit because from the multiplexer measurements i expected a residual of -0,015 ppm/°C on the division ratio and i couldn't see anything with our HP 3456A. That part B is the only one in six that exceeds the 5 ppm/°C tracking spec in the datasheet with a span of 6.7 ppm/°C between measured min and max TCs.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #72 on: July 03, 2019, 06:36:53 am »
Whether selecting the resistors is only practical if soldering / mounting on a board does not change the TC significantly. For PCB use, there are usually also not that many option to choose from. There may be a few jumpers for options but usually only a few, especially if not many of the resistors are unused.

If soldering has only a limited effect it may at least be possible to test the resistors (if the contacts in an affordable socket are good enough) and than decide on the orientation of the chip. This may avoid a worst case (with a suitable use of the resistors) - which could already be quite a gain.
Selecting a suitable resistor combination is much easier with individual resistors - more like hard with an array.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #73 on: July 03, 2019, 08:23:49 am »
You are not asking, but i will tell you how to do it.

The construction looks like the thin film arrays of the HP 3456A. The Nomca arrays were put onto SMD to DIL Adapters with gold plated pins. On the carrier board there are gold plated socket strips. Wiring is underneath the carrier board with short soldered bridges, that i prefer over the plugged configuration bridges of the HP3456A. This construction isn't ideal in the sense that is has short copper tracks and wires and the pins/sockets in the path that are not tracking well during fast temperature changes, but the arrays are not soldered anymore after characterization.

I used a similar construction before with S102 resistors soldered on top of 4 pin modules so that one can use 4 wire measurements for characterization and not solder them afterwards. In that case the doubled pins give some redundance when the parts are on the carrier board afterwards. I also thought about making an array of S102 resistors and see how they compare. Individual TCs will be smaller, but temperature tracking will be worse, so finally the result may be similar.

Regards, Dieter

PS: One may even think about using TSSOP32 adapters for the Nomca arrays in order to have 4 wire connections.
« Last Edit: July 03, 2019, 08:31:58 am by dietert1 »
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #74 on: July 03, 2019, 06:40:29 pm »
The NOMCA arrays are not particularly good for certain applications:

!. Ta2N resistors are rather noisy and has 1/f noise as well, all above the theoretical resistor noise level.
2. While short term stability is not bad, long term stability is questionable especially at low PPM.
3. Shelf stability at 25°C is only fair, tolerance drift at up to .01%, ratio drift up to .002% for the first year, Vishay states no specification beyond 1 year.  If these exhibited good long term stability you can be sure Vishay would have put that spec in there.
4. Ratio matching of 5 PPM/°C is hardly impressive, other types can be below 1 PPM/°C including wire wound.
5. All factors affecting the specifications more or less apply to any resistor type being used.  This requires careful design to manage thermals and unequal heating even on a single substrate.
6. Sensitivity to soldering, per the data sheet, possible significant (and permanent) changes can occur, not only to tolerance but TCR and stability.

In the case of the current discussion on a 7V to 10V boost circuit, any deviation of resistor matching will be multiplied by the amplifier gain, this includes noise and drift of resistor specs over time.  While a NOMCA array may have an attractive price, its specifications for this particular application leaves a lot to be desired.  If you are expecting a long term stability and low noise 10V output over months, these are not the resistors to pick.  The Vishay data sheet is pretty clear on the specifications they do give, NOMCAs are not for this application.

S102 resistors also stink for this application.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #75 on: July 03, 2019, 09:58:03 pm »
Sorry, i would rather prefer learning from industry leaders who have been applying thin film technology in precision measurement for many years.

I got somewhat impressed that 6 $ plus proper wiring of that part gets me a voltage divider with less than 0.5 ppm change of the division ratio during an 18°C oven cycle. Which is less than 0.03 ppm/°C. And this is the worst of six parts and still so good that i can't really measure the residual change. I know that number doesn't appear in the data sheet. Please don't tell me you know how long the adjustment will last if i don't touch the part anymore. I think right now nobody can tell.

Regards, Dieter
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #76 on: July 04, 2019, 02:32:50 am »
Well dietert1, considering that you haven't been playing with this part for very long, you don't know what the longer term drift is and since Vishay doesn't specify longer term drift, you can accept that if Vishay doesn't provide longer term drift, there is no significant improvement over time or they would have said so.  I know quite well the performance limitations of many resistor technologies and I also know very well how well my precision wire wound resistors perform.

Carefully read the NOMCA data sheet from Vishay, the information I quoted is directly from their own data sheet, so no you don't have to believe what I said because it is in their data sheet quite plainly.

As far as resistors go, I have nearly 46 years in the resistor business, I think I know a tad more about the subject than you think.....I have the instruments that can measure resistance better than most anybody here and to a very high accuracy traceable to NIST, my uncertainty is 0.15 PPM on my primary reference.

However, you are most welcome to believe whatever you wish.
« Last Edit: July 04, 2019, 06:10:44 pm by Edwin G. Pettis »
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Statistical arrays
« Reply #77 on: July 04, 2019, 04:46:17 am »
Edwin G. Pettis

Quote
my uncertainty is 0.15 PPM on my primary reference.
Aww, beat me by 0.01 ppm there  ;D

Jokes aside, how possible would it be to order small qty (10-20?) of PWW network for purpose of 7V/10V and such, but with custom values per network? I'd be interested, and probably some others too to make it into a batch? One could find out precise values using calibrated decades like RS925, and then specify final divider, like 7.1426kohm/20.5216kohm as example. Of course, each network's precise value would be different, matched to specific LTZ circuit.  :popcorn:
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 
The following users thanked this post: Theboel, MiDi

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #78 on: July 04, 2019, 05:48:07 am »
Well, if the TDP / NOMCA networks were really that bad in aspect of longterm stability and noise I'm pretty sure they wouldn't be used in Wavetek/Fluke 7000 and 7001 and if TaN resistor networks were bad in general Fluke wouldn't have used them in 732B and 732C and replaced the former single wirewound resistors by resistor network. So please not another flame war on resistor technology.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #79 on: July 04, 2019, 06:22:21 am »
Sorry Mr. Pettys, we understand that you prefer wire wound resistors and you are not at all interested in thin film technology or statistical arrays.
At least one of your statements above was plain false. Analog Devices offers the AD587 reference part, which is well known for its good long term stability (Geller reference). Know what, the 10 V to 7 V divider within that part is based on thin film technology. No wire wound resistors there.

Resistor arrays offer exactly what TiN is asking for: A configurable voltage divider that supports many different division ratios and pretty fine adjustment. And yes, this needs implementation and measurement instead of general wisdom and reading between the lines.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #80 on: July 04, 2019, 02:08:04 pm »
The specs for long term stability (some are shown in the environmental tests) are not that good. However with something like long term aging this is usually only a best guess, nothing easy to measure and check for every unit.  I am not 100% sure one can use the TDP series to directly use for the SMD version. The SMD version is much more sensitive to board stress.
Especially with the DIP versions, Fluke may have done some selection up front, to get better than typical units.

Especially in a cost critical or small size case wire wound resistors may not such a good alternative. Even if only replacing 2 resistors such an SMD array can be cost and size effective. For some cases there may be better alternatives like the MORN series (4 resistor in SSOP8) with better specs.
When using several resistors in a series and / or parallel configuration the relative TC matching is expected to improve (at least on average and in most cases). However one can still have bad luck and get a relatively poor matching.

The resistor excess noise may in some cases become a problem - but at least from the specs (< - 30dBi) the NOMCA arrays are not that bad, despite the small size. For a LM399 based reference the extra noise should not be a big issue - for a LTZ1000 based reference, one may notice some additional noise. For the resistors directly at the LTZ1000 excess noise should not be relevant with thin film resistors.
The TaN Arrays HP was / is using in there DMMs tend to be larger area with thus a chance for lower noise and possible better matching.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #81 on: July 04, 2019, 04:02:58 pm »
My statements were directly based on the Vishay NOMCA data sheet, there was no interpretation involved.  Comparing the AD587's internal resistors to the NOMCA is not valid, they are not identical even if they are using the similar Ta2N resistors which is generally used in most ICs.  That is the same as claiming all precision wire wound resistors are the same which they are definitely not.  Just because I specialize in wire wound technology doesn't mean I don't know how the other resistor technologies work.

The networks you refer to in the the 7000 are not identical to the NOMCA arrays, the manufacturing processes are not the same so you cannot compare the two.

I specifically stated that in my opinion, the NOMCAs are not well suited to a voltage booster circuit requiring low PPM performance.  So far all I've seen is mainly measurements on the arrays and none done on an actual booster circuit, those are two entirely different operating conditions.  Let's see what your measurements come up with in an actual booster circuit first before throwing rocks at the windows.

For that matter, I have made matched tracking TCRs of resistors down to <0.2PPM/°C which has been independently verified over long term period.  No it is not easy to achieve, the physical circuitry details are just as important, if not more important, than the resistors themselves.  Creating a very low tracking TCR depends on a lot more than just the resistors themselves, thermals can be very complex to control and can completely wipe out the low tracking TCRs of the resistors.
 

Offline Magnificent Bastard

  • Regular Contributor
  • *
  • Posts: 139
  • Country: aq
Re: Statistical arrays
« Reply #82 on: July 04, 2019, 05:34:41 pm »
The AD587 (et. al.), as well as the LT5400 are made with silicon-chromium resistors that are passivated with silicon-nitride.  The ratio of silicon-chromium to silicon-nitride is adjusted to minimize absolute TCR.  The Fluke 7000 series uses Vishay TDP networks, which are nickel-chromium elements that are passivated with tantalum-nitride.  Same as above, they tweak the ratio of nickel-chromium to tantalum-nitride in order to minimize the absolute TCR.  NOMCA are made with a 100% tantalum-nitride process, and they are self-passivating.  During the tantalum sputtering process, the percentage of nitrogen in the chamber is controlled to minimize the absolute TCR.  Because of the way they are made; where the resistors are sputtered onto a substrate in a controlled environment; adjacent resistors will track very well in TCR (and hence are valuable as a ratio device).  In resistor networks, the resistors have a tendency to track each other over time because resistor drift is driven by temperature, and the resistors on a network are nearly the same temperature.

Most of the evils that engineers have experienced with networked resistors have to do with the thick film process, which is totally different than thin-film or metal film.  The data sheets for these metal/thin film parts are VERY conservative.  The tracking TCR is related to R1 against any other resistor in the network (over 6-sigma); but resistors that are next to each other (and closer to the center of the network) WILL track each other far better than data-sheet specs say they will; and the networks can be selected for additional needed performance.  Once built and operating, the ratio of such a network will track within 1ppm/a, but usually much better than that.  You DO have to follow some common-sense rules when you are designing such a ratio device.  Even though two networks come from the same batch, they often TCR track very well, but you cannot rely on this; only that resistors within an individual network will track.  You have to make certain that each network in a multi-package divider shares responsibility for both the upper and lower legs of the ratio divider.  The resistors closer to the ends of the network (R1/R8 or R1/R7) are not going to track as well as the resistors closer to the center of the network; and this is all due to package strain.  A statistical array of at least 3 networks will perform better than just one network...  (and so on and so forth...)  You can figure out all of these rules on your own through testing; but don't forget to test for humidity, which can have bigger effects than temperature.
 
The following users thanked this post: doktor pyta, razvan784

Offline try

  • Regular Contributor
  • *
  • Posts: 112
  • Country: de
  • Metrology from waste
Re: Statistical arrays
« Reply #83 on: July 04, 2019, 06:44:15 pm »
As far as resistors go, I have nearly 46 years in the resistor business, I think I know a tad more about the subject than you think.....I have the instruments that can measure resistance better than most anybody here and to a very high accuracy traceable to NIST, my uncertainty is 0.15 PPM on my primary reference.

Well, Edwin, claiming 46 years in the resistors business and having instruments that can measure resistance better than most anybody does not seem to be sufficient at all when it comes to delivering 10k Ohm flat with 10 ppm precision to a customer.

OK, mistakes can happen.
But refusing to deliver a free replacement shipment upon delivering the requested measurement proof is pretty disappointing.

« Last Edit: July 04, 2019, 06:47:58 pm by try »
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #84 on: July 04, 2019, 07:23:33 pm »
While TCs within a Nomca package resemble better than TCs between different packages, in my measurements i have not seen that resistors close to each other resemble better. Neither have i seen that "inner" resistors are matching better than resistors near the ends of an array.

Apparently resistance tolerances within each array are very small, like +/- 0,005 %. This means it is very difficult to tune a Nomca divider for division ratio by using a certain combination of it's resistors. So when making a 7 -> 10 V booster with a Nomca array, additional tuning resistors are necessary to get close. A usual 1% 100 ppm/K MF resistor that one may have around is not good enough. If it gets its TC attenuated by 200 it will still ruin anything tuned to 0,1 ppm/K or better.

The reference booster resistor array of the HP 3456A has lots of different resistor values organized like a binary DAC. They get so close that in the end a usual  potentiometer is good enough for tuning.

Regards, Dieter
 

Offline antintedo

  • Contributor
  • Posts: 45
  • Country: pl
Re: Statistical arrays
« Reply #85 on: July 04, 2019, 07:34:14 pm »
While TCs within a Nomca package resemble better than TCs between different packages, in my measurements i have not seen that resistors close to each other resemble better. Neither have i seen that "inner" resistors are matching better than resistors near the ends of an array.
Thanks for your input, that was my observation as well. Since chip layout of NOMCA is not symmetrical, did you perhaps observe any patterns? For example in my test of 5 samples the resistor between pins 2 and 15 was almost always an outlier.

I will try to eliminate this and re-test, but it has to wait until the weekend.
So much for that - I'm still trying to setup a decent thermal chamber with a controller. However I tested a few more samples with my previous setup. The TCR pattern between resistors was similar but tracking much better. I edited the original post to include the results of 1 more sample that I had time to check for repeatability.

Regarding recent discussion, I was curious about 1/f noise of thin films, especially NOMCA vs larger models. I did a quick test dividing LTZ1000 voltage by 2 and measuring 0.1-10Hz pp noise within 1 minute period.

LTZ1000: 0.15 ppm
noname THT 3.3k metal film: 0.16 ppm
NOMCA 10k, 3 samples tested, 6 resistors used every time: 0.22-0.25 ppm
S102K (refurbished) 11k, 6 resistors used: 0.16 ppm

Additionally, long time ago I found this publication containing 1/f noise tests of various resistor types and models. Seems like a lot of models achieve good performance with almost no 1/f noise, even in SMD 0402 size.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #86 on: July 04, 2019, 09:14:40 pm »
My Nomca TC measurement results.
There is one measurement (R7 in array "B") that exhibits hysteresis. So that measurement may have some problem and be the reason why that part violates the 5 ppm tracking spec. I'm already using the part as a 13:1 divider for a LTZ1000 reference, leaving R7 unconnected.

Regards, Dieter
 
The following users thanked this post: antintedo

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #87 on: July 04, 2019, 11:15:21 pm »
Since Try can't seem to get on with it:

Appears Try is still suffering from a malady mostly of his own making.  The original order was shipped in December, 2016, and while I tried to talk him out of ordering ±10PPM resistors he insisted, note that he requested no additional special handling of the resistors, even so I still guaranteed them to meet transfer standard specs (ESI) which I later sent him a copy of the specs which using his own numbers was still within spec.  As is standard industry practice, the resistors are guaranteed to meet said initial specifications before leaving the factory.  The initial readings were indeed -2PPM and -6 PPM when shipped.

I received this reply from Try on 1-18-2017: “I measured the one you marked 10k -2ppm. It was either spot on or exactly 10k + 0ppm on my 34401A with four wire measurement.”  He did not mention the other resistor and I was rather surprised at how close his meter read it, that should mean his meter also read the other resistor at -6PPM.  I did not hear again from Try until nearly 14 months later when he wrote to say he thought that both resistors had drifted some but could not get a calibrated reading on them, using several nondescript resistors and an out-of-cal (5 years) 3458A…..that’s okay it should be in the ball park at least.  The figures he gave me was somewhere around 20-30 PPM higher than original.  I offered to re-calibrate them for the cost of postage so he would have accurate readings again, he refused and claimed they were out of spec…..well yes they were probably not within ±10PPM by this time but as they were standard processed resistors that was not unexpected plus there was no way of knowing what had happened to them in the ensuing 14 months.

The standard guarantee for a transfer standard is that the standard will maintain it’s given tolerance for 60 days after manufacture assuming it was handled properly.  There was another batch of e-mails at a much later time which mostly just rehashed what had been said earlier.

I believe the offer of re-calibration was sufficient and reasonable, since Try insisted only on replacement a good year and a half later, I would only offer replacements if the resistors were beyond specifications for a transfer standard of which I had sent him a copy of the standard specifications and guarantee.

He also complained about the pricing, a mere $28.38 for ±10 PPM resistors, I think he got more than he paid for.  All he really needed was calibration of the resistors to accurate known values and he’s back in business with calibrated references.  Now if he thought he was going to get resistors that would hold ±10 PPM indefinitely and for that price, I think he was just off his gourd a bit.

I will still do the re-calibration for the cost of postage even now but I cannot be held responsible for his turning down the offer repeatedly.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #88 on: July 05, 2019, 06:27:12 am »
Mr. Petty, considering your story, i understand the measurements try showed above are those same two resistors you mentioned and that they exhibited a difference of 3.7 ppm between each other after 18 months while you determined 4 ppm in the start? So this means the resistors - though worse than expected - were still useful for making a voltage divider, where a possible 40 ppm drift cancels.

I guess the same applies to the Nomca parts i tested. I got those from Mouser two weeks ago and we may probably assume they have been in stock for years. And they still exhibit tracking with 0,6 to 1 ppm (stdev of 8 resistors). Isn't that remarkable?

If there is anybody with a similar set of wire wound resistors from Mr. Petty, what is the stdev after some years? The result will depend somewhat on how the parts were treated/used. And i would really like to know the TCs of those two resistors mentioned above ..

Regards, Dieter
 

Offline MiDi

  • Frequent Contributor
  • **
  • Posts: 599
  • Country: ua
Re: Statistical arrays
« Reply #89 on: July 05, 2019, 11:25:52 am »
A bit off-topic, but seems to fit here:

Could a precision DAC used in "multiplying" configuration be a serious replacement for precision custom resistor dividers/networks in boost amplifier (e.g. AD5781/-91)?

Relying upon Datasheets, these DACs seem to outperform even the best resistor dividers/networks and they are ootb trimmable down to ppm...
« Last Edit: July 05, 2019, 11:48:27 am by MiDi »
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #90 on: July 05, 2019, 01:30:30 pm »
Yes, i was waiting for that. Just combine a AD587 reference with a PT1000 temperature sensor, a humidity sensor, a Cortex M0 microcontroller and some smart firmware to automatically fine tune the reference including support for teach-in. That should result in a near perfect voltage standard without heating and with slow aging. The firmware could even predict and compensate aging drift.

So lets stop putting wire wound resistors into dry boxes, TEC boxes etc. and do some real work instead.

Regards, Dieter

PS: I just have a 10 KOhm econistor in the dry box. Room humidity is and was 32 % for hours, while temperature increased about 1 °C. While from the temperature increase i would expect a resistance change of about 2.2 ppm the part exhibits a resistance change of -49 ppm after drying for some hours. Anybody wants to fill their voltage reference box with silica gel?
« Last Edit: July 05, 2019, 02:13:24 pm by dietert1 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #91 on: July 05, 2019, 04:22:55 pm »
.....
PS: I just have a 10 KOhm econistor in the dry box. Room humidity is and was 32 % for hours, while temperature increased about 1 °C. While from the temperature increase i would expect a resistance change of about 2.2 ppm the part exhibits a resistance change of -49 ppm after drying for some hours. Anybody wants to fill their voltage reference box with silica gel?

Silca gel is not just for lowering humidity. If left in normal air for some time it mainly works to buffer humidity changes - so that rel humidity would change less/slower, even in a not so hermitic sealed case.

With the NOMCA arrays in an epoxy type case - they may react to humility changes too. With wires wound types it really depends on the materials used - nor wire wounds are equal.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #92 on: July 06, 2019, 05:36:23 pm »
This is the data on the SR-1 transfer standards, not only do my resistors exceed these specs but the cost is far lower.  Notice their initial accuracy is ±20PPM and calibration is only accurate to ±10PPM.  You also appear to not understand the definition of initial accuracy: “Initial Accuracy: The specifications stated in the TEGAM instrument catalogs and data sheets are intended as acceptance specifications and are guaranteed for 60 days from the date of shipment.”  You seem to think that initial accuracy is held indefinitely, no transfer standard does that.


http://www.ietlabs.com/pdf/Datasheets/TegamSR1.PDF

Note the initial accuracy is ±20 PPM and long term is ±50 PPM and calibration is only to 10 PPM, according to Try even after 2.5 years they are still within specifications.  He has nothing to complain about and I even offered a free calibration which he refused twice so I think it is time Try moved on.  He never presented any evidence that the resistors were out of spec.  He also bought a lot of other resistors from me and has never complained about those.  My standard off-the-shelf resistor exceeds the SR-1 specifications and for a lot less money and without post operation processing.

Just for your information, some time back Vishay was chasing after me to become a field applications engineer for them, Felix Zandman interviewed me himself, I turned them down in the end for personal reasons.  The job entailed flying all over the country and at that time I could not do that.  Vishay knew that I had the knowledge they needed, they approached me.  Your erroneous claim that I don’t know about Vishay technology is full of hot air.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #93 on: July 06, 2019, 05:43:33 pm »
Come on, this thread is about statistical arrays and not about Edwin G. Pettis, if you don't have anything to contribute please be silent., Thanks.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #94 on: July 06, 2019, 06:08:22 pm »
I'd be happy to Branadic, I didn't start this nonsense, dietert1 and Try interjected all this nonsense so please feel free to jump them first, it is only fair that I get to re-but them.

Measuring resistors on a DVM  is not a valid technique for determining how a given resistor will perfect in a given circuit under actual operating conditions.  Yes you get a bunch of data on how it is performing under test conditions which may have some indirect bearing on the circuit performance but if you want to know how that divider is going to work under actual conditions, make an actual circuit and try it, see what the end result is, then if it isn't performing as expected you can trace the source.  A resistor under actual operating bias and environmental conditions is not the same as hooking it to the front end of a DVM.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #95 on: July 06, 2019, 06:23:21 pm »
Dear Mr. Pettis, from your table i understand that buying two resistors from you to make a voltage divider will result in a division ratio with up to 10 ppm/K on the division ratio. This worst case happens if one resistor has +5 ppm/K, the other one -5 ppm/K and it means up to 180 ppm change on the division ratio in an 18°C temperature cycle.
I think i demonstrated above that a Nomca thin film resistor array, when wired correctly, will give you a small fraction of a ppm/K on the division ratio and an overall variation of less than 0.5 ppm with a 18°C cycle. The thin film array is a much superior solution, about 360 times better. Hard to understand what you want to contribute.

In the meantime i finished drying the 10 KOhm econistor (my example of a qualified wire wound resistor). The final result of the humidity test going from 32% RH down to about dry was -60 ppm after about 24 hours. Only then the econistor exhibited a hysteresis free 18°C cycle with a TC of 2.67 ppm/K. The negative sign of the 60 ppm resistance change is interesting. It means this huge humidity effect is not surface conductivity, but probably carrier size. Carrier gets bigger by water absorption, wires get stressed to be slightly longer and slightly thinner. Carrier shrinks when dry and resistance shrinks as well.
Started a similar test on Nomca thin film array, but drying is very slow. Effect on resistance is -12 ppm after 8 hours, yet only 0.21 ppm on division ratio (with 13:1 divider example).

Regards, Dieter
« Last Edit: July 06, 2019, 06:29:32 pm by dietert1 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #96 on: July 06, 2019, 07:01:46 pm »
With the comments on wire wound resistors a possible important point got slightly hidden. The NOMCA arrays seem to have noticeable noise:
....
Regarding recent discussion, I was curious about 1/f noise of thin films, especially NOMCA vs larger models. I did a quick test dividing LTZ1000 voltage by 2 and measuring 0.1-10Hz pp noise within 1 minute period.

LTZ1000: 0.15 ppm
noname THT 3.3k metal film: 0.16 ppm
NOMCA 10k, 3 samples tested, 6 resistors used every time: 0.22-0.25 ppm
S102K (refurbished) 11k, 6 resistors used: 0.16 ppm

Additionally, long time ago I found this publication containing 1/f noise tests of various resistor types and models. Seems like a lot of models achieve good performance with almost no 1/f noise, even in SMD 0402 size.

This essentially means the arrays are not good enough for a 7 to 10 V step for a LTZ based reference. The resistor noise in this case seems to be about as much as the LTZ, even with using 3 resistors in parallel each.
They may still be OK with a LM399 based reference, or to set the temperature in an LTZ1000 circuit, at least from the noise side. A still open point is aging. Specs here are tricky, as there is no good and easy way to measure upfront and small differences in the protective layer / epoxy could make quite some difference.

Though not specified much different, there are also NOMCT arrays, using NiCr resistors. These may be better with TC matching / noise, but from what I have read so far could show a little more aging.

With wire wound resistors, there are also set available with TC matching better than the absolute TC. Even without this, chances are good that at least the 2 nd order part will compensate. Especially similar value resistors can be made from the same wire spool to give good matching.

With the resistor arrays the TC matching can on average improve with the statistical use of more resistors, but the worst case does not improve that much. Best case there are 4 resistors on each side. This would be on average an improvement by about a factor of 2 (maybe a little more as there is finite number of resistors and thus more gain than the normal square root N). Without individual matching / selecting the expected divider TC is thus more like 2 ppm/K, maybe often better if the initial specs are rather conservative or for a larger temperature range.
 
The following users thanked this post: TiN, Kosmic

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #97 on: July 06, 2019, 07:04:03 pm »
No you are totally incorrect, you are confusing tolerance with TCR, the tolerance of those two resistors were within ±10 PPM as requested, the TCR would have been <1.0PPM°C, probably close to 0.85 PPM/°C, but that is not the tracking TCR, since both were made of the same wire the tracking TCR (ignoring external factors such as thermal differentials) would be likely less than 0.5PPM/°C (conservatively under real operating conditions)).  The tracking is affected by external environmental conditions which I have no control over and yes film resistors would have similar effects even on the same substrate but possibly smaller because they are much smaller physically but they are also affected by factors which have little to no effect on wire wounds.  Tracking is a very difficult specification to work out since it involves much more than just the resistors, each situation is different.

Yes you very well may be getting pretty close tracking on your NOMCA chip under those conditions, I am not particularly disputing your results, the question is how stable is it over time, your one off only shows that this one chip has these current specifications at this time.  The TCR and tracking TCR of my resistors do not shift significantly over time unless the resistors are subjected to extraordinary conditions.  Your calculations concerning my parts are quite incorrect.  I have demonstrated 0.1PPM to 0.2PPM ratio tracking over long term time by one of my customers, that is not to say everyone is going to achieve that low tracking value, it takes effort on the part of the customer as well.

Econistors are not equivalent to mine, their specs are not equal and as stated, all precision wire wound resistors are NOT the same.  If you are going to compare something against my resistors, then use my resistors but stop making erroneous comparisons thank you.

The table I posted is NOT mine, it is the SR-1 transfer resistor data table from ESI/Tegam/IET, my specifications are better than those which I said in the other posting, please read it more carefully.

It is this kind of misinterpretation which causes problems, it would be much better if you asked a question directly instead of posting first without comprehension.  Yes resistors, like most other components are much more complex that most people understand and when you're chasing PPMs, it gets even more difficult.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #98 on: July 06, 2019, 07:18:23 pm »
In general I agree with you, Kleinstein, those statements are accurate.
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #99 on: July 06, 2019, 08:00:56 pm »
In the meantime i finished drying the 10 KOhm econistor (my example of a qualified wire wound resistor). The final result of the humidity test going from 32% RH down to about dry was -60 ppm after about 24 hours.

Hello,

usually I measure time constants for drying of epoxy in the order of 3-7 days at room temperature.
So it would be interesting what happens after the 24 hours.

Lars also stated 0.5-2ppm change for 8E16 resistors for each % relative humidity.
https://www.eevblog.com/forum/metrology/t-c-measurements-on-precision-resistors/msg1379236/#msg1379236

when I compare the humidity in Colorado with that of my region here I have minimum 30% more humidity.

with best regards

Andreas

 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #100 on: July 06, 2019, 09:33:58 pm »
No, Mr. Pettis, it's not April 1st today. If you present a table with limits you will stick to, that's it. We know from TiNs measurements that your resistors are no better than econistors. My econistor humidity test was to confirm what others found before. Yes, the econistor remains in it's dry pack and i will measure it again after a week or so. I think Mr. Pettis is creating a lot of noise without any signal.

I was not measuring one Nomca part but six, a total of 48 resistors. Sooner or later i will also characterize the other 34 econistors we have, looking at their TCs. And i wrote about bridge measurements were i had one 13:1 Nomca divider outside and one inside the TEC box, both running from the same 10 V reference. That was the measurement which resulted in the +/- 0,5 ppm/K limit for a 18°C cycle. So it wasn't a DVM resistance measurement only, but very close to the indended application.

Concerning resistor noise: Is there a link/reference on how those measurements were done? I mean before someone talks us into using parts with 60 ppm humidity dependence in order to avoid some 0,07 ppm extra noise, i'd like to have a closer look. Maybe a Nomca array just needs to be protected from random ventilation. I mean its thermal mass is very small in comparison to its surface.

Regards, Dieter
« Last Edit: July 08, 2019, 02:12:37 pm by dietert1 »
 
The following users thanked this post: 3roomlab

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #101 on: July 08, 2019, 05:32:52 pm »
Dieter,

I'm about to measure some values soon, still preparing everything. Attached one NOMCA16035001 together with a NCP15WF104F03RC NTC on a ceramic board, configured as a 10k, with the board glued to a BPR10101J resistor as part of an oven.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: TiN, MegaVolt, antintedo

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #102 on: July 08, 2019, 05:51:52 pm »
....
Regarding recent discussion, I was curious about 1/f noise of thin films, especially NOMCA vs larger models. I did a quick test dividing LTZ1000 voltage by 2 and measuring 0.1-10Hz pp noise within 1 minute period.

LTZ1000: 0.15 ppm
noname THT 3.3k metal film: 0.16 ppm
NOMCA 10k, 3 samples tested, 6 resistors used every time: 0.22-0.25 ppm
S102K (refurbished) 11k, 6 resistors used: 0.16 ppm

Additionally, long time ago I found this publication containing 1/f noise tests of various resistor types and models. Seems like a lot of models achieve good performance with almost no 1/f noise, even in SMD 0402 size.

The link in this older post is quite good. The measurement for resistor 1/f noise is relatively easy with a bridge from 4 equal resistors, as described in the link. It mainly take a low noise amplifier, some instrument to record the data at a moderately low frequency. The MIL standard uses a slightly different setup, with only one resistor as a DUT and some fill resistor, but its a little more tricky, especially at very low noise.
So a test may be reasonably easy.

From the measurements quoted the NOMCA resistors as a 1:1 divider (3 x 10 K in parallel each) results in noise similar to the LTZ1000 noise in the 0.1 -10 Hz range  (noise power about doubles).

The < -30 dBi noise specs for the NOMCA are not really good. From the divider measurements they don't seem to be much better than specs.
For the temperature setting divider at the LTZ1000 this would still be well good enough, but not for a 7 to 10 V stage.
 

Offline Kosmic

  • Super Contributor
  • ***
  • Posts: 2501
  • Country: ca
Re: Statistical arrays
« Reply #103 on: July 08, 2019, 06:36:12 pm »
Dieter,

I'm about to measure some values soon, still preparing everything. Attached one NOMCA16035001 together with a NCP15WF104F03RC NTC on a ceramic board, configured as a 10k, with the board glued to a BPR10101J resistor as part of an oven.

-branadic-

Nice assembly  :-+

Is it expensive to order some custom ceramic boards like that ?
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #104 on: July 09, 2019, 05:14:17 am »
Yes, this assembly will certainly perform very well. The large resistor will provide distributed heating with small gradients. But this type of work really requires some sturdyness, since it will take a long time, maybe a year or more to learn how useful such a device can be and whether it helps.
I would also like to mention again SEI RTAN resistors that could be used to do something similar, i mean make an array in a thermostat. Those parts appear to be metrology level, too.

The measurements i started on drying Nomca arrays turn out to be much more difficult than the econistors. I have not seen any humidity effect now after several days, which is good news. The temperature effects are much bigger, even with a part ("C") with only 5 ppm/K on average. I also put a 5 KOhm UPW50 resistor into the drypack and did not see a single ppm change after 24 hours. Apparently our econistors are sealed only on one end and open on the other. And their material is somehow softer, while the UPW50 package appears more like usual IC packages - extremely dense plastic.

Regards, Dieter
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #105 on: July 09, 2019, 04:16:30 pm »
Quote
Is it expensive to order some custom ceramic boards like that ?

Can't tell you that as this is a "design" I gave to my colleague to develope the process on a real world application and without the use of individual masks (digital process chain).

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #106 on: July 21, 2019, 09:14:35 pm »
After about 2 weeks in the drypack a Nomca array exhibits a humidity effect of about 25 ppm on the resistance. Yet the division ratio of a 13:1 divider changed by only 0.7 ppm under these conditions.
After 2 weeks in the drypack, some wirewound resistors have gone mad and exhibit changes of up to 220 ppm (econistor) and about 80 ppm (UPW50). Changes of econistors differ by more than 30 ppm, so useless for metrology grade voltage dividers. Hard to believe, need to check more.

I also found that the TC compensation scheme proposed above using a certain combination of resistors is delicate. I mean the TCs i determined using a 18 °C oven cycle (using box method) are useful but may not give the best possible result at room temperature. Currently i am trying to refine temperature measurements using a PT1000 sensor close to the Nomca array in order to separate humidity effects from temperature effects.
I also suspect that self heating may affect the compensation scheme.

Regards, Dieter
 
The following users thanked this post: doktor pyta, The Soulman

Offline d-smes

  • Regular Contributor
  • *
  • Posts: 101
  • Country: us
Re: Statistical arrays
« Reply #107 on: July 22, 2019, 09:12:25 am »
After about 2 weeks in the drypack ...
By drypack, I assume you mean a desiccant material.  Are you just throwing a few pouches in a metal box?  Or completely encapsulating the resistors in a sealed jar or equivalent?  Any relative humidity readings?  I know with insulation systems, there is such a thing as too dry.  Wondering if that applies here.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #108 on: July 22, 2019, 10:32:29 am »
When we got our 3D printer, we bought 4 kg of orange silica gel to make a drybox for filaments. That drybox got about 1 kg in it and the hygrometer has been reading 10 or 11 % rH ever since. That's the minimal reading of the hygrometer (some technical or physical limit). I just put some spoons of that silica gel into plastic envelopes together with the resistors. The envelopes are closed except where the cables enter. I think humidity inside has been below 11 % rH for at least 14 days.
The results are somewhat confusing. During the first 24 hours 10 KOhm, when i was watching them regularly, econistors exhibited a resistance change of about -70 ppm, but now they are about 150 ppm above the initial value. I think i will take them out of the drypack and test whether they go back to initial values or whether they have been damaged. The UPW50 is 80 ppm above initial value after 14 days.

Regards, Dieter

PS: Now i got a Sensirion SHT31 Smart Gadget Development Kit that indicates less than 5 % RH inside of the drypack (PE bag).
« Last Edit: August 07, 2019, 10:19:59 am by dietert1 »
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Statistical arrays
« Reply #109 on: July 26, 2019, 07:40:49 am »
In the nomca pdf, it states the load life drift at 125C (looks like derated = 100mW?) is 290ppm. what should be the expected/calculated load life drift be at 1mW? 2.9ppm (125C)?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #110 on: July 26, 2019, 08:05:56 am »
For the load life it is likely the temperature that is most important So it would likely not make a difference if with current or with external heating.
 

Offline splin

  • Frequent Contributor
  • **
  • Posts: 999
  • Country: gb
Re: Statistical arrays
« Reply #111 on: July 26, 2019, 01:33:20 pm »
In the nomca pdf, it states the load life drift at 125C (looks like derated = 100mW?) is 290ppm. what should be the expected/calculated load life drift be at 1mW? 2.9ppm (125C)?

Take a look at this Vishay technical note 'Drift Calculation for Thin Film Resistors':

https://www.vishay.com/docs/28809/driftcalculation.pdf

Note that the numbers in 2.4, ' Precision Operation Mode' don't seem to be right. Given from 2.1, drift @ 1000 hours, 125C = .25%.

Therfore at 85C it should be 0.25% / (2^((125-85)/30) = 0.1%, not < .05% as shown.
 
The following users thanked this post: 3roomlab, msliva

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #112 on: July 26, 2019, 03:30:12 pm »
When i make a 10V - 7V divider from a 8x 5KOhm array using six resistors (2 resistors as 2.5 KOhm and 4 resistors as 5 KOhm) it will take 14 mW, which gives about 3 °C temperature increase. So i can run the Nomca at 40 °C. If the equations from the vishay paper apply in this low power regime, then i get an estimate of 290 ppm / 12,70 = 23 ppm, where 12.7 = 2 ** (150 °C - 40 °C) / 30 °C.
A change of -25 ppm has been observed in the drypack, so a 23 ppm number appears plausible to me. During the same drypack tests the division ratio of a divider optimized for low TC exhibited a change of 0,3 ppm of the division ratio.
To me it appears that vishay engineers were more or less focussed on durability in high humidity/high temperature regime, where it is easier to see real effects. In the meantime i can determine TC differences inside a Nomca array with a sigma of 0,08 ppm/K per resistor just from ambient temperature changes (delta T = 27 .. 29 °C),  without using the oven and i am getting results consistent with the oven measurements i did some weeks ago. Still no significant humidity effect on the division ratio after 3 weeks in the drypack.

As others noted before: The Nomca 1603 array in the Fluke/Wavetek 7000 reference means others have done these more detailed tests before and the Nomca part was found to be useful. Of course, as of to today there may be other, even better choices of thin film resistors. For example the Vishay PRA datasheet looks interesting. On its first page they claim "Very low noise < -35 dB and voltage coefficient < 0.01 ppm/V".

Regards, Dieter
 

Offline e61_phil

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Re: Statistical arrays
« Reply #113 on: July 26, 2019, 03:56:47 pm »
@Dieter: What do you mean by sigma? The standard deviation of your measurement? Or is it the error of the fit?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #114 on: July 26, 2019, 04:01:47 pm »
The formula in the Vishay note on drift is only an approximation. It may or may not work very well for different types of resistors. Also the extension to much lower temperatures is speculative. Its more like tested from maybe 100 C - 250 C and increasingly uncertain at lower temperatures.

With the epoxy encapsulated parts, there is also another effect. The usual epoxies have a glass transition temperature somewhere in the 100 C - 160 C range. Things change when comparing the range well below (e.g. some 50 K) and above the glass transition temperature. At higher temperature there is some kind of equilibrium that can be reached, well below it is an increasingly slower drift towards a more dense state.
The rate of cool down from high temperature can have some effect - it kind of sets the disorder that is frozen in. Slower cooling from some 150 C (e.g. soldering) would work like a bun in. It's a window some 30-50 K below glass transition that is important. Burn in at higher temperature would cause new disorder, kind of resetting the clock.
Chances are the formula is for the higher temperature part, with the epoxy more or less in equilibrium.

The humidity is expected to mainly effect the epoxy, so there may not be much correlation with the high temperature drift.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #115 on: July 26, 2019, 07:15:42 pm »
By sigma i meant standard deviation. The number given (0.08 ppm/K) is the standard deviation when comparing eight results of the oven measurements with the line fits of room temperature measurements. When i partition my recent measurements, i am getting similar standard deviations. There is no indication that the oven/box method results are different/worse than the ambient temperature results.
My uncertainties mean that the observed 0.3 ppm humidity effect of the divider isn't a real measurement but more like an upper limit.

Regards, Dieter
 
The following users thanked this post: e61_phil

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #116 on: August 10, 2019, 12:14:27 pm »
Here is a nice online app called "Resistor Network Finder":
http://kirr.homeunix.org/electronics/resistor-network-finder/

As the number of resistors in the network is limited to 6 and I want to use a bag of 10k resistors (+/-25ppm/C) only, I added several serial and parallel combinations of 10k to my list of "stocked resistors".

Interestingly, it looks it can find almost _any_ feasible value with such an approach..  :-+

My current understanding is the resistor's TCs within a lot (?) do follow somehow a "distribution" typical for the lot. Or do not?

« Last Edit: August 10, 2019, 01:03:13 pm by imo »
 
The following users thanked this post: doktor pyta, razvan784, ch_scr, Magnificent Bastard

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #117 on: August 30, 2019, 09:05:06 pm »
Got some more measurements on Nomca arrays. First i repeated measurements on the arrays A and F that have been in drypacks for about two months now, with very stable results. The 6 % reduction in absolute TC results from an improved temperature measurement.
From nine more arrays tested now the parts J, L and O exhibit low TCs and low TC spreads.

Regards, Dieter
 
The following users thanked this post: 3roomlab, doktor pyta

Offline doktor pytaTopic starter

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: pl
Re: Statistical arrays
« Reply #118 on: April 19, 2020, 09:31:25 pm »
New candidate found. Should be better than NOMCA:
https://www.vishay.com/ppg?53033
especially PRA100I8 can be found on ebay.
 
The following users thanked this post: Andreas, 2N3055

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #119 on: July 15, 2020, 03:05:04 pm »
From the measurements quoted the NOMCA resistors as a 1:1 divider (3 x 10 K in parallel each) results in noise similar to the LTZ1000 noise in the 0.1 -10 Hz range  (noise power about doubles).

The < -30 dBi noise specs for the NOMCA are not really good. From the divider measurements they don't seem to be much better than specs.
For the temperature setting divider at the LTZ1000 this would still be well good enough, but not for a 7 to 10 V stage.

I've got some fresh measurements that confirm it. NOMCA and AORN, both Ta2N type, are within specs (-30 dB) but not significantly better. They are very noisy compared to most NiCr ones. Bulk metal foil (SMNZ) are better than -70 dB, so practically noiseless as expected.

These are all 10K arrays in Wheatstone bridge biased with +-10V. The bridge bias is reversed at 100 Hz and sampling (3458A, DCV mode, 9 ms aperture) is synchronized to it, then CDS is implemented in software.

The full results of these tests will be available some time next year.
« Last Edit: July 15, 2020, 03:07:13 pm by Castorp »
 
The following users thanked this post: branadic, Echo88, Andreas, e61_phil, splin, doktor pyta, Kleinstein, chuckb, ch_scr, MegaVolt, antintedo, macaba

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #120 on: July 15, 2020, 04:36:11 pm »
Can you eleborate on your test setup a little bit more? What is CDS standing for, correlated double sampling?

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #121 on: July 16, 2020, 06:51:12 am »
Can you eleborate on your test setup a little bit more? What is CDS standing for, correlated double sampling?

-branadic-

Yes, that's what it is. They have a good description in Linear AN96
https://www.analog.com/media/en/technical-documentation/application-notes/an96fa.pdf

In my setup I have two instrumentation amplifiers - INA163 for low impedances (<1 kOhm) and LT1167, selectable through a DPDT relay. Both amps are set for fixed gain of 100. After that there are two extra cascaded stages with G=10 for a total gain of up to 10000. So far I've taken all measurements with this highest gain setting, as the R arrays yield very decent CMRR (the worst one is about 86 dB). The bias voltages come from two Stanford SIM928 sources (isolated, battery-based). The bridge bias is switched through two ADG1404s. There's a bit of extra circuitry that clamps the instrumentation amp inputs during the bias reversal, to prevent large glitches and overdriving. It's all controlled by a small FPGA that also produces the sample trigger signal for the DVM. The R arrays are all soldered on identical small carrier boards that plug into the main board. Everything is covered, shielded, filtered, well decoupled, etc. That's it basically.
 
The following users thanked this post: chuckb, notfaded1

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #122 on: July 16, 2020, 07:17:36 am »
Using AC excitation is a good idea to get lowest noise, as it avoids the 1/f noise of the amplifiers. And gets around chopper stabilized amplifiers.

Is the excitation directly coupled or like in the normal bridges for resistor noise readings galvanic isolated ?

The higher impedance instrumentation amplifiers however look relatively noisy. So there the classical isolated DC excitation with an chopper stabilized amplifier may be still competitive.

P.s. The AD8421 could be a lower noise alternative in the intermediate resistance range.
« Last Edit: July 16, 2020, 07:20:07 am by Kleinstein »
 

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #123 on: July 16, 2020, 01:45:53 pm »
It's all DC coupled.

You're right, for intermediate source resistances there's a bit of amp noise added. Also for very low (<=100 Ohm), but I wouldn't bother getting something quieter than INA163. In the end it's not a big problem - that's just white noise. The interesting part of the spectra is 1/f, which is entirely due to the R arrays (when biased). Maybe in the final presentation of the data I'll RMS-subtract the known white noise contributions... or just explain their origins.
 
The following users thanked this post: chuckb, notfaded1

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #124 on: March 27, 2021, 08:22:38 pm »
Challenge: How to create a 10k resistor, using 5x 10k resistors?

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #125 on: March 27, 2021, 08:57:40 pm »
Challenge: How to create a 10k resistor, using 5x 10k resistors?

Like this?

I'll use the opportunity to give a little update on the excess noise tests. It got a bit uglier than I expected - some parts show considerable batch-to-batch variation, even for batches with recent date codes. But on the positive side, there are new low-noise champions - NOMC (not NOMCA), TDP, HTRN - they stand next to ORN. I'm waiting for a few more samples to test before I wrap this up.
 
The following users thanked this post: TiN, Kleinstein, ch_scr

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #126 on: March 27, 2021, 09:05:49 pm »
Meh, not fair that you already gave a solution, as you are into it :)

-branadic-
« Last Edit: March 27, 2021, 11:21:06 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: doktor pyta, ch_scr

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3207
  • Country: de
Re: Statistical arrays
« Reply #127 on: March 27, 2021, 09:53:47 pm »
Meh, not fair that you already gave the solution, as you are into it :)

Throwing one of the 5 away would be an equivalent solution.
Through R3 there is no current flow if all other resistors are equal.

With best regards

Andreas
 
The following users thanked this post: e61_phil, MiDi

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #128 on: September 07, 2021, 08:22:52 am »
Finally managed to tick this one off the list. I hope it would be useful.

https://arxiv.org/pdf/2109.02448

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #129 on: September 07, 2021, 09:35:13 am »
The test uses AC voltage (square wave) to test the resistance. For most mechanisms this should be OK, but there may be mechanisms that could react different with DC. That is the resistance fluctuations may be directional. It is still an ingenious method to measure the resistor noise to a very low level. AC modulation of the excitation allows using conventional, no AZ type amplifiers at the critical point.

The more standard (old) way uses DC excitation for the bridge and a chopper or similar amplifier with very low 1/f noise. The noise level for the chopped  amplfiers is usally higher and thus less sensitivty. With modern AZ OPs this still works reasonably well to at least detect the more noise ones.

I see no surprise to sometimes get a higher noise index for the higher value resistors. The very high values could use a thinner film and thus less resistor material. In addition the surface is possibly more prone to noise than the bulk.

There could be a difference between using a thinner film as simple block or a thicker film and than a more fine line pattern. This may also change with value. The substate roughtness may have an indirect effect: a rougher surface makes it more difficult to etch finer patterns.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #130 on: September 07, 2021, 09:24:36 pm »
Wonderful paper, thanks Castrop.  :-+
An explaination why some thinfilm resistors on ceramics show less excess noise compared to others could be, that some manufacturer have their ceramics polished before processing them, though an expensive process. Glass has low surface roughness, but bad thermal conductivity, hence why some manufacturers use silicon instead, which is a good compromise between both and a comparable cheap substrate, given that there is a mass production industry for it.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #131 on: September 08, 2021, 08:32:57 am »
Yes, high purity Alumina can be polished well, or it can be glazed. Those are the substrates used by Susumu:
https://global.kyocera.com/prdct/fc/product/pdf/electronic.pdf

It's not so surprising that 96% Alumina yields much noisier resistors. Such substrates are typically used for thick films.

In any case, there must be several factors contributing. It's impossible to separate their effects unless you can fabricate and test many samples of each technology. I strongly suspect that in many cases noise in the contacts dominates, but I really can't prove it conclusively from these measurements.

I have data for some of the parts under pure DC bias. NOMCA and TOMC show the same noise levels as in the square wave test. NOMC, TDP, VHD, SMNZ - they don't show any hidden DC-only noise either, at least to the measurement sensitivity limits. At some point I tried to degrade some low-resistance parts and make them noisy by applying high bias for several days. It didn't work - I guess electromigration needs longer time or higher temperature (or both) to starting showing in the excess noise.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #132 on: September 08, 2021, 09:42:26 am »
From the paper i understand that the excess noise index definition includes another factor 1 000 000, since noise is measured in uVrms. Simply speaking a noise index of -40 dB means something like -120 dB + -40 dB = - 160 dB in relation to the DC excitation. Is this correct? I mean ignoring some factors 2 or pi that will enter when looking at bandwidths etc.

When looking up those tiny MPM parts favored by low excess noise results, i was wondering whether a low power => high resistance solution will always perform better. Recently i have been looking at the schematic of the Advantest R6581A and when i see they used dividers made of 17K and 19K to process reference voltages, without any filter caps, pickup from digital parts of the circuit can be much more than -160 dB. Maybe sometimes low noise design may favor lower resistance parts that are a little bigger to stay cool while handling the extra mWs of power.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #133 on: September 08, 2021, 12:04:55 pm »
Yes the noise index definition include those extra -120 dB to start with, but it is also comparing noise from 1 decade to DC. So there is some slightly arbitrary factor already, it could have been 3 decades or a factor or 2 or e as well.
The excess noise is one part of low frequency fluctuations, but there are also possible thermal effects. For those thermal effects low power consumption helps. Otherwise a higher resistance does not reduce the 1/f noise, it more adds the classic Johnson noise and makes the current noise of OPs more visible.  Higher resistor values make the contacts a little less important.  Averaging with a parallel / series circuit is also effective with the excess noise. So using 2 resistors in series or parallel reduces the excess noise by 3 dB.
 
I think that not using filter caps for the reference amplification is a bad idea in the R6581T circuit. Caps won't help with the very low frequency part like < 1Hz , but they work quite well with the faster part of some 10-100 kHz or so, that can be modulated back to the near DC range with the reference modulation in the ADC. So the ADC may react to higher frequency noise of the reference with a kind of aliasing of the reference noise. Chances are they just did not see this less obvious noise path. It is not a large noise contribution (but the LTz1000 noise is still higher than usual OP noise), but easy to avoid.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #134 on: September 08, 2021, 02:30:00 pm »
So let's say: Resistor excess noise is a consideration for somebody who has reached a precision level of 0.1 ppm and wants to improve. This may apply to several readers here ("voltnuts"). With "bad" resistors like Nomca they will be limited at about 0.03 ppm and with "good" resistors like MPM they will be limited at 0.0003 ppm level - if they ever get there.

I am just trying to put the resistor excess noise discussion and N. Beevs research into perspective. Noise measurements at a -180 or -190 dB level are extremely difficult (near impossible). Don't know if anybody else can do it to confirm the results.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #135 on: September 08, 2021, 03:23:29 pm »
With a relatively normal AZ OP (e.g. LTC2057 for 10 K resistor, AD8628 for 100 K)  on can do the more classical measurement of excess noise with a DC bridge. With some patience (e.g. 10 min taking data) one can see the excess noise down to about a -50 dB noise index  (maybe a little less with more patience and a suitable resistor value).  This is good enough for most applications, except the very demanding ones.
I have done the test with some thin film resistors (Susumu RR 0805 , 4 K - somewhat comparable to the RM networks) to just see a little excess noise, but not really enough to get a number.
The description in the ariticle is very deailed, so it should be able to repat the experiment, at least to the -60dB range. Good amplifiers are better than the Johnson noise - so it is more about patience, thermal stability than a super low noise amplifier.

I see the difference in noise between NOMCA and ORN type resistors with my ADC circuit: with NOMCA resistors the noise is at about 800 nV, with ORN resistors I get something like 500 nV for 1 PLC.   The resistors were the main source of 1/f noise, more than OPA140 or OP7 op-amps.
A rather demanding case for the resistors is something like inverting a 10 V signal with a classic inverter with 2 resistors to set the gain. So 10 V at the resistors. 0.1 to 10 Hz are 2 decades and 2 resistors contributing to the noise with a gain of 1/2. So  with 0 dB noise index one would get 1 ppm of 10 V and thus some 10 µV_RMS for the noise. With a -40 dB noise index this would be 100 nV RMS or some 600 nV_pp. This would still be more than the noise of a standard precision OP like the OP07.

Not many voltage reference are better than 600 nV_pp at 10 V - so often the excess noise is not that important, unless there is bridge of some type (e.g. in a DAC or ADC) were the reference  part is compensated near zero.

The NOMCA resistors are not even that bad - I had similar noise with other thin film resistors.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #136 on: September 08, 2021, 05:50:32 pm »
Refering to my LTFLU experiements I reported large reference noise using NOMCA networks. Own resistor current noise measurements revealed that TOMC has smaller excess noise compared to TOMC, hence why I redesigned the LTFLU board for TOMC networks and its physically larger package. The board with the same LTFLU had much smaller noise then.
With the results now given by Nikolai I could have simply replaced the NOMCA network with a NOMC, which seems to have even lower excess noise compared to TOMC. Maybe I give it a try at some point.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #137 on: September 08, 2021, 06:36:23 pm »
That's not a confirmation of N. Beevs publication, sorry.

And if an ADC with about 20 V input range improves noise from 800 nV to 500 nV, that's about going from 0.04 ppm to below. I think Kleinstein did see the -30 dB noise index of the Nomcas vanish when choosing better parts, but he did not measure the -70 dB excess noise index of the MPMs. That's another factor 100 less.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #138 on: September 08, 2021, 07:16:56 pm »
Measuring to -70 dB noise index is hard and the way with AC excitation would be the way to go.  The description is quite detailed and if wanted one could repeat it. With well matched resistors to test, the quality of the reference souces should not be that critical   (if needed filter reference noise at around the modulation frequency - so make sure the sources don't have much 400 Hz noise). So I don't think the LTZ1000 sources are really needed.

For resistors not too large or too small (e.g. 5 - 50 K range) the -70 dB range seems doable with patience (it may need a day or two to measure - but hey it probably takes a week of more to build up the setup). Thermal stability and the Johnson noise would be kind of limiting. The excess noise gets larger than the Johnson noise only at really low frequencies. So no matter how good the amplifier, it would need time.

A special build circuit (excitation + AC amplifier and ADC (no need for magic here) and µC for control and demodulation) could be relatively simple. Progamming is still quite a bit of effort for a special build "instrument". The amplifier may be usable for more classic tasks too.
The setup is a bit like a good DMS readout, just also for higher resistance.
 

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #139 on: September 10, 2021, 07:15:39 am »
I think Dieter raises some pretty valid points. But there are two aspects here - one is the measurement of very low levels of 1/f noise, the other one is whether this noise would be relevant in any actual application.

The measurement is in fact not that difficult. It should not be seen as a >180 dB dynamic range measurement because it just isn't - it's a measurement of a tiny signal that is amplified. More crucially - the signal can be separated from most of the background noise. That's nothing new - those who do research on thin films have been measuring very low excess noise using lock-in amplifiers and other phase-sensitive methods for many decades. Most of their literature is not relevant for our purposes, because the thin films are exotic types that are far from being part of practical resistors. Also, in these papers they usually don't use the Noise Index but they quantify the results in other ways.

About the practical value of the results - roughly speaking, for certain applications in metrology it could make a big difference whether you have -30 dB or -40 dB. Some colleagues already gave good examples. For some of our applications we need -50 dB to make resistor noise negligible. I doubt it would matter for anyone whether they have -60 dB or lower, because in any practical system there would be other limiting factors (noise of AZ amps, thermal drift, etc.)

In any case, quantitative knowledge is always an advantage. Even if it's just semi-quantitative (like the upper limits). Even if it's just for reassurance.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #140 on: September 10, 2021, 08:37:12 am »
If there is an application that would be so sensitive to the extra noise  to really need super low noier index, it would also allow to measure the noise. If you can not measure it, it also means it has not effect on the application.

It is still from the physical side interesting to understand the reason behing the 1/f noise. It may correlate with other properties, especially aging. Unstable details that can flicker between 2 states with a slightly higher or lower resistance can also on the long run shift more to 1 side and this way cause drift.
 

Offline ramon

  • Regular Contributor
  • *
  • Posts: 143
  • Country: tw
Re: Statistical arrays
« Reply #141 on: September 10, 2021, 10:28:38 am »
The LT5400 are not of NiCr but SiCr material (Tyler Hutchison, design note 502).

The SiCr material has two times more resistivity (sheet resistance) than NiCr. This could explain why they can have 1M ohm resistors while other networks just go up to 100K only. And why they have higher than average noise of LT5400 compared to other resistors (NiCr over Silicon substrate).

There are also some nice documents google from Vishay and from 'Vishay Elecro-Films' (EFI) that talks about different substrates properties and materials (Al2O3, BeO, AlN, Tamelox ...). With tables and numbers about surface roughness, among many other data.
 
The following users thanked this post: Castorp

Offline Castorp

  • Regular Contributor
  • *
  • Posts: 63
  • Country: fr
Re: Statistical arrays
« Reply #142 on: September 10, 2021, 10:52:56 am »
Thanks for the info, Ramon! I was aware of this DN but somehow I missed the film type. I'm not sure who told me it's NiCr. I did ask ADI about the substrate and they confirmed it's Si. I'll fix the film type in the next revision of the paper.

I guess you're referring to this table (or something similar):
https://www.vishay.com/company/brands/electro-films/Substrates.html

Too bad there are no readily available resistor networks on any other substrates, just Alumina and silicon. I've heard about custom ones on sapphire or AlN but I've never seen them. Also high purity Alumina substrates vary quite a bit between the suppliers.
« Last Edit: September 10, 2021, 11:23:19 am by Castorp »
 
The following users thanked this post: ramon

Offline ramon

  • Regular Contributor
  • *
  • Posts: 143
  • Country: tw
Re: Statistical arrays
« Reply #143 on: September 12, 2021, 10:18:42 am »
Yes, that is the table with the surface information. I think that I saw it in pdf format in a technical document about thin film (with many other tables).

Vishay has NiCr on AlN substrate resistors, but no resistor networks yet.

I guess that the other types of substrates are too expensive or require too much special processing. Not suitable for volume/production. So they just offer them as custom (for microwave high-speed, aerospace,  ...). I haven't seen them too, not even as wire bondable chip.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #144 on: September 12, 2021, 07:37:51 pm »
How do those historical HP fineline arrays used in their lab grade meters compare in this context? Are they low noise?

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #145 on: September 12, 2021, 08:05:37 pm »
At least the resistors (inside U180) in the HP3458 must be reasonable low noise (e.g. noise index better than some -50 dB). There is very little extra noise not explained by known souces (to a large part the simple Johnson noise of the resistors at the integrator input and the OP (LT1001) at the integrator.

The circuit with mixed 40K and 50K resitors and the longer integration time (10 PLC) makes it more sensitive to excess noise than my ADC design.

The fine line construction makes it a good candidate for low noise, as the film can be relatively thick.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #146 on: September 12, 2021, 08:59:22 pm »
Are you saying those are thick film resistors?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #147 on: September 13, 2021, 08:03:28 am »
The resistors in HPs fine line technique are still thin film, just a film that can be a little thicker and thus less surface effect. So thick is relative, like 200 nm compared to 100 nm (I  don't know the numbers - just a rough guess)  and than the latteral width smaller due to good lateral resolution in the etching.

Thick film resistors use a different material (usually more like polymer with metal / oxide particles). Due to the production technique with a mask / screen printing they can not at all produce fine lines, but use much coarser lateral structures.

The naming it not really from the film thickneess but the quitestion on how the pattern is defined: thin film starts with a closed film and that uses etching. Think film directly prints the pattern.  So the foil resistors or a PCB are more like thin film technique.   
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #148 on: July 10, 2022, 08:43:02 am »
Is there any way to calculate the "closest divider ratio with max N resistors" (let say with all N resistors equal at the beginning, but finally each one with a resistance value and a TC value) for "ANY possible wirings they can produce"?

Like generating a possible resistor arrangement (especially interesting with none equal resistors), then evaluate the voltage ratio in all created wiring nodes, and picking up the best fits.. And so on with a next new arrangement..

Sounds like as a pretty interesting problem.. Except some theoretical papers talking the total resistance of such sets, I have not found an answer (or application) for the "divider problem" yet (but I am not g..ling for too long, frankly).


« Last Edit: July 10, 2022, 08:48:17 am by imo »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #149 on: July 10, 2022, 08:47:08 am »
I couldn't find anything by now, hence why I wrote this short article:

Resistor ratio dividers based on N equal resistors

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: MegaVolt, iMo

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #150 on: July 10, 2022, 08:53:39 am »
But the number of variants is huge. A colleague of mine says with 8 resistors it is 8! that is 40320, but that is just a guess, imho. With equal resistors some of the arrangements cancel out, what is not the case with none equal resistors. With the given divider ratio the majority will not fit, but the number of close fit arrangements could be still high..

Would be great to develop a brute force solver for that problem..

PS: "N none equal resistors" means an array with resistors you measured before the calculation, as I expect their values differ a little bit. Also it could be an array (or set of two arrays) with quite different resistor values, like the LT5400 series..
« Last Edit: July 10, 2022, 09:47:51 am by imo »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #151 on: July 10, 2022, 11:29:43 am »
With many of the configurations some resistor positions are excangable (e.g. in a series configuration the sequence does not matter) - so the 8! number is not really applicable.

The resistors in the arrays are usually quite close to each other. Often the relative specs are better than 1% and the actual spread may very well be considerably closer. So while the guarantied limit may be less than 1% difference, it may be hard to find actual more than 0.1% difference and when using more resistors there is still averaging with better matching resistors.

So there is little use of meassuring and selecting the right ones for a specific combination.  This would be only a minor deviation from the equal resistor case that comes with the price of individual PCBs for the specific units. Such smaller trims ( < 0.1% range) are relatively easy with a separate resistor. So I doubt this would be worth the effort of individual measurements.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #152 on: July 10, 2022, 11:55:20 am »
There is still an option with arrays like LT5400 and similar. My current understanding is you can mix two separate arrays (each one with good ratio matching TC) such the TC matching ratio stays the same. For example there are arrays with 1k+1k+9k+9k, 1k+1k+5k+5k,  1k25+1k25+5k+5k, etc. you may combine.

PS: anyhow, my first post above was re existing calculator for this problem - such you enter the resistors params and divider ratio and after a couple of seconds/minutes you'll get the optimal setup for given divider ratio..
« Last Edit: July 10, 2022, 12:07:01 pm by imo »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #153 on: July 10, 2022, 12:58:14 pm »
One can combine multiple arrays, but this comes with some difficulties. The matching between different arrays is usually not that great, especially not if from different batches.  So this may add additional chances for drift.

Especially with equal resisstors one may combine them to have the same weight from both arrays. E.g.  combine pairs of resistors from both arrays in parallel of series to get some kind of average. Here one can mix parallel and series. With different resistor ratios this usually no longer works well.  It may also need extra resistors to get equal mixing (e.g. use 2 in parallel and 2 in series instead of 1 resistor to get mixing of 2 chips).
 

Offline Ole

  • Regular Contributor
  • *
  • Posts: 68
  • Country: de
Re: Statistical arrays
« Reply #154 on: September 21, 2022, 06:55:47 am »
I´ve got a question regarding the LT5400´s datasheet:
The datasheet mentions shelf life stability at 25°C, unbiased for 1 Year as +- 5ppm,
is this for the ratio or the absolute values?
*record scratch noise* Hey, you.
Yes, you. Have an awesome day!
 

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #155 on: September 23, 2022, 02:03:32 pm »
I´ve got a question regarding the LT5400´s datasheet:
The datasheet mentions shelf life stability at 25°C, unbiased for 1 Year as +- 5ppm,
is this for the ratio or the absolute values?
Good question. I dont know the exact answer - just guessing
 I think that +-5ppm is for absolute value
 - Looking at "Change in Matching vs Time" - shows +2ppm max @2000h
 

Offline julian1

  • Frequent Contributor
  • **
  • Posts: 719
  • Country: au
Re: Statistical arrays
« Reply #156 on: September 23, 2022, 10:47:25 pm »
Is anything known about the on-die organisation of the 4 resistors in a lt5400?.

Is the (desposition?) layer physically interleaved with each other, to increase temperature matching/coupling? It would seem like a good design/manufacturing strategy.

If a signal feeds through several lt5400 (series and parallel), does it make sense to select a different resistor position for each package. To equalize proximity of all signals to each other, and reduce cross- TCR?



 

Offline antintedo

  • Contributor
  • Posts: 45
  • Country: pl
Re: Statistical arrays
« Reply #157 on: September 24, 2022, 02:28:16 am »
Attached is a poor quality photo of 4x10k LT5400. Individual golden color terminations are visible, there could be as many as 40 resistive elements. Connections seem to be done through at least 2 metal layers, one on top and one under the surface. Resistors are interleaved but not to the fullest extent possible, they are paralleled in groups first. It probably has to do with limited number of metal layers being available. Silicon substrate seems to be used, it did not look like ceramic. I did not strip the passivation layer so no idea about the geometry or whether they are trimmed.
 
The following users thanked this post: Mickle T., doktor pyta, julian1, ch_scr, iMo

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #158 on: September 24, 2022, 08:27:45 am »
I wonder does someone tested NOMCA statistical dividers for long term.
On paper they must shine in this area.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #159 on: September 24, 2022, 09:14:35 am »
The parts i used for my tests in 2019 are still there, one could use them to repeat those tests. Getting precision on the relative resistance values that determine a divider ratio and the TCs was fairly simple. Guess it will take some time though. Other projects here.

Regards, Dieter
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #160 on: September 24, 2022, 09:30:00 am »
I did some test with 2 NOMCA arrays in my ADC. They performed quite good in the TC and were also reasonable stable, as far as I can tell. However the 1/f noise is rather anoying. There is a chance that the processes leading the excess noise can also lead to drift in the long run.  I would prefer the related NiCr version (without the A at the end), that is supposed to have less noise.

AFAIR the specs are also not that great, but with some of the parameters in the datasheet the question is if this refects the actual part performace, the acceptance creterion or just a test limit - so how good are the instruments used to check the parts.
 
The following users thanked this post: miro123

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #161 on: September 25, 2022, 07:16:47 am »
So we should rather test NOMC (not NOMCA) arrays instead. In N. Beevs paper they appeared at a noise index of about -70 dB and eight resistor arrays are available from Mouser. Maybe they exhibit similar TC properties.

Regards, Dieter
 

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #162 on: September 25, 2022, 09:02:41 am »
So we should rather test NOMC (not NOMCA) arrays instead. In N. Beevs paper they appeared at a noise index of about -70 dB and eight resistor arrays are available from Mouser. Maybe they exhibit similar TC properties.

Regards, Dieter
Thanks for feedback. I've just looked at NOMC/NOMCT datasheet. I still don't know what is the better choose for my application
 - NOMC and PBfree NOMCT have Ti2N version Datasheet page 1 Note - Available upon request. Resistance value range and performance differs from passivated nichrome standard electrical specifications on datasheet, consult factory
 - NOMCT subtrate is SI no ised what will be in Ti2N version
 - NOMCA is automotive specified - mostly means quarantined long term parameters, and more resistant in harsh environment.

I have one questions about NOMCA .
-  How bad is the noise performance of NOMCA? If I do match calculation. I come to very low numbers NI=-30..-40dB is still very low.
-  Is there some some data for frequency below 0.01Hz - lets say 1h, 1day or 1week
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #163 on: September 25, 2022, 09:56:59 am »
You can expect the slope to go on for lower frequencies, but effects like temperature variation to show up too, unless you have everything temperature stabilized.

-branadic-



*https://www.eevblog.com/forum/metrology/statistical-arrays/msg3137942/#msg3137942
*https://www.eevblog.com/forum/metrology/diy-high-resolution-multi-slope-converter/msg3392180/#msg3392180
« Last Edit: September 25, 2022, 10:05:40 am by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: miro123

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #164 on: September 25, 2022, 10:38:19 am »
Once more:
According to N. Beev (link to research paper above) NOMCA and NOMC are different in respect to noise index, NOMC give 30 to 40 dB lower noise . NOMC are nichrome film resistors, while NOMCA is a variation of NOMC with tantalum nitride instead.
Today i ordered some of these: NOMCT16035001AT1

Regards, Dieter

 
The following users thanked this post: miro123

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #165 on: September 25, 2022, 01:40:05 pm »
Thanks  Branadic and Dieter,
Sorry that my question was not clear!
I am interested in noise at 1uHz...100mHz = 10e-6Hz...10e-1Hz. Can I draw straight 1/F approximation line to infinity?
I know that the answer is no
I'm not interest from solution that gives me 0.1ppm lower noise. I'm more interested in solution that gives me 5ppm/year better stability.

The paper shows the noise spectrum until 0.1Hz. I assume it comes from industry standard 0.1...10Hz measurements. I understand CERN decision, they are also not interested in any long term analysis, since they calibrate the equipment frequently with stable HP3458

PS: I already build an small oven with stability of 1mK @24h. Still  not enough time to document it. The problem is that multyloop MIMO controller requires manual tuning.



« Last Edit: September 25, 2022, 01:56:52 pm by miro123 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #166 on: September 25, 2022, 01:58:16 pm »
The excess noise can extend quite a bit down. Usually the limit is not from the 1/f noise gettling less, but more from temperature or similar variations taking over.
How far down the 1/f noise extends is a good question and it can depend on the DUT. Even though on a log scale, like the same noise power per decade in frequency, there obviously needs to be a lower limit.  From the physics side 1/f noise would require some kind the memory for the past and the time how long the old state is remembered shoud about limit the 1/f noise. This could be something like the lifetime of defect states involved.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4627
  • Country: nr
  • It's important to try new things..
Re: Statistical arrays
« Reply #167 on: September 25, 2022, 04:12:45 pm »
..
I am interested in noise at 1uHz...100mHz = 10e-6Hz...10e-1Hz. Can I draw straight 1/F approximation line to infinity?
Yes (the slope will be the same in 1pHz to 0.01Hz).
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #168 on: October 13, 2022, 07:27:25 am »
Meanwhile 10x NOMC 8x 5 KOhm resistor arrays arrived here from mouser.
They were put onto SMD adapters with 0.5 mm pins and cleaned.

Using a setup with a Keithley 2700 DVM including a 7706 scanner,
the 2019 NOMCA test jig  (9 channels of 4W resistance with offset
compensation) and an incubator wired to an Arroyo Tecsource 5235
these are the results:

- First diagram shows temperature cycle as seen by TecSource
and by PT1000 sensor included in test jig. There is some temperature
offset and a time lag. Data analysis is based on the temperature
measured by PT1000 sensor. One 18°C to 28°C oven cycle takes about
4 hours, largely unattended.

- Second and third diagram demonstrate resistance and TC
determination by line fit. There are no anomalies like hysteresis
nor higher order TC. Stability and noise of setup appear to be
good enough.

- Table shows results of all 10 arrays tested. There is some
resolution for all data necessary to judge stability of dividers
built with these parts.

- Two more diagrams illustrate scatter of TCs and resistances.
The TCs show a regular pattern while the resistance deviations
are more random. TCs are within +/- 5 ppm/K and TC tracking
within an array is +/- 1 ppm/K. Array 5 is a good candidate for
making a zero TC 10K resistance standard.

Two of the NOMC arrays were put into a drypack to repeat the tests
after some weeks. Meanwhile the 2019 measurements of NOMCA
parts will be repeated.
   
Regards, Dieter
 
The following users thanked this post: doktor pyta, Kleinstein, alm, razvan784, MiDi, MegaVolt, eplpwr, antintedo, miro123

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #169 on: October 13, 2022, 08:40:04 am »
Thanks for sharing.
The TC shape in 4-th graph is similar across arrays R1 and R8 has lover TC than de rest.
Can be an systematic error from the test setup
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #170 on: October 13, 2022, 09:31:08 am »
 "Systematic error" means some mechanism/explanation.
The repetition of my 2019 NOMCA measurements will be a check on systematic errors, as in 2019 a small oven was used instead of the incubator and a HP 3456A with a DIY relay scanner instead of the K2700.
In 2019 i also wired zero TC voltage dividers with NOMCA arrays and they behaved as expected. Another possible check whether one can trust the TC determinations.

Regards, Dieter

 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #171 on: October 16, 2022, 10:24:42 am »
Meanwhile i looked into two possible sources of systematic errors.

First the SMD adapter i am using for the arrays has copper traces of different length. The traces of the outer resistors R1 and R8 are longer than the others and copper contributes with a strong positive TC of 3930 ppm/K. The trace resistance was determined to be about 25 mOhm, so the variation due to different trace lengths may be half of that. Each resistor has this twice (on each of two terminals). So i am getting
25 mOhm / 5000 Ohm * 3930 ppm/K = 0,02 ppm, much smaller than the observed effect.
A similar calculation applies to bonding and frame inside the array, so this may be part of the observed pattern. And then it isn't a systematic error of the measurement, but an explanation why the parts show that pattern.

Second i changed the scanner process to take two measurements for each channel. The standard deviation between the two measurements is about 1.5 ppm, with no systematic difference, except for the PT1000 sensor that shows a shift of +22 ppm due to self heating. This shift is equivalent to 5.6 mK.

Regards, Dieter
« Last Edit: October 16, 2022, 10:42:47 am by dietert1 »
 

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #172 on: October 16, 2022, 08:16:55 pm »
Thanks for sharing.
I want to start experimenting with either NOMCA or NOMC resistor networks.
You have experience with both of then. I have few questions.
How do they compares in terms of
 - TC
 - TCR
 - Temperature hysteresis
 - stability
Thanks
Miro
« Last Edit: October 16, 2022, 08:18:34 pm by miro123 »
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #173 on: October 19, 2022, 07:17:03 pm »
It all depends on the application.
When i started measurements in 2019 i wanted to use those NOMCA arrays to make voltage dividers for LTFLU1 or LTZ1000 references. Then everybody wrote that NOMCA arrays are noisy and not good enough. So i made some references with other parts and they turned out to be good for 0.1 ppm, e.g. after a 5 month power-down.
Then others found NOMC arrays to be less noisy.
With my test setup, the determination of TCs within an array works well, to about +/- 0.05 ppm/K. I can say that the TCs are immune against humidity change (1 week in dry pack at 0% RH).
The resistance measurements look good as well. The NOMCs stay within some ppm. Except my resistance measurements suffer from the contact resistance of the socket. For a 5 K resistor, 1 ppm is 5 mOhm, and the biggest unexpected change i have seen is 3 ppm. Need to order another set of NOMC arrays and measure them using SSOP32 adapters - with the 4-wire ohms connection point on the adapter (behind the socket).
Measured all NOMCAs once more and started to combine the results with those of 2019, to see what happened within two years. Need more time.

Regards, Dieter

 

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #174 on: October 20, 2022, 11:29:25 am »
Thanks,
I've  found few interesting articles. on this subject.
This is what manufacturer says about their products.
https://www.vishay.com/docs/60019/spm.pdf
https://www.vishay.com/docs/49562/49562.pdf

Nichrome Thin Film have typically been used in applications requiring excellent performance but have exhibited catastrophic failure under harsh moisture environments, unless protected in hermetic packages. Tantalum Nitride resistor films have typically been the film of choice but require performance trade offs for precision application
 
The following users thanked this post: doktor pyta, MegaVolt, dietert1

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #175 on: October 20, 2022, 04:40:50 pm »
Thanks. As the VISHAY paper about SPM passivation is dated 2008, we can safely forget NOMCA parts. The datecode of the mouser NOMC parts mentioned above is 2225 and they should be stable against moisture. Those military moisture/temperature tests are rather aggressive and their results describe another regime of operation. For my intended application a 0.1 % resistance change is a failure. I am more interested in 1 ppm resistance changes induced by mild changes in environmental conditions.
The second paper is interesting reading, too. Yet it demonstrates how complicated reality can be and i think sometimes one needs to test parts in their application, in a prototype. That's what i meant above.
Today i ordered some 0.65 mm pitch SMD adapters to get 4-Ohm measurements at the NOMC pins, plus more NOMC parts.

Regards, Dieter
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #176 on: October 24, 2022, 08:30:05 am »
As i am waiting for parts, i unsoldered the pins of NOMC array 7 and wired it as a 0.7 voltage divider to make a TC test, with seven resistors combined for zero TC according to previous TC results. R1 of the array is not used.
Input voltage of the divider is a 14.1 V reference (10x LM399 array). Output voltage of divider is connected to K2700. A simple setup with a divider output impedance of 1750 Ohm. Accuracy certainly depends on K2700 input leakage, but let's try.
Diagrams show log and TC correlation of three 18..28 °C cycles in incubator. There is some modulation with TC estimated as -0.032 ppm/K - certainly compatible with zero.

The intended application of the divider is inside a small TEC oven. Existing prototypes of that oven show temperature variations of about 10 to 20 mK. So the divider TC effect should be negligible:
-0.032 ppm/K * 0.02 K = -0.000 64 ppm

Regards, Dieter
 
The following users thanked this post: doktor pyta, ch_scr

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #177 on: October 24, 2022, 12:05:27 pm »
Great results,
The absence of temperature hysteresis is more appealing for me.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #178 on: October 27, 2022, 04:05:51 pm »
Meanwhile i tested the divider 7 in a larger temperature range 13 .. 43 °C in order to check again for hysteresis and higher order terms (curvature). Curvature was found, hysteresis not. Arrays 8 and 9 have also been wired as zero TC 0.7 dividers and swept in the incubator.

Code: [Select]
Results compared:
Array 7 TC -0,032 ppm/K
    TC(23°C) -0,035 ppm/K (larger temperature span)
Array 8 TC -0,040 ppm/K
Array 9 TC -0,031 ppm/K

Probably the measurement setup causes a systematic shift and the residual TC of these dividers is yet another order of magnitude smaller. One might use this type of divider without oven.

In order to improve my setup i should use a reference with 4-wire output to feed the divider. Currently there is a voltage drop along the cable that feeds about 1.7 mA into the divider inside the incubator, once more with the strong positive TC of copper.

Regards, Dieter
 
The following users thanked this post: miro123

Offline Ole

  • Regular Contributor
  • *
  • Posts: 68
  • Country: de
Re: Statistical arrays
« Reply #179 on: November 01, 2022, 11:14:30 am »
A bit over a month ago branadic posted the comparison in noise for the different resistor networks.

Concerning that I have a question:
In which resistive materials can 1/f noise be expected and in which is it not present?

Thanks in advance,
Ole
*record scratch noise* Hey, you.
Yes, you. Have an awesome day!
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #180 on: November 02, 2022, 03:16:13 am »
The only resistor technology that has not had 1/f noise detected is wire wound resistors, all other technologies have 1/f noise to one degree or another depending on the resistive element(s) used.  While Evanohm (and derivatives) are used in both wire wound and film/foil type resistors, the wire form does not exhibit any 1/f noise that has been conclusively measured while the film/foil types have mostly been verified to have at least some 1/f noise present due to the way the alloy is worked.  All other resistor types have varying amounts of 1/f noise depending on how and what they are made with, in their final form.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #181 on: November 02, 2022, 07:06:50 am »
I fear this statement as given is a myth. Also wirewound resistors exhibit 1/f noise as any imperfection in the atomic lattice will lead to 1/f noise. I agree that with the "MIL-STD-202G, METHOD 308 - Current-Noise test for fixed resistors" to measure resistor noise it was literally immeasurably low back then, but with a bridge configuration and with todays test equipment you can show 1/f noise even for wirewound resistors. Been there, done that.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #182 on: November 02, 2022, 09:00:05 am »
..
Concerning that I have a question:
In which resistive materials can 1/f noise be expected and in which is it not present
,,
Is this question about producing low noise resistor parts?
The diagrams already posted in this thread are in terms of available resistor parts/families. We can get and characterize those parts. You could try and find out the material of parts that were found to be low noise. If they are all made of nichrome, then this could be an answer.

As mentioned in the ADR1399 thread, i am logging a bootstrap/gain stage based on a NOMC resistor array and results are perfect. The 0.7 divider does not add significant noise at the 0.1 ppm level and the TC of the division ratio is less than 0.1 ppm/K. Characterization of long term stability will take some time, of course.

Another approach to testing long term stability is shown in the photo. I made five of those 4 wire adapters with 5K NOMC parts. The idea is to characterize resistance differences within each array to better precision. With the previous test jig the socket contacts were between the 4-wire conncetion and the DUT.

Regards, Dieter

« Last Edit: November 02, 2022, 09:23:31 am by dietert1 »
 
The following users thanked this post: doktor pyta

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #183 on: November 02, 2022, 09:33:24 am »
From the data shown, it is not just the resistive material. One can use the same resistive material with a thicker layer and finer etched pattern to get a longer trace instead of a very thin layer and less pattern to get the same resistance. Expect less noise from thicker film and as one step up the bulk metal foil, that usually is a bit thicker. So not all thin film resistors of the same size are the same. Things can also change with resistor value. Chances are the lower resistor values (e.g. < 1 K) are just a simple layer of suitable thickness with no pattern. The high values (like >100 K) would hardly get away without a pattern.

AFAIK the details of the 1/f excess noise are not fully understood. My understanding is that the bulk metal should not cause much excess noise, but the problem is more with the surface and possible cracks / weak links. With the relative thick material wire wound resistors are naturally low noise, though not per se noise free.  For tunneling contacts it is known that it only takes minmal changes to jump from some 13 K to near infinite or 6.5 K. For me the surprising point is more that some resistors (carbon, but also some thin film) are that noisy. How critical the resistor noise is also depends a lot on the application - often the normal thin film resistors (or even thick film) are good enough and no need to worry. It is rare to require really low noise resistors (like better than -40dB Ni).
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #184 on: November 02, 2022, 01:25:00 pm »
The answer to this question is: "It highly depends...". Many different aspects have to be taken into account. There is high evidence that e.g. surface roughness of the substrate in thin film resistors has an influence using the very same resistive material and process to apply the resistive layer to the substrate, but also by which process the resistive layer is applied to the substrate. You can read through the paper "Measurement of Excess Noise in Thin Film and Metal Foil Resistor Networks" by Nikolai Beev and associated papers referenced to in there.
It's a complex story and while some process and material combinations exhibit lower noise, others show very large excess noise. While wirewound and metal foil resistors in principle can show very low excess noise compared to thickfilm or carbon composit resistors, this isn't true in general and highly depends on how the wire and foil is processed and treated during manufacturing. To some extend you might reach the limit of your measurement setup, but that doesn't mean there is no excess noise at all and with an improved setup you might spot excess noise even in the best resistor available.

You can think of excess noise as a fluidic channel running water through it. The smaller the amount of water (similar to the amount of current through a resistor), the lower the noise due to imperfections (such as stones representing imperfections in the lattic of the resistive material) in the channel. Since we are not living in an ideal world, no resistor technology is perfect. Beside the resistive material also the leads and how they are attached to the resistive material contribute to the overall noise performance of a resistor.

-branadic-
« Last Edit: November 02, 2022, 07:13:35 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 
The following users thanked this post: zrq

Offline Ole

  • Regular Contributor
  • *
  • Posts: 68
  • Country: de
Re: Statistical arrays
« Reply #185 on: November 02, 2022, 02:13:06 pm »
Thanks for the informative answers,
Nikolai Beevs paper was quite enlightening.
The question was originally due to my own interest in noise and my plans to build a low noise amplifier.
Since most of the noise is voltage dependent it wont be that bad for a low noise amp.

Cheers
Ole
*record scratch noise* Hey, you.
Yes, you. Have an awesome day!
 

Offline RandallMcRee

  • Frequent Contributor
  • **
  • Posts: 540
  • Country: us
Re: Statistical arrays
« Reply #186 on: November 03, 2022, 11:59:48 pm »
Quote
Since most of the noise is voltage dependent it wont be that bad for a low noise amp.

No, most of the noise in a low-noise amplifier comes from the temperature noise of the resistors and amplifier in the front end. Doesn't matter what type of resistor.

Consult, for example Motchenbacher, "Low Noise Electronic Design".

Even an online noise calculator like http://dicks-website.eu/noisecalculator/index.html will show the major factors in a noise calculation and resistor type is not one of them.

There are many, many, good low noise amplifier designs out there, hopefully you are well acquainted with them?



 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #187 on: November 13, 2022, 08:00:15 pm »
For the time being i finished measurements of the five NOMC arrays with 32-pin adapters mentioned above (for improved 4-wire ohms). I measured each array four or five times. After the first rounds i cleaned the devices once more in order to reduce leakage currents. Still results are mixed.
TCs are measured to about 0.01 ppm, better than expected.
The relative resistance within each array - that is important for divider stability - appears more difficult. Since the setup is measuring the eight resistors interleaved i was hoping for a similar accuracy, but it seems to be more like 0.3 ppm or so.
I inserted a delay to help offset compensation, yet without obvious improvement.
Will try and repeat those measurements after some months.

Regards, Dieter
« Last Edit: November 13, 2022, 08:01:51 pm by dietert1 »
 
The following users thanked this post: MiDi, MegaVolt

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #188 on: November 15, 2022, 01:11:46 pm »
Meanwhile i found that i forgot to clean flux from the 32-pin socket wiring. Little patience when checking the new 4-ohms wiring to the Keithley multiplexer.
The first tests after cleaning the socket look better. Repeatability on the relative resistances improved to about 0.05 ppm. With a 5 KOhm resistor as DUT this corresponds to a leakage "resistance" of about 100 GOhm.

Regards, Dieter
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #189 on: January 18, 2023, 07:05:18 pm »
After Nikolai has published his paper on Measurement of Excess Noise in Thin Film and Metal Foil Resistor Networks with a version of it accessible via the Arxiv I'm happy to let you know that our paper Review on Excess Noise Measurements of Resistors was today published too. It is OpenAccess.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Dr. Frank

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
Re: Statistical arrays
« Reply #190 on: January 18, 2023, 08:00:25 pm »
Hello branadic,

thank you very much for these very interesting papers & measurements.

 On page 17 of your paper, why is the technology of Alpha Electronics resistors not documented?
I always thought they are also BMF like the VHx and similar, directly from Vishay?

Frank
« Last Edit: January 18, 2023, 08:20:49 pm by Dr. Frank »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #191 on: January 19, 2023, 07:49:42 am »
Hi Frank,

it's not directly included in the datasheet, but if you search through the ultra precision resistor cataloge it is somewhat indirectly given that they are BMF resistors.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline miro123

  • Regular Contributor
  • *
  • Posts: 197
  • Country: nl
Re: Statistical arrays
« Reply #192 on: January 21, 2023, 08:00:59 am »
 Thanks for sharing. I value the incredible job done to test and document all this.
I have few questions
1. Do you belief that  1/Fa noise increases to infinity? - question is related to  page 4 "However, in[15], Horowitz and Hill argue that excess noise increases forever. "
2. How do you remove the resistor TC in equation? I mean every oven introduces some waves /temperature frequency/
3.  Does HP3458 U180 and  W7000 use TaN resistor networks? If yes, why this technology scores so bad in such short term(24h/50uHz)tests

feedback from other forum members are welcome too.
I think that answer to this questions will give us also better understanding of LTZ1000 vs ADR1000 behavior. Mid term vs long term noise
BR
Miro
===========
Here is my input
1. No it does not  goes to infinity. It is based to on my mathematical understanding of Fourier transform and observing the nature. e.g. we have not seen in long term 1/f low for earthquakes?
2. My question was toward TaN resistors. Generally they have higher TC.
3. No idea. I know that many automotive application uses TaN due to tight long term stability requirements.
On the other side I belief that mass production wins in long term. So at certain point the advances in mainstream thin film Nichrome will reach and surpass bulk foil and TaN




« Last Edit: January 21, 2023, 08:31:02 am by miro123 »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
  • Sounds like noise
Re: Statistical arrays
« Reply #193 on: January 21, 2023, 09:53:24 am »
There were several discussions in the past, with opposite arguements, However, there was not final conclusion as far as I know.
Ignoring all effects such as influence of temperature, humidity, pressure, chemical processes, just to name a few of them, we still have to keep in mind that for the age of the universe the spectrum would only go down to about 2.3 aHz, while each decade adds the same amount of power to the spectrum, one year is about 32 nHz.
We simply haven't observed anything long and stable enough to tell and there are too many effects that are way larger than what we want to observe I guess.

The measurements were performed under controlled environmental conditions and we waited a while for the resistors to thermally stabilize after the bias voltage was applied, hence the temperature was stable during the measurement and t.c. of the resistors could simply be ignored. Proper shielding, cables and thermal insulation is required. We had a cookie box inside a styrofoam box plus the stabilized ambient temperature.

As far as I'm aware 3458A and W/F7000 uses NiCr resistors, not TaN. On the other hand TaN or NiCr networks with TaN passivation have the advantage of growing a self-controlled oxid that surpresses humidity influences.

-branadic-
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: Statistical arrays
« Reply #194 on: January 21, 2023, 10:16:11 am »
The resistor excess noise is not only a function of the material used, but also a question of how much. A combination of 4 resistors as 2 in series and 2 in parallel has 1/2 the excess noise of one the resistors alone. The network in the HP3458 is quite large in size. In addition they call it fine line, which would imply a relatively long track and thus thicker film compared to resistors that us a coarse etched structure.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: br
    • CADT Homepage
Re: Statistical arrays
« Reply #195 on: January 21, 2023, 11:02:20 pm »
As far as i remember this thread had some studies about NOMCA arrays until the Beev et al. paper appeared, with about -30 dB excess noise for those arrays, while NOMC arrays measured better, at about -60 dB, that is similar to metal foil resistors. Later we learned that NOMC arrays have a "new" proprietary passivation (Vishay).
Meanwhile there are some results on long term stability, also Andreas had a long term log for a 7-10 V gain stage with a similar array (TDP). Yet there was some doubt concerning humidity. I used a NOMC array for a 7-10 V gain stage with an ADR1399, with sub-ppm stability.
I did not find the NOMC arrays in the new paper.

Regards, Dieter
 

Offline xerbo

  • Contributor
  • Posts: 10
  • Country: gb
Re: Statistical arrays
« Reply #196 on: February 15, 2023, 05:10:07 pm »
Throwing my 2 cents here, an exact 7/10 divider for lower voltage references (i.e. LM399)
 
The following users thanked this post: doktor pyta, mycroft, ch_scr, Grandchuck, Ole

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: Statistical arrays
« Reply #197 on: March 02, 2023, 09:43:26 pm »
The 'myth' is not a myth, if you had read my published article in EDN, it explains the source of wire wound woes.  In this case, any 1/f noise is going to be generated in the flawed, non-welded joints, not in the wire.  You very well may see a bit of 1/f noise but the odds are highly in favor of the 'weld' joint causing the noise.  In high quality resistance alloys, the metal mix is as near perfect as modern technology can produce.

As I have pointed out in the past, the cold-rolled alloy film/foil Vishay uses has imperfections in the matrix from not only the cold-rolling but also by the etching/laser tweaking process leaving very rough microscopic edges which generates 1/f noise in pretty good quantities.  Precision wire alloys are made with a process that leaves little to no 'rough' edges and the alloy mix is highly homogeneous, again leaving little imperfections to generate 1/f noise.

The 'art' of measuring 1/f noise is far from easy especially when there is little to be seen.  There are no amplifiers, sans those in liquid Nitrogen, that have no 1/f noise of their own, separating out which one sourced the 1/f spikes is nigh onto impossible.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf