RayOfLight
New member
Hi all,
Some time ago I decided to use my Thermaltake Toughpower 750W to power my 12V 7W SMD5050 LED strip with the green wire ATX jumper trick. After some weeks, it started to buzz when it wasn't powered up (I mean, plugged but not started [no jumper]).
When I placed it back inside my computer (sometimes you need a computer more than a simple LED strip), It wouldn't start, after confirming that the problem was in the PSU, I put it on my bench and placed a 12V 5W car light bulb on all on its terminals.
So, I jumpered the green wire and tested all the voltages: 12V was working, but 3.3, 5 and -12V made it stop, like some protective circuit triggered. OTOH, the 5VSB light it up, dropped the voltage on the bulb to almost 0 and then recovered back to 5V, all in 4 seconds, very very weird.
So you may be guessing, this guy is going to ask what got wrong on his PSU, NO!, altough it would be cool to know, I must say that it got extremely cannibalized by my dad with no reason and now I have more important questions.
Then I started tracing all the circuit PCB so I'm providing you with this View attachment PSU.pdf
D8 designation is wrong, it's a scottky diode with a different ID, sorry.
Green boxes mean things that are too little for me to see, mostly SMD PNPs, the big one hides an optocoupler that doesn't relate to the main circuit I guess, It could have something to do with the 5VSB side.
On card1 slot, there is a CM6800 IC, pin compatible with ML4800 CCM PFC + PWM combo
On card2 slot, there is a temperature sensor, a PS229 protection IC, fan header and some other little things (and not so little, there is a 5W ceramic resistor)
I know there are many parts missing, but I'm learning how it worked day by day, since the first day, I've learnt how PFC works, how optocoupled feedback works and I hope someday I will be confident enough to plug it to AC again.
I've stared many hours onto that diagram and have some questions:
1) My dad wants to test it without heatsinks and no PFC MOSFETS, I'm sure that no heatsinks and no power consumption is good, but the CM6800 without being able to control PF will work?
2) I haven't completely understood CCM boost PFC, I'll have 325VDC on C38, and I hope it's enough for low loads, but, If I were to place the PFC MOSFETS back , what voltage would I end up with? How this voltage is controlled? the capacitor is only 85ºC 400V rated. If it wasn't for the CM6800, could I end up with 2x the input rectified voltage? where is the voltage limit on a Boost topology?
3) The CM6800 duty cycle is 0 to 50% depending, I guess, on the feedback. Supposing there is no feedback because CARD2 is missing and U3 is not conducting (shuts down the main IC). If I short the U3 output, it should start right up, but, as the U5 optocoupler doesn't have anything connected, the main IC will not sense anything, I've read on forums that the transformer output will get up to 30V, what do you guess? optocouple not conducting means it will have max duty cycle or minimum duty cycle? if it were max, then how many volts could I end up with on the secondary? I know it's just guesses, but I'm very curious.
4) The skottky rectifiers on the output part, STPS40L45CG, are placed in a very akward way, I need to know how they work, I mean, one diode is going to do all the work, but the other? It seems like it is never going to work, look at how it's grounded...
5) The S2 coil is the 12V side, the S1 coil, feeds the 5V and 3.3V rail but I don't know how this board gets the 3.3V out of the 5V as I see no linear regulator (and ofcourse not, as it should provide 30A), so, how would you get that voltage?
6) and -12V?
Thank you very much for reading after all this brick and I'll do any test you propose, only If i can and If it is safe, as I don't even have an oscilloscope nor an isolating transformer, but it will be fun
I you have no knowledge of how it works but you guess something, please reply, any input is welcome!
TL;DR; just some newbie questions on an ATX PSU and how to start it after having removed almost every component.
Some time ago I decided to use my Thermaltake Toughpower 750W to power my 12V 7W SMD5050 LED strip with the green wire ATX jumper trick. After some weeks, it started to buzz when it wasn't powered up (I mean, plugged but not started [no jumper]).
When I placed it back inside my computer (sometimes you need a computer more than a simple LED strip), It wouldn't start, after confirming that the problem was in the PSU, I put it on my bench and placed a 12V 5W car light bulb on all on its terminals.
So, I jumpered the green wire and tested all the voltages: 12V was working, but 3.3, 5 and -12V made it stop, like some protective circuit triggered. OTOH, the 5VSB light it up, dropped the voltage on the bulb to almost 0 and then recovered back to 5V, all in 4 seconds, very very weird.
So you may be guessing, this guy is going to ask what got wrong on his PSU, NO!, altough it would be cool to know, I must say that it got extremely cannibalized by my dad with no reason and now I have more important questions.
Then I started tracing all the circuit PCB so I'm providing you with this View attachment PSU.pdf
D8 designation is wrong, it's a scottky diode with a different ID, sorry.
Green boxes mean things that are too little for me to see, mostly SMD PNPs, the big one hides an optocoupler that doesn't relate to the main circuit I guess, It could have something to do with the 5VSB side.
On card1 slot, there is a CM6800 IC, pin compatible with ML4800 CCM PFC + PWM combo
On card2 slot, there is a temperature sensor, a PS229 protection IC, fan header and some other little things (and not so little, there is a 5W ceramic resistor)
I know there are many parts missing, but I'm learning how it worked day by day, since the first day, I've learnt how PFC works, how optocoupled feedback works and I hope someday I will be confident enough to plug it to AC again.
I've stared many hours onto that diagram and have some questions:
1) My dad wants to test it without heatsinks and no PFC MOSFETS, I'm sure that no heatsinks and no power consumption is good, but the CM6800 without being able to control PF will work?
2) I haven't completely understood CCM boost PFC, I'll have 325VDC on C38, and I hope it's enough for low loads, but, If I were to place the PFC MOSFETS back , what voltage would I end up with? How this voltage is controlled? the capacitor is only 85ºC 400V rated. If it wasn't for the CM6800, could I end up with 2x the input rectified voltage? where is the voltage limit on a Boost topology?
3) The CM6800 duty cycle is 0 to 50% depending, I guess, on the feedback. Supposing there is no feedback because CARD2 is missing and U3 is not conducting (shuts down the main IC). If I short the U3 output, it should start right up, but, as the U5 optocoupler doesn't have anything connected, the main IC will not sense anything, I've read on forums that the transformer output will get up to 30V, what do you guess? optocouple not conducting means it will have max duty cycle or minimum duty cycle? if it were max, then how many volts could I end up with on the secondary? I know it's just guesses, but I'm very curious.
4) The skottky rectifiers on the output part, STPS40L45CG, are placed in a very akward way, I need to know how they work, I mean, one diode is going to do all the work, but the other? It seems like it is never going to work, look at how it's grounded...
5) The S2 coil is the 12V side, the S1 coil, feeds the 5V and 3.3V rail but I don't know how this board gets the 3.3V out of the 5V as I see no linear regulator (and ofcourse not, as it should provide 30A), so, how would you get that voltage?
6) and -12V?
Thank you very much for reading after all this brick and I'll do any test you propose, only If i can and If it is safe, as I don't even have an oscilloscope nor an isolating transformer, but it will be fun
I you have no knowledge of how it works but you guess something, please reply, any input is welcome!
TL;DR; just some newbie questions on an ATX PSU and how to start it after having removed almost every component.
Last edited: