PC Peripherals Power Supply Ratings [Inside look]

[KiD0M4N]

Actually tight regulation has no benfits for Cpu overclocking. See the table below, it shows the max fsb achieved for every psu. Test config included: Prescott P4 3.2GHz @ 3.8GHz, Radeon 9800 Pro video card @ 400MHz/365Mhz, 3 7200rpm hdds(40GB and two 200GB), a DVD-ROM drive, 108MBit wireless network card & two 120mm fans.



The difference between cheap/low-end antec 350 smart blue and the super regulated Antec Trucontrol 550w is less than 2mhz fsb which falls very well within the experimental/testing limit.

Edit: I am sure the senior members might have a few words/ personal observations to post about this phenomenon.

well actually Truecontrol is not that good at voltage regulation. It droops preety bad also, just not as bad as cheap brands. The real benefits are generally experienced when the regulation is achieved on the scales of 0.1V (i.e. a droop of < 0.1V from idle to FULL load.)

and here we are looking at the FSB frequencies. and we also dont know what exact droop was involved in each of the case. so very very hard to speculate anything from a simple image. i am sure if there was a PSU drooping by 0.4 V and one with droop of .05V, we would see a differnnce of 0.1-0.3V in the Vcore required to achieve the OVERCLOCK.

@ blade: i think you missed my point. I mean to say is that the tight regulation will allow the user to reach the same overclock at a lower Vcore.

 

[Blade_Runner]

well actually Truecontrol is not that good at voltage regulation. It droops preety bad also, just not as bad as cheap brands. The real benefits are generally experienced when the regulation is achieved on the scales of 0.1V (i.e. a droop of < 0.1V from idle to FULL load.)

Oh i dunno, take a look at the following multimeter readings for how good the voltage regulation is with the tru control.

Seems ok to me. Also if my memory serves me well, the variance allowed is within 5% in idle state and within 10% when on full load. You can check the ATX specs here
http://www.formfactors.org/developer/specs/ATX12V PSDG2.01.pdf

and here we are looking at the FSB frequencies. and we also dont know what exact droop was involved in each of the case. so very very hard to speculate anything from a simple image. i am sure if there was a PSU drooping by 0.4 V and one with droop of .05V, we would see a differnnce of 0.1-0.3V in the Vcore required to achieve the OVERCLOCK.

@ blade: i think you missed my point. I mean to say is that the tight regulation will allow the user to reach the same overclock at a lower Vcore.

No i din't miss your point but i lack the evidence which proves what you are trying to put forth. Now only if you'll be so kind.................

 

[Anish]

Karan, I think you are getting a bit too touchy on this PSU matter.

We all know that around 16-18A on the 12v rails are enough to power a moderately OC'ed rig with a decent GPU,

And 20+ A for a Single rail PSU for a stock rig with a decent GPU.

Now all this VDroop and all is ok, but isn't it going a bit overboard

Just an observation, not meaning to interrupt the flow of this thread or offend anyone

 

[KiD0M4N]

Oh i dunno, take a look at the following multimeter readings for how good the voltage regulation is with the tru control.

Seems ok to me. Also if my memory serves me well, the variance allowed is within 5% in idle state and within 10% when on full load. You can check the ATX specs here
http://www.formfactors.org/developer/specs/ATX12V PSDG2.01.pdf
No i din't miss your point but i lack the evidence which proves what you are trying to put forth. Now only if you'll be so kind.................

lol man.... here goes..... rofl: first darky then u )

a 10% regulation at load translates to 12V +- 10% which is 12V - 1.2V or +1.2V which agn is 10.8V to 13.2V. Are you telling me a PSU that falls to 10.8 V (while still within BOOK specs) is good? Nahha.... ideally these 2 things are looked at:

1- Variance in a rail under static load.
2- Vdroop under dynamic load.

1 has far greater weightage than 2 (5x say) because that leads to lose of stability due to variation in voltage supplied.

For example,

if a PSU was 12.0V under idle load and went to 11.95V under full load, then it would categorised as excellent where as a PSU which went from 12.6V @ idle to 11.40V under load would be categorized as "highly not recommended." (mind you, still very much within ATX specs.)

ANTECs have a nature to wobble under static load (all Channelwell manufactured exhibit similar traits) (source: http://jonnyguru.com/) and demonstrate a thick line when put under a oscilloscope (5mv-5ms) Which makes them not so good....

@Anish:

Why you generalize man? Be specific... afterall we are dealing with a very inexact science here, so we should try and minimize the confusion by elimination humane diffuse. Hope you get my flow....

Also about going "overboard".... we have different thresolds of knowledge buddy, some ppl just cant have enough (like me ) Maybe thats why they end up at different jobs/posts/salaries in life.

PS: @ blady, yaar give me 1 day, i will compile (actually refind) all the sources which i have gone through to come to my above conclusion. just a bit swamped right now to put together a big post.

 

[x86]

my business what about neo he? :bleh:

 

[Blade_Runner]

lol man.... here goes..... rofl: first darky then u )

a 10% regulation at load translates to 12V +- 10% which is 12V - 1.2V or +1.2V which agn is 10.8V to 13.2V. Are you telling me a PSU that falls to 10.8 V (while still within BOOK specs) is good? Nahha.... ideally these 2 things are looked at:

1- Variance in a rail under static load.
2- Vdroop under dynamic load.

1 has far greater weightage than 2 (5x say) because that leads to lose of stability due to variation in voltage supplied.

For example,

if a PSU was 12.0V under idle load and went to 11.95V under full load, then it would categorised as excellent where as a PSU which went from 12.6V @ idle to 11.40V under load would be categorized as "highly not recommended." (mind you, still very much within ATX specs.)

ANTECs have a nature to wobble under static load (all Channelwell manufactured exhibit similar traits) (source: http://jonnyguru.com/) and demonstrate a thick line when put under a oscilloscope (5mv-5ms) Which makes them not so good....

I guess i am n00b in such matters that why the rofl icon Anyways since you din't get my point, i was asking for evidence which proved that

am sure if there was a PSU drooping by 0.4 V and one with droop of .05V, we would see a differnnce of 0.1-0.3V in the Vcore required to achieve the OVERCLOCK.

Cause i was pretty sure the mobo's power circuitry is responsible for that, poor cooling of it can also lead to such "droops".

Also nowhere did i recommend going for a psu which varies so much i.e. by 10%, still neither you nor i have written the atx specs which says its ok for a psu to do so at full load. Another thing is you've posted that the trucontrol is as bad as cheap psus when the trucontrol has pretty tight regulation. Still we aren't talking about POS psus but some really good brands.

Last but not the least when being picky about stuff you gotta realise we still don't get that many options which our US counterparts get.......Jonnyguru can really afford to be picky !

PS: @ blady, yaar give me 1 day, i will compile (actually refind) all the sources which i have gone through to come to my above conclusion. just a bit swamped right now to put together a big post.

Sure no problems.
PS: does jonnyguru oc his system :rofl: to test the psus. I know its nice to test with a sunmoon psu load tester but thats hardly anywhere near what usual day to day users will subject their psus too. For starters find me a PC which will emulate 100% static load loads. Secondly with a PC/Rig there are thousands of factors involved with testing a psu unlike a psu tester and therein could lie a flaw with this psu testing methodology.