5v rail to high

[Drack]

Hi guys, Merry x-mas
My psu( https://www.badcaps.net/forum/showth...781#post702781 ) is having a litte problem, the 5v rail is getting a bit high(5.280v the max allowed by ATX specs is 5.250v). i Know 30mv is not huge, but i want to Know what do you guys think about it

[mariushm]

Re: 5v rail to high

It's due to the design of the power supply.
It simply expects you'd have more stuff using 5v in your computer but you don't because modern computers use mostly 12v nowadays.

5.280v is nothing to worry about, most components that use 5v will be fine with up to 5.5v

[budm]

Re: 5v rail to high

Also how well is your meter calibrated?

[Drack]

Re: 5v rail to high

I measured with a yaxun dmm AMD with Asrock oc utility both said the sale thats only when i push the 12v rail, i know, its group regulated thats why its that high, i have 2 delta fans on 5v, but when those fans are not connected the 5v rail will go down to 5.256v/5.236, I would put the fans on 12v but they are to loud for my taste at full blast(42db)

[Pentium4]

Re: 5v rail to high

Are they molex fans? What I do with those is just cut the diagonal part off of the molex connector and flip it around, so the fan runs at 5V. Most fans still move a decent amount of air at 5V and are A LOT more quiet. This will also satisfy the 5V rail a little bit because it will be giving it something to do.

[Drack]

Re: 5v rail to high

Sorry I didn't write what I wanted to, my mind was in another place, what I wanted to say is that the 2 deltas are on the 5v rail, and since I put them in the voltage went up, and yes they move a lot of air at 5v one of them is a 110 cfm fan, it's a copy of a delta afb but it works really fine, the other one is a 80mm intel fan Delta guetto mounted on the side panel

[Longbow]

Re: 5v rail to high

Well, the light load theory is a stab in the dark with these supplies. Regulated 5 volt supplies should be within .1 volts. The whole principle of regulating a voltage is that the voltage output remains constant whilst the load current varies. How close are the other voltages, by the way?

[Drack]

Re: 5v rail to high

12v is beetwen 11.768/11.840, 3.3v is at 3.36v, it's an old half bridge and group regulated PSU with 30a rectifiers on all rails, so you can't ask for much but is well built and it works OK for 300w

[PeteS in CA]

Re: 5v rail to high

It's possible that it was set wrong, but the +12V being low like that - close to the spec limit, IIRC - suggests the system load is 12V heavy and 5V light.

[Drack]

Re: 5v rail to high

I know, maybe 4 amps on 5v and like 24/26A on 12v, the minimum allowed by ATX specs is 11.4v so nothing to be worried about

[momaka]

Re: 5v rail to high

It's possible that it was set wrong, but the +12V being low like that - close to the spec limit, IIRC - suggests the system load is 12V heavy and 5V light.

11.7-11.8V on the 12V rail is not that low at all. My 250 HiPro PSUs typically run in that ballpark too with a high-TDP Pentium 4 CPU, like the 89 Watt Prescotts.

What I am more worried about is the high 5V rail. Some motherboards do NOT regulate power going to the LPC/SuperIO chip, and those are sometimes rated for 5.5V max. So staying close to 5V is better.

One easy way to lower the 5V rail is to put some 20-40 Watt car light bulbs on the 5V rail. The increased load should lower it a bit, hopefully, and possibly bring the 12V rail output up.

Another method would be to tweak the feedback in the PSU to lower the output voltages a bit. That way, both the 5V and the 12V rails will get lowered.

And yet another way... though this one is a bit more experimental... would be to change the 5V Schottky rectifier in the PSU with a fast-recovery type. Something like a F16C20 should have a much higher voltage drop and will bring the 5V rail down. Of course, this will also bring down the PSU efficiency and make it run a bit warmer too. But if you are only going to use this PSU with a 12V-heavy / 5V-light computers, then that should not be a problem.

[Drack]

Re: 5v rail to high

Remember it has an 35 transformer so I think that would be too much load for it, and I would like to not disassemble this PSU because if it brokes I don't have a way to buy a new one, I'm only 17 so I can't find a job at this moment and that one was a gift

[Drack]

Re: 5v rail to high

I was watching the PSU and I saw that it does not have a feedback cable for 5v that might be why is to high because it's not monitored
And I don't know why but when I putted the fan on the 5v rail it went up and not down(from 5.256 to 5.280) but that's only at full load on the CPU and GPU. It was the opposite of what I was expecting... (Lol I think)

[cpt.charlie]

Re: 5v rail to high

It's normal, "modern" CPU and GPU take the power of the 12V rail (even graphics cards without 6/8pin power connectors), so, in a group regulated PSU expect a voltage increase at 5V line, some load at 5V line should keep things under control, but you could consider other options like undervolting to decrease the power requeriments.

[Drack]

Re: 5v rail to high

That's what is weird, when I putted more load on 5v the voltage went up, and I mean 5v, so I don't really know what to think, at idle 5v is at 5.236v that's a bit high already, what I can see over here is that if it's not touching 5.5v it should be OK.
Sorry if I'm bothering you guys I just don't want that cheap PSU killing my HDD or mobo
Ps. My Mobo is a ASRock g41m vs3 rev.2 by the way

[Drack]

Re: 5v rail to high

Someone?

[Longbow]

Re: 5v rail to high

Think about it. I would start with the reference, which is typically in the secondary. In your case it monitors the 12 volt line. Don't forget to look at the ripple on the 12v. and 5v outputs. Are you really measuring a d.c. level?

[Drack]

Re: 5v rail to high

I don't have an oscilloscope, and almost anybody in Venezuela have one, I measured with the DMM set to 20v d.c...

[Drack]

Re: 5v rail to high

I don't think ripple is too high, it has 3 caps and one pi filter per rail

[momaka]

Re: 5v rail to high

Also, if you really want the 5V rail to go down, you will need to put a much bigger load on it than just a fan or two. More like car light bulbs that eat up to 3-5 Amps total.

But I guess if everything is running stable, then just let it be then and don't modify anything. I doubt you will get any catastrophic failures.