PDA

View Full Version : H7874-00 power supply



mloewen
August 18th, 2009, 05:50 AM
I have a VAX 4000/300 with a bad power supply - does anyone have a working spare for sale? The system also has a R400X expansion cabinet, and I have it running temporarily by moving the supply from the R400X into the BA440 along with the 3 functioning hard drives, but I want to get both boxes running.

paul
August 18th, 2009, 04:00 PM
Have you considered attempting to fix the PS?

The one in my Personal DecStation 5000/25 simply had one shorted tantalum cap on the 12V output, which of course makes it shut down a moment after switch on.

mloewen
August 19th, 2009, 06:50 PM
Have you considered attempting to fix the PS?

The one in my Personal DecStation 5000/25 simply had one shorted tantalum cap on the 12V output, which of course makes it shut down a moment after switch on.

There are no obviously blown or leaking caps, nor burned areas or parts. I have no service manual.

tezza
August 19th, 2009, 07:38 PM
I've found shorted caps are not always obvious.

I'm still a learner when it comes to electronics but it is my understanding there should be some resistance across a cap. It's a big job to unsolder one end of all the caps to check resistance across each of them, but one way you can reduce the number of suspect caps is to measure the resistance across them in circuit. Although you won't get a true ohms value for the cap, nevertheless, if you get a reasonable resistance the cap won't be shorted.

On the other hand if you get 0 ohms or very low ohms then either the resistance in the attached circuit is low OR the cap is indeed shorted.

Either way, you can eliminate the seemingly good caps, and just focus on the subset that gives low readings. You may have to unsolder these to identifiy the exact one which is shorted, but at least the number would have been narrowed down, perhaps considerably.

I used this technique to diagnosis a faulty disk controller, and found it reduced dozens of potenial caps to just three suspects. Sure enough, one of those was the culprit.

Tez

modem7
August 20th, 2009, 12:26 AM
>> SAFETY <<
Obviously, disconnect power before delving into a power supply. And wait a while because it takes some time for some of the large capactors to discharge. A 10 minute wait should be enough.


I'm still a learner when it comes to electronics but it is my understanding there should be some resistance across a cap. It's a big job to unsolder one end of all the caps to check resistance across each of them, but one way you can reduce the number of suspect caps is to measure the resistance across them in circuit. Although you won't get a true ohms value for the cap, nevertheless, if you get a reasonable resistance the cap won't be shorted.
Tezza. Most in the trade would have an ESR meter in their toolkit. Have a read of http://www.flippers.com/esrkttxt.html for an explanation. You can measure in-circuit (with power off). An ESR meter has saved me many hours of work. The ESR meter I use is shown at http://www.flippers.com/esrktmtr.html and has typical ESR readings printed on the front.

tezza
August 20th, 2009, 02:28 AM
Tezza. Most in the trade would have an ESR meter in their toolkit. Have a read of http://www.flippers.com/esrkttxt.html for an explanation. You can measure in-circuit (with power off). An ESR meter has saved me many hours of work. The ESR meter I use is shown at http://www.flippers.com/esrktmtr.html and has typical ESR readings printed on the front.

Interesting. Thanks for alerting me to these. I see the final sentence says that ordinary multimeters are ok for shorted caps. Of course many caps are not shorted, but are still faulty, so I can see the need for such a device.

Tez

paul
August 20th, 2009, 03:26 AM
There are no obviously blown or leaking caps, nor burned areas or parts.

That's usually the preferred scenario.


I have no service manual.None of us do, especially for power supplies for which the manufacturer would rather sell you a new one. But for someone with the right skills there is often nothing difficult about fixing one. The semiconductors, resistors, and inductors are pretty bullet proof - it's the tantalum and electrolytic caps which age and cause the PS to simply shutdown on overcurrent or excessive ripple.

Yes, it is dangerous if you don't know what you are doing but any local electronic technician should have no problem checking the easy stuff for you at minimal cost.

If the PS fan starts momentarily (after sitting for a while) and then stops it can almost certainly be fixed.

cosam
August 20th, 2009, 03:51 AM
That's usually the preferred scenario.
I don't know - If it's not working, I'd rather open the thing up and find a big, smoking hole in the top of an electrolytic capacitor, or two bare metal legs where a tantalum used to be. Kind of gives you idea of where the problem might be ;-)


None of us do, especially for power supplies for which the manufacturer would rather sell you a new one. But for someone with the right skills there is often nothing difficult about fixing one. The semiconductors, resistors, and inductors are pretty bullet proof - it's the tantalum and electrolytic caps which age and cause the PS to simply shutdown on overcurrent or excessive ripple.
For a lot of DEC stuff, the service manuals (and very detailed ones at that) are in fact very often available via sites like Bitsavers. But I agree that a great deal of failures aren't that hard to fix without documentation.

marmotking
August 17th, 2013, 10:02 PM
Did you ever get your supply fixed? I just switched on my 4000/200 and the supply is dead. Just like that. Grrr... It's been eaten by a gru. I pulled a supply out of my 4000/300 for now because I want to show off this 4000/200 tomorrow at SRCS.

mloewen
August 19th, 2013, 06:57 PM
Did you ever get your supply fixed? I just switched on my 4000/200 and the supply is dead. Just like that. Grrr... It's been eaten by a gru. I pulled a supply out of my 4000/300 for now because I want to show off this 4000/200 tomorrow at SRCS.

No, still on my TODO list, but way down. I found two replacement supplies on Epay to get everything up and running.

ApogeeData
October 9th, 2013, 07:40 AM
I have them with 45 day warranty $ 150 sales@apogeedata.com

marmotking
September 14th, 2015, 02:30 PM
So the one I stole from my 4000/300 is now dead too. These seem to be the weak link, don't they. It shuts itself off after a few seconds. Do it two or three times and it'll keep running after that. I suppose I need to figure out how to fix it or find one for sale.

What I really need to fix this is a test bed. I'm not sure what or how to hook a load on this one or if it can be tested without a load. Would make repair easier and safer than testing in an actual machine.

g4ugm
September 14th, 2015, 11:55 PM
These are not easy to fix. I have one on the bench now, and a second is with a friend who worked on proefessionally repairing such kit. I also have so IBM 3174 supplies which are dead. These are nice because they have built in load resistors and can be powered up without a load.