[volt-nuts] HP 3420B calibration assitance

[Dick Moore]

Hi, volt-nuts -- I picked up an HP 3420B diff. dc meter for *real cheap* and it works pretty well. I've replaced the batteries with AA NiMH units in 4 each 10-batt plastic holders (there was just enough room to shoehorn them in -- the A version is line-power only, the B version is line or batt), and have done a little clean-up and some repairs, replaced some knobs, etc. 

This unit has cal pots to adjust the total resistance of each of the resistor sections for the first two decades of the six decade divider (modified Kelvin-Varley constant-R) string. HP supplied a special test board for this purpose that plugs into a socket on the divider assembly, but the test board is long gone from this unit. The manual copy I have shows a *very* convoluted process for adjusting these pots in the first two decades and this depends on having the test board plus some precision resistors and some connectors.

Is there a more direct way to adjust these pots using a precision variable reference voltage (from the Fluke 5440B that I own) as input? Any help appreciated. This is a very nice meter with an excellent ratio function. Be nice to have it working at or near spec. I do own an HP 3458A to use to look at internal voltages and resistances too. Because all of the string resistors in a decade are always all in series, I can't wrap my head around a procedure to adjust the pots, since they interact....

The general form of the first two decade dividers is as follows: the first decade has five resistors which are switched in and out in combos to make up Ra and Rb, and to maintain a constant 11k, using 5 resistor sections, each with a tweak. The second decade uses 4 resistors, each with a tweak, to maintain 10k total. The dividers use a 4k, two 2k and one or two 1k resistors for the Ra and Rb, each with a tweak, so there are five pots for each resistor of the first decade and four pots for each resistor of the second decade. The remaining four decades are fixed with each having a 10k total R. I formatted the following diagram with a constant width font, so I hope it displays OK... worth a try....

I---------|
          |
          Ra
          |
Ref input |
          |
          |--------O
          |
          |
          Rb       Out to next decade
          |
          |
I---------|--------O

As Ra goes up 1k ohm, Rb goes down 1k ohm, & vv

Dick Moore