Author Topic: T.C. measurements on precision resistors  (Read 253085 times)

0 Members and 1 Guest are viewing this topic.

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #150 on: July 20, 2014, 11:42:58 am »
Hello folks,

By George, those certainly are hermetic resistors, fairly old ones but their specs for these resistors are not very impressive and even less so for hermetic resistors.  Those are all hand soldered, I've done many of those myself over the years.  Brass tubes and usually, but not always, oil filled (those were always fun to solder), all two terminal in there.  Can't tell who may have made them, Fluke plastered their own part numbers on a lot of components made by others.

It is possible I may have seen an early production model because I do not remember seeing hermetic resistors in it.  Fluke could have easily changed the resistors when they went to full production on them.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #151 on: July 28, 2014, 05:40:09 am »
Hello,

I have updated Z201#2 measurements in the Z201 post on page 1.

measured T.C. of #2 is  -0.27 ppm/K which is nearly the "typical" value  of +/- 0.2ppm/K
compared to
measured T.C. of #1 was 0.9 ppm/K

with best regards

Andreas
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #152 on: July 29, 2014, 05:10:54 am »

Humidity sensitivity:

On the other side a volt-nut friend of mine has reported a humidity sensitivity between 5-50ppm for 30% RH change of 8E16 resistors which are similar to the UPW50 wire wound resistors. I hope I will get further information on this topic from him when he is back from his trip.


I got now some measurement values for humidity values:
these values are reversible when humidity changes back and forth.
Time constant for humidity is in the range of days to weeks.

8E16 1K about 10-30ppm for a 25-30% change in rH. (around 25% to 50% absolute)
8E16 100K about 50ppm in the same range.
the trend for wire wound seems to be that for higher ohmic values the humidity sensitivity increases.

S102J 1K + 10K about 30 ppm for a 30% change in rH.
S102J 100R is about 5-10ppm in the same range
So also the trend seems to be lower for lower ohmic values.

So I am wondering wether the measured hysteresis on my S102J has to do something with humidity.
On the other side the time constant is relative fast for a humidity change.

Any ideas?

@Edwin: How do you explain the (reversible) rH sensitivity on the wire wound resistors?
Is the trend for higher ohmic values explainable?

With best regards

Andreas

 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #153 on: July 29, 2014, 07:17:39 am »
Hi Andreas,

Two reasons, the main defect at work here is the so-called weld joints, water molecules are getting into the joint and causing trouble (caution, oxides at work) at the 'weld' joint, it has a similar, although usually smaller, effect than a crimped joint produces.  The secondary defect, which would be of no consequence if the joints were welded, is that the resistor is not sealed completely and allows humidity to sneak inside the resistor.  When the resistor is heated sufficiently, the water evaporates returning the parts value to the original value or close to it.  Repeated exposure to moisture and heat cycles will eventually cause the resistor to drift more and more, possibly out of tolerance if the original tolerance was tight.  A well crimped joint (like the resistors that went to the Moon) can actually stay intolerance for years under the right circumstances.

The enameled wire will have exceedingly little or no response to water as the enamel is very effective at keeping the water out and Evanohm (or its other family members) is not particularly sensitive to water either but a defective weld joint is an open invitation to trouble.  One thing which does help keep these resistors in better health is simply keeping them powered all the time.  That is less likely these days as not so many can afford to keep their equipment running 24/7.

In the case of foil resistors, they are usually welded, that is not the problem, in foil's case, it is the fact that the resistor element is so thin it is very susceptible to water effects as are the very tiny interconnect wires between the element and the outside world.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #154 on: July 30, 2014, 06:06:23 am »

As for foil resistors, the early designs [and even the newer Z-foil resistors] use high-temperature epoxy to glue the foil to the ceramic substrate. 

The new Z1-foil resistors use polyimide to glue the foil to the substrate and to seal the resistor on the top after they are done adjusting it. 


Hello Ken,

from where do you have the information with the epoxy glue (or polyimide).

The only document that I have found speaks from "cement" to fix the foil to the ceramic substrate:
http://www.vishaypg.com/docs/63171/TN104.pdf
On page 5 they speak of about 20 ppm seasonal changes for humidity. (foil resistors).

On the other side: all metal (wire or foil) can be regarded (more or less) as strain gage.
So if the epoxy housing swells due to humidity it may have a influence on the resistor value.

With best regards

Andreas
 

Offline nadona

  • Regular Contributor
  • *
  • Posts: 205
  • Country: us
Re: T.C. measurements on precision resistors
« Reply #155 on: July 30, 2014, 06:51:00 am »
...

The solution for foil resistors is to get them in hermetic packages if this kind of shift is important to your design-- or, just buy less expensive PWW resistors from Edwin that minimize this problem.

So Edwin has less expensive PWW resistors than Vishay resistors?

Ha-ha-ha. That's good, too!
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #156 on: July 30, 2014, 09:02:45 am »
Hi Andreas,

To quote, " On the other side: all metal (wire or foil) can be regarded (more or less) as strain gage.
So if the epoxy housing swells due to humidity it may have a influence on the resistor value."

Whoever (Vishay no doubt) wrote that sentence, doesn't know what they are talking about (and I think I know who it was).  PWW, properly processed, are NOT strain gauges, for that matter, they never were characterized as such, unlike the Vishay resistors.  Yes, there is some strain imparted to the bottom layer of turns on a PWW resistor, this is the quite normal result of winding but that stress (again, if properly made) is essentially removed during the resistor processing.  On the 'other' PWW resistors floating around out there, instability is more often caused by nonexistent weld joints, a not so good seal or maybe just an inept resistor design.  Winding stress in every manufacturer's resistors I've examined has been mostly relieved and is not generally the source of other instabilities.

In the case of foil resistors, they are the most sensitive to stress of all thick. thin and foil technologies, this stress cannot be removed simply by baking the part, in the case of foil, they have had to find materials which minimized the actual stress on the foil byway of its bedding on the substrate.  PWW and Vishay resistors are two entirely different resistors, their manufacturing processes share little if anything in common and mostly have different issues affecting the resistance.

Epoxy does absorb water like many materials, but it depends on just what epoxy and hardener was used, some epoxies will absorb only a couple of tenths of a percent of water during a humidity bath while others can absorb a magnitude more.  While that certainly is a problem for foil resistors as Ken put it, PWW resistors (again, properly made) will have little to no effect from humidity.

I believe it was asked if humidity may have a greater effect on higher resistance values than lower values, that is a qualified yes, in the case of foil, if humidity does get inside of it (and Vishay seems to indicate that it does for non-hermetic) it would tend to have a larger effect on higher values than lower values because of the much finer circuit traces making up the resistor, the water molecules will have less distance to travel between close circuit traces than farther circuit traces.

In PWW resistors (again, correctly made) humidity would not affect the resistance any differently be it high or low.

I do not know if Vishay is using a similar enamel like mine or not, I have not seen any information that gives that detail.  I would have to say, that in the case of PWW resistors, the enamel I use does not absorb enough water to be of any concern.

Since the inception of film resistors (of any kind) moisture has been a problem, carbon, metal, thick, thin or foil, they all have a sensitivity to moisture and that resulted in many different encasements, even glass to stop the humidity's effects.  Plastics, for foil at least, have turned out to be only a partial fix to the problem.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #157 on: July 31, 2014, 07:04:18 am »

I have a love-hate relationship with foil resistors.  Currently, I am hating them.  Tomorrow is another day...


Hello Ken,

I can understand this.
Same resistor similar game two days later but totally different results:

On 27th I had forgotten to switch off the heater during night.
So actually i had a fast cool down from 45 deg C to 25 deg C in the morning.
So most probably a "dry" resistor.

On 29th the temperature was "room temperature" during night.
The result is a strange behaviour for the first "cold" cycle of the T.C. measurement.

At the moment it is rather rainy outside.
So humidity is rather high (65-67% rH) on my hair hygrometer.

So on rainy days you can hear the resistors whine.

With best regards

Andreas
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #158 on: August 03, 2014, 05:15:41 pm »
Hello,

same game two days later on 1st of Aug, similar humidity (room temperature since last measurement) but slow temperature ramp (0.12 K / minute)

So it seems that its not humidity but temperature ramp speed which makes trouble to the Z-foil resistor.

with best regards

Andreas

« Last Edit: August 03, 2014, 05:17:53 pm by Andreas »
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #159 on: August 04, 2014, 01:32:56 am »
Hi Andreas,

Humidity's rate of absorption or egress tends to be a balanced process under the same conditions, note that is the same conditions, when the rate is changed in either direction, so changes the rate of absorption or egress.  As the mass of a foil resistor is small, it has a comparatively short thermal tail, even though the plastic it is encapsulated in has a low thermal conductivity.  In your second test, you changed the rate of flow of humidity out of the resistor by using a very slow temperature ramp, thereby changing the conditions and making the comparison between the two measurements somewhat inaccurate inadvertently.

If the foil resistor does not have an encapsulation defect, the absorption/egress of humidity should be relatively slow compared to the response to temperature.  Obviously, a faster temperature ramp causes accelerated responses in both humidity and foil.

If I am reading your graphs correctly, I believe you are seeing a case of internal condensation effect for the first cold cycle, the condensation being evaporated during the subsequent heating cycle (for the 29th).  The resistor did indeed start off 'dry' because of the heater being left on during the night (for the 27th), therefore you are indeed seeing the effects of humidity on the foil resistor.  It is not a terribly big effect but it certainly is a significant effect.  That is why, in electron tube days, the equipment was left on 24/7 because the carbon resistors sucked water like a sponge and keeping them nice and warm effectively eliminated the effect.

The effect because of a difference between 0.12°C and 0.30°C ramping rate is quite insignificant.
« Last Edit: August 04, 2014, 01:41:52 am by Edwin G. Pettis »
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #160 on: August 04, 2014, 06:51:10 am »

If I am reading your graphs correctly, I believe you are seeing a case of internal condensation effect for the first cold cycle, the condensation being evaporated during the subsequent heating cycle (for the 29th). 


Hello Edwin,

I already had the same idea. But why do I see this only on the "fast" cycle and not on the "slow" cycles (before and after).

With best regards

Andreas


 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #161 on: August 04, 2014, 06:59:16 am »
Hello,

I have updated the Riedon USR2 post on page 1 with first measurement data.

Since the USR2 has some "sticker" (for labeling) its case does not fit easily into the thermal block.
So I had to use some (gentle) force to put it in.
There is some ageing drift (which I already mentioned yesterday after soldering with room temperature).
First measurement indicates T.C. around -1.1ppm/K.
I will have to check wether the mechanical stress of the thermal block has a influence.

With best regards

Andreas

« Last Edit: August 04, 2014, 07:09:03 am by Andreas »
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #162 on: August 04, 2014, 08:48:25 am »
Hi Andreas,

As to your question, "But why do I see this only on the "fast" cycle and not on the "slow" cycles (before and after)."

Let's see if I have the sequence correct, on the 27th, the resistor had been 'cooked' over night so any humidity inside was minimal for the first test.  Rate of temperature change 0.12°C/minute.

On the 29th, resistor had been sitting at room temperature for two days in moderate humidity before testing, resistor showed odd reading.  Rate of temperature change 0.3°C/minute.

On the 1st, the same resistor had been sitting at room temperature and similar humidity, the only 'difference' was the rate of temperature change, 0.12°C/minute.  Readings appeared about normal.

I have a question or two for you.  How many cycles of cold and hot do you do before the resistor rests?  Any particular reason why you chose a different rate of temperature change for the second test?

Throwing out an educated premise, if we assume the rate of humidity absorption by the resistor was roughly similar between testing periods, the faster  delta T may have caused a different rate of condensation inside the resistor, i.e. it got colder faster, therefore more condensation occurred inside before the heating cycle was able to remove it.  With the slower delta T rate, the condensation took longer to form, thereby lagging the temperature change and forming less condensation inside to remove during the subsequent heating cycle.  I am referring to condensation directly on the foil of course; not so much around it, although that can have a small role in the effect as well.

I might add that a resistor 'sitting on a shelf' in a humid environment is considered benign, the usual humidity tests are conducted at an elevated temperature >70°C and 85%-95% humidity for 100 hours or more depending on the test requirements.
« Last Edit: August 04, 2014, 01:53:12 pm by Edwin G. Pettis »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1386
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #163 on: August 04, 2014, 08:53:59 pm »
Question: For what I know IC package mold mass is filled with glas balls, to minimize humidity influence. We use similar materials when doing Foil Assisted [Transfer] Molding process to create IC packages with [optical] interfaces. Is the same true for molded resistors?
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #164 on: August 05, 2014, 03:08:24 am »
Hello Branadic,

To my knowledge, Vishay does not use any glass balls in their resistors unless that is something recently added.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #165 on: August 05, 2014, 06:20:55 am »

Let's see if I have the sequence correct, on the 27th, the resistor had been 'cooked' over night so any humidity inside was minimal for the first test.  Rate of temperature change 0.12°C/minute.

On the 29th, resistor had been sitting at room temperature for two days in moderate humidity before testing, resistor showed odd reading.  Rate of temperature change 0.3°C/minute.

On the 1st, the same resistor had been sitting at room temperature and similar humidity, the only 'difference' was the rate of temperature change, 0.12°C/minute.  Readings appeared about normal.

I have a question or two for you.  How many cycles of cold and hot do you do before the resistor rests?  Any particular reason why you chose a different rate of temperature change for the second test?


Hello Edwin,

the number of cold and hot cycles depends a bit on how much time I have on the day when I am soldering.
I try to do a (fast) cold and warm cycle after soldering. But this does not happen always.

For Z201#2 the history is as follows: (It was a saturday so I had some time).

Soldering on 26th after this a fast heating to 45 deg C
Fast cool down to 25 deg.
Fast cold cycle to 10 deg C.
Fast heat up to 25 deg
Fast heat up to 45 deg (over night)
Where fast means in this case: as fast as my equipment can.
(picture attached)

On 27th in the morning a fast cool down to 25 deg C
from this the measurement beginning with a cold cycle 0.12 deg/minute.
on 28th no measurement just room temperature.

On 29th the measurement (3 cycles) with 0.3 deg / minute.
On 30th+31th no measurement just room temperature.

On 1st the measurement 1 cycle with 0.12 deg/minute.
(pictures already shown in previous posts)

The different rates have several reasons.
with several cycles within one measurement I want to see wether there are some ageing
or hysteresis effects which change over time.
(0.3 deg/K gives up to 3 cycles per day).

With the slow ramp I wanted to see if the hysteresis has something to do with thermal delay.
So if the ramp speed is factor 2.5 slower then a "hysteresis" due to thermal delay (between resistor
and themperature sensor) should also change by factor 2.5.

But since the factor observed does by far change not that much the root cause of the hysteresis
is not (primary) a temperature sensing problem.

I also had different effects on hysteresis with different ramp speeds of (different types of) voltage references,
so I wanted to see if it has also a influence on resistors.
So Z201#2 is the first candidate where I see something unusual.

With best regards

Andreas

« Last Edit: August 05, 2014, 07:02:36 am by Andreas »
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #166 on: August 11, 2014, 11:46:18 pm »
Hello,

I have updated the USR2 section on page 1 with measurement data

https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462305/#msg462305

USR2 #1 has a large ageing drift. (Or is it humidity because it came freshly out of a welded polyethylene bag)

https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/?action=dlattach;attach=105219;image

USR2 #2 the first measurement indicates a T.C. of about -1.4ppm/K
Ageing drift is much smaller than on #1.
But this candidate had 1 week to acclimate after opening the bag.

With best regards

Andreas
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #167 on: August 29, 2014, 07:01:29 am »
Hello,

I have updated UPW50 section on page 1 with several measurements.

https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462300/#msg462300

UPW50#2 0.63 ppm/K @ 25 deg C
UPW50#3 1.85 ppm/K @ 25 deg C
UPW50#4 4.76 ppm/K @ 25 deg C

So #2 beats even one of my Z201 resistors.

To keep track I created a overview from the measurements up to now.

I would interpret the results as follows:
With high sophisticated applications it is worth to measure the T.C. (+hysteresis) within the application temperature range and with some thermal cycling also the ageing drift over some days to further select the resistors and to sort out the "stinkers".
Some datasheet specs seem to be a bit too optimistic when looking at the "typical" values.
Sometimes (often?) the hysteresis is larger when having a low T.C. within a resistor type.

With best regards

Andreas
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #168 on: September 01, 2014, 06:40:13 pm »
Hello,

I have done the first measurement on a UPF50 (molded metal film resistor) from TE.
see metal film section on page 1:
https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462301/#msg462301

Interestingly this device has no visible hysteresis.
So it seems that the hysteresis has nothing to do with my measurement setup
but is within the wire wound and metal foil resistors.
Unfortunately there is no shelf life data supplied with the UPF50.

With best regards

Andreas

 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #169 on: September 02, 2014, 11:35:06 am »
Hi Andreas,

I've been busy working on a resistor order so I've had little time to reply.  The UPF50 data sheet appears to be a 'preliminary' one, likely they do not have any long term figures available as yet but given the environmental test limits, even if this resistor type was twice as good as the maximum test limits, they are nothing particularly special.  With test limits of 50 PPM to 2000 PPM for resistance change, they are about average for metal film, about the only 'special' spec is the 0 ±5 PPM/°C TCR which is quite good for this type of resistor.  Their power range is average, restricted above 70°C, derated to 0 at 125°C.  The shelf life should not be significantly different from any other similar metal film resistor of this type.  I did not see anything in the specifications that would lead me to think otherwise.

I would not be too quick to jump to conclusions about the presence or nonpresence of hysteresis, I have never seen a metal film resistor that didn't have it and this resistor's construction is no different.  One sample is much too small for any conclusions, there may be a valid reason why you did not detect hysteresis.

Ciao,

Edwin
« Last Edit: September 02, 2014, 02:19:07 pm by Edwin G. Pettis »
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #170 on: September 05, 2014, 01:28:25 am »
Hello Edwin,

so you think that shelf life is around 300 ppm / year similar to RC55Y device?

In the mean time I did further measurements on UPF50 #1. (see also on page 1).
https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462301/#msg462301

With fast ramp (0.3K/minute on 01.09.2014) I got some hysteresis on the rising edge after the first cold cycle.
The slow ramp (0.12K/minute on 02.09.2014) showed again low hysteresis.

The 2nd device UPF50 #2 showed some hysteresis on cold cycle (0.12K/minute on 03.09.2014).
Here I did not do a cold cycle after soldering against sample #1.

With best regards

Andreas


 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #171 on: September 05, 2014, 06:28:46 am »
Hello,

for those who did not mention it (like me).
Ham Fest in Weinheim (Germany) is a possibility to meet:

https://www.eevblog.com/forum/chat/eevbloggie-meeting-at-ukw-tagung-weinheim-ham-fest/msg503262/#msg503262

http://ukw-tagung.org/

With best regards

Andreas
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #172 on: September 07, 2014, 07:57:07 am »
Hello,

I have updated the metal film section on page 1:

https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462301/#msg462301

UPF50 #2: the hysteresis gets smaller with each temperature cycle but is still larger than on #1.

Also did first measurement of a RC55Y with 15ppm/K in data sheet.
This turns out as nearly -9ppm/K in actual measurement.
But there is a large ageing drift of 4-5ppm within the measurement.

With best regards

Andreas



 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #173 on: September 09, 2014, 07:24:48 am »
Hello,

I finished measurement of RC55Y #1.
https://www.eevblog.com/forum/projects/t-c-measurements-on-precision-resistors/msg462301/#msg462301

total drift in 3 days around 14 ppm.
So it makes no sense to test further RC55Y.

Attached: Overview of the measurements up to now.

with best regards

Andreas



 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2302
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #174 on: September 29, 2014, 08:10:33 am »
Hello,

currently I am characterizing two new ADCs with capacitive (LTC1043) dividers.
I decided to compare a LTC1043 divider against a high precision resistor divider.
In this case it was a Vishay DSMZ 5K/5K (1:1) divider.

http://www.vishaypg.com/docs/63121/dsmz.pdf

from datasheet
typical ratio tracking T.C. 0.1ppm/K
maximum tracking     T.C  0.5 ppm/K (-55..125 deg C for a 1:1 divider)

Measurement was ratiometric: The VREF of the ADC was used as supply for the DSMZ; canceling out the VREF drift.
The DSMZ was soldered with thin copper wires to a DSUB-connector which I plugged directly to the ADC.

I got 28.3uV change in 31.1 K which is 0.91uV/K or 0.36ppm/K average with respect to the nominal 2.5V at the output of the divider.
There is also a large hysteresis or ageing drift in the measurement.

As comparison a LTC1043+LTC1050 capacitive divider with ratiometric measurement (dividing VREF of ADC).

In this case I got 1.0uV change in 32.2K which is 0.031uV/K or 0.0124 ppm/K @ 2.5V
no hysteresis.

The offset drift of the LTC1043  #1101_004 divider was 3.2uV/33.2K

And the ADC #16 itself measured a gain drift (5V Vref measurement) of -2.1uV/33.6K
and a offset drift of 0.6uV/32.9K

With best regards

Andreas

Edit: attached picture of DSMZ#1 with ADC#16
« Last Edit: October 06, 2014, 05:06:39 am by Andreas »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf