On Wed, 28 Oct 2020 10:24:31 -0000, NY wrote:

How good are UPSs at providing mains that is sufficiently glitch-free,
when switching to and from battery, that computers don't notice?

That's what they are designed to do and there is a cycle or two of
stored energy in the PSU anyway.

we never got round to trying for a couple of years - and when we did,
the battery would not provide mains, even with negligible load, for
more than about a second. The battery was either dead on arrival or else
had failed through lack of use (maybe through being stored with no
charge).

Or simply self discharged. Did you leave it powered up for a day or
two, it might have recovered. If not the batteries can normally be
replaced.

So I've never seen how seamless the changeover is from mains to battery.
I presume they are more sophisticated than a relay, where contact
bounce, even if mains cycles match, may cause problems.

Judging by the clicks and clunks my small APC unit makes on switching
to battery or voltage reduction/boost mode it's simple relays. The
short disturbance isn't a problem.

Do most UPSs run the inverter all the time (without taking power from
it) so it is ready at a moment's notice,

Most cheap UPS's are "off line" in that the power normaly goes
straight through, possibly with some boost or buck done by tap
switching on the invertor transformer. The invertor electronics is
probably running all the time but not used, the battery is of course
maintained.

Other UPS's are "on line". The power out is from the batteries and
invertor at all times and is thus closely regulated and "clean". The
battery "charger" has to be able to provide the full load of the UPs
and charge the battery at suitably high rate to restore the output in
a sensible time after it has shut down for low battery. Ideally it
also needs to switch to a "maintenance mode" when the battery is
charged.

... locked to mains frequency and phase for instant switchover. I
presume the big problem with a UPS is when the supply is switched back
to mains, because the inverter will have been running off a free-running
oscillator that will be a different frequency so there will be a
discontinuity at the point of switchover.

Your over thinking this, at least for the small "domestic" UPS's of a
few kVA. Larger ones, 10's to 1,000's of kVA are a different kettle
of fish all together. B-) The glitch caused by a break before make
relay is not big enough to worry yer average PC or most other kit. If
kit is that sensitive normal glitches on the raw mains would be
giving it problems.

Or do UPSs stay on battery after the mains has returned, tweaking the
local frequency and phase gradually until they are locked to mains and
then switch without a discontinuity.

For small UP's over thinking again B-)

This was a cheap consumer 700 VA APC UPS, so probably one that switched
over,

Yep if that was the "Smart" variation it would be similar as the one
I have. I say similar as mine was manfactured 11/03/98 (is that March
or November I wonder?). 2 x 7 AHr 12 V SLAs easy to replace. It's on
it's 4 or 5th set as until I "got at it" it would cook the batteries
in around 4 years. Current set were fitted after I "got at it" in Mar
14 and are still good.

--
Cheers
Dave.