I have bought some resistors of 0.1, 1.0, 10, 100, 1k, 10k & 100k Ohms. These are in little boxes with 4 BNC connectors spaced 22 mm apart. This sort of thing http://m.ebay.co.uk/itm/231460477406 although I paid a *lot* less. They are designed for use verifying the performance of LCR meters. I don't know what the uncertainty of *impedance* on these are, but they are probably optimised for use to 1 or more MHz rather than DC resistance standards. At least one I bought is supplied with a 120 MHz LCR meter, but the others are possibly designed for a lower frequency. Getting specifications on these is not easy. My HP 4284A LCR meter is at Keysight being calibrated. My aim is to use these resistors as a way of tracking drift in the meter which would indicate it needs calibration. The basic uncertainty of the HP 4284A meter is 0.05%, although it is higher at very low and very high impedances. Is there realistically anything I can do to stabilise these resistors? I don't want to go to the expense of adding Peltier temperature controllers to them. I suspect the fact that they are several years old means that they are as stable as they are going to be, but I just wanted to check if there is any improvement that I could make. I expect thermally cycling these in an oven and fridge would do more harm than good, but perhaps I am wrong. I do have a 3457A 6.5 digit multimeter which has not been in a calibration lab for at least 15 years, although I have measured some 0.005% resistors as a check for gross errors. It would be worth tracking these HP resistors with that too. It is not worth paying commercial rates for calibration of the resistors as getting the LCR meter calibrated at Keysight is not that expensive (£207 GBP). Keysight also agreed to provide me a software option 006 (support for 2 m & 4 m cables), so getting the meter calibrated was especially good value. Dave

The HP calibration set that I used had cheap SMD chip resistors (!) inside, which behave nicely up to several hundred MHz. The overall high frequency performance is determined by the layout and connections which are critical. The resistors drifted a lot as expected and were checked before every calibration with a 3458A. The high frequency behaviour was assumed to be constant (enough) and the small parasitic inductance and capacitance were taken from the HP calibration certificate. 2015-08-16 10.06 UTC+03.00, Dr. David Kirkby (Kirkby Microwave Ltd) <drkirkby@kirkbymicrowave.co.uk>: > Is there realistically anything I can do to stabilise these resistors? I > don't want to go to the expense of adding Peltier temperature controllers > to them. > > I suspect the fact that they are several years old means that they are as > stable as they are going to be, but I just wanted to check if there is any > improvement that I could make. I expect thermally cycling these in an oven > and fridge would do more harm than good, but perhaps I am wrong.