Agilent 6642A Overvoltage protection and other potential issues

[SebastianH]

I recently purchased an Agilent 6642A power supply manufactured in 2012. Generally speaking it's in almost perfect condition. And at the first glance everything seems to work fine--the actual voltage (CV)/current (CC) is absolutely spot on (vs programmed value). The readback might be off by a few mV (I have to check more thoroughly though, could be mainly due to the limited resolution).

But there is one obvious issue: As soon as the overvoltage protections trips (which it does at the programmed voltage) the SCR crowbar seems to engage properly. But there is a residual voltage of roughly 0.72V @ no load and a current in the 20mA range when shorting the output. According to the operating manual the voltage should not be there:

OVP voltage entered is less than the output voltage. This causes the OVP circuit to trip. The output drops to zero, CV turns off, and Prot turns on.

So I checked the relevant circuits (not neccessarily in this order) (Service Manual probably requires registration/mail address): https://www.keysight.com/de/de/assets/9018-01114/service-manuals/9018-01114.pdf):

• +-15 V, +5V Pass, haven't found specific values for RAIL+-, but should be fine
• +15 V ISUP is at ~13,5 V -> Pass
• Overvoltage compare circuit: checked with scope just to be sure - triggers -> Pass
• Output bias/shutdown circuit: receives INHIBIT from secondary micro processor; receives OVSCR from GAN; resulting in SHUTDOWN becoming active; so the output bias circuit is shut down (+-15V Gated measure 0V, and the CONTROL line for the 7 output stages is very close to 0 as well) -> Pass
• SCR circuit: Opto isolator sees OVSCR, SCR gate is driven to ~0.85V

If I simply disable the output (without OV protection triggered) I measure ~+0.7V on both +-15V Gated and 0.5V on CONTROL, but 0V at the output. I don't know what to make of it.

I would appreciate any suggestion what to do/check next.

I also observed that the transistors might not share the load equally. I measured the voltage drop at those 0 Ohm resistor "fuses" (F205, F206 and so on). At full scale (10A) I got something between 60 mV and 80 mV with the exception of XQ305, which was more like 120 mV. It seemed to get a bit hotter also (close to 70°C, whereas all others were more in the 65°C region), so I assume it might not be just a variation in resistance of the "Fuses". Maybe someone can shed some light on whether this is normal or not.

And finally I read somewhere that the power supply considers the programmed CC value in downprogramming operation (second quadrant). Is this actually the case? Mine doesn't do that and I wouldn't neccessarily expect it to do so, because the purpose is not to operate the power supply as an electronic load, but to improve the response for voltage changes.

BTW: If someone has a good solution for the fan noise issue please let me know. I'll probably go with a noctua fan + PWM control.

Thanks in advance for any suggestion!

Sebastian