Tek 7904A rapid ticking - no HT...

Sounds like bad filter caps in pwr supply

The 7904 has an overcurrent protection circuit, which will cause the ticking. This is a good thing.

First step, remove all the plug-ins. Still get the ticking? If yes, you can isolate boards by removing connectors, one board at a time.

I had one of these fail that turned out to be a bad diode on the Z-axis board.

Most likely problem: a bad cap on one of the boards. The tantalums go bad and Tek used them by the truck load.

Good luck.

Steve.

It could also be that one of the high voltage power supplies arching to ground.

Check the HV wires for breaks in the insulation.

The post deflection accelerator is very high voltage so try not to get killed.

I've had that one, too.

A good plan can be to remove each receptacle, one at a time from the PSU output cable socket molding to isolate each rail. Pull the sensing feedback plug first, though, or you can find that a shorted rail will fry one of the sensing bypass resistors. The PSU will regulate, albeit a bit poorly, with the sensing disconnected, because of those same bypass resistors.

The PSU won't run properly with no load at all, BTW.

There are some 100uF (??) electrolytics on the 15V rails on the interface PCB that can go short under operating conditions, but appear OK on low voltage. They are a bitch to get at.

Five'll get you ten that the PSU itself is OK.

ground.

killed.

Is Jim Yanik (sp?) still around?

Thanks

Fred Abse wrote in news: snipped-for-privacy@cerebrumconfus.it:

The 7000 series switcher supplies DEPEND on the sense feedback from the interface PCB. I would not advise trying to troubleshoot with it disconnected.

"Mike Deblis" wrote in news:dpmrhd$6n5$ snipped-for-privacy@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com:

The fast ticking,IIRC,indicates a failure in the PS itself,like a shorted diode,bad electrolytic,open switcher xstr(s).Probably a primary-side failure. I strongly advise reading the serivce manual's circuit description on the PS. If you don't have a svc manual,then I would not attempt repair;it's a very complex series-resonant switching supply.There are under/overvoltage lockout circuits,several current and voltage feedback loops.

Sorry, Jim. I hate to have to contradict you, but, in this case, you are wrong.

Take a look at the 7904 manual (I didn't have an "A" manual handy, so the circuit refs will be different)

Between each PSU output pin, and its associated sensing connector pin, there is a low (ish) value resistor - R1429, R1591, R1499, R1539, R1589, et al.

Those resistors are there for the purpose of ensuring that the PSU will still regulate safely if the sensing plug gets disconnected.

Just pull the two HV drive coax leads out of the back (the front, really) of the PSU. Much quicker, and it leaves everything else connected.

Right. But take care to have a minimum load, since the switcher power supply needs at least appr. 10% loading to regulate/work properly. I don'T know if this is correct for ALL 7904/7904A. But at least in the 6

your mileage may vary ....

hth, Andreas

Yes, 40 watts (ish) is usually good enough, spread over all the rails. The switching drive IC gets its supply from the primary current sensing transformer, so, not enough primary current, not enough volts to supply the chip. Quite why they did it that way, I've never found out. Probably reduces component count. I've run a a few dozen of those PSUs standalone on active loads for fault finding. You really *do* need a variac to work on those puppies.

Yup :-)

Fred Abse wrote in news: snipped-for-privacy@cerebrumconfus.it:

ACTIVE loads? Wouldn't a dummy load be -passive-?? The original mainframe would be an "active" load.

Definitely.There are so many control loops that you need to start them on a Variac with the control loop broken to see what feedbacks are not right. Also,on the early 7904's(below B260K),there were many mods to the circuitry around the control IC.The only complete documentation on them was in the "Mod summary" section of the microfiche.(now long gone)

While at TEK,I made up dummy loads for several 7K power supplies,put them in nice vented enclosures,they had decoupling caps just like a mainframe.They're all sitting (unused) in TEKs DCFO right now,or scrapped. I doubt they knew what to do with them,no documentation.

I use various electronic loads, mostly commercial, some made in-house, for PSU testing. More convenient than finding the right resistor. I can go right up from a few milliamps to 200 amps (obviously not using the same instrument). Useful for testing on varying or pulsed loads, too.

I've got some (probably not all) of that information.

That's the sort of thing I've always *threatened* to do when I get time. Trouble is, I never get time.