PSU repair - help

[Bicurico]

Hi,

I am trying to repair the PSU of an HP 8594E.

So far I replaced a couple of resistors, one diode and the two pulse generators for the mosfets.
I replaced the big fat capacitors, as well as the noise filtering capacitors, as those are probe do explode (Dave made a video of it).

The result is that the PSU turns on and the status leds on the PSU show that all voltages are OK. The spectrum analyzer starts, but after 2-3 seconds the PSU switches off. If I disconnect the power cable and wait 3-5 minutes, I can turn the device on again and it will run for another 2-3 seconds, before the PSU shuts off again.

On the PCB of the PSU there is no visual sign of any damage.

I am now kind of stuck: what could be causing the PSU to shut off?

My next steps would be:

- replace the MOSFET's
- replace the big fat capacitors back again (I could not get premium brand replacements and I wonder if I should put the original ones back - they measure OKis outside the PCB - around 430uF instead of the specified 470uF, but so do the new ones I bought)
-

Please help, as I am kind of struck! Also, I already resoldered all solder points, in case there was a micro-fissure.

I actually have three (!) of these broken PSU's, so I could swap parts around. But doing that randomly seems futile due to the sheer amount of components.

Measuring with live mains is not easy, as the PSU has two daughter-boards (these are OK for sure) and the PSU only turns on when there is a load.

I have been trying to repair this for the last two months and I am frustrated, but I won't give up. So far I learned quite a bit about switching power supplies, but these f***ers are quite complicated.

Another question:

The MOSFETs are Motorolla (I guess) P5N40E and P4N50E.

From what I gathered so far, The P5 means 5A, P4 of course means 4A.
the 40 means 400V and the 50 means 500V.

Can I replace a P4N50E with a P4NK80Z (supposed to be 4A 800V)?

How risky is it to replace the P5N40E with this P4NK80Z? I couldn't get any closer match in the local supplier...

Thanks for all the help, suggestions and the ultimate tip to get this PSU back to work.

Regards,
Vitor

[Bicurico]

Hi,

Yesterday I killed the small PCB on the PSU I was working on. 

In order to test it, I assembled it without using all screws. The power inlet was just hold by one screw and when I had the pSU inside the 8594E I attached the power cord... To make it short: the metallic case of the power inlet caused a short on the small PCB. This PCB controls the standby and power on.

Because I did not understand what had happend, I stupidly tried again the PSU, but with the small PCB from broken PSU #2. Again a short (causing mains in the house to shut down). With this I figure I broke the second small PCB.

After searching this morning in HP/Agilent specific forum, I have come across several posts of people I presume know more about fixing switching PSU's and they all comment on how bad HP's implementation is and how difficult, if not impossible, it is to fix these PSU's of the HP 8590 series. There is no schematics available, hard to find components (some unobtainable) and high complexity.

Apparently most PSU's fail due to this small PCB. This explains why I have been measuring capacitors, resistors, diodes and transistors for the last few weeks and all apparently check out fine. I now am led to belive that possibly a component on the small PCB failed.

Diagnostic is difficult as this is SMD based and unacessible when the PSU is assembled.

My hope is to find someone willing to sell me a working PSU.

Regards,
Vitor

[lordvader88]

No schematics, and now do u still have a 3rd 'working' PCB to have it work, until it shuts down in 2-3 seconds ? On these small PCB's is there some controller IC on it. Maybe some 14 or 16 pin chip ?

Could be some controller IC shutting things down when some current or voltage goes too low/high.


Do u have an oscilloscope ?

1 switching PSU I'm working on, cycles on/off, like clockwork, and it's a controller IC doing, but I never figured out why yet. I'm pretty new myself.

I haven't touched it in 2-3months, time to try again with new stuff I'm learning.

[bitseeker]

Make sure the feedback loop is operating correctly. If the switching controller can't receive feedback within a few seconds, it'll turn off the output. A failed opto-isolator is often the cause. Also, check for shorts on the output side as that will also cause the supply to shut down.

[Bicurico]

I have for now given up on the repair...

It was driving me mad to a point my mind would not stop thinking about it and I started to have trouble sleeping.

Perhaps I resume this in a few  weeks...

[bitseeker]

Yeah, sometimes it's good to give the brain a break and see it with fresh eyes later. When you return, it can be helpful to add some photos.