Author Topic: T.C. measurements on precision resistors  (Read 396327 times)

0 Members and 3 Guests are viewing this topic.

Offline miro123

  • Regular Contributor
  • *
  • Posts: 202
  • Country: nl
Re: T.C. measurements on precision resistors
« Reply #1125 on: February 05, 2022, 07:14:47 pm »
Thanks for your answer Branadic,
I trying to understand what is the problem that solve such circuit in modern day.
As already mentioned I'm beginner, I dont see that is tha problem with initial wirewond resistors with well defined alpha and beta TC

I try to keep to Einstein moto - "Everything Should Be Made as Simple as Possible, But Not Simpler"

That is whay I asking what is the initial problem
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1126 on: February 05, 2022, 07:48:05 pm »
Nummeric correction is possible, but it still need an extra measurement of the temperature and recording the data together with the other reading. It is possible, but is extra effort for the program and may need an extra scanner channel.

Having a very low TC if just more convenient, though it may add a little more drift.

A 3rd option would be a small oven for the resistor, e.g. to keep the resistor at some 30 C if needed.
 

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2018
  • Country: br
    • CADT Homepage
Re: T.C. measurements on precision resistors
« Reply #1127 on: February 06, 2022, 09:26:32 am »
In Juli 2020 i selected wirewound resistors for LTFLU reference gain stages and ordered several lots of DALE RS-02B wirewound resistors from different sources. Those resistors are well made and very stable when used at low power. They exhibit TCs of about +/- 4 ppm/K.
First diagram show various 330R resistors from three lots. They are similar within the same lot. I'd guess resistors in one lot behave similar.as they were made from the same spool of wire. I also combined selected 150R with 180R resistors to get "zero TC" 330R resistors. Anyway - in our LTFLU references they run inside the oven at near constant temperature.
Second diagram shows one lot of 750 Ohm resistors. These ones had their zero TC point within the test temperature range. The yellow one has less than 0.2 ppm/K for 20 to 37 °C. To find these i also bought several lots.

Regards, Dieter
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1128 on: February 06, 2022, 09:59:39 pm »
Attached is a very first TCR measurement.

Unfortunately, a GPIB cable nearby the shielded 4 wire PTFE cable caused some intereference as I could already rule out (see the very end of the dataset with the cable removed), thus the larger "noise".   :wtf:

Nevertheless, I can already see the important part, I still have a little too much (series-) copper, overcompensating the linear negative component. So the resistor needs some more tweaking.

Something more can be extracted: The 0.1K/min slope of the TEC controller is too large and should be decreased, but as that was the minimum the TEC controller can do on its own I used it as is to not having to go the software way of implementing the control loop. Something for the future though.

But I'm getting there and the improvement is already imaginable.

Edit: Repeated the measurement with the GPIB cable rearranged and the interference has vanished.

-branadic-
« Last Edit: February 19, 2022, 06:24:20 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: TiN, Mickle T., Dr. Frank, Andreas, HighVoltage, SilverSolder, bsw_m, ch_scr

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1129 on: February 19, 2022, 09:53:51 pm »
I meanwhile trimmed the copper to the final length, which took multiple runs in the temperature chamber.  The final amount of copper differed quite a bit from the design phase.
I'm about to finish my resistor project, already did the trim of the series resistor to the final value (changed the 2x 2.7Ω in parallel to 3.3Ω||5.6Ω) and will verify the result tomorrow with a final temperature run.

-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: Mickle T., Dr. Frank, Andreas, doktor pyta, ch_scr

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1130 on: February 20, 2022, 07:50:28 am »
Hmm,

on temperature rising the resistance follows the temperature.
on temperature falling there is always some "overshoot" until it stabilizes.

Where does this overshoot come from?
- thermal EMF?
- condensing humidity?

The 3 extra temperature + humidity sensors are environment values outside the chamber I guess.

with best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1131 on: February 20, 2022, 08:30:04 am »
Hi Andreas,

I was thinking about that too and my best guess is that there is a small time lag between the copper on the outside of the brass tube of the hermetically sealed resistor compensating the alpha term and the wirewound resistor on the inside of the tube causing that peaks.

Exact, the 3 extra temperature sensors and the humidity and pressure sensor are monitoring the environment of the lab.

-branadic-
« Last Edit: February 20, 2022, 09:11:41 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 dietert1

  • Super Contributor
  • ***
  • Posts: 2018
  • Country: br
    • CADT Homepage
Re: T.C. measurements on precision resistors
« Reply #1132 on: February 20, 2022, 10:00:10 am »
There seems to be another slow process, since resistance drifts about 0.5 ppm during the stationary temperature states (about 2 hours, "90 cm" diagrams). Would be interesting to see whether/when that drift stops. I have seen Vishay hermetic resistors drift for days after a temperature step. Also i have seen long lasting drift when using epoxy, i mean a year or so.

Regards, Dieter
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1133 on: February 21, 2022, 06:34:39 am »
I guess what seems to be visible as drift is the thermal time lag between copper and resistor. Attached is the final temperature profile, followed by an 8 h stability measurement at lab temperature. As can be seen it takes about 5.5 ... 6 h for the whole thing to equalize.
This time could be reduced by improving the thermal transfer inside the case. I was thinking about filling the case with ceramic balls (3-4 mm in diameter) or ceramic pouder, thermally conductive but electrically insulating.

Edit: Added the initial thermal resistor behavior for comparison.

-branadic-
« Last Edit: February 21, 2022, 06:14:50 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 miro123

  • Regular Contributor
  • *
  • Posts: 202
  • Country: nl
Re: T.C. measurements on precision resistors
« Reply #1134 on: February 21, 2022, 05:30:57 pm »

This time could be reduced by improving the thermal transfer inside the case. I was thinking about filling the case with ceramic balls (3-4 mm in diameter) or ceramic pouder, thermally conductive but electrically insulating.
How ceramics pouder or balls react to moisture? How long will take to go the moisture away?
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: T.C. measurements on precision resistors
« Reply #1135 on: February 21, 2022, 06:42:21 pm »
Perhaps glazed ceramic balls are available in that size?  Unglazed ceramic certainly adsorbs moisture readily on its surface.
Back in grad school, we purchased synthetic sapphire (Al2O3) balls with absurdly tight specs on diameter and sphericity, and they were a beautiful ruby red, but probably too expensive for this application.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1136 on: February 22, 2022, 08:37:10 pm »
In retrospective I come to the conclusion that the beta is overcompensated, so the series resistor of the NTC needs to be slightly reduced. The heatshrink over these series resistors and the NTC adds an extra time lag that should be avoided. So the heatshrink better has covered the solder joints only, not the complete components.

So if I find the time I will rework the build and hopefully get better results out of it. Just for the fun of it and to see what can be reached at the extrems.

-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 TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: T.C. measurements on precision resistors
« Reply #1137 on: February 23, 2022, 01:28:26 am »
Why thermal lag should be avoided? Its a good thing , no? Unless we have different understanding to thermal lag (change of standards output versus ambient T change?).
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1138 on: February 23, 2022, 07:16:41 am »
In principle I agree, a thermal lag between the ambient and the device inside your case smoothes out fast temperature changes. But you don't want a thermal lag between the device inside the case and the accociated temperature compensating components of it. Instead you want all these components at the same temperature. Otherwise you create time constants, that end up in weird behavior, that manifast in this beautiful butterfly curve I've presented. Is it more clear now what I was pointing to?

-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: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #1139 on: February 13, 2023, 08:55:53 pm »
I finallized and measured my 10 kΩ resistance standard based on two selected AlphaElectronics HC 20 kΩ resistors, that are sitting inside a Rolec case with a 10 kΩ NTC temperature sensor and low thermal binding posts. The result is quite usable and even better than L&N 4040B in terms of t.c..

-branadic-
« Last Edit: February 13, 2023, 09:48:39 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: Dr. Frank, Andreas, splin, doktor pyta, alm, miro123

Online guenthert

  • Frequent Contributor
  • **
  • Posts: 706
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1140 on: March 16, 2023, 09:35:01 pm »
Not sure, whether it belongs here, but for contrast/comparison I offer my characterization of a Leeds & Northrup 4030-B resistor.   I have it for a while, but only now got around to hack a small oven (which doesn't perform all that well, but sufficient for this purpose).  Unfortunately I don't have access to my Datron 1271 at this time; my HP 3457A will have to do for now.

gnuplot fit finds:
After 61 iterations the fit converged.
final sum of squares of residuals : 1.12195e-05
rel. change during last iteration : -6.15703e-12

degrees of freedom    (FIT_NDF)                        : 1497
rms of residuals      (FIT_STDFIT) = sqrt(WSSR/ndf)    : 8.65717e-05
variance of residuals (reduced chisquare) = WSSR/ndf   : 7.49467e-09

Final set of parameters            Asymptotic Standard Error
=======================            ==========================

a               = 4.87296e-06      +/- 1.389e-08    (0.285%)
b               = -6.552e-07       +/- 1.932e-09    (0.2949%)
R23             = 100.027          +/- 3.034e-06    (3.033e-06%)


correlation matrix of the fit parameters:

               a      b      R23   
a               1.000
b              -0.894  1.000
R23            -0.147 -0.164  1.000




There's some gain error visible in the discrepancy of the temperature measurements with the NTC and the PT1000, neither of which is calibrated though (someday I'll get a gallium cell ...).  Here I just wanted to know, at which temperature the TC is least.  This seems to be ~ 26°C (23 + a / -2b ~= 26.7), at which both sensors have incidentally a good agreement.
« Last Edit: March 17, 2023, 01:04:03 am by guenthert »
 
The following users thanked this post: Andreas, splin, alm

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1141 on: March 17, 2023, 05:11:21 am »
Hello,

it would be interesting to see also the rising temperature.

This would show wether the resistor has some hysteresis (or if the temperature sensors have not the right temperature = position or a time lag).

with best regards

Andreas
 

Online guenthert

  • Frequent Contributor
  • **
  • Posts: 706
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1142 on: March 17, 2023, 05:10:04 pm »
Hello,

it would be interesting to see also the rising temperature.

This would show wether the resistor has some hysteresis (or if the temperature sensors have not the right temperature = position or a time lag).

with best regards

Andreas

Well, there is the attached image.  However there was a programming error in the driving script causing a shorter then desired setting time, i.e. even greater slope in the 'plateau'.  Also, due to the greater than anticipated spread of measurement results, an insufficient number of measurements were taken (unfit for 'fit' ;-} .  I might redo it someday.

I looked back in my records and the resistance over time remained stable to within some 10 ppm for the last three years (but perhaps the resistor and the 3457A drifted in the same direction?).

If there is hysteresis, I'm afraid direct measurement using the 3457A won't show it.  Perhaps using some clever bridge configuration with one arm at constant temperature (like the pros do it [1]) would do.

(disregard the bump in ambient temperature in the middle of the run: my laptop was running then and blowing at the sensor)

[1] https://www.researchgate.net/publication/337693822_DETERMINATION_OF_THE_STANDARD_RESISTOR_TEMPERATURE_COEFFICIENTS_AND_THEIR_UNCERTAINTIES/fulltext/5de5c724299bf10bc33a7422/DETERMINATION-OF-THE-STANDARD-RESISTOR-TEMPERATURE-COEFFICIENTS-AND-THEIR-UNCERTAINTIES.pdf
« Last Edit: March 17, 2023, 05:28:05 pm by guenthert »
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1143 on: March 17, 2023, 06:40:14 pm »
Mhm,

if you now put the temperature on the X-Axis and the ppm on the y-Axis there should be a large difference at the temperature extremes.
32 degrees -5 vs -10 ppm
21 degrees -17 vs -8 ppm

if I see it right.

The 26 degrees remain stable at +10 ppm
with best regards

Andreas
 

Online guenthert

  • Frequent Contributor
  • **
  • Posts: 706
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #1144 on: March 17, 2023, 07:35:02 pm »
Mhm,

if you now put the temperature on the X-Axis and the ppm on the y-Axis there should be a large difference at the temperature extremes.
32 degrees -5 vs -10 ppm
21 degrees -17 vs -8 ppm

if I see it right.

The 26 degrees remain stable at +10 ppm
with best regards

Andreas

That doesn't seem right. l100.log,2 ends (~32.6°C) with
1678890905: 100.02594300000099Ohm @305.73
1678890914: 100.025840499997Ohm @305.73
1678890923: 100.02601549999899Ohm @305.73
1678890932: 100.02580287500001Ohm @305.7
1678890942: 100.025825Ohm @305.72
1678890951: 100.02595Ohm @305.73
1678890960: 100.02595749999999Ohm @305.71
1678890969: 100.025979000001Ohm @305.71
1678890978: 100.02579300000001Ohm @305.72
1678890987: 100.025840499997Ohm @305.72


while l100.log starts (~32.25°C) with
1678895116: 100.026115499999Ohm @305.38
1678895139: 100.026165Ohm @305.38
1678895154: 100.02606Ohm @305.38
1678895169: 100.0260525Ohm @305.37
1678895184: 100.02602875Ohm @305.38
1678895199: 100.025965Ohm @305.38
1678895214: 100.026074000001Ohm @305.38
1678895229: 100.025937999997Ohm @305.38
1678895244: 100.026075250001Ohm @305.38
1678895259: 100.02602999999999Ohm @305.4
1678895274: 100.0260275Ohm @305.37
1678895289: 100.025902875Ohm @305.39
1678895304: 100.025984Ohm @305.42
1678895319: 100.025979000001Ohm @305.4
1678895334: 100.026002875Ohm @305.43
1678895349: 100.0259225Ohm @305.44
1678895364: 100.02600799999999Ohm @305.44
1678895379: 100.02602375Ohm @305.44
1678895394: 100.026Ohm @305.46
1678895409: 100.02589898437499Ohm @305.45
1678895424: 100.026040499997Ohm @305.46
1678895439: 100.025989625Ohm @305.46
1678895454: 100.02611175Ohm @305.46
1678895469: 100.025955Ohm @305.45
1678895484: 100.0260058125Ohm @305.43
1678895499: 100.02597649999899Ohm @305.46
1678895514: 100.026043000001Ohm @305.45
1678895529: 100.025869000001Ohm @305.47
1678895544: 100.02590196874999Ohm @305.45
1678895559: 100.025955Ohm @305.45
1678895574: 100.026040499997Ohm @305.47
1678895589: 100.025948000001Ohm @305.44


l100.log,2 starts with (~21.3°C):
1678809320: 100.026325Ohm @294.52
1678809350: 100.02639087499999Ohm @294.53
1678809380: 100.026377749999Ohm @294.52
1678809410: 100.0263815Ohm @294.52
1678809440: 100.02634800000101Ohm @294.53
1678809470: 100.026340499997Ohm @294.53
1678809499: 100.02623799999701Ohm @294.52
1678809530: 100.026298984375Ohm @294.52
1678809560: 100.026369000001Ohm @294.52
1678809589: 100.02632875Ohm @294.52
1678812920: 100.0261Ohm @294.44


l100.log ends at a lower temperature, but has around 21.2°C:
1678974324: 100.02589699999999Ohm @294.32
1678974339: 100.02591749999999Ohm @294.29
1678974354: 100.02580987500001Ohm @294.35
1678974369: 100.0259979375Ohm @294.35
1678974384: 100.02589393749999Ohm @294.36
1678974399: 100.02585749999999Ohm @294.36
1678974414: 100.02600196875Ohm @294.34
1678974429: 100.02583799999701Ohm @294.36
1678974444: 100.0260009765625Ohm @294.35
1678974459: 100.0258958125Ohm @294.34
1678974474: 100.026077749999Ohm @294.35
1678974489: 100.026008625Ohm @294.35


Given the spread (and the fact that the temperature is recorded by an Arduino reading a voltage divider using a NTC and a run-of-the-mill metal film resistor), I'd be hesitant to read a hysteresis into that.  More importantly perhaps, while the TC of the 3457A in the comfort zone might be small, it ain't 0.  I'd think if one wants to measure individual ppms, one needs a better set-up (e.g. aforementioned temperature-stabilized comparison resistor).

I use the graphs only for illustration, not in an attempt to extract data as I keep the data files around (wished hardware manufacturers would do the same).  It's just easier to share the graphs, as the data files are on a different host (and the eevblog server doesn't like .log files  ::) .

« Last Edit: March 17, 2023, 08:44:27 pm by guenthert »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf