Corshair RM750 - doesn't allow for board to turn on

[kotel studios]

I think we've found the issue!
5VSB with mainbard plugged in which is OFF dives down to 3.45V! I also hear a high pitched squealing from the PSU (probably 5VSB generation circuit). When the board is turning on by pulling FP low, it skyrockets to 4.2V. Probably limited by the light bulb protection.

Now this is weird since the PSU is fine with spinning up three 3.5" HDDs with it, but when I just plug in a mainboard and short FP suddenly the APFC has issues with the bulb protection.

My idea now is to isolate the corshair's 5VSB rail and inject 5v from my benchtop PSU and see if the PSU is fine powering on an sacrificial board. That way we could rule out bad 5VSB generation. Even simpler idea would be to just put like 800mA-1A load on 5VSB and see if I can power on this PSU.

FP pin with mainboard which is OFF is 2.54V (don't know why the 0.2V decrease ). While pulling FP low it drops down to 0v.

PCB is cleaned, no extra shorts or signs of damage.

But before I do this, does my idea with 5vsb make sense?

[kotel studios]

Okay, PSU can turn on with 0.4mA load on 5VSB.

Switching protection from 120W to 500W also doesn't change 5VSB voltage (stuck at 4.2V). Although now the board stays on and doesn't shutdown.

[stj]

i would check the 5v standby with different loads to see if it's faulty,
it should give stable 5v upto atleast a 2A load
did you re-cap the 5v standby section?

[CapLeaker]

Check ALL capacitors on this unit.

[kotel studios]

I forgot to mention, with the 0.4A load without mainboard, the 5VSB dips to 4.8v :/

It'd be odd how bad caps would appear on an 2013 unit. Sure, they aren't high quality (all secondary are Ltec iirc), but like, it's too new to have bad caps.

Bad caps are very likely. Should I also check those on the vertical PSU output connector PCB (one where ATX24, GPU etc connectors are located)? Those are near impossible to check (need to somehow remove that big trace of tin).

[stj]

you can get bad caps on high frequency circuits like switching psu's in under 5 years these days.
specially as the frequency the things run at keep increasing.
it used to be that switching regulators ran at 25-70KHz,
now some are upto a MHz!!! those arent even suitable for electrolytics, you need ceramics and polymer on those!!

[kotel studios]

New day, and more weird stuff happening!

This time primary capacitor burning through an 10k 1W resistor while being charged up to 320VDC. So I took an 100k resistor to discharge it and what do you know, bright flash and voltage dropping to 75VDC! I DIDN'T EVEN ACCIDENTALLY SHORT IT!

Kotel's curse shining again with it's weird issues only I have...

Primary FETs aren't shorted yet, so hopefully nothing else got screwed (I saw thin gray strands in the air after the arc).

[kotel studios]

So after I've checked all the caps around 5VSB transformer, I found only one 1500uF 16V cap by Koshin which showed as an resistor on my LCR meter (it was reading the value fine if I pulled the cap mid-check, it also took a long while to finish).
Sadly, no major improvement. Only now 5VSB with the mainboard is at 3.7V (0.35V increase from the previous reading). FP pin stayed the same.

Still, I cannot boot up the mainboard.

The one other Ltec 100uF 50v cap checked out fine (read within +/-5% of capacitance), although my LCR didn't give me the ESR value.

Any ideas?

[stj]

unlees you can read capacitance, ESR & Leakage you cant really test a cap

[kotel studios]

Okay, that 100uF 50V cap near 5VSB transformer had been replaced. With one 330uf 25V cap (this one was fine) near the supervisor PCB (the other 220uF 25V was in spec with cap. and ESR and I didn't have any replacements on hand), alongside one 47uF 50V cap on the supervisor PCB itself (4.7uF 50V wasn't replaced for the same reason).
Sadly no improvements. These were basically all of the caps not responsible for main power rail filtering.
Main rails are okay with load that isn't an mainboard.

Any ideas?

[CapLeaker]

Not sure what to think. Are you measuring the caps when they are still hot or warm? Because that affects the ESR reading and a bad cap can look good again as long it it warm. They need to be measured at room temperature.

[stj]

does it use a TNY chip?

[CapLeaker]

I’ve seen these LNK, TNY, TOP etc. switchers do weird things, but not that. Usually the secondary output voltage is pretty much bang on too, but here not so much. Makes me wonder about the feedback circuit.

[Agent24]

I think we've found the issue!
5VSB with mainbard plugged in which is OFF dives down to 3.45V! I also hear a high pitched squealing from the PSU (probably 5VSB generation circuit).

Sounds like you have a fault with the 5VSB circuit - unless the motherboard you plugged in is shorted or something.
If the 5VSB is low, it's probably causing malfunction of the supervisor IC and resulting in the weird problems.

What voltage is the 5VSB when unloaded and the power supply is off?

What happens if you then give it a simple, resistive load, like a 10 Ohm resistor? Any ATX PSU should be able to handle 500mA on 5VSB.

[kotel studios]

Not sure what to think. Are you measuring the caps when they are still hot or warm? Because that affects the ESR reading and a bad cap can look good again as long it it warm. They need to be measured at room temperature.

I give them 5-6 minutes to cool off before measuring them (enough for them to feel room temperature).

As for the TNY chip, nope. Model is (FS)B127H by Fairchild. That's the only 8 pin IC I could find which looked responsible for the switching near the primary side.

As for the 5VSB circuit, 5.10V is when PSU has mains applied with no mainboard as load. When I give it 500mA it dives down to 4.8V.
Mainboard is a known good sacrificial one, can power it on with any of the good PSUs I own.

[Agent24]

I give them 5-6 minutes to cool off before measuring them (enough for them to feel room temperature).

As for the TNY chip, nope. Model is (FS)B127H by Fairchild. That's the only 8 pin IC I could find which looked responsible for the switching near the primary side.

As for the 5VSB circuit, 5.10V is when PSU has mains applied with no mainboard as load. When I give it 500mA it dives down to 4.8V.
Mainboard is a known good sacrificial one, can power it on with any of the good PSUs I own.

You may well have more bad capacitors in the 5VSB circuit that you haven't found yet, or the ones you tested that seem good, aren't really that good.
Have you checked the FSB127H's own power supply? It will have a capacitor smoothing power to Pin 2 (Vdd) that is fed from another winding on the transformer.
If that capacitor is bad you can get problems too.

Otherwise...

Maybe the FSB127H is bad. Those integrated switcher ICs like failing in weird ways, I've noticed, and not always just blowing up.
Sometimes you get overheating and shutdown issues, or power cycling, I wouldn't put it past one to just stop outputting the right voltage.

But a high pitched squeal or whine typically means an overload. Perhaps the 5VSB rectifier is going bad under load.

Another possibility is you already have something overloading the 5VSB standby supply within the PSU itself. Maybe it's just within capacity to supply the current, but when you add the extra load of a motherboard etc, it's too much.
Note that the 5VSB transformer may supply not only 5VSB, but could also generate a second voltage rail for the PWM or Supervisor IC and that rail may be overloaded also.

Now this is weird since the PSU is fine with spinning up three 3.5" HDDs with it,

Not that weird since the HDDs don't load the 5VSB rail, but a motherboard will, and it's the 5VSB circuit that seems to be falling over.

[kotel studios]

I just got an idea to inject extra 5.10V into the 5VSB rail with the board plugged in. That was I guess we could find out if the mainboard will power on while the 5VSB rail is being "assisted" by my benchtop PSU. I would probably need to limit the benchtop to 500mA instead of 1A (this board needs ~700mA 5VSB current when off).

As for the 5VSB caps, all of the electrolytic ones around the transformer were replaced (there aren't any more on the pre-main power generation circuit). There is one small cap near the primary FETs which I haven't checked since it's in a place hard to desolder. Gonna have to desolder the schottky diode heatsink again. It may be responsible for the 5VSB, but it's very likely for the switching of primary FETs.

Another thing could be that the FETs are still in FP case, not TO-220 but that shouldn't be an issue since 3x 3.5" HDDs should in theory draw more than 60W on spinup (while the board without 4Pin CPU PWR connector should draw way less).

Wasn't the FSB127H responsible only for the primary FET switching? I'll solder some probing wires in order to check it.

[stj]

if you inject voltage into a running psu you have a risk of messing with the feedback pin.
you could remove the stanby circuit output rectifier diode and inject power there to see how much current the psu draws.
do it with the psu unpowered.

[kotel studios]

Well, this is odd.....

So I've measured the pin 2 of FSB127H and... the results are questionable...

Without mainboard:
PSU with mains applied - ~4VAC
PSU ON - ~5.3VAC

With mainboard:
Mainboard OFF - ~4.7VAC
Mainboard with shorted PWR_SW (still can't power it on) - ~60VAC

Now that 60VAC is very questionable. Max this chip can handle is 30V according to the datasheet. From what I could read from the typical application schematic, it also uses DC voltage and not AC (I made sure the MM is in the correct mode, in DC mode it spews -0.08V with spikes up to 20V when cutting mains power, AC mode has the results above).
What's even weirder for me is the fact that the PSU can boot without the mainboard with AC Vdd.

I made sure I soldered the probe wire to the Pin 2 and not Pin 7 (attached image is where the probe is soldered on).

[CapLeaker]

Agent24 said, there are weird ways the switcher can fail, but haven’t seen this. So check the power input pin and see if the voltage drops to the switcher. Other than that, either the ultrafast rectifier diode, something is wrong with the feedback, the caps for that 5VSB on the output, maybe there is a starter cap on the primary. There isn’t anything else much… besides the motherboard is drawing too much power.

Read the spec sheet. VDD needs to be 12VDC to start and 6V or below it stops.