power supply runs fine with paperclip, but not with motherboard

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
So yeah, Ive got a little power supply here, unfortunately the only one I have that fits a case I'm playing with. Its a FSP170-60SI, TFX form factor. I have 2 different ITX motherboards that both boot and run with another supply, but will not start with this one. When I do the paperclip the PSU starts fine and has all its voltages, +12 a bit low at 11.2, but I have seen that often enough in the past on supplies that would at least start.

Is anyone familiar with how the pin 16 start circuit typically works? I am reluctant to just throw all new caps at it, it came out of a junk bin after all, trying not to spend money if I can help it.

edit/ I Just hooked up a 600mA load to the 12V rail and it came up to 11.94, so I think that's good.
 
Last edited:

auntjemima

[H]ard DCOTM x2
Joined
Mar 1, 2014
Messages
11,156
So yeah, Ive got a little power supply here, unfortunately the only one I have that fits a case I'm playing with. Its a FSP170-60SI, TFX form factor. I have 2 different ITX motherboards that both boot and run with another supply, but will not start with this one. When I do the paperclip the PSU starts fine and has all its voltages, +12 a bit low at 11.2, but I have seen that often enough in the past on supplies that would at least start.

Is anyone familiar with how the pin 16 start circuit typically works? I am reluctant to just throw all new caps at it, it came out of a junk bin after all, trying not to spend money if I can help it.

edit/ I Just hooked up a 600mA load to the 12V rail and it came up to 11.94, so I think that's good.
Two things.

1. If the power supply boots and then shuts off, maybe it can handle no load, but the initial start load is too much and it powers off.

2. If it doesn't make any indication, I would think it could have an internal short that is only present on when all 20 pins are connected.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Not even a fan twitch when you short the start pins on a mobo, so it's not like too much load. The standby LED on a mobo does come on when you plug it in though.


edit after the fact, in retrospect, the standby LED was dim, this should have been a tipoff where the problem was
 
Last edited:

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Well obviously. What I'm hoping for is some insight into how the pin 16 start circuit is implemented to give me a bit of a head start in fixing it.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
So digging a bit more into how this all works; the chip on a mobo that pulls ATX pin 16 (the green wire) low to start the supply is called the PCH (platform control hub). The PCH is powered by the 5V standby on ATX pin 9 (the purple wire). Since my supply works when pin 16 is manually pulled low, but 2 different mobos fail to pull 16 low, suspicion naturally falls on pin 9. Low and behold, my 5V standby is only reading 3.6V open, and 2.9V when connected to a mobo.

I believe I shall start poking around the standby supply regulation for faults.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Aaaaaand, it's fixed. Get the thing out of the box. Follow the purple wire down to the board. Find 2 badly bulged little electrolytic caps right there, 330uF 10V, the little multi component tester thing doesn't even see them as caps...reports "no, damaged or unknown part". No 330uF in stock so I tried a couple 220uF 16V I had. Bam! 4.99V standby, mobo starts the supply, problem solved, YMMV.
 

auntjemima

[H]ard DCOTM x2
Joined
Mar 1, 2014
Messages
11,156
Aaaaaand, it's fixed. Get the thing out of the box. Follow the purple wire down to the board. Find 2 badly bulged little electrolytic caps right there, 330uF 10V, the little multi component tester thing doesn't even see them as caps...reports "no, damaged or unknown part". No 330uF in stock so I tried a couple 220uF 16V I had. Bam! 4.99V standby, mobo starts the supply, problem solved, YMMV.
I love how you actually fixed it. Love you.
 
  • Like
Reactions: Axman
like this

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Awwwwww, love you too. I'm just glad I was able to end up making a thread that might actually be useful to someone. I don't know if you guys have noticed, but the internet sort of sucks for this kind of stuff these days. All the good old hits are buried under mountains of vacuous "fixya" bullshit and retarded reddit threads that don't go anywhere. Pretty lame that the whole goddamned internet couldn't just tell me to look at the standby 5V when I searched for "PSU works with paperclip but not with motherboard" honestly.
 

CAD4466HK

[H]ard|Gawd
Joined
Jul 24, 2008
Messages
1,606
Aaaaaand, it's fixed. Get the thing out of the box. Follow the purple wire down to the board. Find 2 badly bulged little electrolytic caps right there, 330uF 10V, the little multi component tester thing doesn't even see them as caps...reports "no, damaged or unknown part". No 330uF in stock so I tried a couple 220uF 16V I had. Bam! 4.99V standby, mobo starts the supply, problem solved, YMMV.
I have a old ass 1000w Ultra that will run when paperclip jumped, but not connected to a mobo, since 2008.
Thanks for giving me a great idea on how to fix it!
 

GiGaBiTe

2[H]4U
Joined
Apr 26, 2013
Messages
2,400
So digging a bit more into how this all works; the chip on a mobo that pulls ATX pin 16 (the green wire) low to start the supply is called the PCH (platform control hub). The PCH is powered by the 5V standby on ATX pin 9 (the purple wire). Since my supply works when pin 16 is manually pulled low, but 2 different mobos fail to pull 16 low, suspicion naturally falls on pin 9. Low and behold, my 5V standby is only reading 3.6V open, and 2.9V when connected to a mobo.

I believe I shall start poking around the standby supply regulation for faults.

The motherboard also won't start if PWR_GOOD signal on pin 8 isn't present. This is why you can force a bad PSU on by shorting pin 16 to ground.

Aaaaaand, it's fixed. Get the thing out of the box. Follow the purple wire down to the board. Find 2 badly bulged little electrolytic caps right there, 330uF 10V, the little multi component tester thing doesn't even see them as caps...reports "no, damaged or unknown part". No 330uF in stock so I tried a couple 220uF 16V I had. Bam! 4.99V standby, mobo starts the supply, problem solved, YMMV.

You need to replace all of the capacitors, even the tiny ones buried in and among other components. FSP units are known to use trash capacitors, usually OST brand. I've recapped dozens and dozens of them over the years.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
I have a old ass 1000w Ultra that will run when paperclip jumped, but not connected to a mobo, since 2008.
Thanks for giving me a great idea on how to fix it!
Yeah, since I figured this out I found another one in the pile with bad 5Vsb. It seems like a lot of these older supplies implement a single chip solution as the oscillator/switch that slapps the HV DC across a tiny little xfmr to generate the low voltage that is then rectified and regulated to produce the standby voltage, often some variant of the TNY26x family of 7 pin DIPs, so if your supply has something like that be sure to check for shorted FET outputs there.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
The motherboard also won't start if PWR_GOOD signal on pin 8 isn't present. This is why you can force a bad PSU on by shorting pin 16 to ground.



You need to replace all of the capacitors, even the tiny ones buried in and among other components. FSP units are known to use trash capacitors, usually OST brand. I've recapped dozens and dozens of them over the years.
These were the green TEAPOs, which I have seen failed a lot, really common in monitor supplies.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Tis the tinkering season

makin' a list
checkin' it twice
gonna see whose ESR isn't nice
capacitor claus is comin' to townnnn
 

GiGaBiTe

2[H]4U
Joined
Apr 26, 2013
Messages
2,400
I don't envy you having to crack open that many Antec's and god knows what else.
I think the OEM that made Antec PSUs in that era bought up the whole Fuhjyyu factory, because I don't think I've seen them in anything else. Just Antec and Power Man, which were virtually the same units internally.

And they all used that bitch ass 10x35mm 3300uF 16v capacitor, which ONLY Fuhjyyu made. Everyone else used the more standard 12.5mm diameter for that rating and you had to relocate the capacitor elsewhere because they were all smashed together with zero extra space.

If you want a really bad time, try recapping an Antec Aria PSU. You have to take the whole damn secondary stage out to get to many of the capacitors.

I have nightmares in Antec and Fuhjyyu.
 

jojo69

[H]F Junkie
Joined
Sep 13, 2009
Messages
11,122
Yeah, thankfully easy to unbox, I have a 1U supply with bulged caps in the pile not so much easy to unbox.

I just have to document enough repairs at this point to justify a DigiKey order.
 

Deadjasper

2[H]4U
Joined
Oct 28, 2001
Messages
2,312
I have a 850w Rosewill PSU I bought years ago. It's never given me minutes trouble but I still haven't learned to fully trust it because of it's heritage, kind like Antec these days. :p
 
Top