The 3458A is not the best instrument for resistance transfer measurements, due to its mediocre (unstable) current source, its cumbersome 4W measurement method and because the probing voltage level is too low in some ranges.

The Datron 1281, and its successors, the FLUKE 8508 / 8558 / 8588 are much better, concerning stability and transfer uncertainty for higher resistance values..

The 3458A is as well not specified at all for resistance transfers.

Fluke characterized that for the HP3458A HFL version, i.e. (0.5 + 0.1) ppm for 10kOhm range, (0.1 + 0.1) ppm for the 100kOhm range.

Anyhow, I regularly achieve about 0.2ppm Standard Deviation for my 10kOhm measurements, which is the real transfer uncertainty for my group of six 10KOhm resistors.. I can clearly demonstrate, that the transfer repeatability is on that order of magnitude, which is again an order of magnitude worse than the DCV transfer uncertainty. Therefore, another DCV method would be superior. It is required to do the whole transfer within 20..30minutes, and checking the room temperature and TEMP? for a < 0.2°C change.

If you want to use the 3458A for 10K transfers, you need to follow my recommendations:

- 4W measurements with shielded PTFE cable, use Guard on both sides

- use OFFSET COMP

- resistors must either be characterized for low T.C., or need to have an internal thermometer, to mathematically compensate for its T.C.

otherwise, your transfer from the National lab to your home lab is quite worthless.

- fixed 10kOhm range

- APER 1 (not NPLC 50 or 100!)

- DELAY 1

- use statistics on 16 samples (1 min)

- use MEAN as a valid 8-digit result and StD value for check of stability (StD <= 0.2ppm or 2mOhm required)

For more precise transfers, use bridge configuration, or external current source.

Frank