[ale] OT: U.P.S. Recommendations

[Dan Lambert]

On Wed, 2007-04-11 at 23:15 -0400, James P. Kinney III wrote:
> On Wed, 2007-04-11 at 22:45 -0400, Jim Popovitch wrote:
> > On Wed, 2007-04-11 at 22:02 -0400, Jeff Hubbs wrote:
> > > Aaron -
> > > 
> > > What you probably want are "true online" UPSses - marketroidese for
> > > UPSses that run the load off their inverters all the time.  This is
> > > distinct from the typical APC/Belkin stuff that still has a "burp" when
> > > they switch over from direct to inverter, relying to some extent on
> > > systems' power supplies' tolerance for burps and perhaps even
> > > ferroresonant transformers (meaning that they use transformers whose
> > > electromagnetic circuit is designed with a built-in 60Hz "tuning fork"
> > > effect.
> > 
> > Be warned, "true online" equals full time noise.  Modern home
> > electronics can handle the "burp" USPes without problems, plus you
> > wallet will appreciate it in addition to your ears.
> 
> Hmm. I was drooling over those when I though about the noise factor.> 
> > _______________________________________________
> > Ale mailing list
> > Ale at ale.org
> > http://www.ale.org/mailman/listinfo/ale
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale
> However, I have a backup generator that at the low end of the AC cycle
> it generates some serious crud that a large APC that powers my server
> stack doesn't like. At all. These system appear to be battery backed
> power conditioners. One would handle my issue nicely and it's in a
> closet anyway. Wife made some very unlady-like comments about the sound
> out put from the 1U server. The comments got worse when I explained it
> was one of 4. :)
> > 
> > -Jim P.

If that's the case, James, then your generator needs to be adjusted. It
should come up to speed and stabilize at 60Hz, +/- 1HZ to be correct.
What is generally more of a problem with generators is the AVR
(Automatic Voltage Regulator) sensitivity. Some generators have to have
a minimum of double the capacity of the UPS just to have enough
available energy to handle the inrush when the UPS tries to synch with
their output and return to "normal" mode.

Conversely, some of the different UPS designs are more sensitive to the
fluctuations in frequency that smaller generators produce under load.

There are several different philosophical camps regarding UPS'. Each
brand or design has it's own proponents. Each design has it's own
strengths and weaknesses, too. Some are better in some situations than
others. It's just like anything else, in that there is no panacea.
Everything is a tradeoff.

I have worked in the critical power industry for over 25 years, and have
worked on MGE (and EPE before they were MGE), MPL, Emerson, Gould,
Lortec, Powerware (And Exide before that), Socomec, Best, Liebert, APC,
and probably 20 more brands, and I can tell you for a fact that some of
them are great designs, and some are real junk. The good ones survive
and thrive. The not so good ones, well, they aren't around any more.

I now work for APC/MGE in their 3 phase power systems division. I have
very good friends that work for Powerware, Liebert, and several other
companies. I know their technologies very well, and they know mine. If
chosen wisely, any name brand UPS on the market will do quite well for
99.997% of the situations they are made to deal with. Where the rub is,
is in the fact that more often than not, when someone has problems with
a UPS, it is due to poor selection rather than poor design.

A number of people will give anecdotal evidence based on a limited
number of instances in certain circumstances, and while those are valid
case studies, for every one that someone describes for a given brand or
model of system, I can probably dig up 25 where the same system
performed perfectly.

If you want straight talk about UPS', I'll be happy to help out.
Naturally, if my company makes a UPS that fits your needs, I'll tell you
about it. If something else fits your needs better, though, I'll tell
you that, too.

Dan