+12 volt line stability

[orion456]

I am about to get a new PSU so I used speedfan 4.17 to graph the +12v line fluctuations of my current noname brand PSU as I added more components online. Look at the graph below and I will list the various component changes.

Before A --> FAH on, 50% CPU usage, HSF on medium speed

A --> FAH off, HSF to low, intake fans to low

B --> FAH on

C --> add HSF to high

D --> add intake fans to high

E --> add Prime 95, 100% CPU usage

After E --> back to A

Notice the marked jumps in voltage as components are added or subtracted. This is still within the 5% tolerance but hardly a picture of stability. Normal line noise is around 3%, it appears as load is added the line noise decreases at C but returns as the fans come online. Wonder if my fans are causing problems directly or through the +12 volt line.

--------------------------------------------------------------------

[orion456]

Here is my Vcore going from FAH to idle and back to FAH. That Vcore is fluctuating over 6%....geez.

[orion456]

My +12 volt line going from FAH to idle and back to FAH (50% usage). See how my Vcore changes with the +12v line, so they are pretty much locked together. (this graph is slightly time shifted relative to the above Vcore)

[larrymoencurly]

Speedfan once said that my +12V fluctuated between +6V and +8V every second, all while it was actually rock steady and off by just 1% according to my multimeter. MBM 5 did better but reported the voltage as 10.64V. Moral: trust only digital multimeters. Sears, Radio Shack, and Jameco have cheap ones, including some that can even record measurements or transmit them to the computer,

[blackjackel]

if you use them incorrectly you WILL fry your equipment, and its not hard to do either (fry your stuff)

[orion456]

Speedfan once said that my +12V fluctuated between +6V and +8V every second, all while it was actually rock steady and off by just 1% according to my multimeter. MBM 5 did better but reported the voltage as 10.64V. Moral: trust only digital multimeters. Sears, Radio Shack, and Jameco have cheap ones, including some that can even record measurements or transmit them to the computer,

It's not hard for programs to misinterpret voltage signals, however since the effect is consistent I think it is happening. It's also within specs and that makes sense as well. I will stick a multimeter in there shortly and check to see if the speedfan result can be verified.

I can say that my temperature probe near the CPU produces the same trends as speedfan's temperature probe graph though they differ in reading by 1.5c (which I assume is because speedfan reads the CPU temp and my probe can only read the temps just outside the CPU.

[orion456]

I finally got a digital voltmeter and checked my +12v supply off an empty Molex connector and also off +12v line that enters the mobo. Interestingly the voltages were not the same for any of them. I would have expected the +12v line to be consistent. Also I had a 2nd analog voltmeter and its readings were 0.15v lower.


Location . . voltage range . . FAH on . . FAH+Prime . Fans on . All off

speedfan. . 11.85=>12.28v . . 12.07v . . 12.00v . . . 12.07v . . 12.17v

Molex . . . 11.90=>12.06v . . 11.93v . . 11.92v . . . 11.91v . . 12.03v

Mobo . . . 11.76=>12.02v . . 11.85v . . 11.82v . . 11.90v . . 12.00v

So it looks like SpeedFan reads about 0.17v high, while the unused Molex reads about 0.1v high on load and about 0.02v high on idle.

However, just like SpeedFan showed, my voltage drops by at at least 1.5% on load and this is verified by the Mobo +12v connection which also dropped 1.5%. The SpeedFan fluctuations where not as apparent using the multimeter, however I'm not sure what the response time is for the meter, it might be averaging things out a bit (it looks like it made a reading every second). SpeedFan also said my voltage was fluctuating down to 11.85v and again this was verified by the Mobo readings which were 11.76v.

I conclude, SpeedFan is showing an accurage representation of voltage fluctuations but the actual voltage values are slightly off (in my case of course). Now I await my OCZ Powerstream to see if these fluctuations can be minimized.

[orion456]

Got my OCZ PowerStream and I am all set to hook it up when I discover no BTX adapter. After a few frantic minutes I realize they have changed the BTX plug so that an adapter isn't necessary any more. The BTX cable simply slides apart and voila an ATX cable instead....nice new idea.

Ok so right off I can see that things are way better with my new PSU. For starters my temperature fluctuations are way down. Here's a speed fan output, the top graph is my old psu, the bottom the OCZ. I thought the fluctuations were due to my fans changing speed but that doesn't seen to be the case. So perhaps it's a more stable Vcore.

[Oklahoma Wolf]

I'll be interested to see what the meter says - the old PSU looked to be under a little too much 12v strain for it to handle. Out of curiosity, what did you have before?

[orion456]

Here is my +12v supply OCZ first, then my old PSU which was a Macron MPT-460 (has a 2 year warranty and the +12vs listed as 30 amps). Notice the great improvement in the +12v stability.

[orion456]

Here is Vcore before and after. Again a great improvement in Vcore stability according to SpeedFan. I'll use the voltmeter and check those out soon. The big drop in Vcore happens when the load goes to 50% with FAH being turned on. This must be a characteristic of the p4p800 asus board I have.

The drop in Vcore is not mirrored by the +12v line which is stable with the OCZ but changes significantly with my old supply (see above).

[Oklahoma Wolf]

Macron is down there close to Powmax build quality - going to the OCZ was a good sized upgrade ;)

[orion456]

Oh oh.....with a crackle and a pop....it died. :cry: :cry: :cry:

And I hadn't even got around to using the multimeter yet....damn good thing I would have kicked myself around the block...twice. As it was I just touched the power button and it kinda got hit twice really fast.........

Supplier says I'll have a new one Monday....but sheeshers....

[Oklahoma Wolf]

That sucks :(

Glad to hear you'll be getting a replacement fairly soon.

[orion456]

That sucks :(

Glad to hear you'll be getting a replacement fairly soon.

Yep....funny thing was, when I phoned OCZ and said I had my PSU fail the guy said right away, "Oh, it's a 520." Like he knew right off. Interesting. Maybe a bad batch of capacitors.

[orion456]

Here's my Macron supply in my friends case. Looks like it's very stable at a lower power draw (2.6 p4, no overclocking). I guess that PSU can't give the power that the p4p800, p4, @3.6 needs.

[Fangs404]

i told you i'd do mine (i have a fortron 530w), and here i am. :)

http://img46.exs.cx/img46/5059/12V_fluc.jpg

i updated my sig to accurately reflect my system. i ran in-place large FFTs (prime95) for just over 20 mins, and this is what my graph looked like. pretty damn stable if you ask me. the high was 11.86, and the low was 11.76. i haven't adjusted the pots at all.

[orion456]

i told you i'd do mine (i have a fortron 530w), and here i am. :)

YAY!! Glad you made it back.

Most excellent, less than 1% fluctuation which is just about as good as you can ask for.

Did you try turning off Prime and seeing if the voltage changed in response to a load reduction? I did that for the OCZ and there was no change at all.

Looks like the Fortron 530 you have is a great PSU.

[Fangs404]

YAY!! Glad you made it back.

Most excellent, less than 1% fluctuation which is just about as good as you can ask for.

Did you try turning off Prime and seeing if the voltage changed in response to a load reduction? I did that for the OCZ and there was no change at all.

Looks like the Fortron 530 you have is a great PSU.

yeah, i turned off prime, turned my tornado all the way up for a min or so, turned it all the way down and did the same thing with my tornado while prime was running, and it didn't change any more than the ±0.1V that you can see in my graph. i like my psu. :)

[CandymanCan]

I measure my voltages with a Multimeter and my 12v Stays @ 12.04v Load and Idle. With my other psu which is a Aspire 500w (in the avatar) goes up from 12.37v idle to 11.90v during load

[orion456]

I measure my voltages with a Multimeter and my 12v Stays @ 12.04v Load and Idle. With my other psu which is a Aspire 500w (in the avatar) goes up from 12.37v idle to 11.90v during load

Would be interesting to see what SpeedFan says it does....kinda checking on its reliability at least as far as valid fluctuations go.