EuP/ErP power supplies are a sick joke...

numskull

Distinguished
Jul 24, 2011
7
0
18,510
Hello, does anyone ever check how much power their EuP-compliant power supply is drawing on standby? Of course it uses <1w, doesn't it, because the manufacturer says it does...or does it?

I have been trying to get such an item for my computer and have now been through no less than four power supplies in a row that draw up to 30x the stated power at standby. Yes, you read right, I did say 30 times the declared consumption. What does this mean in real terms? Aside from the environmental effects, every standby watt is worth just over £1 per year on your electricity bill, so over 5 years my psu could have cost me an amazing £150 on top of the purchase price.

So which ones have I tried? First there was a Corsair CX400 that used 11w (although not strictly EuP, and after a while it did sort itself out to around 1-2w); then there was a Corsair CX430 v2, which used 19w when switched off and 31w under no load. I sent two of these back to the supplier. Next came an OCZ StealthXStream2 400w, which used about 20w under no load. Finally today I received an FSP SagaII 400w that uses 9w when switched off and /14w under no load.

So - do you really know how much your beloved EuP psu is costing you?
 
Yeah noticed that too, but power bars have off switches for a reason. It is a combination of Board + PSU. My X58A-UD5(Antec EA650) with EUP on, does about 4-5 watts off.
EUP also kills off usb power when the system is off, so no chargy for devices with the system off.
 

numskull

Distinguished
Jul 24, 2011
7
0
18,510
Well, that's not as bad as 30w from the twin Corsairs I had, but I imagine most people will be thinking like me that with EuP you can forget about constantly switching main power on and off, without wasting a significant amount of power, whereas the reality is that you could be unwittingly spending as much as £30 per year. Not to mention the environmental cost involved and the flagrant breach of the whole 1 watt initiative.
 
It generally helps to actually understand what you are testing and what the claims are before you start making crazy complaints.


First off, exactly which power meter were you using? Many are only accurate to within a few watts so at low load levels their results are totally meaningless as the tool inaccuracy can be greater than the power draw of the device in question. You can also run into an issue if it happens to be measuring the apparent power not the real power because the power factor of the stand by circuits in most PSUs is about 0.5 so knowing the tool extremely well is extremely important in this case.

Secondly, Eup/ErP is for the power supply in an off mode isolated from everything else. This means its sitting on a table with only the wires from the mains plugged into it with none of its connectors connected to a system. If you have them plugged in to a system the motherboard is drawing about 5W of stand by power from the 5VSB rail which is always on as long as the PSU is plugged in and the master power switch is in the on position. Most people who check their "off" power still have the unit plugged into the system which is pulling 5W from the PSU and they wonder why their idle power is so low.
 

numskull

Distinguished
Jul 24, 2011
7
0
18,510
Thanks everybody for your contributions so far. A few points worth mentioning in reply to hunter315:

1) The two Corsairs CX430v2 were tested and found to be faulty. They also made a slight whining noise that suggested a short circuit somewhere in the units. I noticed on the second one that the main circuit board even appeared to be slightly bowed. As they both had the exact same errant power draw I suggested to the seller that they test some more units. Although they did not feed back to me on that, I did notice that the units were no longer available for sale a couple of days later, and they had previously had at least a hundred in stock. Similarly OCZ confirmed that their unit must be faulty and will be carrying out further tests. My original Corsair CX400 also made a lesser whining noise and drew significant power, but sometime later stopped doing so and also stopped whining. I can only surmise here, but maybe a bit of solder from some sloppy manufacturing had bridged a couple of circuits, but then burned off over time. I've only just received the FSP unit so it's too early to have had a reply.

2) I know my power meter quite well as I have tested a number of items with known power consumption, and it gives a reliable reading except for missing the first 2w of power draw. So my 11w bulb reads as 9w, the 20w one reads 18w and if something is drawing 1 or 2w it will read as 0w.

3) I don't totally understand exactly what Power Factor is or how it works, but I'm reliably informed that it has very little effect on the power consumption by the consumer, but is more about how efficiently the supplier is able to deliver the power. This video really helps: http://www.youtube.com/watch?v=dPFKcUxbNuQ.

4) The figures quoted are for when the units were all totally disconnected, so there was zero power output. The 1w figure is for 0.5w at the +5v SB rail which is required for an EuP ready motherboard, so should be less with no power draw. A modern power supply of good spec (EuP, 80+) simply should not be using more than 1w under no load or you have a faulty product that may cost you loads of money in the long run.

5) At the moment I have had four distinct faulty units in a row and I'll keep everybody up to date with my continuing attempts to find a supply that actually fulfils the manufacturer's description. This would suggest that there could be an awful lot of faulty units out there. So if you have access to a meter, please check your own one and let me know...
 
What power meter is it?


Power factor doesnt affect how much customers in North America have to pay for electricity but it does matter for some european countries, and can have a significant impact on power meters depending on the meter.
 

numskull

Distinguished
Jul 24, 2011
7
0
18,510
@hunter315: It is labelled as HO1569 by Proteam UK Ltd. I know it's not a professional tester, but as explained I know it well enough to know that even with the maximum margin of error, the readings it gives me show the units are way out of spec. And the Corsairs that I returned were confirmed as faulty and seem to have all been withdrawn by the supplier as a result.

@Motopsychojdn: Yes, I have an EuP motherboard. Probably I should have made it clearer that I ended up just testing the PSUs with them completely disconnected from the motherboard, i.e. with no load whatsoever. Therefore the 1w is actually a generous provision, since it should be a maximum of 1w at 0.5w +5vSB load, so 0w load should cause less than 1w power draw at the mains.

But the main point is that even under these no-load conditions, all these units I have been supplied with are drawing far more power than the regulations and manufacturer's specs allow, potentially wasting people's hard-earned cash and damaging the environment, and are simply faulty. Some have also drawn considerable power even when switched off at the back of the unit. And I have had 5 duff units in a row, which seems way beyond the bounds of coincidence. I am interested in knowing if I have been seriously unlucky or if a whole load of other people are being affected by manufacturers passing off shoddy, untested units as "energy-saving".
 

numskull

Distinguished
Jul 24, 2011
7
0
18,510
"completely disconnected from the mains" should of course read "completely disconnected from the motherboard", but it won't allow me to edit the reply.
 

numskull

Distinguished
Jul 24, 2011
7
0
18,510
"completely disconnected from the mains" should of course read "completely disconnected from the motherboard", but it won't allow me to edit the reply.

Now I can't delete this - help!