UPS Recommendations

Charles Curley charlescurley at
Fri Aug 27 06:00:32 MDT 2010

On Fri, 27 Aug 2010 01:56:57 -0600
Shane Hathaway <shane at> wrote:

> On 08/26/2010 09:40 PM, Michael Torrie wrote:
> > Generators bring some interesting things to the mix.  As you say you
> > can't have a generator without a UPS.  But the UPS has to be pretty
> > smart when switching back and forth between line, battery, and
> > generator.  Not only does the UPS have to carry the load while the
> > generator gets up to speed, it also has to match phase with the
> > generator.  This doesn't take very long, but you have to have a
> > pretty expensive UPS to do this.  Then when line power returns, the
> > UPS has to carry the load again while it adjusts the phase to match
> > the line phase. We have a nice big, in-line UPS at work that does
> > all this.  Works well if you have the resources.
> So, for a time, the UPS generates power at a slightly incorrect 
> frequency in order to adjust the phase?  That's a cool hack.

Yup. I worked briefly for a company that takes generators, batteries,
etc., builds them into its own UPSs, and then sticks them into back
alleys for cable companies, those green boxes on concrete pedestals
that most people just ignore.

The problem is when power returns. The battery system or the generator
is cranking out AC power, but it may be out of phase with with the
mains. Also, power may not be reliable for some seconds after it
returns, so you have to wait for several seconds between the return of
stable power and when you finally let that power on to the circuits you
are providing power to. So you adjust your phasing from 60 hertz to a
bit above or below that, for long enough to get the two in synch. Only
then do you let power from the mains go to the load.

Another, related, issue is the definition of stable power. Mains power
is normally very noisy; that is why we have surge protectors. After a
power failure it is extremely noisy. You don't want to allow the load
to see that power until it has settled down.

Similarly for the period after a power failure, when you shift from
battery power to alternator.

> I prefer the idea of putting batteries in the computer, though.  It 
> works for laptops, why not for servers?

Laptops are expected to operate on their own, out next to the swimming
pool, off the mains, for hours. Servers are not. Different mission.

Also, having hundreds or even thousands of batteries in a data center to
replace instead of only a few is a maintenance nightmare; no thanks!
And add in the recycling issues. Also now you have hundreds or thousands
of charging circuits, power conditioners, etc., for the batteries, and
that is more expensive that a few large charging circuits, etc.


Did you notice the publication date on that?


Charles Curley                  /"\    ASCII Ribbon Campaign
Looking for fine software       \ /    Respect for open standards
and/or writing?                  X     No HTML/RTF in email    / \    No M$ Word docs in email

Key fingerprint = CE5C 6645 A45A 64E4 94C0  809C FFF6 4C48 4ECD DFDB

More information about the PLUG mailing list