APC BE550G Tear-Down

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780

If that was the intent, then UPS manufacturers wouldn't be specifying "Max 330W" or "Up to 330W" on a 550VA UPS, explicitly stating a ceiling on how much power output is officially supported. Also, if you look at manufacturers' runtime graphs and tables, they define 100% load as equating to the wattage rating. Anything beyond that is off-specs.

Being able to draw 500W from a 550VA UPS when the manufacturer specifies 330W max is more of a coincidence: stepped approximation UPS cannot distinguish between VA and W when the load is behind a diode bridge - no reactive power is ever going to come back through that bridge. I'd write that off as one of those things that you may be able to do but aren't supposed to.

One clear indication of that is the battery wires: in consumer UPS I have looked at so far, they are already on the thin side for the UPS' wattage rating even on an intermittent basis. Ex.: 80A through #10 battery wires for the SMART1000? Good thing its battery only lasts three minutes at 600W.
 

g1abhi

Expert
Ambassador
Jun 13, 2015
958
0
11,360
Continuing on my prev post, i am pretty sure my APC UPS is ultra shit quality , because i am running a skylake i5 6500 with a GtX 960 and 2 drives (1 SSD) . no matter what i do its max load can never cross over ~200W in Gaming. I had filed complaints to APC regarding this and they sent in some replacements , ( although they just kept sending the same UPS with the same model no, repacked again and again as fresh piece ) but none was able to handle the load. I honestly want to sue APC but i feel lazy :( .
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780

What is your PSU's brand and model? Crappy PSUs may not have enough hold-up time while under load to keep going until the UPS kicks in - relays can take close to 8ms to close and some PSUs don't hold even that long even though the ATX spec states 16ms as the minimum.

As Nuke said, there is also the possibility of a PSU with an APFC circuit which has a particular distaste for stepped approximation, in which case no amount of swapping one stepped approximation UPS for another will ever solve your problem. In that case, you can either buy a new $60 PSU that will work fine with your $100 UPS or buy a "pure sine wave" UPS for which your only option below $200 is Cyberpower.
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780

If you actually knew what apparent power was, then you'd know it is entirely possible for a power source to have different limits for in-phase power, VARs and overall VAs since the out-of-phase power products get returned to the source while the in-phase power and transmission losses are gone for good.

If you know so much about apparent power, then you would also know that the power factor is split between the displacement factor (phase shift of the fundamental) and distortion factor (everything else) which are only defined for a sinusoidal waveform. A stepped approximation UPS has 25-40% of THD depending on dead-time, so even if the fundamental is in-phase, your overall power factor can only be worse than 0.75 since PF = cosφ * dF with dF = (1 - dS / S), dS being the distortion power.

I wrote this from memory. If I had written this based on Wikipedia, I wouldn't have known about the distortion factor.
 

Slobodan-888

Reputable
Jul 17, 2014
417
0
4,860
@Daniel Sauvageau

After reading a bit on UPS power rating, I have found that apparent power rating (in VA) is a marketing bollocks. The actual maximum output power is real power (in W), because it is (wrongly) assumed that power factor is always 0,6.
Once again, marketing department prevailed over the engineering department, and fake specks are given. Same bollocks as TDP for CPUs...
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780

Power factor in the sense of traditional AC distribution is completely meaningless with a stepped approximation waveform and the waveform's own 25-35% THD already guarantees a power factor below 0.75.

If UPSes were actually designed to deliver watts up to their full VA rating, their runtime charts wouldn't end at the wattage rating. Running the UPS beyond its official wattage rating is like overclocking: you can do it to some extent but there are no guarantees.
 

g1abhi

Expert
Ambassador
Jun 13, 2015
958
0
11,360

I have a Seasonic SII 620W http://www.newegg.com/Product/Product.aspx?Item=N82E16817151096.

 

g1abhi

Expert
Ambassador
Jun 13, 2015
958
0
11,360

The UPS shuts off with a red light indication of overload , i am using the old APC be650y-in for now , still works better than newer cheap grade UPS offered by APC. I cant find any information about the waveform , but they both were entry level UPS and definitely not sine wave inverter ones.
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780

He said he got multiple same-model replacements and got the same result with all of them, so the same defect across multiple units seems unlikely.

~200W PSU output / 80% efficiency = 250W add an LCD and possibly other accessories, the total may climb above 300W. Depending on how the S12-II behaves when it receives a stepped approximation input, it could very well trip the UPS' battery or output over-current protection early.

My PC is low-specs compared to the vocal THG average but if I let FurMark + CPU Burn run for a few minutes, HWInfo will occasionally catch the UPS reporting my modest system peaking at 300W from an average of 250W. (That includes a USB hub, a handful of USB-powered devices and my main 24" display.)

Some GTX960 would be capable of tripping a 330W UPS all on their own: http://www.tomshardware.com/reviews/nvidia-geforce-gtx-960,4038-8.html

Look at the total graphs (last image), they show the Strix peaking at 350W, that's 420W at 80% PSU efficiency even before adding the rest of the system. None of the others break 250W. Still, add 70W for the rest of the system, a 30W LCD, PSU losses and you are over 400W.
 

g1abhi

Expert
Ambassador
Jun 13, 2015
958
0
11,360
thanks that did clear something up. Peak usage might really exceed things up. I am using a APC old software suite to check my current usage with the OLD UPS, BE-650 , it runs somewhere around 180-203W in current status during gaming. I believe this usage includes the PSU extra wattage loss due to efficiency. Correct me if i am wrong. Maybe Its not able to catch the spikes produced , and its just showing Avg usage over a few seconds of time.
I did once see 235W over the duration of a minute.
sadly i tried even using the CPU only with UPS , and offloading the monitor , still UPS shut down instantly. IS there something else i can do ?
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780
Yes, power seen by the UPS would include all losses downstream from it and the UPS' current limiting would happen at a much finer time scale than what the controller can report: to prevent a gross overload from destroying the inverter's FETs, over-current conditions need to be detected within hundreds of microseconds while the reporting interval is closer to once per second.

Without recording the current vs voltage waveform at the time where the UPS shuts off, it is impossible to tell exactly what is happening. If the S12-II's APFC circuit shuts off on non-sinusoidal input instead of going into boost regulator mode, the UPS may not like the heavy inrush current from that 470uF primary capacitor getting topped off by a stepped approximation output: during that 5-6ms of off-time between pulses, the capacitor will discharge from ~160V to ~130V and the sudden rise to ~160V will create 10-15A repetitive peaks. From the 12V UPS battery's point of view, those will become 120-150A peaks. The UPS shutting down is preferable to blowing its non-replaceable fuses or FETs.
 

Daniel Sauvageau

Honorable
Aug 12, 2014
313
0
10,780
For powering a gaming system during actual gaming through a small power outage, I'd recommend going with an UPS that uses a 24V battery pack if you can afford one: having twice the battery voltage means about half as much battery current for a given load.
 
Status
Not open for further replies.