I know there are a few threads and a lot of into on the web on modifying atx to get higher outputs. But since many atx are different in designs (and schematics are hard to find) I'am pretty lost. Here is some pictures to start with:

I have desoldered all the wires for the power and ground since they where in the way of pictures + it's easier to work. Will solder back some new thick wires if i get this to work. The PSU have two TL431 on it, which is the voltage reference. My guess is that one of them is for 5v standby and the other one for the 12v rail but i'm not sure which one. I'm having a hard time locating the feedback resistors for the TL431. Any idea where to start? The ics on the supply are: DWA108-A, UC3843B, TNY267PN. and then two TL431, thanks!

 

Higher output as in more wattage, more voltage or more amperage?

Also, will it be running as multi-output configuration or a dedicated single output?

Without that info, it's just a speculated guessing game on our end.

 

Higher output as in more wattage, more voltage or more amperage?

Also, will it be running as multi-output configuration or a dedicated single output?

Without that info, it's just a speculated guessing game on our end.

Sorry for the lack of info, higher output VOLTAGE. I forgot to say that. And i only want a single output. i.e be able to adjust the 12v rail from maybe 1 - 15 volts or similar. Cheers.

 

Study the UC3843. It is the brain. Usually there is a resistor IIRC that you could tweak.

Be careful with high voltages AC DC at the input.

 

Study the UC3843. It is the brain. Usually there is resistor IIRC that you could tweak.

Be careful with high voltages AC DC at the input.

Thanks for your answer. This is the datasheet: https://www.fairchildsemi.com/datasheets/UC/UC3843.pdf

Do you know which of the pins i should be "tracing"? I have the 8 pin version.

I also found this: http://danyk.cz/at_atx_en.html

I guess that Forward converter supply (single MOSFET) mod would work for me, But i don't really understand octocouplers.. so i'm a bit confused.

 

Thanks for your answer. This is the datasheet: https://www.fairchildsemi.com/datasheets/UC/UC3843.pdf

Do you know which of the pins i should be "tracing"? I have the 8 pin version.

In the datasheet there is a test circuit. It shows voltage-setting signal going into pin No. 2.

Prior to start altering anything:

a) draw the circuit on paper, identifying ALL associated parts to make sure you deal with safe voltages.

b) Also identify where Vref comes from (probably one of those TL431).

c) Measure its value.

d) Check if the divider justifies the voltage you measure at pin 2.

I cannot give further advice on this and keep in mind you are playing at your own risk.

 

Bear in mind that the cooling fan of the unit runs from one of the +ve supply rails (often 12V), so if you vary that voltage you may need to make changes to the way the fan is powered or its speed is controlled.

 

In the datasheet there is a test circuit. It shows voltage-setting signal going into pin No. 2.

Prior to start altering anything:

a) draw the circuit on paper, identifying ALL associated parts to make sure you deal with safe voltages.

b) Also identify where Vref comes from (probably one of those TL431).

c) Measure its value.

d) Check if the divider justifies the voltage you measure at pin 2.

I cannot give further advice on this and keep in mind you are playing at your own risk.

Thank you, and of course, i will be very careful. Especielly if i follow the tutorial in the link since the octocoupler and the UC IC is on the primary side.

Bear in mind that the cooling fan of the unit runs from one of the +ve supply rails (often 12V), so if you vary that voltage you may need to make changes to the way the fan is powered or its speed is controlled.

Thanks for the tip, Then i will at least know the problem why the fan is spinning like crazy if i get this to work. The DWA IC on the chip is the one that controlls the OV protection/OC protection and prolly the fan too. so i can look at that later.

 

Thank you, and of course, i will be very careful. Especielly if i follow the tutorial in the link since the octocoupler and the UC IC is on the primary side.

Thanks for the tip, Then i will at least know the problem why the fan is spinning like crazy if i get this to work. The DWA IC on the chip is the one that controlls the OV protection/OC protection and prolly the fan too. so i can look at that later.

Wish I could feel that sure... Could the electronics inside the fan, sustain an overvoltage? Maybe it would keep running but not faster and then...nada.

 

Wish I could feel that sure... Could the electronics inside the fan, sustain an overvoltage? Maybe it would keep running but not faster and then...nada.

just to be safe i can use a resistor between the fan and v+ when testing. i tested the fan myself and it can handle up to 18v if i limit the current.

In the link the guy said "Moreover, there is a link between the RC reference and cathode,". Cathode i get, but does anyone knows what he means with RC reference?

 

just to be safe i can use a resistor between the fan and v+ when testing. i tested the fan myself and it can handle up to 18v if i limit the current.

In the link the guy said "Moreover, there is a link between the RC reference and cathode,". Cathode i get, but does anyone knows what he means with RC reference?

Please show the full paragraph. Are you talking of the fan or back to the Vref or Vout?

 

Please show the full paragraph. Are you talking of the fan or back to the Vref or Vout?

Sorry i meant the fan itself. The red wire that goes from the fan and then into a jack in the main board.

 

I've done that mod once. I just stripped all non 12-volt circuit related components off the board and then changed the voltage divider network that the 431 uses to read the 12-volt output through a pot to make it an adjustable 12 - 15 volt power supply.

It had no problems running at near the max wattage rating of the power supply like that but the first time it got shorted the whole line side switching system gave up on life.

Your 12-volt fan should have no issues with running at 15 volts though. I've done it many times.

 

I've done that mod once. I just stripped all non 12-volt circuit related components off the board and then changed the voltage divider network that the 431 uses to read the 12-volt output through a pot to make it an adjustable 12 - 15 volt power supply.

It had no problems running at near the max wattage rating of the power supply like that but the first time it got shorted the whole line side switching system gave up on life.

Your 12-volt fan should have no issues with running at 15 volts though. I've done it many times.

That's nice! Exactly what i wanna do, any tips for me on how you found the divider network? Been looking with my multimeter for two hours now, I can't find any resistor that goes from the cathode/or the Reference pin of the TL431 to any resistor on the board (except a few SMD ones under the board)

 

That's nice! Exactly what i wanna do, any tips for me on how you found the divider network? Been looking with my multimeter for two hours now, I can't find any resistor that goes from the cathode/or the Reference pin of the TL431 to any resistor on the board (except a few SMD ones under the board)

Typically the TL431 and the feedback circuit optocoupler are powered and sense off of the 5-volt power rail so that where you need to concentrate you poking around.

The optocoupler should have its power feed through a resistor from the 5-volt rail then out to the TL431 IC which should have one pin to the main ground plane of the DC side of the power supply.

The one remaining pin of theTL431 should be attached to the midpoint of a two resistor voltage divider where one resistor goes to the ground plane and the other goes to the 5 -volt rail.
The how and where of that layout is up to you to find but once you do it should just be a matter of cutting the circuit that feeds from the 5-volt rail and adding the right resistance for it read from a tap off of the 12-volt rail.

A clear picture of the board form the top down and the bottom up of the DC output side would be helpful on our end so we can see the component layout and the circuit traces separately.

 

Typically the TL431 and the feedback circuit optocoupler are powered and sense off of the 5-volt power rail so that where you need to concentrate you poking around.

The optocoupler should have its power feed through a resistor from the 5-volt rail then out to the TL431 IC which should have one pin to the main ground plane of the DC side of the power supply.

The one remaining pin of theTL431 should be attached to the midpoint of a two resistor voltage divider where one resistor goes to the ground plane and the other goes to the 5 -volt rail.
The how and where of that layout is up to you to find but once you do it should just be a matter of cutting the circuit that feeds from the 5-volt rail and adding the right resistance for it read from a tap off of the 12-volt rail.

A clear picture of the board form the top down and the bottom up of the DC output side would be helpful on our end so we can see the component layout and the circuit traces separately.

Thank you! I will continue looking and post more and better pictures tomorrow. Right now i need to get some sleep

 

Make sure the voltage ratings of the electrolytic caps on the output are high enough to withstand comfortably the higher voltage you intend to create.

 

(except a few SMD ones under the board)

They are resistors, aren't they?

Please keep in mind that from here, I can do little to retrieve the actual schematic of that board.

 

Typically the TL431 and the feedback circuit optocoupler are powered and sense off of the 5-volt power rail so that where you need to concentrate you poking around.

The optocoupler should have its power feed through a resistor from the 5-volt rail then out to the TL431 IC which should have one pin to the main ground plane of the DC side of the power supply.

The one remaining pin of theTL431 should be attached to the midpoint of a two resistor voltage divider where one resistor goes to the ground plane and the other goes to the 5 -volt rail.
The how and where of that layout is up to you to find but once you do it should just be a matter of cutting the circuit that feeds from the 5-volt rail and adding the right resistance for it read from a tap off of the 12-volt rail.

A clear picture of the board form the top down and the bottom up of the DC output side would be helpful on our end so we can see the component layout and the circuit traces separately.

Here are some more pictures:

Red = 5V
Yellow = 12V
Black = GND

 

Make sure the voltage ratings of the electrolytic caps on the output are high enough to withstand comfortably the higher voltage you intend to create.

Yes i will replace the caps if i get it all to work. The caps for the 12v rail seems to be rated for 16v. So at least they can handle a few volts more.