Power Consumption of UPS units


olympia

Recommended Posts

I am using an APC Smart-UPS 1500. I was curious about the power consumption of the unit itself, so I measured today with a kill-a-watt and I was actually shocked to see it is sucking 60W in idle with nothing connected. It was also fully charged at the time of measuring.

 

I am wondering if this normal... I thought it should be max. 15-20w.

...or is this something with my unit?

 

Another, maybe silly question: is the UPS using more power as the batteries are wearing out? I.e. could the issue be due to weak batteries (I can imagine maybe they need continous charging or something)?

 

Link to comment

Was that meter measuring 60 watts or 60 VA?  It sounds more like the latter. 

 

Some kill-a-watt style power-meters do not take the effect of the power-factor of the load into account.  Other users here have had similar issues with strangely high off-state power readings when taking measurements of their servers.  Unless the UPS was charging the batteries at the time (and you say that it wasn't) then this looks like the VA reading, and I would guess that the actual wattage was probably about 10 -15 watts for a UPS rated 1500.

 

Checking my APC BackUPS 1000 RS just now, it reads about 8 - 10 watts when idle with nothing connected and batteries charged.  If I set my meter to VA, then I see about 21 - 22 VA, and the power factor reported is about 0.44.

 

As jumperalex pointed out, some units can do a great deal better.  I have read that some of the CyperPower UPS units have a quiescent power draw in the 2 -3 watt range.  I intend to try one the next time I lose a UPS (or if a replacement works out cheaper than the batteries :) ).

 

Link to comment

I'm very surprised at that level -- 60 watts ... or even 60VA is FAR too much.

 

The unit almost had to be charging the batteries when you measured that.    I'd be inclined to leave it plugged into the Kilo-Watt for a couple hours with no load on the UPS and then see what it shows.

 

By the way, are you sure there was in fact no load -- i.e. was anything plugged in to the "surge only" plugs?  The best way to ensure NO load is to have nothing plugged in to the UPS.

 

Link to comment

I've got 6 UPS units ... 2 Belkin 1500VA, a couple APC 1000VA units, and a couple CyberPower units (1000VA and 1500VA) => but all are on systems I run 24/7, so I don't want to shut everything down to measure the idle loads, but the next time it's convenient I'll do that for at least the APC and CyberPower units [the Belkin units aren't available anymore].  But I did at one time check the APC unit that's connected to my backup server, and if I recall correctly it was only adding 2-3 watts to the total power draw.

 

I do have a new CyberPower 1000VA Sine Wave unit ordered that's due to arrive tomorrow, so just for grins I'll measure the idle draw before I put it in service (after it's had time to charge the batteries).

 

Link to comment

Thank you for all feedback so far!

 

You were right, it is actually 60VA, my meter is not featured with wattage measurement and since I am not deeply in electricity, I always considered "VA" as "watt".

 

My UPS is the older model (SUA1500I), so it is probably not equipped with latest green technologies, still I find 60W(VA) is indeed FAR too much.

 

...but I don't think it actually do charge the batteries. If I am not mistaken in charging mode the UPS activates the fan and get a lot more noisy. What actually it does if it really charges. When this happens, then it draws 110-130VA.

 

...or is it possible that there is a low charging mode, when the unit actually doesn't turn on the fan, but still charging?

Link to comment

I suspect that there are 2-stages of charging ... a faster mode for when the batteries are discharged lower than some specific percentage; and a slower, less intense mode for when they're getting pretty close to a full charge.

 

The fan is likely thermally controlled -- when the charging circuit heats up, the fan kicks in (it will probably also come in if the source of heat is the inverter ... so if you're on battery power and drawing a fairly high load you'd likely hear it kick in then as well).

 

As I noted, you could test this by running it without a load and seeing if the power draw drops to a lower level after an hour or two.

 

Link to comment

I've two UPS at home.

 

APC SmartUPS 750 LCD (about 5 years old) : idle with nothing connected - 42-46W floating

Cyberpower ValueUPS 600 (just over a year old) : idle with nothing connected - 20W exactly

 

The APC's efficiency is terrible at no/low load.  Once load goes up the efficiency is much better, it goes up in to the mid-90% once you've got a few hundred watts of load on it.  The Cyberpower is a "GreenUPS", and seems to have similar efficiency with no load compared to high load.  I've no idea how that works, but it does seem to.

 

Really, like power supplies, you need to size your UPS correctly - too big and it's way out of it's efficiency sweet-spot. 

 

The APC also hums annoyingly, though it is much better since I replaced the two leaking caps on the board.

Link to comment

I suspect that there are 2-stages of charging ... a faster mode for when the batteries are discharged lower than some specific percentage; and a slower, less intense mode for when they're getting pretty close to a full charge.

 

The fan is likely thermally controlled -- when the charging circuit heats up, the fan kicks in (it will probably also come in if the source of heat is the inverter ... so if you're on battery power and drawing a fairly high load you'd likely hear it kick in then as well).

 

As I noted, you could test this by running it without a load and seeing if the power draw drops to a lower level after an hour or two.

 

Last night I tried what you suggested and the idle power draw dropped from 60VA to 41-45VA overnight without a load, so it certainly seems like an issue with the batteries. However, I wonder where would it end up with new batteries. I feel 40VA is still a lot. Dilemma is now whether if to buy new batteries and live with whatever I end up or try to look up some good deal for another, greener UPS, e.g. the CyberPower mentioned here.

Link to comment

40VA may be a lot, but it is very much dependent on the power factor of the load.  If the UPS has a poor power factor when not powering anything and when its batteries are charged, you might be looking at a real power consumption of only 10 watts or so, and it is the real watts figure that the utility company will be measuring. 

 

 

Link to comment

... edit: I don't even think they actually have a 220v model for Europe, do they?

 

No, I don't think they do.  Didn't realize you were in Europe.

 

... It's surprising they haven't made a 220v model, as it's a VERY popular tool that I'm sure would sell very well in 220v markets

 

 

Link to comment

Really, like power supplies, you need to size your UPS correctly - too big and it's way out of it's efficiency sweet-spot. 

 

Hmmm, does this mean that the 10% load what I have when all HDDs are down in my unRAID server is an issue for me in terms of this increases the consumption of the UPS itself?

Link to comment

Really, like power supplies, you need to size your UPS correctly - too big and it's way out of it's efficiency sweet-spot. 

 

Hmmm, does this mean that the 10% load what I have when all HDDs are down in my unRAID server is an issue for me in terms of this increases the consumption of the UPS itself?

Well "issue" is strong word there. While you might be getting less efficient as you pull less power through the UPS, you aren't pulling more power out of the wall. That would imply a rate of decrease in efficiency I cannot imagine any UPS possesses.

 

Try the math. Imagine 80% efficiency at 150watts load, which is 187.5watts out of the wall. Now spin down all your drives, and stop all cpu load, for a 50watt load. At the SAME wall draw of 187.5 you need a drop in efficiency to 50/187.5 = 26.7% I can't imagine any UPS gets that bad.

 

Nor should you be considering buying a new UPS tomorrow for its low-load efficiency unless your payback period is shorter than the expected life of your UPS. For that you're going to need a guess at expected life, expected wall draw over that life, and your electricity rate. I'm gonna bet that you should just wait until the UPS dies and THEN spend the premium (if it even is a premium at that point) for a more efficient UPS. Technically you should again be making a pay-back calculation but I think its safe to say it would probably be worth it at that point to buy the "efficient" model.

Link to comment

I wouldn't be at all concerned about the UPS.  Note that any efficiency differences in the inverter for various size units only come into play when there's a load on the inverter -- i.e. when the unit is on battery power.

 

The rest of the time (i.e. 99+%) the unit is simply serving as a superb surge protector, but is simply passing the A/C power through to the load.    The monitoring electronics that is active during that time is effectively the same whether it's a 500VA unit or a 1500VA unit.

 

The exception is with very high end units that are always using the inverter to provide total isolation from the incoming power.  These are FAR more expensive than anything you're going to buy for a home (or even a small datacenter). 

Link to comment

I do find the statement "The APC's efficiency is terrible at no/low load" rather comical. At no load the efficiency of every UPS is 0%.

 

Seriously, you need to measure the watts to have any real clue about what the load costs you in electricity.

 

Gary gave the right basic idea. If the unit is a switching unit and the monitoring and battery float charging electronics draw say 20W then you take that part of the wall draw to calculate the efficiency of the UPS. 100W from the wall with 20W to the UPS and 80W to the load means 80% efficiency.

 

But, the units that use the inverter full-time are not that expensive. Every small data center should be running that type.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.