VA & watts in ATX PSUs

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
There is an article in next month's (Feb 2009) issue of Australian PC
User Magazine which compares the real power and VA consumption of
several ATX PSUs under various load conditions.

Some PSUs appear to incorporate APFC while others do not.

A Huntkey 80Master 700W PSU consumes 100W and 114VA when idling in
WinXP desktop mode. However, in standby (S3) mode, with power applied
to system RAM, the figures are 5W and 44.5VA. In the "off" state the
numbers are 3.3W and 43.4VA.

The other PSUs also have similar results for the standby and off
states.

Why are the VA figures so high?

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Re: VA & watts in ATX PSUs

"Franc Zabkar"
Quoted text here. Click to load it

**  How completely ***  IDIOTIC **  posting a question involving

 the precise technical details of a magazine article that has

NOT   YET   BEEN   PUBLISHED   !!!
-----------------------------------------


BTW 1 :

How many time have I lectured  YOU on the issue of the high rms values of
peaky current waveforms ??

BTW 2:

What value C is across the active and neutral in the EMC filters??

A 1 uF cap draws 75mA  ( = 18 VA  )  but  zero watts .



......    Phil





Re: VA & watts in ATX PSUs



Quoted text here. Click to load it

The light power load results in a very short conduction period / angle of
the AC input rectifier bridge which results in the high VA (apparent
power) figure.

Graham


Re: VA & watts in ATX PSUs
On Fri, 02 Jan 2009 21:41:13 +0000, Eeyore
composed:

Quoted text here. Click to load it

I'm trying hard to understand why a PSU that incorporates APFC would
have such a bad VA rating when doing nothing. I can only assume that
APFC is not applied to the +5VSB standby rail, or that it is difficult
to design an APFC circuit that can accommodate a wide range of loads.
In order for a real power of 3.3W to appear as 30VA or more, the PF
would have to be around 0.1. According to my calculations, this is
feasible, but only if the conduction duty cycle is very low,
approximately 0.005 if we assume a square current pulse centred on the
voltage peak.

I don't have access to any PSU circuits with APFC, but here is such a
circuit in an LG TV:

 
http://www.users.on.net/~fzabkar/MB-042C/PSU-UL.jpg

There is an 0.22uF EMI suppression cap across the mains terminals.
This would contribute about 4VA to the figures. I wonder how *it*
behaves at standby. BTW, I realise that it is not just a matter of
adding the VA contributions of each section of the circuit.

AFAICS, anything connected to the APFC side of the supply would appear
as a resistor. So an ideal PSU would have an "active" resistor in
parallel with an EMI cap.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Re: VA & watts in ATX PSUs

"Franc Zabkar"
Quoted text here. Click to load it



**  Forget pommy Graham  - he is a  total ASS  !

Read my reply  -  it is mainly due to having EMC filter caps of several uF
directly across the AC supply .

So the AC current is near sine wave, 90 degrees out of phase, in standby
mode.



......  Phil








Re: VA & watts in ATX PSUs



Quoted text here. Click to load it

It definitely does ? Yes it must I guess for C tick.


Quoted text here. Click to load it

Not at all unlikely. It doesn't have to because it falls below the 75W IEC
61000-3-2 (IIRC) limit in that situation.


Quoted text here. Click to load it

Also quite likely. Read an app note on the subject. I could speak VOLUMES on
the subject. IEC1000-3-2 as it started life had to be withdrawn PDQ when
certain industry sectors (including ME) pointed out certain appliances could
never conform.


Quoted text here. Click to load it

Pretty stunning isn't it ?


Quoted text here. Click to load it

No great surprise. APFC is noisy as hell.


Quoted text here. Click to load it

Not at all sure I follow your logic there.

Graham


Re: VA & watts in ATX PSUs
On Sat, 03 Jan 2009 04:30:37 +0000, Eeyore
composed:

Quoted text here. Click to load it

I should have included the results for loaded conditions.

The PF for Windows XP desktop idling was 0.877.

For Crysis (3D game) at 1600 x 1050 pixels it was 0.964.

Other non-PFC PSUs returned Crysis figures of 0.6.

Quoted text here. Click to load it

AFAICS, if the PF is unity, then the APFC circuitry is making the load
look like a variable resistor, ie current in phase with voltage.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Re: VA & watts in ATX PSUs

"Franc Zabkar Stubborn WOG Idiot "


**  Forget pommy Graham the autistic  -  he is a  total ASS  !

Read my reply -  the poor PF  is mainly due to having EMC filter caps
of several uF directly across the AC supply .

So the AC current is near sine wave, 90 degrees out of phase, in standby
mode.




......  Phil





Re: VA & watts in ATX PSUs

"Eeysore"
Quoted text here. Click to load it

  **  So this one is  DEFINITELY  an APFC model.

http://www.gatumdale.com/webshaper/store/viewProd.asp?pkProductItem51%2


Quoted text here. Click to load it


**  ROTFLMAO   !!

The  * WHOLE  BLOODY  POINT *     of having an active PFC supply is that it
does  NOT   draw spikes of current from the AC supply   !!!!!!!!!!!!!!!


YOU  CRIMINALLY   STUPID

 POMMY   SHITHEAD  !!!!!!!




......  Phil




Re: VA & watts in ATX PSUs



Quoted text here. Click to load it

In standby mode PFC isn't required because the power level drawn is below the
75W threshold as defined in IEC61000-3-2 regarding harmonic currents and the
threshold for (serious or any ?) correction.

It happens to be a standard I became VERY familiar with since it would (as
originally draughted) have effectively caused ALL audio power ampliers of a
decent wattage to HAVE to use SMPS supplies with PFC. NO more toroids etc,
unless you had PFC on the secondary side !

I was one of the first to notice this and alerted John Woodgate at PLASA
immediately. Curiously IEC1000-3-2 1st Edition strangely went out of print just
after we got a copy and its very existence was formally denied. There was
something else in too that was IMPOSSIBLE.

I could ramble at great length on this one - the bureaucratic 'fixes' required
to make the standard workable were typically Byzantine in typical EU manner
when producing the equivalent ENs.

Graham


Site Timeline