UPS Games

[SSD]

I was looking through my syslog and noticed the following 2 log entries ...

 

Jun 25 14:28:26 Tower apcupsd[2038]: UPS Self Test switch to battery.

Jun 25 14:28:34 Tower apcupsd[2038]: UPS Self Test completed: Battery OK

 

Interesting to see that it would play this little game that COULD crash the server if the UPS battery were failing.

[bubbaQ]

No it wont.  If the batter is bad, it will switch back w/in one half AC cycle, and that is enough for a transformer based PSU (like in a PC) to continue just fine.

[SSD]

No it wont.  If the batter is bad, it will switch back w/in one half AC cycle, and that is enough for a transformer based PSU (like in a PC) to continue just fine.

 

Interesting.  This seems like a particularly detailed tidbit of information.  How did you come to know that?  Is it true of all APC UPSes?  Other brands?

 

I had another question about the UPS.  Last summer we had a power outage and I wound up shutting down my server shortly afterwards.  The UPS was plenty juiced and I decided to take it to my bedroom to run a fan.  The fan sounded funny all night long, like it wasn't getting enough power.  Once the power came back early in he AM, the fan started spinning at normal speed.  This is a pretty good size Honeywell fan (15 inch blades I think), but on low it doesn't blow that hard or draw much electricity.  I was surprised that the UPS didn't fully power it with no sweat!  BTW, the UPS has powered my unRAID server and a second SageTV server for 20+ minutes after a power failure without a problem.

 

Any ideas why the fan gave the UPS a problem?

[jetkins]

I had another question about the UPS.  Last summer we had a power outage and I wound up shutting down my server shortly afterwards.  The UPS was plenty juiced and I decided to take it to my bedroom to run a fan.  The fan sounded funny all night long, like it wasn't getting enough power.

 

Any ideas why the fan gave the UPS a problem?

 

From APC's KnowedgeBase:

APC's Back-UPS line is designed for use with Computer-type loads only. They are not designed to be used with motor loads such as fish filters, air conditioning units, space heaters, vacuum cleaners or any other machinery.

 

One potential issue is the inrush current draw for a motor. It can easily overload a Back-Ups UPS model. Secondly, these models of UPSs output a Stepped-Approximated sine wave when on battery. This waveshape, while ideal for computer-type equipment, is not particularly compatible with most motor loads. The result may be that motors run very slowly, erratically, or not at all when the UPS is on battery.

 

To protect this type of equipment, APC recommends first determining the appropriate size UPS for the motor load in question. Ensure that the Volt Amp rating of the UPS is sufficient to handle the start-up inrush current draw of the motor. Secondly, select a UPS which outputs a Pure Sine Wave when on battery, such as a Smart-UPS (minimum of a SUA700), Matrix-UPS, or Symmetra Power Array

[Joe L.]

You can read about the output waveform from various types of UPS here:

http://www.tomshardware.com/forum/248245-28-tomshardware

 

Most inexpensive UPS do not output a sine-wave similar to what normally exists from our outlets.  Instead, they provide either square-waves, or stepped square-waves. (sometimes, the marketing folks will call these "near sine-wave"  Trust me, they are not sine-wave...)

 

Any device you connect that is highly inductive (translation, has a transformer, or windings of copper wire around a metal core) will do poorly with anything other than sine-waves.  This is why your fan did not work properly.  If you had a APC Smart-UPS, it probably would have done just fine as the output is pure sine-wave.

 

For many consumer electronics items these days, and certainly any computer power supply I've seen in the past 20 years or so, the power supply is a "switching" supply.  They first rectify the incoming AC, convert it to a high-voltage DC, then switch it on and off at a very high frequency using a much smaller transformer at a much higher efficiencies than older supplies.    Since the input is converted to DC first, they do not care what the waveform looks like.    If your device has a big heavy power transformer, or a motor, only use a sine-wave UPS.

 

Joe L.

[bubbaQ]

If your device has a big heavy power transformer, or a motor, only use a sine-wave UPS.

 

Utter claptrap.... an urban legend bread by UPS manufacturers.

 

When you feed a square wave to a transformer, you get a nice sine wave out the other side, shifted 1/2 cycle... it doesn't directly "hurt" anything (except efficiency, which of course has to be accounted fo as it generates heat, and reduces capacity).  It is less efficient, and will result in the transformer getting hotter, but that's it.  Indeed, from the secondary side of the xformer in an properly sized system, it is almost impossible to tell whether the input was square wave or sine wave.  Xformers will also move (vibrate) more with a square wave, and that 60 Hz hum is annoying.

 

Xformers are like dashpots on doors... they store energy (for 1/2 A/C cycle)and restrict the flow of it to smooth out perturbations... they are resistive to rapid changes in voltage.  The worst part for a xformer when fed a square wave is when it powers on, as there is not yet any energy stored in the xformer to work with, so it looks like a short.  Once you have 1 full AC cycle of square wave, you are OK.

 

Squarewaves are bad for some regulating devices, such as to feed TO the UPS itself, and for things like studio flashes and motors because those devices act as a dead short or close to it at startup. I used to design power supplies for studio flashes (usually diode/capacitor trippler networks).... I know what I'm talking about.  Inductive motors are like transformers that turn on and off as each coil passes a brush.  So the stored energy effect of the coils is dissipated during rotation to the next brush.... that's why square waves are crap for motors.

[SSD]

The UPS I am using is the APC Back-UPS XS 1300VA LCD.

 

It says it produces a "Stepped approximation to a sinewave ".

 

Should I avoid running a fan off this thing?

[bubbaQ]

As a matter of practice... I would not run a fan on it unless it was an emergency.  If you do, I would not use a fan with more than 10% of the UPS rated load.

 

Now if you have a nice big isolation transformer (for example, Isobar made these... they could boost voltage in a brownout and protect equipment from all kinds of nasties) put it in line b etween the square-wave UPS and the fan, and you will be rocking.

[jetkins]

Now if you have a nice big isolation transformer (for example, Isobar made these... they could boost voltage in a brownout and protect equipment from all kinds of nasties) put it in line b etween the square-wave UPS and the fan, and you will be rocking.

 

While the load may be happy in that situation, how happy is the UPS going to be feeding its square or stepped-sine output into such an inductive load?

[bubbaQ]

After the first have AC cycle, it will be perfectly happy.  The difference between a fan and a transformer is that the fan, as each coil passes the brush it is completely discharged, and then when it touches the next brush, is a short again to the UPS.... and over and over.

[lionelhutz]

I highly doubt the fan in question has brushes.....

 

The waveforms on Tom's Hardware Guide aren't quite correct. The stepped output I have seen has the same peak as the sinewave but typically has narrower pulses. Basically, a UPS is designed for switched mode power supplies where the AC is recified. The power supply will end up with the same DC operating voltage as long as the UPS waveform has a peak voltage the same level as the peak of the utility sinewave supply. Heck, you could feed about 170VDC into the power supply and it should work. There is dual rated AC/DC equipment in the industrial world but this generally isn't seen in the consumer world.

 

At any rate, I would expect that the RMS voltage out of the UPS was lower than the RMS voltage of a sinewave and this lower voltage caused the fan to run slower. The different sound was likely due to the distorted waveform causing a different magnetic hum in the fan.

 

This makes me think of something else - how well do the power factor regulated supplies work on these types of UPS's?

 

Peter