On 4 June 2014 09:16, Paul Sladen <[email protected]> wrote:
> On Wed, 4 Jun 2014, Paul Thornton wrote:
> > A UPS that drops the load as soon as the mains power fails
>
> Have you load-tested the batteries? If not, get a drop-tester or
> other large dump load, and see what happens to the voltage when you
>
I see that the batteries in this case have been ruled out, however, I feel
this is worth saying as the list may not be very power-electronics oriented
and the batteries are the usual cuplrits, being consumables :-
(for smaller UPS batteries, 12vVolt SLAs) 12v 50W car headlamps make
convenient loads as long as you don't knock them while hot (small vibration
eg: when driving are tolerable), but bear in mind that they get bloody hot
and a China Syndrome is hard to stop in a panic once the furniture is
alight.
If the battery banks can be disconnected and physically isolated easily,
testing of those with moderate loads (don't abuse the batteries just to
"test" them) at least should be relatively simple (with due regard to
explosion risk caused by gassing off) if rather dangerous for the untrained
(*)
For higher-voltage banks of 96V (nominal) or more, I'm a great fan of 300w
and 500w linear Halogen bulbs used in parallel networks (the "old
fashioned" ones that used to be common for outdoor floodlights) because
they are enclosed so reduce (but do not remove) the risk of catching fire.
They can be combined is series or parallel if you know what you are doing,
as can cheap mains halogen heaters, 400w or 600 per bar (so-called
"infra-red" versions of the old 1/2/3bar electric fires). With 40 or 50 of
these heaters (shop around in summer you can get them for 6 quid a piece in
quantity, they are handy for DR/BC work in wet/damp or unheated
cold-standby alternate sites) you can test quite a lot of large battery
banks under reasonable load for a minute or two to get an idea of how the
banks or individual cells are coping with discharge into
approximately-known loads (derived from the average power consumed, rathr
than an attempt at calculation from measured cold resistances **)
For real Heath-Robinson loadsand serious tests you can buy a few hundred
meters of resistance wire and make up a large spaceframe to have a variable
load connected by adjustable clamps. This usually means evacuating the room
to due to the touch-hazard and having someone sensible stand by with a 5kg
or larger CO2 extinquisher but gives great flexibility at
infinitely-variabe low resistances and high powers. The characteristics of
good quality resistance wires are plotted on graphs so you can then make
good estimates of resistance at any given temperature/current/resistance
(all three of which will be dynamic during the test, remember). you need to
be careful with your theory to take this approach but I find that a
planar-spiral-wound frame around a meter square, winding on both sides, is
about the best compromise of portability and useful loads. I keep one for
high resistances and a few different ones for low resistances that I can
arrange in series or parallel to handle most of the kit I tend to see.
Combine the above with a notebook and voltmeter (DC amps being difficult to
measure with cheap equipment and no formal training - the limit for cheap
kit is about 5amps on a 40 quid multimeter for any length of time, ("10
Amps" it says printed on the meter casing but normally with a very
prolonged cool-down period in the small print of Chinglish instructions -
avoid anything that has an unfused 10A range!) and you have a nice way to
prove large batter banks are in workable condition when removed from the
UPS.
Your main point of reference is to buy one single new battery of the same
type (or as close as you can get, datasheet-wise). You then perform the
same test on that when fully charged to the same cell voltage and compare
the reading with one of the the old ones.
If nothing else, this sort of approach gives you the chance to watch
Frankenstein at work (from a safe distance of course)
(*) a competent power electronics engineer or possibly a very experienced
industrial electrician (you're looking for a 50-year-old guy with scars and
a pocket full of pens and slide rules, not a sparky with only 5 years
experience or merely wiring stuff in) should be able to perform basic tests
for you. You don't actually need a UPS specialist per se, because they
don't use any voodoo magic, though I admit they need experienced hands,
eyes and brain.
(**) Incandescent lamp filaments and resistive heating elements have
dramatically different hot and cold resistances and (splitting hairs for DC
batterys) dynamic impedances (inrush current changes and rise-time effects
mixed with delta-T resistances) to surges due to any inductive effects
(mainly relevant only in spiral-wound heating elements, which can make the
first second or so rather hard to comprehend, so voltage readings need to
be taken after a few second when the load has stabilised.
--
sent via Gmail web interface, so please excuse my gross neglect of
Netiquette
