UPSs and Active PFC PSUs.

[Lennart Sorensen]

On Fri, Dec 03, 2010 at 11:40:34PM -0500, Scott Sullivan wrote:
> In the recent UPS thread a major question came up about Active PFC PSUs  
> and UPSs.
> I've take some time to research the issue and provide a little clarity  
> at the best of my ability.
>
> # My Current Setup
> I recently got an APC 1500VA Back-up UPS myself.
> I have two older PCs (both doing server duty), network gear connected  
> and a single 15" LCD attached.
> Currently they draw 105W of the 850W the UPS can provide, run time is an  
> estimated 53 min.
> One of the machines has a 350W Active PFC PSU.
> I HAVE tested this load with a simulated power failure. (I pulled the  
> plug).

What brand/model of power supply and what efficiency is it?

> # Definitions:
>
> Active PFC:
>     "Active power factor corrector" is a power electronic system that  
> controls the amount of power drawn by a load in order to obtain a power  
> factor as close as possible to unity.
> http://en.wikipedia.org/wiki/Active_power_factor_correction#Active_PFC
>
> Plain Language:
>     As far as I've wrapped my head around it, it adjusts it input  
> current to reduce the wasted power that non purely resistive load (like  
> that of a computer) presents.
>
> # Why is Active PFC not good on some UPS models?
>
> The general gist has been that Active PFC should not be used with a UPS  
> that does not Supply a pure sine-wave which is most consumer UPS models.
> Anecdotal evidence to support this claim include:
> - UPS shutting off when switching to battery with a Active PFC PSU
> - Strange Humming or Resonant noises while on battery power.

It seems that some high efficiency power supplies have issues unless
they receive the sine wave they expect.  Not all designs are as picky,
but if you don't want to go by trial and error, you simply make sure to
get a sine wave output UPS for an active PFC power supply.  If you are
willing to take your chances and experiment, you may get lucky.

Essentially a stepped sine wave looks like power is going up and down
in voltage many times per cycle, causing the power supply to adjust its
PFC hundreds if not thousands of times per second.

> # What APC has to say on the matter.
>
> http://www.apcmedia.com/salestools/RMUZ-7DTKRC_R1_EN.pdf
>
> Active PFC *CAN* be used as long as you have enough spare Wattage from  
> the UPS.
> This is because Active PFC compensated for the micro loss of power  
> during battery switch over by Drawing more current. The consumer UPS  
> models respond to this load by shutting down when it exceeds their  
> operating limit.
> According to APC's documentation, 500W UPS and greater is about the  
> magic number that will cause this failure reaction. YMMV.
>
> As for the noise, that just the Active PFC working harder compensate for  
> the Stepped Sine Wave that these consumer units create. It does not  
> appear to actually be harmful to the PSU but is annoying.

Apparently some power supply companies believe it is harmful.  Also very
annoying.  Some power supplies will in fact turn off when it happens,
not just the UPS.

I know the corsair power supply I have requires a sine wave input.
Corsair has explicitly stated that, and people who have tried cheaper
UPSs have found the power supply simpl turns off.

> == Disclaimer ==
> I am only sharing my experiences and details I have collected. I can not  
> be held responsible for how this information is used. Any mistakes or  
> omissions are my own.

-- 
Len Sorensen
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists