Bad Supplies: linkage of +5V & +12V

[relkin67]

I've got a couple dodgy power supplies & I've observed that +5V and +12V seem to be correlated but at opposite extremes of the ATX specification. In one I read +5V=5.3 and +12V=11.5 and in the other (truly bad) one +5V=5.6 and +12V=10.8.

This is just for my interest, mind you, but I'm curious if this is the normal modus operandi for poor designs and, if you could indulge my curiosity, why this is the case?

[starfury1]

Re: Bad Supplies: linkage of +5V & +12V

well from what I understand you will find that there is a relationship.

Regulation being tied (Referenced) to the +5 Volt rail, so what happens there will effect the 12 Volt rail

Exactly how this is done I don't know off hand.

I do remember round here there is some info on it

if this is only done with poor or cheap supplies I dont know either for sure

Supplies tend to take a basic form depending on the era and I guess most of the cheaper ones would follow the same script so to speak

Some would probably have some changes to improve things like protection etc

But I dont think you wold see anything major in design till you really get up to the expensive ones build by supply manufactures that know there game and charge you heaps.

anyway if you what a read on the psu, try here

Anatomy of Switching Power Supplies

http://www.hardwaresecrets.com/article/327/1

here as link to a schematic

http://spblan.narod.ru/bp/shema/iwp300a2.gif

do have another just cant find the link but Ive posted around this site before
it uses the chip refereed to in the article above.


sure others should have some good input

HTH cheers

[linuxguru]

Re: Bad Supplies: linkage of +5V & +12V

The sense voltages from +5v and +12v are divided down using a resistive divider, then summed to get a feedback voltage that's compared against a reference voltage. If one of the rails drifts higher (due to a bad cap, for instance), the PWM controller adjusts the duty cycle so that the feedback voltage remains constant - this usually means that the other rail has to drift lower. Usually, if +5v drifts higher, +12v will drift lower, and vice versa. How much it can drift apart depends on the cross-regulation imposed by the coupling between the windings of the secondary inductor (which are wound on a common toroid).

[relkin67]

Re: Bad Supplies: linkage of +5V & +12V

Are just the O/P filtering caps [one of the] causes for voltages to rise or can others be guilty (such as the myriad of smaller electrolytics or perhaps even the primaries)?

[999999999]

Re: Bad Supplies: linkage of +5V & +12V

The primaries will not cause this. The myriad of smaller electrolytics generally won't either, unless one was shorting and it happened to be on the rail that's too low, but with a small cap to pull down a rail that much, great internal heat should have vented it already or made it burst if no vent is present.

Generally the problem is as linuxguru mentioned, if the particular supply's true current output (ratio) is not very well matched to the current demands of the system. In other words, it seems most likely you have supplies biased towards more 5V current but systems not so biased in their usage.

One potential way to differentiate between these factors is to run the supply with no load, or at least minimum load needed to keep PSu running and stable. For example a hard drive and an optical drive connected (then short PS-On to Gnd to make it run), you can measure voltages in this scenario to see if they're closer to what you'd expect. If there were a bad cap shorting to pull down a rail, it would still be doing so to a measurable extent in this test, unless a matter of changes in temp exacerbating the problem..

[gonzo0815]

Re: Bad Supplies: linkage of +5V & +12V

I would second, what is already said. Cross regulation is on of the problems most PSU`s have, even if they have no cap related problems.

I once used a FSP ATX 2.0 OEM PSU, which was intended for 12v powered AMD K8 systems with an Asrock K7S8XE+, a purely 5v powered SKT A board.
The 12v was going straight up to about 13v, as this psu was not intended for a low 12v load and a incredibly high 5v load.

The systems was running over a week until i noticed this issue. I was very happy, that my poor Maxtor hdd has survived this.
From the datasheet i learned, that this PSu`s OVP would trip at 15v.

If a cap can`t hold the voltage stable due degradation, this would signal the controller chip that more power is needed.
But this can lead to any thing from high to low rails, depending on the regulation circuit and on which rail the badcap is.