In an earlier thread entitled ‘Wanted, 24-way PCB plug’, I made a reference to building a test rig for testing PC PSUs.
In one post, by way of a response,
Mike Newcombe wrote:
”When you complete the Project, please post details including some photos.
I have long thought that a voltmeter alone is insufficient and only good as a guide, especially when testing PSUs etc.”
O.K.: here it is!
First, I would like to state what it isn't!
It isn't an all singing, all dancing universal tester for every make, style, design and build of every PC PSU known to exist! So it has various limitations - and that is because it was primarily designed to test PSUs that are installed in various PCs here, plus PSU spares, and the occasional PC that is brought to me with the request "Can you fix this PC, please?"
Second, the cabinetry and build: that's none to brilliant either. That's because this item began as a 'development' on a roughly assembled breadboard - and it slowly morphed into the finished article. Although I try to make my projects look as professional as I can, since it will only ever be used by me, I am settling for that compromise. In brief, it does what I want it to do.
"If you have a suspect PSU, all you need to do is fit a 'known good' one, so why do you need this item?" is a question that could easily be asked. My response to that is that that Q. begs the question: how do you
know that that PSU is a 'good one'? It might be for a different type of PC; it might have an intermittent fault; it might even damage the host PC!
etc., etc.
So, a few comments on its design.
PC PSU types: there are many. This item is designed to cater for those with a 12-pin + 12-pin Molex mobo connector plus the auxiliary 4-pin connector for the CPU. (It will probably be suitable for testing the older 10 + 10 pin PSUs, but this has not been tried at time of writing). The d.c. power connectors for hard drives etc. are always powered from the same rails as those feeding the mobo in a PC, so no provision is made in this test rig for those connectors.
PC PSUs can have different internal arrangements for their + 12 v. rails. If there are 3 yellow wires emanating from the PSU box, it might have only one +12 v. source or may have three. And those three may have separate sources and may have separate over-current protection on each. (Example: the PSU for the HP xw4600 has one source and three O/P rails with separate over-current sensors for each rail). This test rig is designed for use with PSUs featuring 3 separate +12 v. rails, but can also be used for other types.
This rig features fixed loads for the –12 v. and the +5 v. stand-by rails. The load on each is controlled by switches: high, low and off.
Each input voltage rail is fed to each load by fast-acting fuses.
The rig’s front panel has a ‘PSU ON’ switch and an accompanying LED. This is the power-on line from the PSU that is connected to 0v. to switch the PSU on. (Note: some PSUs do not seem to use this – the PSU for the HP xw4600 being one example).
Reliability tests.
With a suitably-rated PC PSU, the rig was set up for 9 amps on each + 12v. rail; 9 amps on the +5 v. rail and 9 amps on the 3.3.v. rail. The rig was then left running for several hours in a room of ambient temperature of 23 deg. C. After that time, the rig was found to be performing normally.
Note the ‘tunnel’ fitted over the heat-sink at the rear of the unit. That combination of heat-sink plus tunnel is fan blown and makes a decisive difference in the temperature of that heat-sink, compared to not having that tunnel and fan fitted.
If you've waded through that lot, thanks for reading it.
I expect various questions & comments will now arise. I am not aware of any serious shortcomings in this item's design, but will appreciate all feedback - and will respond accordingly.
Four photos below; four more in the next post.
Al.