Look at my 5v rail !!

[YviE]

We're having really hot weather here, outside it's about 45°c and that's not directly under the sun .....

Look at the 5v rail :

http://www.theforumisdown.com/uploadfiles/0103/Superlow.jpg

Is the hot weather making my PSU to function less efficiently?
(room temps approx. 32°c)

Haven't experienced any crashes and for now the CPU is at 12.5 x 166Mhz. NB gets too hot and i get BSOD.

[bigben2k]

I can't say it enough: if MBM shows that a rail is out of spec (more than 5% off), then what you have is a reason to double check the voltage with a meter. MBM will not guarantee you an accurate reading.

The mobo's power connector can drop a fraction of a volt, all by itself, so keeping it clean always helps.

[rogerdugans]

I am with bigben2k on this: check with a multimeter and verify voltages on the 5 and 12 volt lines. If they are vastly different than what is reported in software (including bios) check that your atx connector is clean and tight.

All that said, it IS entirely possible that the heat is affecting your psu or the voltage regulation circuitry on the mobo.

PSUs have fans to keep things cool because too much heat affects them; one of the fairly common mods around is putting heatsinks on the voltage mosfets on the mobo: some get higher readings, others just get more stable voltages at the mobo.

[wannaoc]

Stupid question: What should the low/high and average be?

[Oklahoma Wolf]

Double check with a meter - my system hits lows of 4.50v and is always stable (meter shows 5.03v at the PSU).

[emericanchaos]

supposively the limit for your rails is 10% change in either direction. 4.5 is exactly 10% off of 5 so you're pushing it. my PSU never goes under 4.8. currently reads 5.04.

i'm running a true power 550 and my board is an nf7-s rev 2.0.

[Oklahoma Wolf]

Originally posted by emericanchaos
supposively the limit for your rails is 10% change in either direction. 4.5 is exactly 10% off of 5 so you're pushing it.

My PSU always shows 5.03 at the multimeter load or idle with the occasional flicker to 5.04 - the low 5v reading in BIOS/MBM is a function of the motherboard circuitry only, thus my usual recommendation to check the PSU rails manually with a good quality multimeter. Since I'm an electronics tech, I have 3 good multimeters (well, one's a Radio Shack so maybe not so good that one ;) )

Incidentally my board seems to have no control over vcore - I may send it back soon. It's stuck at 1.85v, so I used that as an excuse to push the old Duron to 1.4 - my 5v changes with how hard I push the processor indicating the AK79G-Max runs the proc off the 5v.

[YviE]

Originally posted by Oklahoma Wolf
my 5v changes with how hard I push the processor indicating the AK79G-Max runs the proc off the 5v.

My mobo does the same. The higher the CPU clock speed/voltage, the lower my 5v reading (and the higher the 12v reading ....... that's what you get when you don't have a True Power PSU )

[Oklahoma Wolf]

Hehehe - Truepowers are nice :)

I should clarify - my 5v changes in BIOS only - on a meter the PSU (this is the OEM version of the Antec pp412x BTW) rarely moves from 5.03 though back when I had my old MSI KT133 board it read 5.07. Overclocked or not, on this board 99% of the time I get 5.03 and 12.33 at the PSU. Haven't checked the 3.3 in a while but even in MBM that one's stable.

I really am questioning the wisdom of getting this board - the onboard TV out is nice, but the CPU multipliers and vcore are not adjustable - that was important to me :(

[YviE]

Originally posted by Oklahoma Wolf
I really am questioning the wisdom of getting this board - the onboard TV out is nice, but the CPU multipliers and vcore are not adjustable - that was important to me :(

Yeah, nowadays it's out of question getting a mainboard without adjustable multipliers, vcore, etc.

[wannaoc]

I guess its time for me to get a new PSU, my 5v never goes over 4.85. :(

[YviE]

Yeah, i think i'm gonna get a Fortron 350W with 12cm fan.
I've heard a lot of good stuff about this PSU. So i guess it would power the system in my sig quite well.

[emericanchaos]

the 350 is a good unit. it's basically the same as the 400w which explains why the 350 can do over 400w. the 5v on the 400 is rated 8a higher. other than that it's the same.

[YviE]

Originally posted by emericanchaos
the 350 is a good unit. it's basically the same as the 400w which explains why the 350 can do over 400w. the 5v on the 400 is rated 8a higher. other than that it's the same.

8a, that seems a lot to me. Does the 400W really deliver those extra 8a?

Do the 350W & 400W units have the same construction (same heatsinks, capacitors etc)?

[q149]

Always use a multimeter if you want to be sure. NO voltages in my bios or any program are correct and some are sooo off it's not even funny.

[larva]

Originally posted by YviE


8a, that seems a lot to me. Does the 400W really deliver those extra 8a?

Do the 350W & 400W units have the same construction (same heatsinks, capacitors etc)?

The 400W model does make more 5V for sure. It's just that most systems load the 12V most heavily and that is the same between them. Since your system is hard on the 5V you may want to spring for the 400W model, but it is pricey and there is no 120mm fan version. The 350W model is still very capable, but there is no doubt the 400W version has a stronger 5V rail.

[Gage8]

n00b question...where do you check the 5v rail with the multimeter?

[Cjwinnit]

Originally posted by Gage8
n00b question...where do you check the 5v rail with the multimeter?

Connect the multimeter between a red wire and a black wire from the PSU. Wires from a molex is fine.

[Gage8]

i see, the molex connector carries both 12v & 5v...learning every day.

[Cjwinnit]

Originally posted by Gage8
i see, the molex connector carries both 12v & 5v...learning every day.

Yep. That's why you can do the "7v mod" trick. Connect a fan to 12v and 5v to make it quieter :)

[q149]

Or 5v. If you want to measuare 3.3v jam the multimeter in the orange wire on the end of the ATX connecter thing.

[thorilan]

hey i got an idea . someone make a guide using pics of a multi meter and simple laymens terms to show people how and where to check voltage
( yes i know that you can use the molex but not many people know how to check 3.3v power and it can also help teach people the ideas behind 7v mods)

here is a link that will help people understand a little better . its an antec true power link but it gives a few bits of insite into what to look for in a psu

http://www.antec-inc.com/info_DIYArticle1.html

[YviE]

Good idea! I'm one of those that doesn't know how to measure all power lines....

Anyway, i unplugged the power connector and tried to clean it. It wasn't dusty or anything, but when i plugged it back in, my 5v reading was 0.3v higher (in the bios). I haven't checked with a multimeter yet, but this is always better than before.

[bigben2k]

Originally posted by wannaoc
Stupid question: What should the low/high and average be?
As per SSI specifications, all voltages have to remain under 5% from their nominal values.

So for 12v, it can vary between 11.4 and 12.6
for 5v, it can vary between 4.75 and 5.25
for 3.3v, it can vary between 3.14 and 3.47

Any deviations, especially on the low side, will lead to lock ups and/or reboots.

Deviations on the high side will start to fry things.:eek:

Now it's normal for the voltage to fluctuate, as the load increases/decreases, so if your "low" reading is a little bit off, it's probably OK, as long as your system doesn't lock up.

If your average falls outside of these specs, it's time to take out the multimeter. If it's still out of spec, it's time to replace (or tweak ;) ) the PSU.

[bigben2k]

Originally posted by YviE
Good idea! I'm one of those that doesn't know how to measure all power lines....

Anyway, i unplugged the power connector and tried to clean it. It wasn't dusty or anything, but when i plugged it back in, my 5v reading was 0.3v higher (in the bios). I haven't checked with a multimeter yet, but this is always better than before.
Good show!


The 3.3 line can be measured directly at the mobo's power connector.