Several weeks ago I was looking through the electronics for sale section on CraigsList and found someone who was selling a lot of used test equipment, every piece with some problem. Along with many other pieces of test equipment I ended up with about seven HP3478A, one with a very defective display that became a parts donor for the rest. Two or three had dead batteries and gave the ‘uncalibrated’ message on power up, and a couple of missing or blown input fuses, a broken power switch, and a missing power button cap. There was one that had been dropped and although there wasn’t any physical damage the connector between the main board and the front panel was broken and that took some time to repair.
I have some stable power supplies and resistance standards and I also have an HP 3457A 6.5 digit meter that I have great confidence in. After I repaired the problem(s) with each meter I could use my voltage standard (EDC 330, 7 digits) to set zero and all the multiples of 3 for each D.C. voltage calibration point. If there was any small disagreement between the EDC reference and the HP 3457A, I adjusted the reference to agree with the HP 3457A, and that was the method I relied on to set the calibration points on all the HP 3478A meters. I followed the same procedure for resistance using an ESI 6 decade box and a couple of very high value resistors to calibrate all the resistance ranges, always using the HP 3457A as the final word. I used the same procedure for A.C. and D.C. amps. After running through all the calibration steps I could exit the calibration mode and every repaired meter passed self test and worked properly. Checks at random points of other references and standard resistors indicated that the meters were within specs. Although it isn't exactly a true calibration I am quite confident in the results.
I was pretty please that I got all but one of them working and that they were within driving distance, and they were only $20 each.