PC Peripherals Power Supply Ratings [Inside look]

Status
Not open for further replies.

KiD0M4N

Forerunner
Hi guys,

Last couple of 10days I have been scrouging forums looking to demystify and throw some light upon the power ratings written at the back of the PSU. I know, nothing beats actually getting the PSU and testing it out, but it does save a lot of time when you are able to make intelligent choices without having to "mis"invest in poor/midiocre hardware.

So here goes:

Usually power supplies are rated separately for their different rails. The usual rails one finds on current PSUs are +5, +3.3, +12V1, +12V2, +5Vsb, -12V. -5V, +12Vsb are unusual and not a requirement, but are sometimes put in by PSU manufactures as speciality. Also in case of single rail PSUs, +12V2 is absent and only a single +12V rail is present. Now to the crux of the matter, the rail ratings:

I will approach this with a example:



The above is the exact power ratings of a ANTEC TRUEPOWER II 550W EPS12V PSU.

The most important line in the above table is the last line,

Code:
* +5V, +3.3V, +12V1, 12V2 maximum output 530 Watts max.

The PSU is rated for 550W, and the 4 rails (+5, +3.3, +12V1, +12V2) are rated together at a max of 530W. So that leaves 20W for the the 2 other rails (+5Vsb and -12V) and we see that calculating the wattage of the +5Vsb and the -12V rails, we get approx 22W (5V x 2A = 10W + 12 (not -12) x 1A = 12W = 22W) Its not a mathematical calculation error by the part of Antec, its a very valid specification table (more on that later.)

Now, coming to the main 4 rails, we see that my calculating the wattage by multiplying the rails and their amperage we get:

5V x 40A + 3.3V x 32A + 12V x 19A + 12V x 19A = 200W + 105.6W + 228W x 2 = 761.6W

Now, either some people are very weak in their maths over at Antec or something weird is going on. Surely they mention that max power draw on the 4 rails is 530W and the calculated value comes up to 761.6W? Well the truth is a bit complicated, and to really appreciate it we need to look at some numbers:

*assume that the 2 low power rails, +5Vsb and -12V are loaded to max and lets take them out of picture*

The MAX ampere ratings given below the rails are the MAX that can be draw from that particular one at any given time, keeping in mind the fact that other rails will have lesser power available to them.

Code:
RATINGS:

[COLOR="Red"]5V: 40A, 3.3V: 32A, 12V1: 19A, 12V2: 19A[/COLOR]

5V: 40A, 3.3V: 32A, 12V1: 19A, 12V2: 19A, [COLOR="Red"]NOT OKAY, LOAD = 761.6W > 530W[/COLOR]

5V: 40A, 3.3V: 32A, 12V1: 10A, 12V2: 10A, [COLOR="Red"]NOT OKAY, LOAD = 545.6 > 530W[/COLOR]

5V: 36.8A, 3.3V: 32A, 12V1: 10A, 12V2: 10A, [COLOR="Green"]OKAY, LOAD = 530W <= 530W[/COLOR]

5V: 40A, 3.3V: 27.2A, 12V1: 10A, 12V2: 10A, [COLOR="Green"]OKAY, LOAD = 530W <= 530W[/COLOR]

5V: 17A, 3.3V: 20A, 12V1: 18A, 12V2: 13A, [COLOR="Green"]OKAY, LOAD = 524W <= 530W[/COLOR]

5V: 20A, 3.3V: 20A, 12V1: 18A, 12V2: 13A, [COLOR=Red]NOT OKAY, LOAD = 539W[/color]

5V: 10A, 3.3V: 13.3A, 12V1: 18A, 12V2: 18A, [COLOR=Green]OKAY, LOAD = 526W <= 530w[/color]

The above computations should make it preety clear to anyone that the max ampere ratings mean "jack" if they arent backed up by some real estate power (read Watts) from the PSU.

Hope the above has helped some people in understanding the mystery behind the ratings, because it took me a good while to actually get some solid facts rather than just presumptions.

Regards,

Karan
 
  • Like
Reactions: 2 people
hunt3r said:
Nice info there Karan. Repped. Still havent understood most of it tho :P so will read more into it later.

thnx for the rep man..... but maybe i am seeing things..... 3 other people had repped me before you and now they have disappeared? :huh:
 
Add reputation.

See on the Top right of every post.

Theres one Balance thingy and a Exclamation mark thing.

Thats + Reputation Point spread and Bad Post report.
 
Err Karan, you cannot use the P=VI formula and multiply straight away.

If you do that, NeoHE's 12v rails alone come to 612W ;)

There's something called efficiency also. Maybe Chaos/Deejay will explain what I mean ;)
 
Anish said:
Err Karan, you cannot use the P=VI formula and multiply straight away.

If you do that, NeoHE's 12v rails alone come to 612W ;)

There's something called efficiency also. Maybe Chaos/Deejay will explain what I mean ;)

rofl.... get ur facts clear m8.... hv i multiplied P=VI and "added" anywhere. you missed the catch word.

also efficiency has "jack" to do with output side, its how efficiently the PSU is able to convert input power to output power.

for example: say your "l33t" PSU is giving 500W (say X) of real power, then let it actually consume "Y" amount of power from the 230V (R.M.S. mind you) input line.

Y is always ">" X because nothing is 100% efficient. Efficiency usually ranges in the range of 60-80% range. If it were 100% efficient, there wouldnt be any heat generated, would it.

So here efficiency = X/Y

Mind you, I am not talking about Power Factor Correction efficiency. Just plain PSU efficiency.

Regards,

Karan
 
Any PSU's available locally that have dual rails and minimum 22A on each rail? Unfortunately, the Antecs don't.
 
^^ No. The highest offered is by the Zebronics 640W, 20A/20A.

@ Karan - Ahh ok, will research a bit more and get back to you m8 :)
 
Blade_Runner said:
Powersafe 600W has dual 12v rails with 22a each on both rails.

Blady.... sorry to correct you but thats the MAX you can draw from either rail ONE at a time, that too assuming its not grossly overrated. From what I saw of Darky's result, the power regulation is preety run-of-the-mill with the Powersafe 600W (Vdroop from 12.2V to 11.8V, a .4V delta, not good)
 
Karan said:
Blady.... sorry to correct you but thats the MAX you can draw from either rail ONE at a time, that too assuming its not grossly overrated. From what I saw of Darky's result, the power regulation is preety run-of-the-mill with the Powersafe 600W (Vdroop from 12.2V to 11.8V, a .4V delta, not good)
Oh np, but i'd really like to know more about this.
 
Blade_Runner said:
Oh np, but i'd really like to know more about this.

:)

Ok lets see.....

suppose you get a PSU with 2 rails each rated 24 A (which again is total BS because the concept of Dual Rails was introduced by Intel saying that a maximum of 240VA (thats 12 x 20A) should be allowed per rail) and a combined max wattage of 400W on the 12V rails, we have:

400/12 = 33.33A total on the 12V rails.
So the combined amperage of the 2 rails is a mere 33A instead of 22A + 22A = 44A.

Also, we must take into consideration the cross loading requirements which certain PSU designs impose. For example, to get stable 12V(s) readings, we must put increasing loads on the 5V rail as we draw more power from the 12V rail. Otherwise, the 5V shoots up and the 12V rails fall below regulation.
 
Karan said:
:)

Ok lets see.....

suppose you get a PSU with 2 rails each rated 24 A (which again is total BS because the concept of Dual Rails was introduced by Intel saying that a maximum of 240VA (thats 12 x 20A) should be allowed per rail) and a combined max wattage of 400W on the 12V rails, we have:

400/12 = 33.33A total on the 12V rails.
So the combined amperage of the 2 rails is a mere 33A instead of 22A + 22A = 44A.

Also, we must take into consideration the cross loading requirements which certain PSU designs impose. For example, to get stable 12V(s) readings, we must put increasing loads on the 5V rail as we draw more power from the 12V rail. Otherwise, the 5V shoots up and the 12V rails fall below regulation.
Oh i was under the impression that Intel introduced the dual rail concept since the International Safety standards called for not more than 240VA on any wires to reduce the risk of a fire hazard. This called for the 20A current limiter that you speak of. However i donot know if there is a current limiter with the particular Powersafe model in question since most Psu tend to have it for the single 12v source. Also intel had removed the 240va limit, You can check the list of approves psus here
and as you can see many of them exceed it.

Another thing is i was under the impression that cross-loading occurs if the psu is group regulated. Now i am unaware abt if the powersafe is non-group regulated or not so i wudn't comment about it.
 
Hey Karan,well even if there is a drop,i dont think it matters much to me,all i know is it hold pretty well when Opty @ 2.7GHz n X1900XT @ OC'ed speeds.

Its rock stable...
 
:hi: darky....

well stability differs a lot with these drops. for example, if you had a lessed vdroop you would be able to push higher clocks at lower voltages, because when the rail drops, the regulated voltages also droops a tiny bit, because of which u need to give it a higher voltage upfront, so that even if it droops, it stays above the req voltage.

hope u agree :D
 
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.
 
Karan said:
:hi: darky....

well stability differs a lot with these drops. for example, if you had a lessed vdroop you would be able to push higher clocks at lower voltages, because when the rail drops, the regulated voltages also droops a tiny bit, because of which u need to give it a higher voltage upfront, so that even if it droops, it stays above the req voltage.

hope u agree :D

Well karan i never saw any stabilty issues.
 
Status
Not open for further replies.