To charge the 110Ah battery bank I built, I need a power supply that can provide at least 10A at 14.6V. Since I have many old ATX power supplies lying around and the 12V rails of these power supplies are more than capable of providing 10A, I decided to modify one such power supply for using as a 4S LiFePO4 battery charger.

However, this modification is not as straightforward as just tweaking an output adjustment potentiometer. All ATX power supplies have some kind of voltage supervisory circuitry for monitoring the output voltage to prevent an over-voltage situation should a fault develop somewhere in the power supply. Once the output voltage drifts outside the allowed margin (typically 5% to 10%) the power supply would be automatically shutoff to prevent damage to connected components.

So to modify the a standard 12V power rail to 14.6V, we will need to disable the supervisor circuit. This is where it is getting tricky as different PSU manufacturers use different voltage supervisory techniques and there are at least a dozen dedicated ICs (e.g. TI’s TPS3511, ON Semiconductor’s NCP4350, etc.) for this purpose. But once you have identified the power supervisor chip, it is fairly easy to disable it by disconnecting the power good pin and pull it permanently low (or high depending on the chip).

Some power supplies however, do not use dedicated power supervisor chips but rely on some other techniques for over-voltage protections instead. In the power supply I have (a rather generic DURO 400W power supply manufactured in 2003) for example, over-voltage protection is achieved by pulling the deadtime control pin on the PWM IC TL494 high when over-voltage is detected which effectively increasing the deadtime to 100% and thus disabling the output transistors. This technique is commonly used with PSUs that use TL494’s for regulating the output voltages.

To disable the over-voltage protection, we can simply cut the trace from the DTC pin (pin 4 of TL494, see schematics below) and put a 3.3K resistor (R4) between the DTC and ground. The value of the resistor is not critical. Because the DTC pin is no longer connected to the voltage comparator circuitry that monitors the output rails, we can proceed to adjusting the output voltage.

The output voltage is determined by
\[V_o=V_{REF}\left(1+ \frac{R_1}{R_2} \right)\]

Since we only care about the 12V rail voltage, we can simply lower the value of R2 by paralleling a resistor R3 till the desired output voltage is reached. For this PSU, I used 15K for R3 to achieve the 14.6 output voltage.

And here are a couple of pictures showing the added resistors to the PCB. In the picture to the right, you can clearly see the trace connected to pin 4 of TL494 was cut. In the picture the lead of the 15K resistor seemed to be nearly touching pin 2 of the chip, this is due to the angle of how this picture was taken. In reality the lead is pretty far from that pin. If you are worried about accidentally shorting out the pins, you can use some heatshrink tubing here.

To prevent battery backfeeding into the power supply when it is powered off, I added a Schottky diode (the two diodes in the same package are paralleled together) at the output.

Here is a short video explaining this power supply modification.

Be Sociable, Share!

6 Thoughts on “Modifying a Computer ATX Power Supply For Higher Output Voltage”

  • Hi Kerry, I have an old ATX P4 power supply JIL-2406 and could not find the circuit diagram from the google search. And I wanted modify to get 13.8V or say higher than 13V. Could you please see if you could guide me or show me the way for this very ATX power supply?

    Chinghem Chingtham

  • Hi, I messaged you from your video of this on TY, but found your site and thought I would expand a bit here. I believe I have this same PS. I was hoping to modify it this coming week, but instead of going higher, I want variable output from 3-12 V. I have two questions:

    1) I assume this can be done by using the above equation Vo=Vref…. and having a fixed resistor and a potentiometer for the range needed, correct?

    2) Would I still have to cut and use the 3.3K resistor to ground for the DTC pin, or can I leave overvoltage protection intact since I don’t want to exceed the maximum anyhow? I would also use the 12V rail as you have so overvoltage might not be an issue like if I used 5 or 3V rails.

    Thanks, in advance, for your response.

  • Hello, Just to follow up on my initial questions as I have now modified the supply:

    1) Yes, simply following the equation gets the range needed.

    2) No, I did not need to cut the trace the the DTC pin so overvoltage protection is intact.

    Just to note that my desired range was 3-12V and not higher than 12V, so it works. Just thought I would share my follow up.

    • Hi Jon, sorry I didn’t get a chance to respond to your original inquiry, I receive a ton of emails every day… Anyway, if you are adjusting downwards only, you can just add a pot in series with R2. I have not tried what was the lowest voltage you could get but you may not be able to go down to 3V as there is a limit on the PWM circcuit. You don’t need to disable the overvoltage resistor in this scenario.

  • I have PC power that the voltage is controlled by TPS3510P IC, it has total of 8pins, I have downloaded the data sheet, I want to adjust the voltage to 13.8v, am already having 12v 14AmP, the circuit is different from the one you posted here

Leave a Reply

Your email address will not be published.