Adapting a Computer PSU

[Mark Hennessy]

Just to say that I haven't forgotten about this. I have located one of these power supplies, and it's in pieces on my desk at work.

I have found that there's actually a 3rd opto-isolator that's not obvious from the photo. Actually, now that I know it's there, it's not that all well hidden really. It's to the left of the transformer for the standby supply, and it's doing the voltage feedback for that. So that means we have two mystery opto-couplers now, not just one. I have theories, but no chance as yet to properly reverse-engineer the circuit to verify them. Give me a couple of days...

[Red to black]

One is possibly for the shutdown/sleep/stby command on a PC PSU, I have to say though that I did not realise this newer generation of PC PSUs were designed this way, it is only looking at your prior posts Mark in this thread, interesting stuff, thanks

[Crackle]

I have been occupying myself with making a box to house the power supply.
It is made from some 5mm ply left over from a door that I panelled.
The round opening is for the PSU fan and the narrow slots on the sides are for air intake.
The large window in the front is where the meter for volts, amps & watts will go, and the on/off switch.
The output binding posts will be fitted through the ply below the meter.
Its currently in primer waiting on final testing for size & fit when I get the meter and the aluminium panel the meter will be mounted on.
       
Mike

[Crackle]

Hi Mark
I was just wondering how it was going.
Thanks
Mike

[Mark Hennessy]

It's been in pieces on my office desk for a couple of weeks at least. Pressures of work means that progress has been disappointingly slow. This power supply is unusually complex, and not all of my initial assumptions were correct. I'm still not sure I know what the TL594 is actually doing. I have to return the PSU in working condition at some point, but if the supply wasn't needed, I would desolder the daughter boards and reverse-engineer them more thoroughly, but as it is, it's really quite hard to follow the circuit. It looks like - and I could be wrong, so don't hold me to this - they shut down the 12V standby supply when the main converter is running to save energy (the 12V SB output is supplied from the main converter rather than the separate SB converter when not in standby). I've read rumours of that happening in modern supplies, but HP were ahead of the game if that's what they're doing with this relatively old PSU.

It looks like the presets are related to the current sensing, which tallies with your earlier experiments. If I ever discover what resistors you need to change to alter the 12V output, I'll let you know, but until then, I wouldn't drill the hole in the front panel for the pot. But then, I don't think my question about supply voltage (how much practical difference does it actually make?) was answered, so in the absence of that, you'll have to forgive me for not giving it a huge amount of priority for the sake of a dB or thereabouts.

[Crackle]

Hi Mark
I estimate the transceiver is putting out 3 watts at 12v, at 14v I estimate it would be 4 watts.
This small increase of 1 watt is amplified by the gain of the linear amp, so it can produce about 50 watts (at 14v) rather than what I estimate is about 35 watts at 12v.
I do appreciate that you are a very busy guy.
Thanks
Mike

[ppppenguin]

It may sound like a lot of watts but it's still only 1.5dB difference.

[Crackle]

That may make the difference to getting over the background noise or being lost in it.

Mike

[ppppenguin]

Under extreme conditions, yes possibly, just about. 1dB is reckoned to be the limit of what you can discern.

[Mark Hennessy]

OK - so that's roughly what I guessed. Thanks...

My curiosity remains piqued, so I'm pretty keen to get to the bottom of how this works, but there's every chance I'll have to give it back before that happens.

One of the frustrating problems is the SMT transistors don't have markings that lead to positive identifications. Some aren't recognised at all. There's been a fair bit of having to figure it out based on the surrounding circuitry, which isn't ideal. Just look at how much circuitry is driving the main switching transistors - it's looking like it's a discrete control scheme. Having had another look over lunch, I'm increasingly sure the TL594 is just controlling the fan. Kudos to whoever did the PCB design - they did a good job of fitting it all onto a single-sided layout, but it's not easy to reverse-engineer.