who where wrote:
On Tue, 08 Dec 2009 13:14:30 +1100, Sylvia Else
wrote:

who where wrote:
On Mon, 07 Dec 2009 14:09:42 -0700, D Yuniskis
wrote:

Hi,

I don't have anything other than "gut feel" to resort to
as evidence but it sure *seems* like most UPS designs
EAT batteries!

We rarely have outages here. So, there is no real
draw on the batteries in our UPS's. So, they should just
be on float charges.

Yet, the batteries seem to wear out far too frequently, IMO.

I could possibly understand regular outages cycling the
batteries too deeply.
That's not what kills most UPS batteries.

And, chargers too aggressively
replenishing them (after all, a UPS that quits because its
battery wasn't fully recharged earns a bad reputation for
its manufacturer).
That, but also an unreasonably high float voltage, is what kills them.
We've seen literally dozens of UPS' from assorted manufacturers with
failed batteries - baked dry, swollen, cracked cases. On fitting new
batteries and checking the charger behaviour, we note that the
recovery charge rate is definitely "too aggressive" (motive onvious)
and the eventual float charge voltage was definitely destined to kill
the cells within a year or so.

The only time we were happy that the cells were within the
manufacturer approved envelope was ... on DISCHARGE.

But, I can't see how even periodic
battery tests (performed by the UPS itself) could be the
problem -- unless there is something wrong with the
approach being used?

Can anyone with firsthand knowledge shed some light on this?
I.e., what *is* the UPS doing to/with the battery when it
is not being used to supply the load?
Both my old and my new SOLA UPS use a float voltage of 13.5 volts (near
enough), which is at the low end of the battery manufacturer's
recommendation (13.5 to 13.8). I get about three years out of batteries
in the old UPS. Too soon to say about the new one.

That takes care of one of the killer parameters - float voltage. But
at what rate do they recharge after a serious discharge?

I haven't checked that. My AVO meter apparently has a significant
voltage drop on its current range which messes up the result

Sylvia.