T.C. measurements on precision resistors

[Andreas]

  - 2 pieces of Teflon tubes where the resistor leads are inserted

Hello,

from where did you get this?

With best regards

Andreas

[mimmus78]

  - 2 pieces of Teflon tubes where the resistor leads are inserted

Hello,

from where did you get this?

With best regards

Andreas

I get from eBay. They are used somewhere in 3D printers. If you search for Teflon tubes you should find it.

Inviato dal mio Nexus 6P utilizzando Tapatalk

[mimmus78]

STRANGE RESISTOR

Well on the quest for finding a 1K resistor with +2ppm/°K TC for my LTZ1000 reference I found this one?

This resistor has quite good TC measured at +0.3ppm/°K circa and it has gone a dozen of thermal cycles on it in the past few days. If you look at fist cycle of today it almost recovered all resistance variation when it was at 65°C (30°C to 65°C was only 2ppm apart).

On this UPW50 resistors a recovery after a positive thermal shocks always cause the resistor to drift down if TC is positive or drift up if TC is negative. I think this drift is caused by temperature stress on the resistor ... can be this drift used to evaluate stability of the resistor?

PS: first hour on chart it seems the resistor has negative TC but this is just the meter heating up.

[mimmus78]

Another UPW50 resistor, this 120 ohm give me more consistent result.
Still you can see a 17 ppm change from fist and last reading at 30°C, but curve is less pronuced.
At least it seems drift/hysteresis is less evident after fist thermal cycle.

This TC includes also TC of the two leads (8 mOhm total) that was calculated to influence total TC with +5% circa.

   step:temp 000:25 - avg.: 120.059312 ohm (1473  samples)
   step:temp 001:30 - avg.: 120.062963 ohm (1414  samples) total +30.40 ppm - +6.08 ppm/K°
   step:temp 002:65 - avg.: 120.083843 ohm (729  samples) total +173.91 ppm - +4.97 ppm/K°
   step:temp 003:30 - avg.: 120.062053 ohm (1437  samples) total -181.45 ppm - +5.18 ppm/K°
   step:temp 004:65 - avg.: 120.083235 ohm (1193  samples) total +176.43 ppm - +5.04 ppm/K°
   step:temp 005:30 - avg.: 120.061325 ohm (1361  samples) total -182.46 ppm - +5.21 ppm/K°
   step:temp 006:65 - avg.: 120.082729 ohm (684  samples) total +178.28 ppm - +5.09 ppm/K°
   step:temp 007:30 - avg.: 120.061072 ohm (1393  samples) total -180.35 ppm - +5.15 ppm/K°
   step:temp 008:65 - avg.: 120.082460 ohm (758  samples) total +178.14 ppm - +5.09 ppm/K°
   step:temp 009:30 - avg.: 120.060839 ohm (1135  samples) total -180.05 ppm - +5.14 ppm/K°

[Andreas]

When I see this:

how is the drift of your K2001 during measurement?
What happens if you leave the resistor temperature slightly over room temperature?

Just to distinguish if the drift comes from some creeping epoxy or from some warm-up effects on the K2001.
I also think you would get different results if you would cycle temperature around room temperature.

With best regards

Andreas

[mimmus78]

Well I tested 10k range and it was stable ...  but I didn't tested yet every range. Resistors generally recover most of the drift after one day resting at ambient temperature.

Warm up effects? I don't think so. During week end I leave meter turned on and the room has AC to keep temperature in 1°C. Resistors measured during week end shows the same behaviour.

I measured one of this Epoxy/not hermetically sealed 10K Vishay metal foil 0.005 percent resistor and I have seen no histeresys there. Only the TC curve was a little bit wonky but it's seems normal with metal foils resistors.

More investigation needed. I will receive some resistors from Edwin in a few days. I'm very curious to see what I will find.

Inviato dal mio Nexus 6P utilizzando Tapatalk

[mimmus78]

Just finished the software for automatic TC ... bad me now I have to port it form K2001 to 3458a.



Inviato dal mio Nexus 6P utilizzando Tapatalk

[PartialDischarge]

wow, this is a hell of a thread and I just discovered it today, time to start reading

[mimmus78]

An this is another UPW50 measured with a 3458a ... drifting effects after power cycle is still there and are identical to what measured with K2001.
So it's almost impossible that this effect is due to meter heating or drifting.

[acbern]

It would be interesting if you repeated this with less of a cycle with the exact same device, say only +/-15K. This way one could see if the hysteresis is linear with delta T or if there is some acceleration at higher delta T.

[Andreas]

Hello,

when I compare this with my own UPW50 measurements the behaviour is quite different:
https://www.eevblog.com/forum/metrology/t-c-measurements-on-precision-resistors/msg462300/#msg462300

I usually have a "lag" between temperature change and resistance. This gives the typical hysteresis opening around the middle of the temperature extremes. Over night the resistance returned around to the middle of the opening.

On the other side I never had temperatures above 45 deg C.

With best regards

Andreas

[mimmus78]

So this is another experiment with this resistor.
This experiment was done 24h after the previous. In this first experiment the resistor has drifted 12ppm down. After 24h of resting it has recovered 11ppm returning back almost at the same point of start.

The second experiment consists in three thermal cycles with low temperature delta, 20°K from t.amb and two with 40°C of delta.

Results are with first three thermal cycle, the ones with limited temperature delta there was only a drift down on 1ppm in first cycles.
With the high temp cycle the drift was perfectly the same as with the previous day.

EDIT:
I just noticed the temperature of the room changed a little bit during the night ... even if the measurements are consistent this test must be repeated.

READINGS:

  step 00 - temp. 25 °C - avg. res. 10005.033472 ohm (0122 samples)
  step 01 - temp. 30 °C - avg. res. 10005.236821 ohm (0098 samples) total +20.32 ppm - +4.06 ppm/K°
  step 02 - temp. 40 °C - avg. res. 10005.597550 ohm (0100 samples) total +36.05 ppm - +3.61 ppm/K°
  step 03 - temp. 30 °C - avg. res. 10005.223934 ohm (0084 samples) total -37.34 ppm - +3.73 ppm/K°
  step 04 - temp. 40 °C - avg. res. 10005.601928 ohm (0065 samples) total +37.78 ppm - +3.78 ppm/K°
  step 05 - temp. 30 °C - avg. res. 10005.225218 ohm (0080 samples) total -37.65 ppm - +3.76 ppm/K°
  step 06 - temp. 40 °C - avg. res. 10005.602340 ohm (0110 samples) total +37.69 ppm - +3.77 ppm/K°
  step 07 - temp. 30 °C - avg. res. 10005.225215 ohm (0036 samples) total -37.69 ppm - +3.77 ppm/K°
  step 08 - temp. 65 °C - avg. res. 10006.284930 ohm (0081 samples) total +105.92 ppm - +3.03 ppm/K°
  step 09 - temp. 30 °C - avg. res. 10005.190542 ohm (0026 samples) total -109.37 ppm - +3.12 ppm/K°
  step 10 - temp. 65 °C - avg. res. 10006.281207 ohm (0036 samples) total +109.01 ppm - +3.11 ppm/K°
  step 11 - temp. 30 °C - avg. res. 10005.181792 ohm (0043 samples) total -109.87 ppm - +3.14 ppm/K°

[mimmus78]

when I compare this with my own UPW50 measurements the behaviour is quite different:
I usually have a "lag" between temperature change and resistance. This gives the typical hysteresis opening around the middle of the temperature extremes. Over night the resistance returned around to the middle of the opening.
On the other side I never had temperatures above 45 deg C.

This was partially confirmed with my last experiment. At 40° drift was of just 1ppm. When temperature was raised at 65° I had the other 4ppm.
Anyway just noticed the temperature of the room changed a little bit during the night ... I will repeat this test one more time.

[d-smes]

When temperature is stepped to the higher set-point, there is a downward drift in resistance.  This is especially noticeable when stepped to 65 deg C.   What would happen if resistance was allowed to stabilize at 65C before the next step down in temperature?

[mimmus78]

What would happen if resistance was allowed to stabilise at 65C before the next step down in temperature?
As far as i understood the most of the drift happens in the first 30 minutes at high temperature ... sub sequential thermal cycle cause drift but this became less evident cycle by cycle.

Need to do another experiment with long resting times. Something like this should be appropriate:

• 30 minutes at 25°
• 30 minutes at 30°
• 6 hours at 65°
• 12 hours at 30°

Up to now I have measured this things:

 1. when the temperature is stabilised after a temperature change there is always some drift
 2. this "drift" is proportional the the dt
 3. drift is negative when temperature dt is positive (and viceversa)
 4. when temperature goes back to "ambient temperature" this drift is still there and measurable
 5. resistor with lower TC generally drift lesser than the ones with higher TC
 6. most of the drift is recovered (or it drift back) after many hours of resting at "ambient temperature"
 7. this drift effect reduces much with many therm cycles

Points 6 and 7 are known facts and observable in data but I still need to confirm with appropriate experiment that is not a multimeter drift or some changes in the measurement methods I made in last days (changed leads and probes from cheap mini grabber to Pomona golden plated ones).

[Andreas]

Hello,

I think it is better to solder the connections instead of using grabbers.

with best regards

Andreas

[EmmanuelFaure]

*ugly grabbers

[mimmus78]

Hello,
I think it is better to solder the connections instead of using grabbers.
with best regards
Andreas

OK I can give it a go ... may I solder 4 copper wires to the leads and than grab in the outside box where temperature is more stable?


Just realised in 3458A you can screw wires directly to the banana plug ... still have to get accustomed to it.

[mimmus78]

Another day of 3458a logging resistor.

As by Andreas suggestions this time I used copper wire (from ethernet cable) soldered to the end of resistor leads. I also used a clamp to reduce overheat of the resistor.

Measured resistance of the DUT in free air before to start tests. Temperate 21°C +/- 2°C circa (no way to measure it).

   - DUT R with pomona leads 12K00596
   - DUT R directly to binding posts 12K00592 (observed heating/cooling process when in the binding posts)
   - DUT R with soldered ethernet copper leads 12K00597

I also measured the 10K resistor inside my MF calibrator before to start test and at the end of the test to be sure multimeter doesn't drift.

   - start of test: 10.000061 ohm
   - end of test: 10.000059 ohm

At the end of cycle I added also a log period at 25°C (steps 10 and above). Every step has same temperature and lasts 15 minutes, by this I can get resistance value averaged every 15 minutes. If you really force a bit readings you can notice a slight drift down because 25°C is still higher than normal temperature.

I give it also a run 30/65/30 without offset compensation just to check if anything changes but also this tests give me the same results.

Here are all data and charts.

READINGS - 4W - SOLDERED COPPER ETH LEADS - OFFSET COMP ON (DELAY 0.5s):
===================================================================================================

  step 00 - temp. 25 °C - avg. res. 12005.837343 ohm (0026 samples)
  step 01 - temp. 30 °C - avg. res. 12005.656371 ohm (0088 samples) total -15.07 ppm - -3.01 ppm/K°
  step 02 - temp. 40 °C - avg. res. 12005.198936 ohm (0101 samples) total -38.10 ppm - -3.81 ppm/K°
  step 03 - temp. 30 °C - avg. res. 12005.612139 ohm (0054 samples) total +34.42 ppm - -3.44 ppm/K°
  step 04 - temp. 40 °C - avg. res. 12005.195729 ohm (0102 samples) total -34.68 ppm - -3.47 ppm/K°
  step 05 - temp. 30 °C - avg. res. 12005.607506 ohm (0042 samples) total +34.30 ppm - -3.43 ppm/K°
  step 06 - temp. 65 °C - avg. res. 12003.806769 ohm (0076 samples) total -149.99 ppm - -4.29 ppm/K°
  step 07 - temp. 30 °C - avg. res. 12005.559735 ohm (0076 samples) total +146.03 ppm - -4.17 ppm/K°
  step 08 - temp. 65 °C - avg. res. 12003.794508 ohm (0065 samples) total -147.03 ppm - -4.20 ppm/K°
  step 09 - temp. 30 °C - avg. res. 12005.531690 ohm (0026 samples) total +144.72 ppm - -4.13 ppm/K°
  step 10 - temp. 25 °C - avg. res. 12005.740309 ohm (0018 samples) total +17.38 ppm - -3.48 ppm/K°
  step 11 - temp. 25 °C - avg. res. 12005.743185 ohm (0056 samples) total +0.24 ppm - +0.24 ppm/K°
  step 12 - temp. 25 °C - avg. res. 12005.742822 ohm (0051 samples) total -0.03 ppm - -0.03 ppm/K°
  step 13 - temp. 25 °C - avg. res. 12005.742047 ohm (0051 samples) total -0.06 ppm - -0.06 ppm/K°
  step 14 - temp. 25 °C - avg. res. 12005.743028 ohm (0050 samples) total +0.08 ppm - +0.08 ppm/K°
  step 15 - temp. 25 °C - avg. res. 12005.742188 ohm (0051 samples) total -0.07 ppm - -0.07 ppm/K°
  step 16 - temp. 25 °C - avg. res. 12005.742950 ohm (0051 samples) total +0.06 ppm - +0.06 ppm/K°
  step 17 - temp. 25 °C - avg. res. 12005.741966 ohm (0049 samples) total -0.08 ppm - -0.08 ppm/K°
  step 18 - temp. 25 °C - avg. res. 12005.740597 ohm (0050 samples) total -0.11 ppm - -0.11 ppm/K°
  step 19 - temp. 25 °C - avg. res. 12005.740345 ohm (0051 samples) total -0.02 ppm - -0.02 ppm/K°
  step 20 - temp. 25 °C - avg. res. 12005.740984 ohm (0051 samples) total +0.05 ppm - +0.05 ppm/K°
  step 21 - temp. 25 °C - avg. res. 12005.740076 ohm (0050 samples) total -0.08 ppm - -0.08 ppm/K°
  step 22 - temp. 25 °C - avg. res. 12005.740010 ohm (0050 samples) total -0.01 ppm - -0.01 ppm/K°
  step 23 - temp. 25 °C - avg. res. 12005.738172 ohm (0051 samples) total -0.15 ppm - -0.15 ppm/K°
  step 24 - temp. 25 °C - avg. res. 12005.739758 ohm (0050 samples) total +0.13 ppm - +0.13 ppm/K°
  step 25 - temp. 25 °C - avg. res. 12005.739483 ohm (0050 samples) total -0.02 ppm - -0.02 ppm/K°
  step 26 - temp. 25 °C - avg. res. 12005.739246 ohm (0051 samples) total -0.02 ppm - -0.02 ppm/K°
  step 27 - temp. 25 °C - avg. res. 12005.739165 ohm (0051 samples) total -0.01 ppm - -0.01 ppm/K°
  step 28 - temp. 25 °C - avg. res. 12005.738020 ohm (0051 samples) total -0.10 ppm - -0.10 ppm/K°
  step 29 - temp. 25 °C - avg. res. 12005.737698 ohm (0050 samples) total -0.03 ppm - -0.03 ppm/K°
  step 30 - temp. 25 °C - avg. res. 12005.738319 ohm (0051 samples) total +0.05 ppm - +0.05 ppm/K°
  step 31 - temp. 25 °C - avg. res. 12005.738173 ohm (0051 samples) total -0.01 ppm - -0.01 ppm/K°
  step 32 - temp. 25 °C - avg. res. 12005.738530 ohm (0051 samples) total +0.03 ppm - +0.03 ppm/K°
  step 33 - temp. 25 °C - avg. res. 12005.738807 ohm (0050 samples) total +0.02 ppm - +0.02 ppm/K°
  step 34 - temp. 25 °C - avg. res. 12005.738572 ohm (0051 samples) total -0.02 ppm - -0.02 ppm/K°
  step 35 - temp. 25 °C - avg. res. 12005.737621 ohm (0051 samples) total -0.08 ppm - -0.08 ppm/K°
  step 36 - temp. 25 °C - avg. res. 12005.735601 ohm (0051 samples) total -0.17 ppm - -0.17 ppm/K°
  step 37 - temp. 25 °C - avg. res. 12005.715372 ohm (0067 samples) total -1.68 ppm - -1.68 ppm/K°

READINGS - 4W - SOLDERED COPPER ETH LEADS - OFFSET COMP OFF:
===================================================================================================

  step 05 - temp. 30 °C - avg. res. 12005.513999 ohm (0160 samples)
  step 06 - temp. 65 °C - avg. res. 12003.726170 ohm (0119 samples) total -148.92 ppm - -4.25 ppm/K°
  step 07 - temp. 30 °C - avg. res. 12005.495170 ohm (0093 samples) total +147.37 ppm - -4.21 ppm/K°

[Andreas]

Hello,

so I see no more overshoot with soldered wires.
Is that correct? (althought a different resistor value).

with best regards

Andreas

[mimmus78]

In reality I cannot say this 100% ... my idea is that using Pomona leads was enough to solve this
error. A couple of resistor when I was using Pomona leads also didn't showed overshoot.

But in last tests I also changed resistor from UPW50 to ULTROHM from Edwin. Maybe it is just
Edwin's ones that does not have this effect ...

I need to test a couple more resistor from Edwin to build first LTZ1000 reference and than I will
make another RUN with the most overshooting resistor in the UPW50 pool. Only this can confirm
if Pomona leads or soldered leads solved this problem or not.

This stuff really takes forever ...

[mimmus78]

Many resistors later ...

Now I can say Edwin resistors doesn't have this overshoot effect. I tested a total of 15 resistors from Edwin and none of them showed this strange overshoot.

As soon as I back to UPW50 I have this overshoot effect again. I'm now sure it's just resistors type and construction causing it. Also a Vishay foil resistor doesn't have this overshoot.

I speak with Edwin and he find strange this end cycle hysteresis. I will make some more tests in next days as  I'd like to find how many days it takes to recovery to initial value and try to isolate what is causing this.

Inviato dal mio Nexus 6P utilizzando Tapatalk

[Andreas]

Hello,

For me its not clear: what effect do you mean with "end cycle hysteresis"?

For some effects it would be also better to have the temperature as X-axis additional to a time diagram.
And a interesting question would be if the behaviour would be the same if you also had a
"lower than room temperature / 30 deg C" point in your measurement.

I personally blame a part of all hysteresis effects to the viscosity of the epoxy used to seal the resistors.
And another part on air humidity effects to the epoxy. (the epoxy is swelling when taking humidity from environment).
Air humidity is very tricky as the time constants for rH changes are usually long.
I have observed 3-7 days time constant for changes around 55% rH on a DIP8-package.
The viscosity seems to be more a effect "over night" or several hours.
On resistor measurements I have often restarting the measurement value somewhere in the "middle" of the hysteresis opening when repeating the measurement next day.

with best regards

Andreas

[mimmus78]

>> For me its not clear: what effect do you mean with "end cycle hysteresis"?

"end cycle hysteresis" is the resistance change measured after one or more thermal cycles.

If you get the following example you can see that at beginning (step 2) resistance was 20011.70 ohm circa.
After the temperature going up and down two times resistance was 20011.39 ohm circa (step 47).
So the resistor has lost 15ppm after thermal cycling.

Code: [Select]

|  # | °C | avg. resist. - sam. | abs ppm - dt  | ppm/°K - dt  |
| -- | -- | ------------------- | ------------- | ------------ |
| 00 | 25 |  20011.229886 ? 021 |               |              |
| 01 | 25 |  20011.236979 ? 019 | +000.35   +0  |              |
| 02 | 30 |  20011.703395 ? 022 | +023.31   +5  | +04.66    +5 |
| 03 | 30 |  20011.703168 ? 019 | -000.01   +0  | +04.66    +5 |
| 04 | 35 |  20012.125043 ? 021 | +021.08   +5  | +04.22    +5 |
| 05 | 35 |  20012.120768 ? 019 | -000.21   +0  | +04.17    +5 |
| 06 | 50 |  20013.388645 ? 022 | +063.36  +15  | +04.22   +15 |
| 07 | 50 |  20013.374047 ? 019 | -000.73   +0  | +04.18   +15 |
| 08 | 50 |  20013.363585 ? 020 | -000.52   +0  | +04.14   +15 |
| 09 | 30 |  20011.460479 ? 019 | -095.09  -20  | +04.75   -20 |
| 10 | 30 |  20011.463921 ? 019 | +000.17   +0  | +04.75   -20 |
| 11 | 30 |  20011.472355 ? 020 | +000.42   +0  | +04.72   -20 |
| 12 | 60 |  20014.381990 ? 021 | +145.40  +30  | +04.85   +30 |
| 13 | 60 |  20014.352014 ? 021 | -001.50   +0  | +04.80   +30 |
| 14 | 60 |  20014.339421 ? 019 | -000.63   +0  | +04.78   +30 |
| 15 | 30 |  20011.333014 ? 021 | -150.21  -30  | +05.01   -30 |
| 16 | 30 |  20011.346870 ? 020 | +000.69   +0  | +04.98   -30 |
| 17 | 30 |  20011.356837 ? 019 | +000.50   +0  | +04.97   -30 |
| 18 | 30 |  20011.361145 ? 020 | +000.22   +0  | +04.96   -30 |
| 19 | 30 |  20011.368689 ? 019 | +000.38   +0  | +04.95   -30 |
| 20 | 30 |  20011.375579 ? 019 | +000.34   +0  | +04.94   -30 |
| 21 | 30 |  20011.379270 ? 020 | +000.18   +0  | +04.93   -30 |
| 22 | 30 |  20011.384615 ? 020 | +000.27   +0  | +04.92   -30 |
| 23 | 30 |  20011.390463 ? 019 | +000.29   +0  | +04.91   -30 |
| 24 | 30 |  20011.391643 ? 021 | +000.06   +0  | +04.91   -30 |
| 25 | 30 |  20011.399790 ? 020 | +000.41   +0  | +04.90   -30 |
| 26 | 30 |  20011.401468 ? 019 | +000.08   +0  | +04.89   -30 |
| 27 | 30 |  20011.403340 ? 020 | +000.09   +0  | +04.89   -30 |
| 28 | 30 |  20011.406747 ? 019 | +000.17   +0  | +04.88   -30 |
| 29 | 30 |  20011.413715 ? 020 | +000.35   +0  | +04.87   -30 |
| 30 | 30 |  20011.406163 ? 019 | -000.38   +0  | +04.89   -30 |
| 31 | 30 |  20011.410265 ? 020 | +000.20   +0  | +04.88   -30 |
| 32 | 30 |  20011.412707 ? 014 | +000.12   +0  | +04.87   -30 |
| 33 | 30 |  20011.411674 ? 023 | -000.05   +0  | +04.88   -30 |
| 34 | 30 |  20011.412410 ? 021 | +000.04   +0  | +04.87   -30 |
| 35 | 30 |  20011.415800 ? 019 | +000.17   +0  | +04.87   -30 |
| 36 | 30 |  20011.417324 ? 021 | +000.08   +0  | +04.87   -30 |
| 37 | 30 |  20011.418275 ? 020 | +000.05   +0  | +04.87   -30 |
| 38 | 30 |  20011.417811 ? 019 | -000.02   +0  | +04.87   -30 |
| 39 | 30 |  20011.417795 ? 020 | -000.00   +0  | +04.87   -30 |
| 40 | 30 |  20011.415679 ? 019 | -000.11   +0  | +04.87   -30 |
| 41 | 30 |  20011.411685 ? 020 | -000.20   +0  | +04.88   -30 |
| 42 | 30 |  20011.412853 ? 019 | +000.06   +0  | +04.87   -30 |
| 43 | 30 |  20011.406068 ? 019 | -000.34   +0  | +04.89   -30 |
| 44 | 30 |  20011.404360 ? 020 | -000.09   +0  | +04.89   -30 |
| 45 | 30 |  20011.398074 ? 019 | -000.31   +0  | +04.90   -30 |
| 46 | 30 |  20011.392920 ? 020 | -000.26   +0  | +04.91   -30 |
| 47 | 30 |  20011.392633 ? 003 | -000.01   +0  | +04.91   -30 |

This is result from my python script.
The TCR tracing consists of more than 50 steps/measurements points (this run was interrupted at the 47th step).
Every step last 10 minutes circa.
Resistance measurement are made every 30 seconds with K2001, 10NPLC with offset compensation, 4W connection with Pomona leads. So there is 20 measurements circa per each step.
(I back to K2001 as my 3458a is occupied to log the LTZ1000 reference).

1st column is the step number
2nd column is temperature set point
3rd is averaged resistance of n samples
4th is ppm difference and temperature difference referred to previous step
5th is ppm/°K and temperature difference referred to the previous step where temperature set point was different. In other words ppm/°K is always calculated respect to the previous step with lower or higher temperature.

[Alex Nikitin]

I suppose this is the right thread to post my results. Here we have the Ohmite RX-1M (glass encapsulated hermetic) 10G resistor, mounted in an alloy Hammond box with BNC connectors and measured by Keithley 617 electrometer with temperature varied from 21.5C down to 13C. I was expecting much larger tempco, here it is not easily measurable, with deviation around 150ppm p-p and the measurement noise around 50ppm p-p, over 8.5 degrees temperature drop in about an hour time, making the tempco less than 20ppm/C in that range of temperatures. I will leave the resistor warming up overnight and see how it behaves with temperatures higher than the room, up to 45-50C. Vertical scale for the resistance is 0.005% (50ppm)/div, the temperature is measured by a thermocouple fixed to the resistor's enclosure.

Cheers

Alex