5 volt rail at 4.65 volts; wtf?

[qviri]

According to Speedfan, my 5 volt rail oscillates between 4.71 V and 4.65 V. The minimum recorded was 4.57 V, max 4.84 V, mean is 4.68 V. Seeing as the 5% tolerance limit would be 4.75 V, I'm kind of concerned. The 3.3 V and 12 V rails are within 3% of rated.

The PSU is a mATX Seasonic SS-200SFD, fanswapped with a Zalman ZM-F1, which in turn is turned down quite low using the board's fan header and Speedfan. However, neither the exhaust air nor the area immediately the PSU feels hot.

I have a socket A CPU and its power is drawn from the 5 V rail. Then again, it's a 45 W XP-M, undervolted and -clocked, and most of the time it's idling.

On the other hand from the weird voltage values, I just hit nine and a half days of uptime, and previously I've had it up for three weeks or something, so it really doesn't sound like I'm having problems.

I haven't checked with a multimeter yet. Is Speedfan misreading that voltage somehow?

[JazzJackRabbit]

I haven't checked with a multimeter yet. Is Speedfan misreading that voltage somehow?

It might be, although in my experiece on-board voltage sensors are pretty accurate within 0.1 of what I measured with real voltmeter. If you want you can certainly check it yourself with voltmeter. Although I doubt it will change anything? What are you going to do if it's really at 4.6? I've had much lower on 12V line with pico PSU and surprisingly my computer survives. I've only had one shutdown and I'm not really sure if that was PSU fault, could have been anything.

[Redzo]

SpeedFan might be wrong. Onboard sensors are rarely accurate. And that 5v rail changes are normal.

[Techno Pride]

According to Speedfan, my 5 volt rail oscillates between 4.71 V and 4.65 V. The minimum recorded was 4.57 V, max 4.84 V, mean is 4.68 V. Seeing as the 5% tolerance limit would be 4.75 V, I'm kind of concerned. The 3.3 V and 12 V rails are within 3% of rated.

The PSU is a mATX Seasonic SS-200SFD, fanswapped with a Zalman ZM-F1, which in turn is turned down quite low using the board's fan header and Speedfan. However, neither the exhaust air nor the area immediately the PSU feels hot.

I have a socket A CPU and its power is drawn from the 5 V rail. Then again, it's a 45 W XP-M, undervolted and -clocked, and most of the time it's idling.

On the other hand from the weird voltage values, I just hit nine and a half days of uptime, and previously I've had it up for three weeks or something, so it really doesn't sound like I'm having problems.

I haven't checked with a multimeter yet. Is Speedfan misreading that voltage somehow?

same behaviour here.

but i'm on a Bluestorm.

It's got something to do with the voltage regulators. I don't know the technical details, but it's got to do with too much power draw on the +5v line and too little load on the +12v line.

you can *fix* this with a mobo like the Abit NF7 (uses +12v for cpu) or get a power hungry video card.

[Elixer]

If you can run Prime 95 for 24 hours without an error than don't worry about it. Who cares if it's the proper voltage so long as it works. I have doubts about the accuracy of the sensor. Check the voltage in Bios.

[Gojira-X]

qviri: What you describe is a "feature" of all socket A's, 370's, slot A's, Slots 1's and VIA cpu's.

Techno pride is correct in that the boards draw alot of power from the 5V rail. The main reason for this is that the original ATX PSU spec specified alot more current capacity on the 5V rail than current ATX12V and and ATX 12V 2.0 psus can supply.

It is for this lack of capacity that is the reason for the shift to higher rated psus.

In order to sort theis problem (which was made worse when i installed a geforece fx 5700ve, I upgradded first from a 300w Mercury (Kobian) psu to a 400W Mercury (Kobian) psu to the final psu, a 420W Xilence Power (Levicom) unit.

[Techno Pride]

If you can run Prime 95 for 24 hours without an error than don't worry about it. Who cares if it's the proper voltage so long as it works. I have doubts about the accuracy of the sensor. Check the voltage in Bios.

I've been running my computer on a fanswapped Bluestorm for over a year.

I've seen the +5v line go down to 4.6v, +12v line going up to 12.6v. All verified with multimeter.

My computer has not blown up yet.

[qviri]

What are you going to do if it's really at 4.6?

Well, I was planning on reprimending it sternly

Thanks for the replies, everyone. I was a bit worried I was killing my PSU with insufficient airflow. I guess that cross-loading is taking its toll, even though this PSU is ATX1.3. Pretty much the thing running off 12 volt is one 3.5" drive, and whichever motherboard parts are powered off it -- probably 1.5 amps if that.