HP 8640B dead supply

[_rushin_or_draggin_]

Hey all,

I just got a HP 8640B and FedEx seemed to have tested the shipping box integrity. It probably doesn't help that I saw on my cameras the delivery guy dropping it on the ground! Anyway, upon powering it up I have all zeros across the board, the meter is pinned to max no matter which mode is selected, the "Reduce FM Vernier" light is lit up even though the FM is off, and I have a -20V supply that reads 600mV. I am getting good voltage into the board, the fuse is blown, temporarily sorting across the fuse results in the "Reduce FM Vernier" light to shut off but comes right back on once the fault is restored, and I get about -700 mV once I do short the fuse. I'm guessing the crowbar circuit at the output is doing it's job? I tried to connect the board out of circuit and shorting the fuse results in 6 mA about. I have pulled a few boards out and reseated them hoping maybe something got jarred loose, but still nothing. I have not taken all the boards out and reseated them, mainly because it requires taking the metal boxes apart and I am trying to save that as a last resort.

In the eBay listing everything was working fine. There were numbers that changed on the display, the meter was not pinned, it all seemed to be in working order. I am at a loss as to what could be the issue. The rectifier board was fine (I checked it but I also checked the voltage going to the -20V board, and the transistors, thyristors, and regulators seem to be fine. I get the same input and output resistance for both the -20 and the +20 volt supplies so I am starting to think something deeper in the instrument could be wrong.

How would y'all proceed? I'd rather get some help and do more testing before I start unplugging random boards to see if I can find the fault. I may check the main filter capacitors and see if they are failing since I have no RF output and no counter. I may even check the oscillator and see if it's working.

Thanks all!

[ardiesse]

When the 8640's plugged in and turned on, does the power transformer make a growling or groaning sound?  This is the symptom of shorted rectifier diodes or a shorted power supply rail.  I'd be inclined to remove the PCB with the rectifier diodes and test each diode separately.

Good cooling airflow is critical to the operation of the 8640.  If the fan doesn't turn, the instrument will overheat.

I'd recommend downloading a service manual and following the troubleshooting tree for power supply failure.

[_rushin_or_draggin_]

No transformer sounds and every diode tested perfect. I have the service manual, just haven't made it that far yet. I was convinced something on the -20V supply board was bad but it all tests fine.

[_rushin_or_draggin_]

The service manual says:

"Remove A22 Assembly. Measure resistance from A17xA22-1 to chassis ground"

Is it implying the trace shown in the image circled in red (towards the bottom of the screenshot?

[ardiesse]

Yes.  Removal of A22 allows you to determine if the fault is in the instrument, or in the -20V regulator on A22.  Alternatively, you can measure the resistance from the negative terminal of A17C2 to chassis ground, if that's easier.  If the short to ground remains when A22 is removed, the fault's elsewhere than the -20V regulator.  At this point you can remove cards that have a -20V supply one by one until the short to ground goes away.

The -20V regulator will resemble the +20V regulator closely.  It will look and operate like a regulator for positive voltages, with the exception that the output of the series pass transistor will be connected to chassis ground.  The regulator "common" rail will be at -20V.  HP was fond of this trick with power supply regulators.

[wn1fju]

Here is something to check:

There is a metal screen on the back panel that covers a bunch of pass transistors.  It is somewhat easy for that screen to be pushed in so that it touches one or more of the transistors, especially if it was mishandled during shipping (as your unit was).  The result is a short to ground.  The cure is simply to pull the screen out a little.

At least that is what happened with my HP 8640B when the display read 0.0000 and the unit stopped working.

[_rushin_or_draggin_]

wn1fju if I could take you out to dinner I would. That was the exact issue. Multimeter in amp mode put across the blown fuse is giving 215 milliamps and the display is fixed and the meter works perfectly. I’m still going to go through the service tree and check everything, but that was the main issue. Thank you so much!

[_rushin_or_draggin_]

Alright, I’ve had time to go through the service manual and have found a few issues (namely in the PLL lock circuit) and I am currently working on them. A few weird things I’ve noticed are 1. The level meter will sometimes not respond, but works fine in the AM and FM modes. After a few times turning the output level vernier it will kick in. Another error 2. is tuning to the higher portions of each band, the display shows all zeros when on the internal counter, but double checking with an oscilloscope and the external counter input, it’s still outputting fine. For some reason the internal counter starts counting back down once I get towards the last 7 MHz of the top end and then once I get to the last 6 MHz it shows all zeros. The external counter works fine like I said. Could this be a binary divider or something inside the counter unit itself? I figured I’d ask before I take the whole counter unit out of the machine to debug it.