Most of folks here revolve around making magical 10.000000V references, so for a change let's explore into more difficult murky waters of resistance measurements.
Unlike 10 VDC, this kind of metrology is way less forgiving and more expensive to maintain for hobbyist, so it will be interesting to gather the data in this thread.
Everybody, come join, this is not limited to ivory tower club with stashes of 3458A/8508A/SR104s.
DisclaimerAll test data posted here just for education purposes and provided AS IS, and valid only to specific conditions and equipment used. Use with care, it may apply to your meter, or more likely - may not.
This thread is also NOT a measure of resistance standards testing or doing most accurate resistance measurements on them, but actually using already characterized resistance standards to establish DMM DUT performance.
Even very fine DMM like Keysight 3458A, Fluke 8508A or Datron 1281 are NOT the best tools to perform ppm-accurate resistance measurements. So this thread hopefully will help to answer few questions I have :
- What is actual tempco of resistance measurement function of the common DMMs? We know the manufacturer spec, but it's interesting what are typical values
- ACAL performance functionality tests on meters that have it
- What is annual resistance drift stability of common DMM? This will be challenging for most of hobby-nuts here, as one must have calibrated with known aging rate resistance standard.
- What is impact of the cabling/wiring errors and different connection types? Do we have to use PTFE-insulated cables for even 10KOhm?
- When we need shielded cables/boxes for high-resistance measurements? 100K range? 10Meg?
- When thermal EMF become a problem? 100 ohm? lower?
- Other tweaks and hacks to get better stability in DMM resistance measurement.
If you have performed measurements to evaluate these aspects on your DMM, I'd love to see that shared.
For starters, I have received my 1 Ohm/10 KOhm standard few days ago, and launched a test to estimate temperature coefficient of most calibration labs workhorse, HP 3458A.
So I connected DMM and resistance standard using DIY Belden shielded twinax with copper spade-lugs, to run over changing ambient temperature conditions.
Resistance standard is ovenized and is essentially zero tempco, measured by MI 6010B bridge in airbath +18 to +29C at
Process Instruments. This makes the test much easier, eliminating need of building stable large airbath to keep resistance at constant temperature, while DMM temperature changed with ambient.
I ran 7 batches of tests over remote GPIB, using two 3458A meters.
"Green" meter is the one in question, connected to SL935 10K. May 2018 Calibrated value of SL935 is
9999.9755 +/-0.33ppm."Blue" meter is second set, connected to SR104 located in same room. 2016 Calibrated value of SR104 is
10000.0011 +/-1ppm.Each gap in line represent ACAL ALL procedure (860 seconds), ran simultaneously on both DMMs.
Red line is resistance standard oven temperature, measured by another DMM. SR104 temperature element is not measured this time.
Artifact stability vs temperature chart :
Based on this test, we can see next interesting points:
* Green HP3458A have tempco on 10K range ~1.9 ppm/K. (given assumption that SL935 have zero TCR and no other errors impact).
* Blue HP3458A have tempco on 10K range is ~0.7 ppm/K (given assumption that
my SR104 have zero TCR (which is not correct, this 104 is -0.9ppm at +28c) and no other errors impact).
* Setting APER 1 + OCOMP + DELAY 5 provide faster sampling rate but measurement noise is ~1.5 ppm peak-peak over hour period.
* NPLC200 + OCOMP + DELAY 3 is superslow (1 samples takes 130 seconds to perform, because meter does DELAY after each subsample step, Dr.Frank explained this before) but gives ~0.4ppm peak-peak noise.
Next step is to plot pretty graph and establish annual resistance drift on both of my meters, involving more error sources. Both of my DMMs are self-calibrated using transfer standards from USA lab. Green 3458A calibrated in February 2017. Blue 3458A calibrated in January 2017.