Modifying 12V PSU, how do I identify the +12V Return Sense?

Question

I'm trying to build a 13.8V/14.4V @ 30A power supply, to charge a lead acid battery bank.

For the purpose, I've read that some server PSUs are excellent for the purpose. They are typically able to provide high currents at 12V, and have sensing circuits that can be fooled into supplying a higher voltage, typically up to 14.2V.

Here is an example: https://www.rcgroups.com/forums/showthread.php?1005309-A-simple-high-quality-12Volt-100Amp-Power-Supply-Part1/page107

I've purchased a Dell PS-2521-1D, which was advertised as AA23300, and I've read that they should be the same. It can provide 12V at 44.5A.

Unfortunately, after opening the case, I see that the PCB is a completely different design, yet the pin-out has the same AA23300 layout.

I've tried to identify the voltage sensing circuit by using the multimeter to test for short circuit between the +12V and any of the chip pins. I didn't find any connections direct connections.

The MOSFETs have ~26V on the drain side when powered on, and 12V at the source side.

The gate of the MOSFETs are all connected with a short circuit (0ohm), to the pin in the pictures below. This is the only pin connected to the gates of all the chips.

I've looked at the chips, but the protective coating makes it difficult to read the names.

The names of both chips is EZ22601 with an "ST" logo. But I didn't find anything on Google.

The PSU provides [email protected] and [email protected]. Perhaps these two chips are for regulating those two voltages?

Is this the correct voltage sensing pin?

In the other tutorials, they seem to just slap on a different voltage. But since the resistance is ~0 ohm, wouldn't it be necessary to isolate the pin from the 12V sense?

---

UPDATE: I've identified some additional information.

Answer 1

I would suggest:

• Feeding a low voltage into the 12V output and probing for points that are always equal to (or at a fraction of) the applied voltage.
• Finding how these points are connected to the output; they will likely be connected via a divider.
• Somehow tweaking the ratio (i.e. by cutting the line and inserting another divider) and seeing how the circuit reacts. It should bump up the output voltage as you divide the feedback.

Answer 2

There is never a direct connection between 12V and driver chip, since they operate often on 2.5V reference and need a voltage divider, so you cannot use a multimeter to test short circuit.

You could however use an oscilloscope and display on one trace the bare 12V output and on another trace test every pin: the sensing one will have a different scale but should have the same time-behaviour of the corresponding sensing pin. From there it should be easy to find the voltage divider.

Keep in mind that there could be a filter in the sensing path between 12V and chip pin, besides the voltage divider, so you may find a cleaner trace on the sensing pin.

You could also look blindly for a thin PCB trace from the 12V output going back to somewhere near the chip and add a small resistor in it, it should cause an increase of the final voltage. If it doesn't, it's the wrong trace :)