A bit of help with this Delta for HP PSU (DPS-575AB)

[Adrian_]

Found this fried PSU into a rather nice XW6400 that got donated to charity in non working condition.The primary is good but the damage on the secondary appears rather extended, do you think this can be fixed and if yes how would you go about it?

[R_J]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

That just looks like some heat damage from the 5 watt resistors on the other side of the board, just try resoldering them. but that's not likely the problem if the power supply is not working

[Adrian_]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

It's not starting at all, despite the primary side looking all good. However if you zoom a bit the photo of the back of the pcb it looks to me like the DNA 1002D chip is fried.
Is there any quick way to test it using a digital multimeter?

Aso what could have caused the 5W resistors to heat up like that? The motherboard seems to be working OK with another PSU.

[R_J]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

How do you know the primary is good?
Here is some info on the chip, It does'nt look "fried" It seems to be used on the secondary.
It may be just the lighting in the picture but the connections in the top right of the bottom view don't look that good
https://www.badcaps.net/forum/showthread.php?t=1229

Does this power supply have a green led? is it on or flashing?

[Adrian_]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Yes it does have a green led but it doesn't turn on when plugging the psu to the main with the atx connector disconnected. (as it should) Also the fan isn't spinning at all.
I believe it's the flash making those connections look bad but I'll double check tomorrow and report back

Thanks for the chip info

[Adrian_]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Well, it turned out to be one of those cases where following the "blood trail" is more then enough

After looking at how discolored was the pcb on the back of those 5w resistors I decided it would be a good idea to check the components in that area first. And the most exposed appeared to be those four 0 ohm SMD resistors, probably used as shunts during the automated production. 30 seconds later I have discovered that one of them had departed to a better world, where currents flow unrestricted . Attempting to remove it resulted in the immediate delamination of about 1cm of trace, so you can imagine how cooked the PCB is. Lucky enough the whole area was super-easy to follow so I carefully cut the trace and soldered about 1cm of wire to replace it plus the shunt.

Needless to say the PSU started happily in test mode after that, green led and everything.

Now the question is what made the 5W resistors cook the PCB and the shunt like that? Is this a design flaw? Or was the PSU overloaded so bad for a long time? The workstation in cause appeared to have been equipped normally but of course I have no clue if some smart fellow didn't tried to run a 300watt video card on it's 12A PCI-E rail or something like that...

[PeteS in CA]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Check the fan (though I would expect Delta to use their own, which are good quality).

I wonder whether having two 5W resistors that close together, with glue, and with airflow impeded by nearby components (I'd be worried about that green electrolytic capacitor next to one of the resistors) and output harness wires was a good thermal design choice. If there is space I'd suggest removing the resistors, extending their leads a cm or two, and then reinstalling them. That gets them up into the airflow a little and adds thermal resistance between the resistors and the PCB.

[momaka]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

I would also recap the 5VSB rail, if it was dead prior to this. Also recap that green cap next to the hot resistor, because it looks quite discolored, and I'm pretty sure its probably starting to develop high ESR at this point (or will soon).

[Adrian_]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Thanks guys

The fan appeared to work as intended during the self test, but I will check if it revs up under load as well.

That green cap next to the 5W resistor looks indeed like it had a hard life, I'll see if I have a suitable replacement.

If this PSU wasn't overloaded and those two 5W resistors managed to cook the PCB so well on their own while the unit was working as intended I'd say this is a clear case of design flaw. This PSU was used on a XW6400 workstation that was quite expensive 11 years ago (over 2000$ in the most basic configuration with just one cpu https://www.cnet.com/products/hp-wor...-33-ghz/specs/)

[momaka]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

In that case, it may be worth finding what those resistors are connected to. If they are just dummy load resistors on some voltage rail, you can probably move them elsewhere in the PSU (like very close to the fan) with long wires, or possibly even remove them completely (some PSUs will remain stable without their dummy load resistors, as long as you put a reasonable load on each rail). I do that on a lot of 300W HiPro PSUs that have a 75-Ohm resistor on the 12V rail, which tends to cook all of the caps around it and then the PSU stops working, of course.

[Adrian_]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Those workstations could be equipped with either one or two CPU's (socket 771) and have separate rails for each CPU plus two more rails (one for RAM, PCI and so on and one for PCI-E plus IDE/SATA) . My guess is that those 5W resistors are offering some load for the two CPU rails.

I was thinking just to extend their leads so there is some space between their base and the secondary PCB.

[momaka]

Re: A bit of help with this Delta for HP PSU (DPS-575AB)

Like I said, just see what they are connected to. If it's directly connected on a voltage rail (i.e. between 5V and ground or 12V and ground, or something like that), then you may be able to completely get rid of them, depending on what rail they are connected to. If it's on the 12V rail, you can most definitely ditch them.