Author Topic: Getting 13.8V from a server PSU?  (Read 11856 times)

0 Members and 1 Guest are viewing this topic.

Offline mrfluxTopic starter

  • Contributor
  • Posts: 23
  • Country: bg
Getting 13.8V from a server PSU?
« on: December 19, 2016, 09:02:25 pm »
I'm trying to build a 13.8V/14.4V high-current power supply, to charge a lead acid battery bank.

I've bought two Dell PS-2521-1D (AA23300) which can provide 44.5A 12V and 4.9A 3.3V.

I saw a tutorial which modified this PSU model (AA23300) to an adjustable voltage 13.8-14.2.

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

However, after opening the case, I saw that the PCB design is completely different, yet has the same specs and pin-outs.
But I assume that I can still use the same method as in the tutorial.

Do you have any suggestions as to how I should go about this?

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



I've tried to find the +12V return sense:

With my multimeter, I believe to have identified the PWM pin:
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.



And perhaps the feedback circuit as well for the +12V return sense:

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




With a bit of further analysis, I found that:





Question: How should I go about this?
Should I disconnect the +12V pin in the top?
And put a potentiometer in-between so that I can drop the voltage a bit, such that the input reads 12V when it is actually 13.8V?
 

Offline ekyle

  • Contributor
  • Posts: 24
  • Country: us
Re: Getting 13.8V from a server PSU?
« Reply #1 on: December 21, 2016, 07:50:33 pm »
There should be a voltage divider connected to the sense lines where the middle of the divider connects to the chip in some way. The chip compares this voltage to a reference voltage. Changing the values of the resistors or putting a pot in series with the middle of the voltage divider would allow the voltage to be changed. Also keep in mind, some power supplies have an over voltage "crowbar" that will either short the output or shut it output down if the output voltage goes out of tolerance. One might have to deal with that as well.
 

Offline jpittner

  • Contributor
  • Posts: 14
  • Country: cz
Re: Getting 13.8V from a server PSU?
« Reply #2 on: December 22, 2016, 12:48:57 pm »
Hi,
I have converted several PC power supplies in the 450-500W range to give 13.8V instead of 12V.
Start from the controller chip, usually in newer supplies it is something like SG6105, which
has a comparator input for the control loop, and then protection inputs for 3v3, 5v, 12 rails. Follow the traces from the comparator input which forms a net conencted by various resistors to 0V and all the other rails - measure current through the resistor to 0V and then disconnect the 3v3 and 5v rail resistors and replace the 12v rail resistor with a newly calculated divider. The follow the traces of the chip's protection inputs for 3v3, 5v, 12 rails and apply voltage divisors shifting the inputs back to acceptable ranges. You can also try to just connect them to gnd, with some chips undervoltage will just cause the power good signal to be false, but the controller will work anyway (this might be dangerous though, as overvoltage will not be detected).
Jiri
 

Offline mrfluxTopic starter

  • Contributor
  • Posts: 23
  • Country: bg
Re: Getting 13.8V from a server PSU?
« Reply #3 on: December 25, 2016, 11:42:25 am »
Hi Jiri,

Thanks for the instructions. I've tried for half a day, but with little success.

Any ideas for a better server PSU which is cheap and can easily be modified to at least 14.2V, or preferably up to 16V?
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8248
  • Country: fi
Re: Getting 13.8V from a server PSU?
« Reply #4 on: December 25, 2016, 01:46:48 pm »
To build a battery charger out of a computer PSU, you also need to modify the current limit to work in unlimited Constant Current mode instead of the typical CC with fault limit (hickup or latch mode). Depending on the controller IC, this can range from very easy to fairly difficult.

Without a proper current limit, it still could work if you use oversized PSU on a weak battery, so that the current never goes above the limit and the PSU is in CV mode all the time. In a pinch, adjustable voltage works, you can crank up the voltage manually while measuring current.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf