HP / Agilent / Keysight 54621D MSO Repair

[JWag]

https://www.keysight.com/en/pd-1000000809%3Aepsg%3Apro-pn-54621D/portable-mso?cc=US&lc=eng

I found one of these devices really cheap near me. The reason it was so cheap was that it didn't work (fortunately I already knew this). 

Symptoms:
Turning on the power switch (it is a switch, not soft power), results in nearly all of the control panel LEDs lighting, the cooling fan spinning up, and nothing else. No display, no clicking relays or beeps or any other life signs.

Troubleshooting so far:

• Voltages: all OK and within spec. I'd hoped it was bad PSU caps, but everything looks nice.
• Lithium battery: dead, flat, nothing. It manages a few dozen millivolts, but this collapses with almost any load applied.
• Heat: Using thermal imager and fingertips, nothing gets absurdly hot.
• Carbon: No parts seem to have burned/exploded.
• Oscillators: The 50MHz and 14MHz oscillators are fine. I don't have anything to check the 200MHz ECL one.
• Video signals (suggested by "service" manual): nothing. Guess that explains the lack of display.
• Serial output: a little burp on power-up, but nothing readable, let alone diagnostic.
• U300 (MC68EC020FG25) CPU: 25MHz clock is present. Reset line pulses low (active) for a few microseconds at around 740Hz. Uh oh.
• U340 (ADM691 supervisor): Does not appear responsible for the resets. The noninverting RESET output is always low, but the inverting RESET (with weak pullup current source) is on the same net as the CPU reset and shows the same waveform. Watchdog outputs remain inactive.
• Component Removal: Unplugging the display, front panel, GPIB module, and floppy drive does not result in any improvement

Future steps:

• I've taken pics of the board and will look for other possible sources for the active RESETs. I'll probably be manually beeping out connections.
• It's entirely possible the CPU is asserting its own reset, possibly due to a bad flash. Unfortunately these are annoying wide TSOPs, which I'm not really equipped to read or write.
• I have access to a 54622A. Should be similar, and maybe I'll find something by comparing the two.

I'd be grateful for any suggestions!

[oPossum]

With a dead battery you will have to do a power up reset.

"To reset the entire oscilloscope, press and hold any softkey and switch the
oscilloscope power on. Release the softkey when the display appears".

Give that a try.

[JWag]

With a dead battery you will have to do a power up reset.

"To reset the entire oscilloscope, press and hold any softkey and switch the
oscilloscope power on. Release the softkey when the display appears".

Give that a try.

Thanks! Unfortunately, it didn't work. Presumably this should only take a couple of seconds, correct? I tried a few times, and gave up after about 15-20 seconds.

I wasn't sure if this required a good battery or not (I haven't replaced it yet), so I tried with and without a current-limited 3V source attached. I even checked the softkeys to make sure they were sufficiently low-Z. I picked the one that went to about 1k with reasonable pressing force.

[JWag]

I operated the 54622A using the 54621D front panel, and verified that all of the knobs, buttons and softkeys work on the '21D panel. Also using the '21D cable, so those two parts are probably good.

The '22A boots fine without the floppy drive (but throws an error once booted) so leaving that out while I troubleshoot the '21D is a good policy.
The '22A boots fine with the GPIB module that came with the '21D, so that component is probably not bad.
Next: try to boot the '22A with the '21B's PSU and display to make sure they're ok