HP/Agilent E3610A calibration - negative and non-linear meter reading (solved)

[CatNinja]

My E3610A meter readings were always slightly off (negative at 0v, about 30mV low at 5V), so I tried calibrating it as per the Service Manual.

However, there is only 1 adjustment pot for the voltage, which only allows you to adjust it to be accurate at the recommended 15V. After calibration, I still get a negative reading at 0V, but worst of all, an error that is non-linear across the range, ranging from -0.02V at 0V, -0.03V at 3.30V, +0.05V at 9V, to 0V at 15V (calibrated for this voltage). Looking at the forums, it looks like a lot of people have the same issue with negative zero voltage reading, and the recommendation is just to live with it since it's just a built in power supply meter.

The display driver is a Microchip TC7107, and it's auto-zeroing, which is why there is no pot for zero adjustment. It's used in a lot of low-end DMMs, so it should be a lot more accurate than what I'm seeing. I don't see any obvious bad components in the display either.

I finally found a thread that describes this type of multimeter https://www.eevblog.com/forum/testgear/old-fluke-multimeters/875/ and it described how sensitive these chips are to board contamination. So I cleaned the display board thoroughly with IPA even though it looked clean already. I didn't want to dismount and soak it, so I just brushed on IPA and wiped it off several times. It's now dead accurate (within 1 digit, measured with a 34401A) across the entire range, and reads +/- 0.00V at zero.

I hope this helps someone else with this power supply. It's my favorite by far, especially now that the meter reading is accurate.