Problems with 12V and 5V lines on a PC ATX supply

Sorry for the mention of sorry the crosspost, that was for the (rather dead) electronics groups :-)

On el cheapo SMPS the primary supply is the 5V whose feedback controls the conversion transformer. The 12V supply is derived from the same transformer so if the 5V is not breaking into a sweat not enough current is available in the transformer secondary to get the 12volts up to regulation.

Ok, I do have a degree in electronics, but it was a very long time ago. Can you go into more detail?

Can't be arsed. Have you heard of Google?

You already know the answer, put a load on the 5 volt line and all will be well.

Yip, tried it, not the sort of thing that seems to have been discussed much. I found a discussion on stackexchange, but all anyone said was "The ATX standard requires the power supply to be within 10% of the correct voltage, at ANY loads". So clearly my cheap PSUs are not adhering to the spec, and assuming I'll take 5V from them aswell, but I want to know why this happens and why anyone would design a PSU in such a weird way. The mark space ratio to feed the 12V line and to feed the 5V line have to be seperate, to allow for varying loads on each as the computer runs. So how on earth could they be linked?

I know the workaround, I want to know WHY this makes it work.

Maybe because its not well designed.Switch mode supplies tend to be designed with a minimum load in mind in my experience. Brian

AS I said, its all part of the fact that its switch mode, and the regulation is not a classic series regulator, its a feedback loop controlling the oscillation of the switch mode part, its all interconnected. I first encountered this weirdness many years ago when a pc I had had some tacky bright leds all down both sides of the front, making it look terrible and for me making my failing sight rebel. So I thought, simples, unconnect them. It was then that the pc decided it either would not start correctly or would not shut down an and lots of tother things went weird. Reconnecting the leds and all was well again. In the end I hid the leds in the case. Out of sight out of mind. I expect the spiders enjoyed it. Brian

I can understand it going off with a high load, but zero or low load is just weird.

That's weird. Surely the LEDs weren't the only thing drawing power at that voltage?

There's a group sci.electronics.design that still has life. They might like the diversion from arguing the merits of the electoral college.

The scottish wanker loaded the same question to all the electronic groups at the same time. Playing off one set of groups against the other.

Playing off? WTF are you talking about? Since I don't inhabit those groups, I don't know which have any traffic or anyone who can answer, hence I asked in all the relevant ones.

I shall try it.

You are asking all the same f****ng questions as you have done here - and amazingly getting all the same answers. But still you persist.

Of course I'm asking the same questions in different groups, to perhaps find someone who has the intelligence to answer them.

Did you get your "Does a parrot's foot conduct electricity?" post answered on the s.e.design group?

I understand alright - that designers are morons.