Sort of OT: UPS battery protection?

[KenK]

I have a UPS protecting my computer's power with three 12v 7a batteries.
Recently while I was away for a few hours the power went off and evidently
the UPS ran the batteries down and they failed. I had to replace them.
Perhaps they were about due for replacement anyway but I'll never know for
sure.

Anyhow, anyone aware of a timer of some sort I can use on the UPS so after
ten minutes (or whatever) after the AC goes off it turns off the UPS (or
cuts power to omputer)? If I'm there I'd shut down long before ten minutes.

Suggestions?

TIA

 

[Zaphod Beeblebrox]

I have a UPS protecting my computer's power with three 12v 7a batteries.
Recently while I was away for a few hours the power went off and evidently
the UPS ran the batteries down and they failed. I had to replace them.
Perhaps they were about due for replacement anyway but I'll never know for
sure.

Anyhow, anyone aware of a timer of some sort I can use on the UPS so after
ten minutes (or whatever) after the AC goes off it turns off the UPS (or
cuts power to omputer)? If I'm there I'd shut down long before ten minutes.

Suggestions?

TIA

Many UPSs come with or have available software to load on the computer
that allow you to do exactly that after connecting the UPS to the
computer with either a serial or USB cable. What manufacturer and model
UPS are you using?

 

[J. P. Gilliver (John)]

I suspect so (why three BTW?). Virtually all UPSs _will_ shut down when
the battery voltage gets too low, to prevent just that kind of damage.It'd have to be something that detected the input power going off. I
can't think of an external such device; it'd not be hard to design, but
I suspect nobody makes one.

Many UPSs come with or have available software to load on the computer
that allow you to do exactly that after connecting the UPS to the
computer with either a serial or USB cable. What manufacturer and model
UPS are you using?

Wot 'e said. They set the computer to shut down when things are getting
close. But you have to load the software, and connect the UPS to a
serial or USB port, for this to work - the software tends not to work as
a timer, more on getting a "things are getting close" signal from the
UPS.

 

[Paul]

I have a UPS protecting my computer's power with three 12v 7a batteries.
Recently while I was away for a few hours the power went off and evidently
the UPS ran the batteries down and they failed. I had to replace them.
Perhaps they were about due for replacement anyway but I'll never know for
sure.

Anyhow, anyone aware of a timer of some sort I can use on the UPS so after
ten minutes (or whatever) after the AC goes off it turns off the UPS (or
cuts power to omputer)? If I'm there I'd shut down long before ten minutes.

Suggestions?

TIA

The UPS should stop the inverter from operating,
when the battery reaches 10.5V. You can check this
while doing the calibration run the UPS documentation
may suggest.

It's actually more complicated than that. There
are at least two factors.

1) The electrochemistry equation for the battery, has a
temperature term. The open circuit voltage varies with
temperature. Failure to account for this, can lead to up
to a 50% error in capacity determination. If the lead-acid
battery operates over an extreme temperature range, the
threshold used should take the temperature into account.
Perhaps the value is 10V at one temperature, 9.5V at
another temperature, and so on. (somewhere around
10.5V being the "deep discharge" threshold value.)

For a normal home user, the temperature range in the
house is reasonably well controlled. But the UPS could
be stuffed into an unheated garage or hot attic.

2) The lead acid battery has internal resistance. It is a
relatively small 12V battery in your UPS, from which
you'll be drawing 25 amps during offline operation.

To give a practical example, from my car. When the starter
motor was failing, I got the following measurements. We'll
pretend I measured 12V (nice round number) under no load.
Using a clamp-on DC ammeter, the peak current was 150 amps.
Terminal voltage dropped to a minimum of 9V. Both meters used
have peak/minimum detection capability, which detects the "worst"
values on a run. 12V-9V / 150A = 0.02 ohms internal battery
resistance. That's a rough value for my car battery. The SLA
in the UPS is a smaller battery design, and the resistance will
be higher than that.

So you can't just measure the battery voltage while it's under
an extreme load, and conclude what the charge state is. If I
had used a 10V cutoff in my car, the starter motor would have
been disconnected before getting to the 9V "stall" point.
The 9V value does not imply the battery is flat, but that
the diffusion limit of the electrolyte has been reached, and
the battery impedance is becoming an issue. If I was designing
a cutoff circuit, it would notice the 150A of current, and
use R*I to compensate for the extra drop.

One would hope the UPS takes some of this into account. Even if
they set the threshold to 12V, instead of 10.5V (zero capacity
versus deep discharge point), that would be better than
nothing.

The UPS should have done that on its own. If it fails to disable
the inverter at the 10.5V point, then the UPS is defective and
should be repaired. You won't get much battery life if it
runs the batteries flat all the time. No lead-acid chemistry is
going to like that. Golf cart or marine batteries ($$$) are designed
for that sort of operation (can sustain a limited number of
deep discharges), but car batteries aren't very good at that
at all. The car battery won't put up with that too many times.

I doubt they intended the SLA in your UPS, to be flattened. Maybe
a $50 UPS would do that, but not the next price tier up. That
should have some monitoring.

I was shocked at the circuit board design in mine. No microcontroller.
Just a ton of "jelly bean logic". I couldn't see any measurement
capability as such, implying simple minded voltage comparator-type
thresholds were being used. Most of the board might have
been the logic to run the timer feature (sends alert to
computer when only 2 minutes of battery remain). So my design
appears to be too old for anything clever inside it.

*******

The reason batteries have a cutoff value, is there are multiple
cells, and if one cell runs flat before the others, it can become
reverse biased and lose its plating. Some battery chemistries,
if a single cell is operated all by itself, it might actually
tolerate operation to zero volts. It's when multiple cells
get into a fight, and the "weak loser" gets de-plated, that's
why the battery must be shut off before it gets to that point.
We assume a 10.5V reading, implies one of the cells of the
six of them, is flat and about to be de-plated. Something like
that. They might actually have equal voltage across them,
but, not for long. The cells cease to match, after years of
use, and one of them becomes "a loser".

Summary: The UPS should be protecting your batteries for you.
If it is not doing that, seek warranty repair, or,
replace it. While a timer sounds like a fun project,
you may not find exactly what you want, because
with a properly working UPS, there'd be no need for it.

And while I could find a battery cutoff circuit for
sale, it wasn't rated for 25 or 50 amps or whatever.
And the design looked super-crude (no temperature or
load compensation).

Paul

 

[Bob F]

I suspect so (why three BTW?). Virtually all UPSs _will_ shut down
when the battery voltage gets too low, to prevent just that kind of
damage.
It'd have to be something that detected the input power going off. I
can't think of an external such device; it'd not be hard to design,
but I suspect nobody makes one.

That's part of most UPS's.

 

[philo]

Many UPSs come with or have available software to load on the computer
that allow you to do exactly that after connecting the UPS to the
computer with either a serial or USB cable. What manufacturer and model
UPS are you using?

Correct!


BTW: If the batteries were 4 or 5 years old they were due for
replacement so no great loss there. If they were any good they should
have recharged.

 

[Paul]

When you left your home, couldn't you turn off your PC to avoid a such
incident? I believe such incident was/is not difficult to foresee.

There is a data cable that runs from the UPS to the computer,
that can tell it to shut off. Mine uses RS232, while newer UPS
use a USB cable.

In WinXP Control Panels, Power Options, UPS tab, there is a provision
for accepting the message from the UPS box, and initiating a shutdown.
Works for certain brands and models of UPS. Mine happens to be supported.

That removes most of the load from the UPS, but it doesn't necessary
stop the inverter on the UPS. Only if the UPS had "no load" sensing
and it turned off the inverter, would you have a complete
solution (for when you're not home). An inverter with no load,
still wastes battery electricity, and eventually the battery
will run flat. Only a low battery voltage cutoff on the UPS control
board will stop it entirely.

Some battery chemistries are much more sensitive about the
battery condition. They don't fool around with Lithium based
batteries. They have both low battery cutoff, as well as
precise charging circuits, to avoid starting a fire. If a
Lithium battery has too low a terminal voltage, the charger
will refuse to charge it. At least with lead acid, there is
no reason to fear that situation (as long as the charger
is current limited). I notice there are some models of
UPS (for Japan) that are now using Lithium rather than Lead Acid
(the traditional SLA battery for UPS). I personally wouldn't
want Lithium, as they wear out faster. I managed to get ten
years from my UPS SLA battery. That would never happen with
a Lithium chemistry (using laptop batteries as a predictor).

Paul