[TriLUG] UPS Recommendations?

[Aaron Joyner]

It's been a while since I've had my hands dirty with an oscilloscope,
troubleshooting generators, UPSes and power supplies... so take my
advice with a grain of salt.

In my experience, I've had UPSes *themselves* be grumpy about unclean
upstream power, but never had power supplies be grumpy about unclean
power.  At the worst I would expect things to be slightly less
efficient due to the side effects of low power factor with less-clean
sine waves... but nothing that's going to impact you measurably in
home use.

I'll clarify what I mean about grumpy UPSes.  We had some older APC
UPSes which were sensitive to the quality of the incoming power.  If
the since wave wasn't nice and well... sine... :) ... they'd trip and
refuse to accept the incoming power, running off battery until they
died.  This never happened with utility power, it only happened with
our relatively cheap generator when it was under heavy load.  Because
it was a low-mass natural gas powered unit, it produced a rather
"dirty" sine wave.  With an oscilloscope you could visibly see where
the two cylinders of the motor fired, causing the engine to speed up,
and where the low mass of the rotating engine slowed between those
firings.  There were also some amusing harmonics, further dirtying the
incoming sine wave.  All told, this made some of the UPSes *very*
grumpy.  :)  Those that had a button on the back, allowing us to dial
down the sensitivity to incoming power worked great.  Choose the most
forgiving setting and they worked fine.  We just swapped out those
which did not have that setting.

The moral of the story, if you're running a big datacenter, spring for
a big diesel generator with a lot of rotating mass, which overcomes
this problem... or don't be careful what UPSes you use.  As with any
emergency preparedness: Test, test, and test again.

Aaron S. Joyner


On Sun, Jan 16, 2011 at 8:28 PM, Joseph Mack NA3T <jmack at wm7d.net> wrote:
> On Sun, 16 Jan 2011, Randy Barlow wrote:
>
>> I've read that typical inexpensive UPS's (I think we should call them LIPS
>> - Less Interruptible Power Supplies), don't actually produce sine waves, but
>> approximations of sine waves, often with stepping or triangulation when they
>> switch to battery power.
>
> They used to produce square waves. Now they produce modified square waves
> (hope this ASCII art survives mailing)
>
>    ---
>   |   |
> ---     ---     ---
>           |   |
>            ---
>
> this is a better approximation to a sine wave and clocks at twice the rate,
> making it easier to build.
>
> A peak reading AC voltmeter (which is many meters) will not show 120V for
> non-sinusoidal waves like this, even though the RMS voltage is 120V. Best
> test is to try a 100W light bulb.
>
> A square wave device like a UPS is not designed for inductive loads (eg
> motor running under load). I don't expect anything will blow. You just may
> not be able to deliver any power to the motor. I don't know much about this.
>
>> I also read that such signals can be damaging to the UPS and the power
>> supply.
>
> not true.
>
> I've been using the output from switched power supplies (the one's
> generating the above waves) sine the 60's and never blown a thing attached
> to them. The only people saying this, I expect, are the makers of sine wave
> power supplies.
>
>> It seems that Pure Sine Wave UPS's cost a whole lot more.
>
> they're enormously more difficult to build.
>
> Joe
>
> --
> Joseph Mack NA3T EME(B,D), FM05lw North Carolina
> jmack (at) wm7d (dot) net - azimuthal equidistant map
> generator at http://www.wm7d.net/azproj.shtml
> Homepage http://www.austintek.com/ It's GNU/Linux!
> --
> This message was sent to: Aaron S. Joyner <aaron at joyner.ws>
> To unsubscribe, send a blank message to trilug-leave at trilug.org from that
> address.
> TriLUG mailing list : http://www.trilug.org/mailman/listinfo/trilug
> Unsubscribe or edit options on the web  :
> http://www.trilug.org/mailman/options/trilug/aaron%40joyner.ws
> TriLUG FAQ          : http://www.trilug.org/wiki/Frequently_Asked_Questions
>