low rails on PCPC Quad Silencer 750

S

Somnambulator

Gawd
Joined
Apr 13, 2008
Messages
856
let me start out by saying this PCP&C 750 was almost spot-on with my prev Abit IX38 QuadGT under same conditions.

now with the Formula Maximus, 12v rail is steady 11.82, 5v rail is 4.825 with a .5v ripple, and 3.3 is a steady 3.25. all recorded under full load via OCCT 2.0.0a and 2.0.0b.b1.

idle rails, as reported than the sometimes less than accurate Asus PC ProbeII, show as 11.87 12v, 4.87 5v, and 3.25 3v3.

even if these are considered within specification, they are rather low, especially when considering the 5v ripple, and that these were within +- .05 for 12v and +-.075 5v on the prev abit board.

OCCT shows relatively 0 vdroop under load.

is this a common occurrence w/ this asus board? is the board def? or is this well within normal specs for this supposedly excellent supply?

one reason i ask is b/c ive been having trouble getting this quad 100% stable at 3.6ghz even when it was 24hrs prime stable under the abit until the PWMs fried (twice). it failed Prime after 8.5 hrs at 1.41 vcore which was the same as the abit was stable. i did add another 2 sticks RAM when installing the asus, could the added load be stressing the supply or board too much?

thanks
 
Somnambulator said:
let me start out by saying this PCP&C 750 was almost spot-on with my prev Abit IX38 QuadGT under same conditions.

now with the Formula Maximus, 12v rail is steady 11.82, 5v rail is 4.825 with a .5v ripple, and 3.3 is a steady 3.25. all recorded under full load via OCCT 2.0.0a and 2.0.0b.b1.

idle rails, as reported than the sometimes less than accurate Asus PC ProbeII, show as 11.87 12v, 4.87 5v, and 3.25 3v3.

even if these are considered within specification, they are rather low, especially when considering the 5v ripple, and that these were within +- .05 for 12v and +-.075 5v on the prev abit board.
Click to expand...

I am not sure where this terminology has come from but I am seeing it more and more. What you are seeing is not ripple (being software its really just a guess at voltage regulation).
 
Yep. Software is untrustworthy. Until you take a quality multimeter to it, you know next to nothing about the output. Unless it's showing /huge/ fluctuations and accompanying instability, there's no real reason to assume there's anything wrong with the supply.

It's much more likely a BIOS issue, or the added stress on the MCH by adding another 2 sticks of RAM (known to lower max OCs very often), or the vdroop being higher than you're seeing.

I'd guess it's a combination of more RAM and the fact that pretty much /no/ board actually has 0.00 vdroop. No vdroop under load is even /less/ likely.

To test, take out 2 sticks of RAM (the new ones, ideally) and see what happens. If that doesn't prove to be enough to get stable, bump vcore up to 1.45 or even 1.5v in the short term.


Different OC results with different boards are nothing new. I couldn't get my quad past 3.1Ghz with an Abit IN9-32X MAX, it would do 3.5Ghz with an XFX 680i LT, and it's now stable at 3.8Ghz with my XFX 780i SLI.
 
The on-board voltage sensors are NEVER accurate. I had an Intel mobo that said that my 12V rail was only 11.5V when it was actually 12.01V (when measured with a voltmeter).
If you read professional PSU reviews (like on jonnyguru.com or hardocp) you will notice that they use REAL testing equipment and they never even mention the motherboard sensors in their reviews.
A fully loaded system is not a proper load for a PSU so if you read a review that says that they tested a PSU by running some stress apps on it and plugged in 4 hard drives and 10 fans, just ignore it...

BTW, I have 2 systems using Asus boards (the current has a Maximus Formula like the OP) and on both the voltage sensors report wrong voltages and more ripple than what I have measured with a digital voltmeter.
 
silent-circuit said:
Yep. Software is untrustworthy. Until you take a quality multimeter to it, you know next to nothing about the output. Unless it's showing /huge/ fluctuations and accompanying instability, there's no real reason to assume there's anything wrong with the supply.
Click to expand...
PurpleDiamond said:
BTW, I have 2 systems using Asus boards (the current has a Maximus Formula like the OP) and on both the voltage sensors report wrong voltages and more ripple than what I have measured with a digital voltmeter.
Click to expand...
You can't measure ripple with a DMM, you'd need a oscope for that...
 
Menelmarar said:
You can't measure ripple with a DMM, you'd need a oscope for that...
Click to expand...

I said nothing about ripple, I said output. The OP thought his rails were low. He also thought at least one was fluctuating ~0.5 v. You'd likely be able to see such a large change with a DMM. Yes, as Paul already said, ripple is the wrong word for what he was talking about. I never said anything about ripple.
 
Menelmarar said:
You can't measure ripple with a DMM, you'd need a oscope for that...
Click to expand...

I know that a DMM is not the right tool for measuring ripple. What I meant was that the sensors showed a fluctuation of about 0.5V peak-to-peak on one of the rails while the DMM showed barely a change of 0.1V.
 
PurpleDiamond said:
I know that a DMM is not the right tool for measuring ripple. What I meant was that the sensors showed a fluctuation of about 0.5V peak-to-peak on one of the rails while the DMM showed barely a change of 0.1V.
Click to expand...

...and there can easily be a 0.05 or higher margin of error with a cheap DMM.
 
ok, thanks. looks like i need to experiment with more of the BIOS settings. this asus words half the settings differently than the abit IX38 and many OC sites and guides. i have a quality DMM i have for other purposes but i didnt know software was so innaccurate. maybe i will test with that. if i cant get this 100% stable i'll remove the extra 2 DIMMs and retest.

i used ripple because it's what OCCT's graphical output logs call it.

my main concern was a def board, since i bought it open box, or a dieing power unit, since i had a power outtage i wasnt home for and not sure how my battery backup responded.

oh one more thing: "relatively 0 vdroop" was meant as no noticable vdroop. OCCT showed the vcore steady between 1.41 and 1.43 (set as 1.425). this is with Load Line Cal enabled.

PS. does anybody know what the "built in" OCCT sensor is? i know it pulls CPU etc info from an implemented CPUID via CPU-Z, but cant find any info on the internal temp and voltage senors.
 
OCCT pulls from the same Winbound type chip all software does.
 
You must log in or register to reply here.
Back
Top