14-pin ATX Power Supply pinout

[noccy]

I'm back on track again after a few weeks of illness and madness -- those two are unrelated, so *no* mental illness The little control box that is to host the connections have been completely redone. After hooking it up, still without a load, this is what I end up at:

http://media.noccy.com/watch/powerbox-fail - Video is in webm (vp8) format so use a HTML5 capable browser

After a little testing I removed the 300 ohm resistor and went straight on the +3.3V AUX standby power line.

There is an initial spike (that goes past 5V) and the voltage then slowly drops. What would you suggest to use for putting a load on the PSU? Is a high ohm resistor sufficient or should I try to dig up something like a fan or similar?

 

[(*steve*)]

For the 5V line, I would try a 5W 33R resistor, but you could experiment with other loads first to try to find the smallest load which would keep the power supply in regulation.

The load that needs to be applied varies with power supply design, some needing several watts, others needing none at all.

Who knows? That 300 ohm resistor you just removed may be enough.

 

[MattyMatt]

Just so you know, the 14 pin connector isn't really a full blown standard, its more a like a shared proprietary kind of thing.... (IBM used to do this with a few model power supplies... and almost everyone does it with there server power supplies)

http://pinouts.ru/Power/compaq_power_14p_pinout.shtml

This is the pinout it should be... Gotta love Compaq! (not sure if you had this one or not)

As far as the load goes, I watched your video, couldn't really see the numbers, but I get what you were getting at... I can tell you this... the boards that this took, were really picky... and honestly I would start with the 300ohm resistor, may want to go as high as 500, then start dropping down incrementally. Sorry I can't offer much more advise.

MattyMatt-

 

[noccy]

Hey Matt, and thanks. I got that pinout earlier in this thread tho, and that's what it's all been based on And the fact that it was a Compaq PSU is the reason for converting it into a lab supply (as I don't have anything else that it would work with anyway)

I started with a ridiculously high ohm resistor (a few kohm) but with no luck. I'm gonna try with a lower one today, but I don't know what pins to put a load on either. Right now I've only tested between one of the brown and a black lead. Should I try a few other leads or do you think that the +5V/GND leads are treated as one "segment" in the supply?

 

[(*steve*)]

All of the +5V and all of the gnd leads will be connected together.

There are multiple leads to handle the high current.

 

[MattyMatt]

Agreed with Steve, all of the rails in your power supply should be linked together.

 

[noccy]

So I've linked all the +5 lines together, through a resistor, and back to ground. The same thing with the +3.3V lines. Not through the same resistor tho. Gonna try firing it up tomorrow, right now some other important stuff is getting in the way. Got some new pictures for you guys tho so you can see the progress

Can't begin to explain how much I appreciate all the help I'm getting here

Regarding the grounding, can I hook that to the GND lines from the supply or should I just connect it to the case of the PSU?

 

[MattyMatt]

Both the ground leads (should be all of them) and the case should be grounded to the same ground, you can make sure of this with a continuity test or a basic resistance test, you should be getting VERY low resistance (if any) when ground leads are tested against case... there is probably even a lead that goes from all (or most) of the grounds, right to the case... do some checkin' around inside there

Happy hunting, and good luck on the Power Up Test

 

[noccy]

Okay, good news! I am still alive The bad news (thread-wise, not for my sanity) is that I threw out the old Dell PSU and replaced it with a proper 160W ATX PSU, you know, the kind that powers on when you short the green lead with ground, and stays on. So now I have a proper setup for +12 VDC, -12 VDC, +5 VDC (Variable with pot), and +3.3 VDC. So far, it works!

What I am attempting now is to light it up a bit. For this I have three possible sources of power; The purple line delivers +5V standby power, the gray line delivers +5V from what I can tell also like the standby but indicates power being OK, and the red line delivers +5V while the PSU is powered on.

The box is supposed to have 6 leds in total; 1 indicating that the device has standby power, 1 indicating that the power is off, 1 indicating that the power is on (redundant, I know, but I drilled the holes already to fit a scavenged 3-led setup, which I have now discarded), plus 1 led for each set of lines (+5V, +3.3V, and +12V/-12V) indicating that there is something drawing power from it.

In this setup I am not quite sure what resistor values to pick, as well as what power source to use where, and through what transistor (which I assume I will need). I believe the line status can probably be sorted with a resistor and a transistor, bringing just enough power through to make it tap power from an available +5V line, but that still leaves me with a lot of questionmarks.

Until I can figure that out, hopefully with some assistance from you guys, I probably own the geekiest fan power supply in the neighborhood. It's amazing how much a good old 8cm 12V fan can do in the summer heat

Thanks in advance.