How to build a $20 47A power supply

[BrianG]

Well, that's weird. I did a quick search and the stickers I saw didn't say that. It should work fine then.

If you have a positive and negative supply rated for the same current, you can hook up your load across the two; load+ to supply+ and load ground to supply-. You see that all the time with regular power transformers: you'll have your main outputs, but then also a center tap. If you use the center tap you get a + and - supply. But if you just use the main outputs (not center tap), you get the full voltage. RadioShacks transformers are like this. Like their 12v models; they're rated 12.6v, but if you use the center tap you get +6.3v and -6.3v.

[BrianG]

Originally Posted by snellemin

I figured I post this up for you guys that want an affordable 12.5V 47A capable power supply to power your chargers.

The model number is power supply in question is 321632-001, DPS-600PB HP 575W Power Supply Proliant DL380 G4. All you need to do for the PSU to "Turn on", is short 3 connectors. Basically two hot wires to ground. No load resistor is needed.
When you look at the back of the PSU, you'll find 4 blade plugs. The two on the left are ground and the other two on the right positive. In between the blade plugs, you'll find a set of 12 pin connectors. You'll see a pin that is shorter than the rest, which I call pin 6.

1 2 3
4 5 6<---short pin
7 8 9
10 11 12

Connect 6 and 10 to 8(ground). I believe pin 5 is ground as well. Note that the PSU fan will turn on when you connect it to the AC line. The fan will speed up, once you short the 3 pins together. I opted for plugs to the pins, but you can use a switch instead. You can use servo leads as well.

Solder your power wires to the blade pins and have fun charging.

I picked one of these up for $25 shipped and it came in today. And of course I opened it up as soon as I got home. Bad news is that there is no adjustment pot inside; I even looked under the heatsinks. The good news is that I was able to boost the output to 12.95v. Not much, but still some.

All I did was create a voltage divider consisting of a 220 ohm resistor and a diode. Tied the + output to the resistor, the other end of the resistor to the anode of the diode, and the cathode of the diode to the - output. I then tied a wire from the diode/resistor connection to pin 5, which feeds ~0.7v to pin 5. Looks like pin 5 is not a ground, but a feedback line to boost the output if there is voltage drop due to heavy current. This resistor/diode trick just tricks the PS into thinking it needs to output higher voltage.

[snellemin]

Good stuff Brian.

[Metallover]

I decided I would go about building a 24v power supply. There's a guide on hooking up two power supplies in series here.

I got my power supplies from ebay yesterday and wired em up today. Now I have a 24v 1150w power supply that I paid $26 for.

The final unit is 13x7" and the output measures 25.2v. If I put the leads of my multimeter on each power supply, I get 12v between the two. I haven't touched it yet, and I don't know if it will shock me if I touch it.

I cut some pegboard to mount the power supplies on.


Testing - The power supply on the right is wrapped in clear packaging tape.


I mounted the power supplies to the pegboard with some bolts, M3 I believe. I then used some unshrunk heat shrink (1" wide) and duct tape to try to insulate the screws.


And my awesome duct tape job.


25.2v output under no load.

[snellemin]

That is some serious power supply you got there mang. I'm now thinking to do the same, but with both 12V and 24V capability. Good Job!

[Metallover]

Thanks! When hooking the power supplies up in parallel is it the same procedure as when hooking them up in series? Just change the output leads?

And I just realized you save about $200 in making one of these power supplies!

[BrianG]

Little update on boosting the output for those interested. I ended up using a couple of resistors to make the voltage divider instead of a diode; a 210 ohm and 20 ohm. This gets me ~13.2v. I was able to get to 13.6v, but the supply sometimes wouldn't turn on or stay on. 13.5v worked fine, but I brought it down to 13.2v just for a little extra wiggle room.


First, the stock power daughter board. There is a blue and purple wire which I cut. One is a 5v line, the other is a -12v line.



Then, I cut away where the 5v and -12v lines were to make room to run the wires for the voltage booster resistors. I soldered pins 6 and 10 directly to ground to turn the PS on. Even when these aren't connected, the PS is on somewhat so I figured I'd just hardwire them and use the mains plug as a power switch. I then soldered wires to a ground point, +12v point, and to pin 5 (v adjust pin).



On the other side of the daughter board, I attached those three wires to the resistors and hotglued the resistor pack to the PCB. I did have to remove the bolt lug from the case so the resistors would fit.



Here is a simple schematic to show how to arrange the resistors. The power ratings of the resistors, especially the 210 ohm one, should be 2w or more.



As you can see, I soldered the binding posts directly to the blade plugs. The soldering job is good, it just looks crappy because as I soldered one side, it would drip down. Since the pins themselves weren't being used, I ground them all down.

Here is another pic showing what I know of the pins:

[BrianG]

Another little update for this little PS unit.

I got another one of these and was doing my voltage boost mod when I accidentally discovered the ground connection for the sense/feedback circuit is not even needed. After doing some more playing around, a 1.5k ohm resistor between the main 12v terminal and pin 5 is all that is needed to boost the output to 13.2v.

BTW, the max voltage you can get out of this is around 13.5v. At 13.6v, the PS shuts down.

Also, for those that think the fan is a bit loud on 12v: If you run this PS at 25A or more for extended periods (>1/2 hour), the air coming out the back is pretty toasty. I figure even if this unit is 90% efficient, 35A @ 13.2v is 528w which will produce 53w of heat. That's quite a bit, so if you do turn down the fan, I wouldn't go too far.

[snellemin]

Wow there Brian. You like to eek out every little bit of power out of that PSU heh?!

[BrianG]

lol, why not? Gotta make the most out of $25!

[florianz]

hi there,

after I converted an other pc p/s last weekend (quite easy), I have now two p/s for my chargers, which is good.

finally I got one of those, but wonder if anyone knows about them, I got one of them for 2,5 euro...



with a maximum of 75 amps on 12v I will be safe for the next years.

thanks
florian

[Nard Cox]

Those look badass!!
930W / 75.7A ... watch out LiPo's :D

Have you searched on RC Tech? They have several of these server PSU conversion threads.

[florianz]

thanks for that hint, I'll have a look

got them on egay, used, private seller, I hope they will work...

[snellemin]

This is a good thread. I am glad to see that I'm not the only fool out there looking for some true clean dirt power source for a charger.

[BrianG]

At some point it gets a bit overkill to have such a large 12v supply unless you have multiple chargers to power at the same time. Many (if not all) chargers over 300w require a supply with higher voltage (like 24v) to be able to use it to its max output. Just a little food for thought.