Hacking an ATX power supply, take two
by Antony Chazapis
I was not very keen on playing around with another ATX power supply so soon (first attempt was a few months ago), but when I came across another unit about to be thrown away, I just couldn’t resist. This was a much newer design, manufactured by Tagan, model TG530-U15, rated for 33A at 12V.
First I had to find out why it wouldn’t even power up. One of the PCBs inside, which I later identified as implementing the PFC (Power Factor Correction), had a blown-up capacitor. A new one solved the problem. Next, I employed the same technique as before to raise the output voltage to 13.8V. I found the regulator chip, the pin that samples the output voltage, and replaced the resistor in series with a much larger trimmer. I initially set the resistance to the value replaced, powered up, and gradually adjusted it until I got 13.8V out. Finally, I soldered back the corresponding fixed resistors, to make the change permanent.
I tested the power supply using 12V/20W lamps I had around the house. However, the results were rather disappointing. The first four lamps caused a drop of 2.5V-3V each to the output voltage. With 80W the voltmeter read 12.7V-12.8V, which remained so even when adding two more lamps.
My conclusion from all this is that tricking the regulator into boosting the output is not enough. Some suggest that it is necessary to rewind the output transformer among other changes (read more here and here). Maybe next time… Until then, this one goes back to my parts bin.