[Tfug] UPS sizing

[Bexley Hall]

Hi, John,

--- On Thu, 1/1/09, johngalt1 <johngalt1 at uswest.net> wrote:

> > WorldCare recycles lots of these, pretty regularly.
> > Of course, the batteries are usually toast...  but,
> > then again, you *expect* to need new batteries anyway.
> 
> Too bad WC doesn't offer these UPSes sans batteries in As Is
> condition for people to restore. Guess they have sufficient
> funding..

WC's problem is time & space.  It takes time for someone to
sort through all these things.  And, space to store them (in
the *hope* that someone might come along and buy it).

I've advocated "untested" as a philosophy for putting things
in the store -- and some sort of marking on the device that
makes it quite obvious that "you bought it, you *keep* it"
(folks seem to think "All sales final" applies to all sales
*except* their own... "But this doesn't *work*..."  "And
what part of 'All sales final' do you not understand?").
I've also tried to suggest something more along the lines
of the UofA surplus auctions (clearly, participants therein
*know* better than to try to 'return' something  :>  )

<shrug>
 
> > (I have been tempted to tear some of them open and see
> > how hard they are charging/floating the batteries...
> > it seems like you should be able to get more than a
> > few years out of a set of batteries  :<  I suspect
> > they err on the side of charging "quicker" rather than
> > "safer"  :<  )
> 
> From the data sheets I have read, the best way to reduce
> life is excessive discharge for a UPS application.

Depends on the battery technology.  Overcharging and overDIScharging
are both bad.  Sort of like diesel engines -- there's a sweet
spot where they want to run.

I think the problem on the charging side is you run the risk of
internal heating and outgassing.  I suspect the vendors figure
users want quick recharge "to be ready, again" (since we tend
to be impatient creatures).  It could also be that they aren't
floating the batteries properly.  I see lots of people replacing
batteries that have never really had to carry a load!  :<

> > I've been looking at smaller units -- 500-600VA.
> 
> To me this means cheap unit - low voltage - small SLA pack.
> Destined for failure soon.

>From looking through the stuff that comes into WC, it
seems like most UPS's are discarded because the batteries
need replacement.  I haven't found any that were genuinely
*toast* (but I haven't been actively pursuing this, either)

I think UPS's are like tape backups were -- people feel
that they "must have" them... then, see very little actual
*use* for them once they have purchased them.  And, the
idea of shelling out $30 - $100 for a new set of batteries
("But I never even *used* the thing!") seems ludicrous.

<shrug>

I used to have a 1.5KVA unit here (Tucson) and used it more
to run the TV during the few outages we actually had (since
an outage that occurs when I am not using a computer means the
UPS's role is superfluous, for that original purpose)

> > With hold times just long enough for a quick shutdown.  I had
> > some of the beige 1000's (1200's?) but they were still a bit
> > too large.
> 
> Too large in what way?

Take up too much floor space.
 
> > (If I'm going to buy large, then I want to
> > be able to use marine batteries in the damn things instead
> > of the little/expensive gelled electrolytes!)
> 
> I wouldn't try to use wet cell batteries in a UPS
> designed to charge SLA batteries. Nor, would I use an automotive
> charger on SLA batteries.

Correct.  You disconnect the charging circuit completely.
Marine (duty) batteries are better as they are intended
(er, "more accustomed to") deep discharge than a regular
automotive battery.  If you further reduce the sg of the
cells, you can increase battery life significantly.

> From my interest in Solar Power, consider the following
> battery / charging info...
> 
> http://www.windsun.com/Batteries/Battery_FAQ.htm#Gelled%20electrolyte
> 
> Be sure to see the part about marine batteries.

I'll have to chase the link down, later (I'm getting DNS errors,
now).

Note, though, that the type of use seen in a solar application
is considerably different than in a UPS (fewer cycles in UPS,
*hopefully*!  :> )

> Warning: I did locate the link using Google by the way...
> 
> > How do your loads compare to the sizes of the UPS's
> > that you are spec'ing?  Are you specing large for greater
> > hold times *or* for higher loads?
> 
> Why would you spec large for hold times?. I say buying large
> for additional hold time is probably a misapplication. Just

Larger (KVA) UPS's also tend to be better designed to manage the
thermal load they experience.  A smaller UPS running closer to its
rated (KVA) capacity is probably also running a lot closer to
its *thermal* handling capability.  If you are expecting to
have to run something for long periods of time, you want as
much margin (thermal) as you can afford.

_As_a_loose_rule_of_thumb_, a 1KVA UPS will carry a given load
twice as long as a 500VA UPS (i.e., the batteries are sized
proportionatey larger).  So, you're buying increased up time
with that extra (KVA) capacity.  (yes, I recognize this is
a more expensive way to buy capacity... it has its tradeoffs).

> use appropriately sized (WRT power). Then size for runtime
> using UPSes with modular connectors (easy to add external
> packs) or ones designed for external batteries.
> 
> Anecdotally, it seems to me that battery life is much longer
> in UPSes with additional storage in parallel and UPSes not
> excessively discharged.

The problem I have seen in the circuit topology of external packs
is getting the right circuit topology so the pack is not *truly*
in parallel with the internal pack.  I.e., you want it to be treated
(charged) as a separate battery pack *and* used as a separate
(independant) power source so that the problems in the internal
pack (shorted cell, etc.) don't bias the charger or load down the
external battery pack.

I had an old HP UPS (rebadged from Astec, IIRC) that did this
well.  The external pack (60V) had its own controller, etc.
built in.  I.e., instead of just a *pair* of conductors to
the external pack, there were 5 or 6 wires to allow it to
operate truly independant of the internal pack.  (of course, you
can design a two wire circuit to work similarly but there's
more effort involved in the design as those conductors have to
serve several concurrent purposes)

--don