APC Back-UPS 280B faulty - looking for ideas

[Jeff Wiseman]

I have a APC Back-UPS 280B that has stopped working correctly.
I've had experience in troubleshooting basic electronics and
since the layout of this thing is primarly discreet components
(only one small IC if I remember correctly), I believe that I
might be able to fix it. However, I can't find anything at all
about it on APC's website or anywhere else (schematics, user
guides, nothing).

Symptoms are as follows:

In testing it using a 40 watt light as a load, the output voltage
across the light is normally about 125 VAC. When the plug is
pulled, that voltage drops to about 85 VAC.This was how it was
behaving when the problem was first discovered (power outages
caused computre to go down), so the internal settings were all as
they came from the factory.

Initially suspecting the battery, I checked the voltage on the
battery in the normal running (i.e., charging) mode. Voltage on
the battery was somewhere around 13+ VDC and when the plug was
pulled dropped only slightly to somewhere around 12 and 1/2 VDC
and held steady. I interpreted this as the battery and the
charging circuitry of the 280B were basically good.

The setting on the potentiometer for the output voltage of the
inverter section was physically set originally at about 2/3 full
scale. With a 40watt load and the unit unplugged (output voltage
at outlet around 85 VAC), I turned the pot up to full scale. This
resulted in the output climbing to around 119 VAC.

Although I can (just) get the output up to a "working" level, I
would not trust it, it's stability, or load handling since
something has obviously gone wrong with it. I do not own or have
ready access to a 'scope so I could not examine the output
waveform or frequency (I own a Fluke 27 Multimeter)

It would be nice to have this working again as I do have an
application for it, however:

1) Paying someone to fix it would cost as much as buying another one

2) I can't justify buying another one as I don't have the money
for it

3) If it is only a component or cold solder joint that has gone
bad, it should be very cheap for me to fix, however:

4) I have no schematics or info on this to help locate where the
problem is so I'm a little at a loss, not being familiar with
which levels and where to check.

If I can't fix this I'm pulling the battery and trashing the UPS
which would be shame since I'm sure it is a simple fix once the
problem is located. Any assistance, ideas, or even places I could
look to find literature on this UPS would be appreciated. Thanks!

- Jeff

 

[dkuhajda]

Sounds like a weak/bad battery.
You did not post that you performed the proper test on the battery.

Find out the MAXIMUM current output available under UPS mode of operation.
Place a suitable load on the output, calculate that your computer is not
overloading the unit. Most UPS units will NOT handle both the computer and
the monitor for very long. Then pull the plug while monitoring BOTH the AC
output voltage AND the battery voltage at the SAME time.

FYI it is normal for the meter reading voltage to drop to below 90 volts
when the UPS detects a no load condition. 40 watt light bulb is simply not
large enough to test the output, since most computers alone have 300 watt
power supplies.

UPS sealed lead acid batteries last ALMOST 2 years before they should be
replaced. Occasionally one will last a bit longer, but for full back up
performance, replace them every two years.

David

 

[Jeff Wiseman]

Sounds like a weak/bad battery.
You did not post that you performed the proper test on the battery.

No, I didn't. However, from the fact that the AC output dropped
to 85 volts but the battery still held at around 12.5-12.75
without really sagging led me to believe it was probably ok.

Find out the MAXIMUM current output available under UPS mode of operation.
Place a suitable load on the output, calculate that your computer is not
overloading the unit. Most UPS units will NOT handle both the computer and
the monitor for very long. Then pull the plug while monitoring BOTH the AC
output voltage AND the battery voltage at the SAME time.

This is what I did (i.e., monitoring both the AC output and the
battery level) although not with a computer for a load. This is
only a 280 VA unit not intended for really large loads. I was
planning on using it for my wife's CPAP machine since we
occasionally have power outages at night. The 40 watt lamp I used
was just convenient.

FYI it is normal for the meter reading voltage to drop to below 90 volts
when the UPS detects a no load condition. 40 watt light bulb is simply not
large enough to test the output, since most computers alone have 300 watt
power supplies.

So there not only is a maximum load but also a minimum load for a
given UPS? I could understand there being a voltage foldback for
a no load condition depending on the design of the unit, but a 40
watt bulb pulls 1/3 amp. That isn't totally trivial as the UPS is
only rated for a bit over 2 amps max (depending on application of
course). Seems like an extremely narrow operating range.

So you're saying that if I put a 150 volt bulb on this (over half
the maximum rated load), the output voltage should increase?
(assuming the battery was good)

UPS sealed lead acid batteries last ALMOST 2 years before they should be
replaced. Occasionally one will last a bit longer, but for full back up
performance, replace them every two years.

Yes, I have no idea how long it was in there (I got the unit from
a friend) so it could be that old. However, I still would expect
a bad battery to not be able to hold a voltage of over 12.3 volts
under load even if that load is 1/7 of the maximum load rated for
that UPS. This is a fairly lightweight unit and using it as a UPS
for a 40 watt load did seem plausible.

- Jeff

 

[Asimov]

"Jeff Wiseman" bravely wrote to "All" (28 Aug 04 17:52:25)
--- on the heady topic of "APC Back-UPS 280B faulty - looking for ideas"

JW> From: Jeff Wiseman <[email protected]>

JW> I have a APC Back-UPS 280B that has stopped working correctly.
JW> I've had experience in troubleshooting basic electronics and
JW> since the layout of this thing is primarly discreet components
JW> (only one small IC if I remember correctly), I believe that I
JW> might be able to fix it. However, I can't find anything at all
JW> about it on APC's website or anywhere else (schematics, user
JW> guides, nothing).

JW> Symptoms are as follows:

JW> In testing it using a 40 watt light as a load, the output voltage
JW> across the light is normally about 125 VAC. When the plug is
JW> pulled, that voltage drops to about 85 VAC.This was how it was
JW> behaving when the problem was first discovered (power outages
JW> caused computre to go down), so the internal settings were all as
JW> they came from the factory.

JW> Initially suspecting the battery, I checked the voltage on the
JW> battery in the normal running (i.e., charging) mode. Voltage on
JW> the battery was somewhere around 13+ VDC and when the plug was
JW> pulled dropped only slightly to somewhere around 12 and 1/2 VDC
JW> and held steady. I interpreted this as the battery and the
JW> charging circuitry of the 280B were basically good.

JW> The setting on the potentiometer for the output voltage of the
JW> inverter section was physically set originally at about 2/3 full
JW> scale. With a 40watt load and the unit unplugged (output voltage
JW> at outlet around 85 VAC), I turned the pot up to full scale. This
JW> resulted in the output climbing to around 119 VAC.

JW> Although I can (just) get the output up to a "working" level, I
JW> would not trust it, it's stability, or load handling since
JW> something has obviously gone wrong with it. I do not own or have
JW> ready access to a 'scope so I could not examine the output
JW> waveform or frequency (I own a Fluke 27 Multimeter)


I recall that most multimeters are calibrated with pure sinewaves and
their internal precision rectifiers measure the average value of the
waveform. Thus depending on the shape of the BUPS's output waveform the
reading may be inaccurate. The meter may indicate a correct reading
from the normal AC line but then read low when the BUPS is running.
I think an AC RMS meter would be needed for a proper calibration.


JW> It would be nice to have this working again as I do have an
JW> application for it, however:

JW> 1) Paying someone to fix it would cost as much as buying another one

JW> 2) I can't justify buying another one as I don't have the money
JW> for it

JW> 3) If it is only a component or cold solder joint that has gone
JW> bad, it should be very cheap for me to fix, however:

JW> 4) I have no schematics or info on this to help locate where the
JW> problem is so I'm a little at a loss, not being familiar with
JW> which levels and where to check.

JW> If I can't fix this I'm pulling the battery and trashing the UPS
JW> which would be shame since I'm sure it is a simple fix once the
JW> problem is located. Any assistance, ideas, or even places I could
JW> look to find literature on this UPS would be appreciated. Thanks!

JW> - Jeff

I would try loading the BUPS to a significant percentage of its
ratings and see how the output voltage is affected. If there isn't
much change, say less than 10%, then I would suspect the reference
voltage or the feedback divider chain. Resistors do go open circuit,
capacitors get leaky, solder joints and connectors can go bad, etc.

A*s*i*m*o*v

.... Transformer designers take turns doing it.

 

[Jeff Wiseman]

"Jeff Wiseman" bravely wrote to "All" (28 Aug 04 17:52:25)
--- on the heady topic of "APC Back-UPS 280B faulty - looking for ideas"
JW> The setting on the potentiometer for the output voltage of the
JW> inverter section was physically set originally at about 2/3 full
JW> scale. With a 40watt load and the unit unplugged (output voltage
JW> at outlet around 85 VAC), I turned the pot up to full scale. This
JW> resulted in the output climbing to around 119 VAC.

JW> Although I can (just) get the output up to a "working" level, I
JW> would not trust it, it's stability, or load handling since
JW> something has obviously gone wrong with it. I do not own or have
JW> ready access to a 'scope so I could not examine the output
JW> waveform or frequency (I own a Fluke 27 Multimeter)

I recall that most multimeters are calibrated with pure sinewaves and
their internal precision rectifiers measure the average value of the
waveform. Thus depending on the shape of the BUPS's output waveform the
reading may be inaccurate. The meter may indicate a correct reading
from the normal AC line but then read low when the BUPS is running.
I think an AC RMS meter would be needed for a proper calibration.

I believe that the Fluke 27 does read true RMS (it was a $300
meter 14 years ago). The fact that the 40 watt bulb also goes
extremely dim when the plug is pulled also reinforces my believe
that the reading is likely reasonable.

I would try loading the BUPS to a significant percentage of its
ratings and see how the output voltage is affected. If there isn't
much change, say less than 10%, then I would suspect the reference
voltage or the feedback divider chain. Resistors do go open circuit,
capacitors get leaky, solder joints and connectors can go bad, etc.

When I get the chance I'll try a 150 watt bulb (that's over half
the UPS's max rating) but I still find it a little difficult to
believe that a 280 VA UPS wouldn't work properly to protect a 40
watt load but I'm not real familiar with the normal operating
conditions of these. I have a Tripplite 550VA UPS that will
continue to properly supply just a simple 14 watt flourescent
light when it is unplugged with no side effects (although its UPS
type may be different than the APC I have).

And although circumstantial, I'm still convinced that there is a
non-battery related failure in the APC. If I could track down a
schematic or something it would really help.

- Jeff

 

[dkuhajda]

FYI if you plan on using it on any kind of motor or inductive load, you
need a different UPS. The APC units for computers are not designed for use
on any kind of motor or coil, in fact they even state they should not be
used on a heavy resitive load.

Check APC's web site for the specifications of that model.
David