Noticed a bug with my DP832 today (FW 01.14) while calibrating it. It's possible this has been covered elsewhere, but some searching didn't turn up anything, so I thought I'd share both the problem and the solution.
My output voltage on channel 2 had been about 8 mV low for a while, so I decided to calibrate it today using a modified version of bson's script and my DM3068. Everything was fine until I was checking various voltages after the calibration. That's when I noticed that the DP832 had lost its zero on channel 2, by which I mean it was putting out -86 mV instead of 0 V when it was off. All readings at 100 mV and higher were fine, but once the output setpoint was below about 50 mV, the actual output dropped below 0 V. That's a problem because DP832 channels that are "off" are really just set to "0 V" -- or in my case, -86 mV.
Channels 1 and 3 were -0.5 mV and 0.6 mV, respectively, when they were "off".
I confirmed the problem by wiring my 34401A in parallel with my DM3068. The two meters agreed about all of the readings.
I calibrated again, calibrated a third time, power cycled the DP832, manually calibrated from the front panel, and so on. Nothing helped. The 0 V point remained stubbornly below zero, roughly -60 mV to -100 mV depending on the calibration run.
At that point, I looked through the offsets between the DAC calibration point voltages and the actual voltages. On my DP832, all of the DAC cal points below 5.2 V were off by about 430 mV; said differently, the actual voltages were about 430 mV lower than the nominal cal voltages. The only exception was the 0.2 V cal point, which was only 286 mV low. That made me think: what if I just ignore the actual reading during the 0.2 V cal point measurement and instead tell the DP832 that the measured value was 430 mV low, i.e., -230 mV? So, I patched the cal script to do that, and completed another calibration cycle.
That worked! My channel 2 voltages were accurate to with a mV or two almost everywhere, even with output voltages down in the low tens of mV. Zero was only slightly off: about 4 mV. Close enough!
To recap: I fixed my non-zero output problem on my DP832 by giving the instrument a fake value for the 0.2 V DAC cal point.