Enthusiast Power Protection: Four-Way 900 W UPS Roundup

[Onus]

Hellwig, a true square wave would probably work as you say, but the "modified" sine waves don't simply cross the zero-line, they stay there for some fractional period of time.
I have a SG-650. It does not like the modified sine wave of a Back-UPS 750 at all, and shuts off right away. I had to get a SUA750 for it, and that works fine.
I was REALLY disappointed that these units, particularly the Cyberpower, were not tested with a PSU like that to see if they really worked. Also, why no scope shots of their waveforms, at low, moderate, and heavy load? This article gave a small taste, but I hope was only part I of II (or more).

 

[Crashman]

[citation][nom]hmp_goose[/nom]Soooo you have nothing to say about the TripLite atoll? Dead last? Distant second? Muddled mess? Nice part for not-our-application?[/citation]I'd go with answer number four, which is covered in the page before the conclusion[citation][nom]aldaia[/nom]Wouldn't be more cost effective and energy efficient to use the same unit to both feed our components with DC and the battery?[/citation]That sounds like the way things are done in Notebooks. I've never seen anything like that for desktops, it seems like you'd need a 12-15VDC based power supply and a bunch of external parts to make it work.[citation][nom]g00ey[/nom]I've always been wondering about the possibilities to replace the small batteries in those UPSes with standard car batteries or deep cycle (marine) batteries. Since the batteries in the UPSes are standard 12V lead cells it shouldn't be a problem and this would be a cheap way to keep the computer alive for days without external power.[/citation]Well, hours anyway. I have converted used UPS's into low-cost power inverters, but I'm not sure they could handle the task of charging those big batteries. I also used an old 900VA unit with two 12V "hobby" batteries that were twice as thick as the original 6V units, for several years, because the hobby batteries were less than $10 each (two) and the replacement UPS batteries were over $40 each (four).[citation][nom]jtt283[/nom]Hellwig, a true square wave would probably work as you say, but the "modified" sine waves don't simply cross the zero-line, they stay there for some fractional period of time.I have a SG-650. It does not like the modified sine wave of a Back-UPS 750. and shuts off right away. I had to get a SUA750 for it, and that works fine.I was REALLY disappointed that these units, particularly the Cyberpower, were not tested with a PSU like that to see if they really worked. Also, why no scope shots of their waveforms, at low, moderate, and heavy load? This article gave a small taste, but I hope was only part I of II (or more).[/citation]The lab used here does not have an oscilloscope, and none of the available ones had the ability to interface a PC to record the waveform. All the companies were asked if they'd like to loan one, and all declined. This was particularly unfortunate for testing the Cyberpower unit.
As for the sensitivity of various power supplies to waveform, the site couldn't find a list and had to take a guess as to which one to use!

 

[Onus]

They were asked if they'd like to lend one, and declined? Oh well. Perhaps in Part II.
I read somewhere that the Delta-built units are particularly sensitive to waveform. That would probably apply to some of the Earthwatts models as well. Hmmm, more material for Part II...

 

[bjprice]

In re: Integrated UPS/PS. These were available years ago but didn't resolve the power requirements of CRTs. LCD monitors would need a way to attach to a PC's UPS/PS and most likely need something other than the ~3VDC to 12VDC needed by a PC.

 

[JackNaylorPE]

Was disappointed that there was no reference back to this article to see how these compared with the OCZ reviewed last year

http://www.tomshardware.com/news/ocz-ups-battery-backup,7489.html

Available at Amazon for $300 w/ purse sine wave power.

 

[Crashman]

Was disappointed that there was no reference back to this article to see how these compared with the OCZ reviewed last year

http://www.tomshardware.com/news/ocz-ups-battery-backup,7489.html

Available at Amazon for $300 w/ purse sine wave power.

Discussed this with OCZ actually, and the answer was...unfortunately, that product is EOL.

 

[wotan31]

I don't care much for APC products. They set their float voltage too high, which cooks the batteries. 13.4 volts is correct for a 12v lead acid battery while APC uses 14.1 volts. You'll find yourself replacing batteries twice as often as with other brands. No thanks.

@hellwig, it isn't a true square wave from the less expensive UPS models. It's a "modified" sine wave. Like a square wave, but with very rounded off corners. With audio equipment, there is a noticeable difference - you get loud buzzing and humm from the speakers with the "modified sine wave", while the true sine wave is nice and quiet. For a computer, it may or may not make any difference, I don't know.

 

[hmp_goose]

[citation][nom]Crashman[/nom]I'd go with answer number four, which is covered in the page before the conclusion[/citation]
The line about "white-noise generators"? Not as insightful as the judgment the other units had gotten . . .

 

[Crashman]

[citation][nom]mutarasector[/nom]Personally, I recommend getting one of the (unfamiliar brand step wave) unit.[/citation]And you work for...
Your persistence has paid off! I've reported your messages as probably spam, on units that have the same step-wave problems that this article was looking to avoid.

 

[Crashman]

[citation][nom]hmp_goose[/nom]The line about "white-noise generators"? Not as insightful as the judgment the other units had gotten . . .[/citation]How about adding "it was quiet enough not to get bashed, and noisy enough not to get recommended"? I thought that was implied, but perhaps not.

 

[g00ey]

In a word, "wrong" was the right answer. Ignoring everything else, for the simple reason that charge management isn't where it needs to be; you will destroy the UPS trying to charge and float anything behind 2x or 3x the normal capacity. There are vendors (apc is one of them at the entry level position in the market) that offer expandable runtime UPS solutions; if you are going to be relying on your backup and not just using it as a hobby project (there is nothing wrong with that) you are going to need a properly engineered solution.

I understand that it in a sense may be "wrong" to do this even though I wish that the poster who made that comment motivated his statement and were more specific. However, the fact is that there are quite a few testimonials from people who have taken an old retired UPS and put some new sealed or unsealed deep-cycle batteries into it claiming that it has been working fine for years and successfully protected them from many power surges.

The "complaints" I have read about are the following: I've read some comments that the electrolyte tend to evaporate in unsealed batteries. The reason for this is that unsealed batteries are designed for a lower maintenance/float voltage than sealed gel/lead acid cells are. This was not perceived as a big issue as the batteries kept working as normal "for 5 years and still going strong" as long as one kept the cells filled with electrolyte by supplying distilled water.
Most people reported (in the forums that I've read) that their UPSes are able to keep the larger batteries fully charged or afloat as you call it. But after a power surge the recharger unit has problems recharging the battery. The most common comment about this is that "the batteries take a really long time to get fully recharged again". Some people say that they temporarily disconnect the batteries and recharge them using a normal recharger whenever that happens.

Most UPSes have some kind of a battery health monitor. Some designs however are too "smart" when it comes to replacing the batteries with batteries that have larger capacities for reasons I stated above in a prior post. Different ways to overcome this have been suggested. One suggestion is to modify the firmware of the UPS by changing the charge cycle times. There is a community somewhere working on an open-source version of the APC firmware allowing one to make such modifications.

Sure it could be debatable whether it is right or wrong to modify a UPS in such a way but I wouldn't go as far as to say that you are "destroying" the UPS by doing so.

 

[hangfirew8]

[citation][nom]hellwig[/nom]Sorry, I'm not buying the Sine Wave vs. Square Wave and zero-power state argument, pure PR BS (I've seen power strips promise this same Sine-Wave miracle technology, wonder what the excuse is there).[/citation]

Yeah I had to laugh at that sales pitch too. The "zero component" non-problem was actually a big improvement over the old square wave UPS's that had no zero component.

 

[hangfirew8]

[citation][nom]g00ey[/nom]I understand that it in a sense may be "wrong" to do this even though I wish that the poster who made that comment motivated his statement and were more specific.[/citation]

g00ey, I don't see your idea as "wrong", obviously a decent experimenter can hack something together that works. I have done minor hacks such as replacing 5AH batteries with 7AH in little power-strip-like UPS's. The resulting recharge current requirements might strain the circuitry in a full discharge recovery situation, but not too much, so far no problem.

You've identified some of the caveats such as charge voltage, an AGM charger by itself will never bring a lead-acid battery to full charge. Keep in mind that even "sealed" lead-acid batteries have some hydrogen gas venting, and should be isolated from consumer indoor type electronics.

Auto starter type lead-acid batteries have a very low internal resistance on recharge, Marine deep-cycle a little more resistance, I could see a small consumer UPS tripping the circuit breaker (or frying something) attempting to recover a lead-acid battery from a deep discharge. This would be a fruitful area for experimentation.

 

[TheKurrgan]

[citation][nom]nebun[/nom]wrong[/citation]
Hes not "wrong" persey... The statement of deep cycle marine would net him days of run time depending on how many deep cycle he used.. 6 20 pound 2 volt for example, tied together would net about 3 days at 400 watt draw using an inverter, cant see this being too different. As to a single 12V of any sort, hours -- not days will be the result. A car battery will go for around 5 hours depending on the size and condition. Also UPS devices that are modified to hook directly up to the cars power system (say for making a decent inverter instead of buying one) will NOT work. They are incapable of dealing with the raise in voltage produced by the alternator, and will shut down.
If you need large quantities of time with out AC power, here is a recipe:
1 800 - 1000 watt DC -> AC Inverter.
1 cheap preferably used small gas trimmer
1 1000 CCA Duralast battery
1 87 amp alternator from an old car
1 pulley
1 small V-Belt
1 15" high flow fan (for the trimmer)
150 ft of 20 amp drop cord.
5 gallons of gasoline
5 bottles of 2 stroke oil.
This will keep you gaming for about 6 hours before you NEED to start the trimmer with a new battery. Trimmer + Alt will totally carry the load and STILL be able to charge the battery with a max 800 watt inverter. Trimmer is a little on the noisy side, but a muffler can be attached (lawn mower mufflers are ok)

 

[sblantipodi]

I hoped to see also the BR1500G series from APC in the article.

 

[shadowmaster625]

Average price of $380? You guys are smoking crack again.

 

[_rau_]

[citation][nom]jtt283[/nom] Also, why no scope shots of their waveforms, at low, moderate, and heavy load? This article gave a small taste, but I hope was only part I of II (or more).[/citation]

I ask the same thing, why no scope shots of the waveforms??? I'm very curious on the "Adaptive Sinewave" thing. I wonder how the wave looks like, if it looks closer to a true sinewave or closer to a stepped sinewave.

 

[Crashman]

[citation][nom]_rau_[/nom]I ask the same thing, why no scope shots of the waveforms??? I'm very curious on the "Adaptive Sinewave" thing. I wonder how the wave looks like, if it looks closer to a true sinewave or closer to a stepped sinewave.[/citation]Right, you quoted the question without reading the answer, now thumb me down and look for the answer in the thread.

 

[_rau_]

I hoped to see also the BR1500G series from APC in the article.

This review focused on UPSes that outputs a true sinewave or something close to that, which is not the case for the BR1500G.

I've read your message on the APC forum, the issues your're having with APFC PSUs and you APC BR1500G. I think the Cyberpower Adaptive Sinewave is the answer until other brands like APC do something similar.

 

[_rau_]

Right, you quoted the question without reading the answer, now thumb me down and look for the answer in the thread.

Sorry, crashman!

I'm new on this forum, and english is not my mother tongue, but I confess I haven't read all the thread! Sorry once again!

 

[g00ey]

There are a few cheap oscilloscopes out there. I've seen scopes that can go up to 50 MHz and be connected to a computer for about $200 including software. It is when you want an oscilloscope that can take measurements in the gigahertz range it gets expensive. But we are only interested in the 50-100Hz range so an oscilloscope in the kilohertz range would provide enough detail to assess the quality of the alternating voltage a UPS brings.

Also, the sound card of the computer can work as an oscilloscope and measure frequencies up to 22.05kHz/24Hz (for 44.1/48kHz sound cards - see the Nyquist rate in the sampling theorem) or 48kHz on better sound cards that support 96kHz sampling rates. You just use the Mic or Line In input of the sound card together with some oscilloscope software. I'm quite sure that there are freeware/open-source bundles available for this purpose. The oscilloscope data could then be processed in Matlab, Adobe Illustrator or even Microsoft Excel if you want the charts to look better and be easier to read. What you need in terms of additional hardware is some kind of a probe that can take 120/240V since that kind of voltage will obviously fry the sound card if not the whole computer. The high voltage probes that are available for professional oscilloscopes are not very expensive and are likely to work with the sound card using a proper BNC adapter. Otherwise you can also build one yourself. What distinguishes a high-voltage probe from a regular one is that it is coupled with a megaohm resistor. There should be instructions about it out there.

 

[Crashman]

[citation][nom]g00ey[/nom]There are a few cheap oscilloscopes out there. I've seen scopes that can go up to 50 MHz and be connected to a computer for about $200 including software. It is when you want an oscilloscope that can take measurements in the gigahertz range it gets expensive. But we are only interested in the 50-100Hz range so an oscilloscope in the kilohertz range would provide enough detail to assess the quality of the alternating voltage a UPS brings.Also, the sound card of the computer can work as an oscilloscope and measure frequencies up to 22.05kHz/24Hz (for 44.1/48kHz sound cards - see the Nyquist rate in the sampling theorem) or 48kHz on better sound cards that support 96kHz sampling rates. You just use the Mic or Line In input of the sound card together with some oscilloscope software. I'm quite sure that there are freeware/open-source bundles available for this purpose. The oscilloscope data could then be processed in Matlab, Adobe Illustrator or even Microsoft Excel if you want the charts to look better and be easier to read. What you need in terms of additional hardware is some kind of a probe that can take 120/240V since that kind of voltage will obviously fry the sound card if not the whole computer. The high voltage probes that are available for professional oscilloscopes are not very expensive and are likely to work with the sound card using a proper BNC adapter. Otherwise you can also build one yourself. What distinguishes a high-voltage probe from a regular one is that it is coupled with a megaohm resistor. There should be instructions about it out there.[/citation]I hadn't even thought about using the microphone input...brilliant! Yes, it would need a very high resistor, but that certainly SOUNDS cheap.

 

[g00ey]

I'm not sure if I would use the Mic port since it is rather sensitive but the Line In would do just fine. It is designed for signals with the voltage of 500mV to 2V so I would put at least a 10MOhm resistor on the probe circuit. The input impedance of the Line In ranges between 10kOhm and 47 kOhm on computer sound cards according to what I've read so I would put a "short-curcuit" resistor that has a resistance that roughly matches this impedance so as to protect the Line In. I would also make sure that the resistors can handle at least 5-10W. This kind of a circuit is called attenuator circuit and there are pre-built units for sale. For this kind of application a -24dB attenuator would be suitable.

Instructions on probe design can be found
here. It is a recommended read even though you would decide to buy a "professional" high-voltage probe or attenuator.

It could be worth mentioning that although I've seen implementations of this kind, I have never tried these things myself.

When I come to think about it, the sound card will probably only handle positive voltages. I seriously doubt that -2V would damage the sound card but the voltage will likely be outside its measurement range. I would solve this problem by using a small 1:1 transformer between the probe and the Line In and connect the Line In side of the transformer in series with a 1V power source. It is preferable that the power source is variable at least between 0.5 and 1V. This transformer is just as an extra precaution, I don't think it is necessary, you just have to make sure that its response lies well below 50Hz. When using this isolation transformer in such an application it is recommended to let the "short-circuit" resistor of the attenuator match the impedance of this transformer instead of the Line In. Just make sure that the peak voltage of the output/Line In side is below 1V to avoid clipping.

 

[Kewlx25]

I have an APC 1500 back from 2002. It's almost 9 years old, still on the original battery. It has seen two lightning strikes that made the lights go crazy bright and blew out other devices, a couple of surges that made the light go bright, a few sags, and numerous power outages. Still getting 15-20min run times.

My first UPS was a Cyberpower 1200, but the first time I unplugged it to test, it started to smoke and I couldn't get it to turn off. A dark warped smoking spot of plastic started to form near the power button. That's when I returned it(Same day) and got my APC. My APC cost twice as much, but I'm happy with it.

It was probably just a defective unit, but come on. A short in the power button? really? It's like buying a new car and your horn doesn't work. It's that one thing that shouldn't ever not work.

Left a bad taste in my mouth with CP.

 

[aragond]

Really disappointed some of the more affordable UPSes weren't included in the review. (900 pages of review later... laments Author) We seemingly can't get that CP downunder, but it would have been nice to compare it to the Eaton-PW-5110 series.
Still, a nice review was enjoyed.