Precise resistors for 34401A repair/calibration

[std]

Hi guys.
I need to repair/calibrate HP 34401A, resistance measurement.
It has several ranges.
I would like to consult what values/accuracy ​​of precision resistors I need to buy?

You can test with a resistor at the beginning of the range or in the middle of the range. Which one to choose?

[Neganur]

https://www.keysight.com/us/en/assets/9018-05613/service-manuals/9018-05613.pdf

page 60 is the chapter for calibration procedures. I think the resistors used (edit: for performance verification) are

100 Ω
1 kΩ
10 kΩ
100 kΩ [2]
1 MΩ [2]
10 MΩ [2]
100 MΩ [2]

[coromonadalix]

you do know  you will need ultra precises resistors, 6.5 digit resolutions     is not a toy  no offence       Basic accuracy: 0.0035% DC, 0.06% AC

if not the meter could be rendered  kinda useless,  if it works well   send it to cal facilities


you do have Cadock   resistor network, some have the full values of 1 10  100 1000  etc ...     but very costly, even say  a single resistance in the 0.0010%  can be as 20$ usd or more

check the meter specs and you see in ohms what it means

good luck

[robert.rozee]

as i recall your 3440a had problems with low resistance ranges when you got it due to an existing fault, and that you carried out repairs; in the thread discussing this there was some suggestion that the meter's resistance calibration values may have been adjusted by a previous owner to compensate for the existing fault.

may i suggest that you sit down with a selection of non-precision resistors and an adjustable bench power supply, and carry out a 'sanity check' on your meter's resistance ranges? i reckon the below should work, and have checked the behavior of my own 34401a when set up accordingly.


your 34401a can measure voltages with an accuracy of around 0.0035%, and can measure currents (on the 100mA range) with an accuracy of around 0.05%. so with a little care you should be able to verify the resistance ranges with relative ease to somewhere approaching the 0.05% level. just be wary of heating up the resistors you are measuring!

pick a resistor (lets say 100 ohm), and connected it in series with the PSU and 34401a so you are measuring the current (100mA range, MANUAL ranging). ensure that one lead of the 100 ohm resistor is connected to the COMMON terminal. now adjust the PSU so that you see a current through the resistor of about 100mA. this will require the PSU to be set to around 10v. the resistor will now be dissipating around 1 watt, so, as said above, be wary of heat affecting its resistance.

now switch the 34401a to DC voltage and without moving any of the existing wiring apply a probe from the input (V/ohm) to the junction between power supply and resistor. this will allow you to measure the voltage drop across the resistor. at this point you know (a) the current through the resistor, and (b) the voltage drop across the resistor. from these two value you can calculate the resistor's resistance.

repeat with other resistor values, building up a chart of results. you can use the 10mA and 100mA ranges as they both use the same burden resistor. current flows through K102 (function switching schematic, page 9-9), and once selected the state of K102 should not change unless you switch to the 1A range.

you can also use pairs of resistors: for example a physically large 90 ohm resistor in series with a smaller 10 ohm resistor (the value lower resistor connected at the common terminal) to get around a PSU with some minimum voltage it can be set to.

while this is not a "calibration standard of testing", it will enable you to build up some approximate level of confidence (or lack thereof) in the existing calibration of your 34401a.


cheers,
rob   :-)

[Conrad Hoffman]

You can certainly do a sanity check as described above, but I've found, given a single known resistor, that getting more than two decades away with accuracy is very difficult without an 8.5 digit meter. You essentially lose a digit for every decade you go. I know NIST, in the past, referenced everything to 1 ohm Thomas resistors, but my underground lair always seems to come up short, compared to NIST.

[coromonadalix]

sure sanity checks can be done,   but  you don't call that calibration ever ever  as written earlier