Last night, while feeling particularly good about my shiny new computer, I decided to test out the UPS it's attached to. This is an 825 AVR Cybperpower unit, the first time I've ever bought a UPS that isn't APC.

I made sure both the CPU and GPU were running full tilt, drawing something just north of 400 watts (including the monitors), and unplugged the UPS and expected to get a beep, a sad noise from the speakers (they're not plugged into the battery side of the UPS), and a computer that continues to happily run as if nothing had happened. The UPS is rated for 450 watts, so there's some room to spare, although not a lot.

What actually happened was I got a beep, a sad noise from the speakers, and everything on the battery side of the UPS power-cycled. That was not good, and not at all expected.

Next I tried it with the GPU idle and the monitors plugged into the non-battery plugs, so the computer was drawing about 200 watts. Same result -- the computer and network switch both power cycled.

Repeated again with the computer at idle (about 125 watts) and everything transitioned to battery power just fine.

I've had this UPS for about 8 months and never had a problem before.

Somewhat annoyed, I fired off an email to Cyberpower tech support complaining about how I'd always been happy with APC, Cyberpower's products are obviously inferior and defective, and their mothers wear combat boots. I won't hear anything until Monday; their tech support only works weekdays.

Assuming I'll need a new UPS (and kind of wanting to get a larger one anyway) I started looking around at both APC and Cyberpower UPSs and learned something new. This is important information for anyone who uses a UPS with their computer.

Less expensive consumer UPSs produce simulated sine wave outputs that step up and down rather than having smooth sine wave outputs.

More expensive consumer UPSs (and probably ALL business UPSs) produce smooth sine wave outputs.

For the most part, this doesn't make a difference. At least it didn't use to. It matters now.

Modern computer power supplies are more efficient than ever, and also incorporate something called "active PFC". I'm going to skip the electrical engineering lesson (you can look it up if you're interested), but active PFC is a good thing, and more and more power supplies are using it.

Active PFC and stepped sine waves do NOT play well together. If you get a new computer, or a new power supply, it's possible it might not work with your old UPS. If you buy a new UPS, you should make sure it produces a true sine wave output, not only because you're current computer might need that, but because your next computer is even more likely to require it.

I just ordered this UPS. It's got twice the capacity of the old UPS and has a true sine wave output. It's about the same price as APC's 1500 VA Back-UPS Pro, but the APC does NOT have a true sine wave output and would have the same problem as my old UPS. To get an APC UPS with a true sine wave output I think you have to get their Smart-UPS products, and those are targeted at businesses and are a lot more expensive.

____________
My lucky number is 75898⁵²⁴²⁸⁸+1

That's great to know - i was thinking of buying one myself. Let us know how the new one works. Btw, does it show how much power is being used? like a kill-a-watt does?
____________

275*2^3585539+1 is prime!!! (1079358 digits)
Proud member of Aggie the Pew

That's great to know - i was thinking of buying one myself. Let us know how the new one works. Btw, does it show how much power is being used? like a kill-a-watt does?

Along those lines, here's what my old UPS is reading as I drain down the BOINC queue in preparation for the challenge start in 30 minutes:

(This is with just the computer and a very green network switch plugged into the battery side of the UPS.)

4 cores running PPSE-LLR:

220 watts

3 cores:

198 watts

2 cores:

184 watts

1 cores:

162 watts

idle:

130 watts

____________
My lucky number is 75898⁵²⁴²⁸⁸+1

Along those lines, here's what my old UPS is reading as I drain down the BOINC queue in preparation for the challenge start in 30 minutes:

(This is with just the computer and a very green network switch plugged into the battery side of the UPS.)

4 cores running PPSE-LLR:

220 watts

3 cores:

198 watts

2 cores:

184 watts

1 cores:

162 watts

idle:

130 watts

130W idle ???

____________
Best wishes. Knowledge is power. by jjwhalen

130W idle ???

This is an overclocked Q6600, isn't it?

A couple of misconceptions there.

Those readings are WATTS, not VA.

That computer is the new Core i5. The cpu itself shuts itself down very nicely when idle, but the rest of the computer is another story. (I've seen the CPU drop to below 1 watt when idle.) The motherboard, disk drives, optical disk, etc. are all drawing power. Also, computer power supplies aren't very efficient at low power. And don't forget I have a network switch on there too. It's very power efficient, but it's still drawing 5 watts or so. It all adds up.

(To be honest, there's a lot of power saving options on the motherboard I could probably use, but mostly it's either going to be crunching, or I'll put it to sleep, so idle power isn't a big deal.

The old 460 and the new 580 are similar technologies, so power draws are comparable. The 580 is about twice as fast and draws about twice the power. It consequently runs a lot hotter.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

Okay, but the german magazin c't released a Haswell based system config with an idle power of ~11w without external gpu...
If you change the config from here, you should end up with max 20w at idle. Adding an optical drive with typical 15w, two 1TB hard drives with 7.5w each instead of the used SSD and your GTX580 with an idle power of ~30W from http://www.xbitlabs.com/articles/graphics/display/geforce-gtx-580_5.html, you should see between 70 and 80w at idle state on the UPS.


What was the idle power consumption of your overclocked Q6600 with GTX460?
____________
Best wishes. Knowledge is power. by jjwhalen

Okay, but the german magazin c't released a Haswell based system config with an idle power of ~11w without external gpu...
If you change the config from here, you should end up with max 20w at idle. Adding an optical drive with typical 15w, two 1TB hard drives with 7.5w each instead of the used SSD and your GTX580 with an idle power of ~30W from http://www.xbitlabs.com/articles/graphics/display/geforce-gtx-580_5.html, you should see between 70 and 80w at idle state on the UPS.


What was the idle power consumption of your overclocked Q6600 with GTX460?

I just heard back from Cyberpower's tech support today, and they confirmed what I had learned on my own. The "simulated sine wave" UPS's are not compatible with active PFC power supplies, and the power supply (and potentially the computer components) can be damaged by the UPS.

As far as I can tell, most consumer grade UPSs (including all APC consumer UPSs) produce simulated sine waves, so we really need to be careful about purchasing UPSs.

Cyberpower, fortunately, does have a consumer line of true sine wave UPSs that aren't absurdly expensive.

Here's the email from Cyberpower. Note that everything they're saying matches what I had been able to determine independently about this problem.

Your computer uses a power supply that utilizes Active Power Factor Correction (Active PFC) to improve efficiency. Power supplies with active PFC may experience incompatibility problems with a UPS that does not provide pure sine wave power output when the power supply switches from AC power to UPS battery power. As a result, when a computer system using a power supply with active PFC is attached to a non-sine wave UPS, the system may shut down when it switches to battery power. Also, if the power supply continues to operate, it may produce a humming or high pitch noise while running on battery. This humming indicates the power supply is operating beyond specified tolerances and may damage the power supply.

For computer systems using power supplies with Active PFC, we recommend that you protect your equipment with a UPS that provides sine wave output. We make a full line of Pure Sine UPS models that meet most technical and pricing requirements.

You can view our Pure Sinewave models on our site via the following links for more information:
http://www.cyberpowersystems.com/products/ups-systems/pfc-sinewave-series.html
http://www.cyberpowersystems.com/products/ups-systems/smart-app-ups/pp-series.html
http://www.cyberpowersystems.com/products/ups-systems/smart-app-ups/paragon-series.html

Just wanted to add to the discussion and point readers to X-bit labs as they do their PSU tests using an APC smart UPS that is stepped rather than a true sine wave. Some PSUs will work okay with it (even some with active PFC), and some will not. It is a nice place to start if you are looking to pick up a new PSU unit.


EDIT:
This information for the APC website may also be useful.

Cyberpower also has a Windows 7 desktop gadget that shows various things, including watts. As far as I know, the consumer level APC UPSs do not provide that information. They can tell you the battery state, but not the power being used.

That's not true, APC PowerChute software shows the power being used and actual input voltage, even for consumer level UPSs - personally tested so far on APC Back UPS Pro and Back UPS RS & CS. I got my first APC UPS more than ten years ago and it had those features even then.

Also, I currently have an APC UPS with step-approximated sine-wave output (Back-UPS Pro 1500, BR1500GI) and it works fine with my Active PFC PSU (Seasonic X-650,SS-650KM) - the battery kicks in normally every time the power goes out or if there is electrical noise.

Another brand that I think is underrated and surprisingly affordable is Liebert - they are a huge player in data center power systems, and I didn't realize they have several lines of really good smaller UPS's as well - and they are also pretty price competitive.

This ‘input’ power should be the basis for sizing the UPS, so as not to Overload the UPS. This can be calculated by taking the PFC power supply’s rated output power and multiplying it by 1.25

UPS just delivered the new UPS to my door. (Yes, I've been waiting to be able to say that. I'm easily amused.)

It works as expected. The battery came about 1/3rd charged, but the UPS is working and is able to power the computer running at full tilt.

It's physically smaller than I expected, is actually kind of nifty looking, and the USB charging ports on the front save me from having to plug in a charger somewhere.

Unlike APC UPSs, this comes with the batteries connected, but it won't power on until it's connected to utility power for the first time.

Be warned that if you buy a UPS (any brand) from Amazon, you can not return it. Amazon doesn't accept returns on UPSs. If that's important to you, buy from someone else such as Newegg. On this particular item, at that exact moment in time, Amazon and Newegg were charging the exact same price. Also, depending on where you live, you may need to pay sales tax from one retailer but not the other.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

You need to make sure that the watts rating of the UPS is at least as high as the watts ratings of the PSU

Not quite. The watts rating of the UPS needs to be as high as yours PC power drain at maximum load irrelative to the watts ratings of the PSU. But on the other side, the more watts rating of the UPS the longer the battery run-time.

An Energy Star 4.0 compliant power supply has to be more than 80% efficient. For example, if a attached power supply is delivering 600W output power, its ‘input’ power can be as high as 750W. .
This ‘input’ power should be the basis for sizing the UPS, so as not to Overload the UPS. This can be calculated by taking the PFC power supply’s rated output power and multiplying it by 1.25

The power consumption may be easy metered by the watthourmeter.

Michael Goetz, you are right, but I mean another thing. To measure power you need powermeter, of course. But if you do not have it then you should purchase it. And the powermeter in an UPS is not very accurate. But there are watthourmeter in almosl all houses which one measures energy and "produces" electricity bill. Sorry, my English is sad, but I try to explain. This watthourmeter produces a watt hours, and it has sufficient accuracy. So if you measure the watt hours and a time during which this watt hours was consumed, and then do some mathematical manipulations, you obtain watts that your PC consumes. Of course, to use this method you should turn off all electrical appliances in the house except the computer during the measurement.

Okay, but the german magazin c't released a Haswell based system config with an idle power of ~11w without external gpu...
If you change the config from here, you should end up with max 20w at idle. Adding an optical drive with typical 15w, two 1TB hard drives with 7.5w each instead of the used SSD and your GTX580 with an idle power of ~30W from http://www.xbitlabs.com/articles/graphics/display/geforce-gtx-580_5.html, you should see between 70 and 80w at idle state on the UPS.


What was the idle power consumption of your overclocked Q6600 with GTX460?

The disk drives are black "performance" drives, not green, so they're probably a few more watts.

Don't forget that the power supply has about a 25% overhead on top of what the computer is actually using. (If it's 80% efficient, that means it's using 25% more power than it's supply to the computer components.)

I think the q6600 was topping out around 350 watts, but that was with the 460 which uses less power but was including the two monitors (which are about 35 or 40 watts combined.)

We're getting WAY off topic here -- this is supposed to be a warning that many consumer UPSs won't work with a lot of new computer power supplies.