Author Topic: Project KX : DIY calibrator / reference source/meter  (Read 78488 times)

0 Members and 2 Guests are viewing this topic.

Offline DiligentMinds.com

  • Frequent Contributor
  • **
  • Posts: 999
  • Country: us
  • What,  ME Worry !?!?!?
Re: Project KX : DIY calibrator / reference source/meter
« Reply #225 on: August 05, 2015, 02:09:46 AM »
Take a look at Vishay NOMCA.  After high-humidity burn-in at 70C for 90 days, the absolute value drift will be practically zero.  The ratio tracking can be almost as good as an LT5400, but these parts are cheap.  This is touching on proprietary information, but connecting the resistors in various clever ways can result in very stable divider networks that track very well [much better than spec]...  Sorry, can't go into more detail, as I've said too much already...
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #226 on: August 05, 2015, 02:12:36 AM »
No problem, going with them anyway, seems best cost/performance solution so far, and available too.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Online ManateeMafia

  • Frequent Contributor
  • **
  • Posts: 522
  • Country: us
Re: Project KX : DIY calibrator / reference source/meter
« Reply #227 on: August 05, 2015, 03:01:12 AM »
Nice find on the NOMCA. I was looking at using the ORN based resistors for an ovenized application. Comparing the two I can see the NOMCA would be a better choice for moisture protection.

I am guessing that connecting every other resistor would improve temperature tracking or am I missing something obvious?
 

Offline plesa

  • Frequent Contributor
  • **
  • Posts: 893
  • Country: se
Re: Project KX : DIY calibrator / reference source/meter
« Reply #228 on: August 05, 2015, 03:06:38 AM »
I did not noticed big material difference between NOMC (Nichrome) and NOMCA (Tantalum Nitride). I found that Tantalum nitride network is used in Fluke 7001 voltage reference.They only changed technology name to  "Taitanium-Nitrate" ( I have no clue why, to protect IP? :-)
Added info about ageing trends ( compare NiCR and TaN)
« Last Edit: August 05, 2015, 06:35:12 AM by plesa »
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #229 on: August 18, 2015, 03:19:57 AM »
First try on resistor tempco measurement:



So far no good, TEC temperature control go out into wild. Need rethink and redo box and improve insulation to solve this, and likely play with PID settings.

Controller = Raspberry Pi tied to GPIB dongle.
TEC control = Kei2510 SMU
Measurement = Kei2002
Temperature sensor = Honeywell HEL-705 RTD
TEC = some random chinese 24W 40x40mm.
« Last Edit: August 18, 2015, 02:11:59 PM by TiN »
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 1332
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #230 on: August 18, 2015, 04:56:46 AM »
Hello,

where is the temperature sensor?
For temperature control I usually use (minimum) 2 temperature sensors.

One at the point of interest (poi). (near resistor)
One at the heater/cooler.

with the heater sensor I do a (fast) P-controller against the heater setpoint.
with the poi sensor I try go get the poi - setpoint.
this is done by adjusting the heater setpoint offset temperature
(usually 3-4 deg C) with a slow I-controller.

with best regards

Andreas
 

Offline commie

  • Regular Contributor
  • *
  • Posts: 233
  • Country: gb
Re: Project KX : DIY calibrator / reference source/meter
« Reply #231 on: August 18, 2015, 05:38:17 AM »
So far no good, TEC temperature control go out into wild. Need rethink and redo box and improve insulation to solve this, and likely play with PID settings.

Hi TiN,

Your 'oven' you trying to control is an integrator, correct? If you want quickest response,with your controller,  you need to turn off integration(I) and up the differentiation(D) time, so that a tuned circuit is implemented.Your gain setting is way too high, turn it down a little. :popcorn:
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #232 on: August 19, 2015, 08:16:27 PM »
Alright, will try.

Had a chance to work on ADC a little, finished routing digital section. Analog still need some schematics learning and placement...

xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #233 on: August 27, 2015, 03:04:16 PM »
Was looking around web, and discovered Applicos ATE system. It's similar to NI VirtualBench and Agilent's PXI boxes, but more targeted for high-precision stuff.
One of their applications are for testing DAC/ADC's performance.

Worth to mention, their reference source module using LTZ1000A



And it's worth to take a look of their other modules, as they show photo of each module and block diagram. Good overall design concept for modularity.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #234 on: December 22, 2015, 03:48:56 AM »
Perhaps you wonder how Xmas looking for volt-nuts? Well, happy X-mas here, got some happy toys to play with.
Last four PCBs for LTZ's will find their brains soon.



These are no usual LTZ1000CH's, I warn you :) Perhaps one can get a hint from a date codes, which is 15 week 1990. These references are well conditioned and aged already, so it will be very interesting to see how these perform compared to my barely aged LTZs, bought from LT in 2014 and 2015.



One interesting thing, these old chips don't have gold flash on bottom package can, unlike all recent ones. Perhaps packaging was changed, as LTZ1000ACH in HP 3458A A9 PCBA, made in 1993 is with gold-plated bottom already.
Why that's important? Well, that was one of reasons how we recognized fake LTZs from china before. Old genuine can be without golden plating as well, as it turns out ;)



Pair of National Semi's LM399 1994 and 1997 to juice it up a bit as well... Aged and ready to go...

And that's just warming up...



100K precision wire-wounds and VHA518 100K 50ppm foil? :). Few other values WW as well.



SFERNICE's 30K 1%, 120R 1%, pair of 40.9600K 0.005% S102C in plastic mold packages.



More wirewounds, 10R, 100R, 300R, 354R, 2K, 1.5K, 20K, pair of 10Ks. They are beautiful, in black shiny epoxy. I wish to have some of that epoxy to encapsulate some of Fluke wirewounds I have....

This is great collection for some tempco testing and comparison experiments I'm working on. But hold on, tempco testing? How can one measure temperature coefficient without good temperature sensor?
Here you have it, YSI 44031 precision thermistor network:



If you click on bigger photo, you will see whole label, which warns you to heatsink leads when soldering. That's a usual rule for sensitive precision resistors, to avoid hysteresis and shock to element during assembly!



And to couple all this precision gear with your DMMs, make sure to use good pure copper terminals, just like these heavy stuff.
Require bit of cleaning, but that's no problems. Bigger ones are really heavy-duty, would be great for some decent amps.

I dig out my non-precision weight and compared them together:





There is cheapie brass plated terminal with 35.5 gramms as well... Golden spade lug with 14 grams is actually local Taiwan shop one, but it's real copper with gold plating.

Also I will have one of PCBs for LTZ assembled with this nice beryllium copper gold plated TE AUGAT socket. It clamps pins really tight. Will be a good test to see if quality sockets make any difference on stability or tempco ;)



Have few others as well..



And last , but not least - L&N 12-position rotary switch, just similar to ones used in my recently received 4-decade resistor box.



All contacts silver plated, with just 1 mOhm contact resistance and 1pF of capacitance. Will be a great switch for muxing reference outputs.

So thanks for Santa, feeling like a child which just got nice beefy candy.
« Last Edit: December 22, 2015, 04:04:47 AM by TiN »
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 1332
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #235 on: December 22, 2015, 04:12:29 AM »

Also I will have one of PCBs for LTZ assembled with this nice beryllium copper gold plated TE AUGAT socket. It clamps pins really tight. Will be a good test to see if quality sockets make any difference on stability or tempco ;)


This will give thermocouples with your tin-plated LTZs.

Sockets are generally no good idea.
On a LM399 I had around 4-6 ppm difference depending on how the reference was placed into the socket.

With best regards

Andreas
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #236 on: December 22, 2015, 04:18:34 AM »
That's what I think, but it would be interesting to get some measurements. Also can be fun to try mitigating thermocouples with double-socketing (LTZ tin pin -> top gold-socket -> bottom gold-socket -> tin plated socket pin -> PCB), perhaps it would cancel out a bit.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 2580
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #237 on: December 22, 2015, 05:30:06 AM »
The coating of the sockets/pins is not that important for thermal EMF. The inportant part is the material where most of the temperaturegradient is and where the lowest resitance is. There just is not much temperaturedifference in the µm gold layer.

The difficult part could be that the thermal contakt is not that predictabel / stable. Thus the temperature drop in the kovar wires may change over time. So if you really want to go with a socket, keep the wires long to keep the termal current low. 
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #238 on: December 27, 2015, 07:39:00 PM »
Since I got one of my original LTZ modules running today, I thought it would be interesting to let you all know long-term stability.

11 month ago I had all my LTZs running for 2 months, without any magic burn-in procedures or anything fancy.

Here's reminder for data from back then:



After that references were hanging around in homelab, migrating from one storage box to another. One was sold to zucca, one got it's new home in HP 3245A.

Module I powered on again is one from CH3 data, it is:
Rev A01 PCB, LTZ1000ACH, LTC2057, 13K/1K heater - 7.1367872 VDC(median) reference

Now very same CH3 module being measured by K2001, reading 7.1367822 VDC (median measurement over 6 hours + 0.00005 V offset (i added it to match graph on 3458A scales) ) Te

That gives us 7.1367822 / 7.1367872 = 0.7 ppm. To be exactly fair, it's too good to be true, as there are many other variables involved, but the fact of value matching so close after almost a year pretty neat for me, with all uncertainty involved.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2043
  • Country: us
Re: Project KX : DIY calibrator / reference source/meter
« Reply #239 on: December 27, 2015, 08:19:51 PM »
I'm curious, how much has the MV106 drifted in that time?
If you own any North Hills Electronics gear, message me.
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #240 on: December 27, 2015, 09:25:56 PM »
MV106 on contrary was on 24/7.
Median of 100 readings now : 7.1304862 V

Compared to recorded median 7.1305224 V (same 7.13050 knob dial setting), it's 5.06 ppm lower.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 1332
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #241 on: December 28, 2015, 09:20:30 PM »
Hello Illya,

got something which has a long journey together with my Keithley 2000 from Magnus this year.
now I found the nearly perfect mating connectors for this PCB.
MultiContact SA403

Available e.g. here (in Germany).

https://www.conrad.de/de/messerleiste-stecker-gerade-stift-4-mm-messing-multicontact-sa403-1-st-730880.html

with best regards

Andreas


 

Offline branadic

  • Frequent Contributor
  • **
  • Posts: 545
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #242 on: December 28, 2015, 09:33:25 PM »
Andreas,

I could have told you. I've made an equal board together with a DIN connector a few weeks ago for an 1/10 PT100 temperature sensor
There are types with gold plating available at RS. Sad is, they neither come with a screw nut nor with a washer.

branadic

Prema 5000 | Prema 5017 SC | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Keithley 181, Prema 5017 SC, 34401A, 34410A, HDO6054, Keysight 53230A and other goodies at work
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #243 on: December 28, 2015, 11:47:59 PM »
Those PCBs looks familiar to me...
Pity it's brass :) But neat types they have there, never thought about applications like conductive frames they show  :)
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline branadic

  • Frequent Contributor
  • **
  • Posts: 545
  • Country: de
Re: Project KX : DIY calibrator / reference source/meter
« Reply #244 on: December 29, 2015, 12:48:30 AM »
I have brought this one from RS. This ones are gold plated brass.
Prema 5000 | Prema 5017 SC | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Keithley 181, Prema 5017 SC, 34401A, 34410A, HDO6054, Keysight 53230A and other goodies at work
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #245 on: February 09, 2016, 04:26:09 PM »
Since 3458A is in business now, I got bit more time for other stuff.
Today and tomorrow we will look into my LTZ1000A module temperature stability.

To do so I will use simple DIY temperature chamber, based on next setup:

* Metal cast box with LTZ1000A KX module inside. This one is LTC2057, 13K/1K heater, LTZ1000A (24 week 2014).
* noname TEC module 127060-40 is fixed on bottom of metal box. TEC specs: 55.9W, 12V 6A, 127 couples
* Metal box with reference is located in styrofoam box with cover.
* Aluminum heatsink is on TEC's hot side, with little 60x60mm 5V 60mA fan. Heatsink and fan are protruding on matching cutout in styrofoam box at bottom side.
* Whole assembly standing on pair of heatsinks from Rb units for free airflow.

Power:

* LTZ module powered by 12V 4.5Ah SLA battery
* LTZ module onboard temperature sensor MAX6610 powered by 4xAA NiMH batteries
* TEC powered by Keithley 2510 TEC SMU
* Fan powered by Keithley 2400 SMU

Monitoring:

* 3458A sampling LTZ direct output, NPLC200, AZERO ON. Shielded LAN cable is used between LTZ and DMM. Shield connected to LTZ's GND and DMMs Guard terminal.
* Keithley 2002 sampling MAX6610's temperature sensor output and calculating PCB temperature. Sensor is located on bottom side 5mm away from LTZ chip.
* Keithley 2510 is set to +24.000°C , with PID settings: 110, 0.07, 1.5
* All instruments are talking to Raspberry Pi via NI GPIB-USB-HS and this Python app
* Pi uploads CSV data onto my FTP every 10 minutes
* D3.js plots live graph on dedicated web-page

Photos of setup:





Pi lives in ABS box together with Trimble GPSDO module:



Graph for now:



RAW CSV data

I'll have this bodge run for at least 8 hours, and then I'll set 2510 to next temperature step.
Eventually we should get multiple steps in LTZ's voltage , which can allow to calculate module tempco. Since LTZ is ovenized, it would be indication of tempco of resistors/PCB/opamp impact on reference output.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Theboel

  • Regular Contributor
  • *
  • Posts: 158
  • Country: id
Re: Project KX : DIY calibrator / reference source/meter
« Reply #246 on: February 09, 2016, 11:03:12 PM »
Hi Tin,
Are You use Your Rubidium osc for support the box ?
could You find something more .................. 8)
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #247 on: February 16, 2016, 04:11:56 AM »
I think some resistor-nuts might hate me for some recent sniping.. Got some bits and pieces from Fluke 720A. Make me wish having whole thing as well.
Beware precision-nuts  :rant:!

Eh, anyway...

More photos and some measurements.





 :popcorn:
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2043
  • Country: us
Re: Project KX : DIY calibrator / reference source/meter
« Reply #248 on: February 16, 2016, 06:55:50 PM »
Nice score with those resistors. You are correct with the N/P tempco markings.
If you own any North Hills Electronics gear, message me.
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 2437
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Project KX : DIY calibrator / reference source/meter
« Reply #249 on: February 23, 2016, 05:31:56 AM »
Ok, last overkill bits for power supply arrived.

Transformers from Fluke 332D/335D will do just fine, I bet.





These beasts are heavy. Think I need upgrade my chassis, as it may not handle these beasts safe enough. High-voltage one will be used for 1kV final output stage.

Resistors..



Binding posts...



Now I'm set on primary and power things, so let's finally build something, shall we?
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf