14-pin ATX Power Supply pinout

[noccy]

Hi all,

I am trying to put an old Compaq power supply to use as a lab power supply. The problem however is that it is in fact a Compaq supply. I've been going over pinouts and found lots of information on 20 and 24 pin ATX connectors, but this specific one only has 14 pins plus an added 4 totalling 18 pin.

What is of relevance for me is being able to switch it on and off (which according to the pinout info for 20/24 pin is the green wire) and the clean power status indicator (which is supposed to be the grey wire which is absent)

Anyone that has got a clue on how to hack this bugger so that it is usable for my experimentation?

- Chris

 

[davenn]

obvious question.... does it switch on as it is ?
maybe being an older PSU it didnt have/need a clean power status. ?

Dave

 

[noccy]

It worked when it was fitted inside a Pentium 4-based computer. The motherboard however was covered in leaking capacitors so I decided to chuck the computer salvaging all usable components including the power supply.

As for it working outside of the computer, I still have not tried. It only has one green lead, and I hope that it is the correct one for powering it up. If so I intend to hook the +5V, +3.3V and +12V to three groups of lab connectors together with a mechanical power switch (linking the PS_ON line with GND).

My plans go as far as adding a two color led to it indicating that it is live and also powered up (if I can find the 5VSB line, normally purple).

Is there any risk of serious electric shock from any of the leads going out of the power supply? If not I will ambush it with the soldering iron tomorrow to test it out. Ideally the power supply will be mounted under my workbench with the lab connectors, power switch and led in an external enclosure (12x7x3.5 cm) atop the workbench.

 

[davenn]

no, no electric shock worries, not at 5V and 12V

but just rememvber you dont want to short cct any of the rails to ground there will be some serious sparks. the +5 and +12 rails will have considerable current capabilities

Dave

 

[Resqueline]

If you're unsure if the green wire is PS_ON or +5VSB then try to short it to ground with a 100 ohm resistor. If it's +5VSB it won't burn, if it's PS_ON it'll still turn on.

As long as you hook it up to a grounded outlet using a grounded cord it should be safe. Ungrounded these PSU's tend to have an unpleasantly high leakage current.

 

[noccy]

Ah, thank you guys. I shall test it tomorrow. The outlets here are not grounded tho. But I should be OK as long as I don't touch ground right? Have received a few zaps from my computer through the days when I've accidentally put my feet on the (grounded) radiator while touching the case.

 

[Resqueline]

Hehe, birds perching on high tension power lines survives, don't they..
The leakage current is not lethal (unless a lightning strikes), but you might consider adding a grounding wire from the PSU to that nearby radiator, for comfort.

 

[(*steve*)]

Yeah, I've had a few SMPS units in the past that gave quite a tingle if you touched their output.

They never inspired me with confidence.

(I have one that I must measure one day. If it were not for the fact that I'm still alive, I would swear it is floating at mains potential.)

 

[noccy]

Haha that's not too reassuring, the grounding wire is definitely a must. I don't intend this hack to be my *last* project

Any ideas on how to implement a standby/power led to this death trap? And also, would a plastic case perhaps be handy for enclosing the actual psu in?

 

[Resqueline]

If there's no +5Vsb wire you might still be able to draw a couple of mA from the PS_ON wire without the PSU turning on. This is sufficient for a high efficiency red LED.
Test and make sure the PSU is also able to turn off with this parasitic load present.
Then wire a green LED from +5V for use as a power on indication. If you want to use a single red/green LED make sure it's a common cathode type.
I wouldn't bother with an extra plastic case.

 

[(*steve*)]

I am trying to put an old Compaq power supply to use as a lab power supply. The problem however is that it is in fact a Compaq supply. I've been going over pinouts and found lots of information on 20 and 24 pin ATX connectors, but this specific one only has 14 pins plus an added 4 totalling 18 pin.

Did you eventually find a pinout?

I just googled for "Compaq PDP117p pinout" and found this.

Of special note is this:

Power-On on Compaq REQUIRES high signal to switch on!

Lots of valuable info on that page.

And http://pinouts.ru/ is a really good site for things like this. It was one that had slipped my mind.

 

[noccy]

Thank you, I failed to dig up a proper pinout myself. I am picking up the box and lab connectors today so fingers crossed I will get going tonite or tomorrow.

Could you please explain what is ment with "requires high signal to switch on" and how it can be applied to this PSU? What line am I supposed to snatch the voltage from in order to power it up? Is there an easier way than building the referenced inverter?

I apologize for all my newbie questions, but I am still at the rank of hobbyist when it comes to electronics Trying my best to learn tho

 

[(*steve*)]

TCould you please explain what is ment with "requires high signal to switch on"

It means you have to connect it to +5V rather than ground.

and how it can be applied to this PSU? What line am I supposed to snatch the voltage from in order to power it up? Is there an easier way than building the referenced inverter?

You connect PSON with +5VSB to turn it on.

This appears to be 4 and 5 (white and green). Try it using a 100 ohm resistor first.

pin 4 (white) should have +5V on it when the power supply is plugged in, but powered down. (Measure it against one of the black (ground) wires).

 

[solarpanellessons]

I remember doing a project like this about two years ago

 

[noccy]

I have now rounded up all the accessories needed including lab plugs and a nice black box for the connections. So I'm ready to get to work

This appears to be 4 and 5 (white and green). Try it using a 100 ohm resistor first.

Out of curiosity, what made you jump to 100 ohm? Reason I'm asking is the only ones I managed to dig out of my kit is 300 ohm and 50 ohm.

 

[(*steve*)]

Try 300 first. If that fails, try 50. 100 was a value I pulled from my... imagination.

You just want to have a value low enough to trigger it, but high enough not to damage anything if you're wrong

If 'm right, you will only need a momentary contact to start the power supply up.

Make sure the fans in the power supply start up as well.

 

[noccy]

Try 300 first. If that fails, try 50. 100 was a value I pulled from my... imagination.

You just want to have a value low enough to trigger it, but high enough not to damage anything if you're wrong

If 'm right, you will only need a momentary contact to start the power supply up.

Make sure the fans in the power supply start up as well.

Good news! The green lead measures +5V with power applied, and the supply (and the fan) powers up nicely with the 300 ohm resistor. So so far so good. Now to get the power and standby leds in there.

The power led I can pull from one of the brown leds (+3,3V), and as my selected power switch got two states, it should be safe to put the standby led between green (+5V) and ground when the switch is in the off state. Am I right?

About to start fitting everything in the box now. Some of the leads going from the psu are massively thick, in some cases too thick to fit the solder holes of the connectors. Is it safe attach thinner leads to the connectors and then joining them with the thicker leads using solder and shrink wrap tubing?

Finally getting somewhere with this project. Thank you for helping me get this far

 

[(*steve*)]

Remember that you will need a resistor in series with your LED or it will probably give a single bright flash...

Also, once the PSU has powered up, measure the voltage on all the outputs. You may need to place a load on one or more of them to allow the outputs to regulate correctly.

Measure the current used to turn the power supply on (insert a multimeter in a mA mode in series with the 300 ohm resistor. If there is a very low current (under 1 or 2 mA) then you can try without a resistor in place. It is most likely that a resistor is not actually needed (but was a safe way to start)

 

[noccy]

Made some progress and also encountered some problems. The progress can be seen in my Picasa album; https://picasaweb.google.com/cvagnetoft/PowerBox

For the problems tho - it seems the power supply doesn't want to stay on (with the 300 ohm resistor). Possibly due to the aforementioned lack of load on any of the outputs. Also wondering if the +3.3V aux lead (brown) might be more suitable for powering it up?

There is also a spike of voltage on the +5V outputs I hijacked for the sockets you can see in the pictures as I hook up the power to the psu. This spike lingers for quite a while and it seems it doesn't ever really drop to zero.

I haven't got around to adding the leds yet. Still digging for suiting resistors, I have burnt (flashed) enough of those

 

[(*steve*)]

Do you have any load on the 5V rail?

Try placing a load on it. At least a couple of Watts.

It is quite possible that the PSU requires the signal to be maintained to keep the power supply on.

Do you have 3.3V standby power? or just 5V?