Primary Thermistor Question

[JEWilson]

Have a PSU which is exhibiting a possible primary thermistor issue.

First, when room temp is cold, supply will generally start
When ambient temp in room is a little warmer, the psu will
often fail to start. Unfortunately, these start conditions
are never consistent.
A hard reset often will not help the PSU to POST the system either.

As the system mobo has been recapped - I am sure the problem
does not lie there.

By referring to start, the PSU will always spin up but the PC will
not POST in line with the observation above.

I suspect this is an inrush problem with the primary.
To isolate the issue, opened the PSU, an Enermax EG651P-VE FMA.

There is no thermistor I can identify per se.
What are present, are a power resistor in a ceramic package
mounted vertically adjacent but not bonded to the primary HSF.
Further inspection finds another 5 band resistor (what appears to
be wirewound power passive) which is bonded with an unknown glue or
equivalent to the base of the primary heatsink. This is 65.9kR 0.5% device.

Am I missing something here, the problem is consistent with an
inrush negative coefficient problem thermistor in the primary
or possibly an alternative problem with the startup resistor or,
am I referring to the same thing?

Advice appreciated

[i4004]

Re: Primary Thermistor Question

"A can of cold spray or a heat gun will be useful to track down the bad
component but it could be a frustrating search."

http://www.repairfaq.org/REPAIR/F_mo...ml#MONFAQC_010

i think the only thermistor is the one that controls the fan, if the psu has that feature at all.

[JEWilson]

Re: Primary Thermistor Question

Y

There is a bead thermistor tied to the largest inductor at the O/P
with a cable-tie and is obviously used for fan control.

[999999999]

Re: Primary Thermistor Question

If I understand you correctly, you are saying the PSU is failing to start when the room is warmer, but the PSU is still starting, it is only the motherboard which is failing to POST when it is warmer.

If the above is correct, I suspect an intermittent contact issue on the motherboard. Something like a cracked solder joint (could even be practically impossible to find if it's under a BGA chip like northbridge), board crack has broken a trace, or even a surface mounted component itself has cracked. Could be mechanical contacts like the PSU connector, memory slot to module contact, or any cards in their slots.

Small changes in room temp won't significantly effect a thermistor to the extent that if room is warmer it would become a factor, if the room were warm enough to make it do anything it would just pass more current sooner, a higher inrush current could eventually wear out some part in the PSU but wouldn't cause PSU to run, output stable voltage still but system then didn't POST.

A low value power resistor can often be used instead of a thermistor, it shouldn't be a problem. Actually since it uses a resistor it would have even less of an insignificant impact based on what the room temp was.

I'm wondering about some other details of the troubleshooting. Have you tried another known working PSU, swapping it in place of the Enermax? That would be the thing I'd try next if you haven't already, after measuring the Enermax's output voltages assuming as I wrote at first that the PSU does stay turned on but it's the system that isn't posting.

I'm not trying to place blame but I would place odds on the board becoming stressed causing a crack or solder bond breaking when it was taken out of the system to be recapped, or rough handling of the system and/or the caps failing caused more damage to the board than just caps failing. Thus, besides trying another PSU I would expect it is the board that has to be replaced instead of PSU repair, modification, or replacement. Possibly the bad caps allowed peak voltages high enough to damage some plugged in part instead but ultimately unless I misunderstood what you meant then I don't think the problem has anything to do with resistors or thermistors in the power supply.

Once I had a board (Asus A7N8X nForce2 skt.A) with a somewhat related problem, it seems the southbridge wasn't making good contact and I could only start the system consistently when it was cold by pre-heating the southbridge with a hair dryer.

Since you have the opposite need to cool areas, the can of cold spray that i4004 suggested might be useful, or when the room is cold you could heat the PSU mildly with a hairdryer while it is propped outside of the mostly closed system case (leaving side panel off just enough to get the PSU leads in or put leads in through the rear PSU mounting hole) so that the PSU is heated but the rest of the system isn't. See if it then is more likely to POST with only one part or the other heated.