S
Sylvia Else
My long serving Sola UPS 325, the 750VA version, recently died on me.
Having spent considerable time and effort, and some wasted money, on
repairing it, I thought I'd post some notes about it here, so that
they'll be archived on Google Groups, and someone may benefit from them.
Others might want to add some comments.
The symptoms were that the UPS was abruptly almost completely dead. No
power was coming out of it, and the front panel was dark. The only
reaction was that a relay would click when the mains power was applied
or removed.
This seemed odd, in that even if its PSU part had failed, it has its own
battery backup. Conversly, if the battery backup part had failed, it had
a PSU. This conundrum was subsequently resolved by the realisation that
it is impossible to power up this unit from the mains unless it has a
battery connected with at least some charge in it. Without that, the
internal relays are configured so that the mains power is not connected
to anything useful.
I removed the cover. Before doing this, you should not only disconnect
the mains, but, less obviously, the internal battery. This is because
there is some possibility of shorting things out while removing the
cover. You'll need a Torx 9 (or possibly 10) bit to remove the screws.
Removing the circuit board involves unplugging a lot of leads. It's best
to note what goes where (or as I did, simply photograph it). Note that
one of the leads is plugged into the centre of the board. This lead,
together with one other plugged into the edge of the board, has a clip
that needs to be pushed away from the plug to release it. Most of the
other leads use spade connectors. You'll also need to thread the battery
leads through the slot that gives them access to the battery
compartment, since these leads are soldered to the board.
It was fairly quickly apparent that the 40A fuse has blown. (A 1000 Ah
version may have a different fuse, or two of them). This fuse is of a
type used in vehicles, and is easy to source, but is soldered directly
to the board. Removing it is reasonably straight forward if one first
cuts it in half (bit by bit, using wire cutters). When replacing the
fuse, make sure you're generous with the solder - these joints have to
carry a lot of current.
However, the fuse wasn't the only thing that had failed. I found that
two of the four power FETS mounted on large heat sinks were short
circuited, which certainly explained the failure of the fuse.
Removing the existing FETs is best achieved by cutting them off, so
their leads can be removed individually. Also I found that clearing the
holes so that new FETs could be fitted was best achieved by applying the
soldering iron to one side of the board, and a solder sucker to the
other. Knees are useful for holding the board.
Once the FETs have been removed (leaving the heat sinks in place), and
the fuse replaced, and all the leads reconnected, you should be able to
power up the unit on its batteries (by pressing the large button on the
front panel). If it's connected to the mains, it should start up in its
normal non-backup mode, otherwise you'll get a continuous beeb
indicating an error. Wear protective eye gear while doing this - you
don't want bits of liquified metal in your eyes if things go wrong.
With the mains connected, you may want to check that you have a
reasonable AC voltage between the two thick leads connected to the heat
sinks. Note that the peak to peak voltage is about 48V. This is
potentially dangerous. And, of course, you've got mains voltages anyway
on the board. So be careful! Note that these thick leads are so thick
that you cannot get a meaningful resistance reading from them with a
normal multimetre, so don't bother.
On the 750VA model it's about 32V AC RMS betwen the heat sinks. You can
also check between each heat sink and the third thick wire from the
transformer to the circuit bord, which on the 750VA model gives 16V AC
RMS. The primary purpose of these tests is to verify the integrity of
the transormer windings to which they're connected. If the transformer
has shortcircuited (which would probably trip the circuit breaker), or
open circuit, then it's certainly time for a visit to the tip.
The voltage across the larger two wire plug leading from the transformer
was about 27V AC RMS.
The voltage across the smaller two wire plug leading from the
transformer seems to be zero, despite the UPS functioning. It appears to
be some sort of sense circuit.
If the unit powers up once the FETs have been removed, then replacing
them should allow it to perform its inverter function again.
Now, this was where I made my biggest mistake. I replaced just the two
that appeared to have failed, powered the unit up, and was rewarded by a
flash of light, two new destroyed FETs and another fuse that I had to
remove and replace. Indeed, the two FETs that had appeared OK were also
now clearly damaged. Clearly, what I *should* have done was to replace
all four. Doing that subsequently made the inverter part work.
You may want to backtrack from the circuits that supply the FET gates
(through a resistor) to where there are two bipolar transistors that
drive the gates, and check them. In my unit they were OK.
At this point, I thought I'd finished. But after I'd run down the
battery while testing the inverter function until the unit turned itself
off because of battery depletion, I discovered that it wasn't charging
again. This turned out to be because the LM317T regulator attached to
the third large heat sink had also failed. Presumably it was taken out
either by the failure of the first two FETs, or by the more dramatic
event when I tried it after replacing the two FETs.
However, replacing the LM317T regulator restore the unit to full service.
There's a large diode (looks a bit like a power transistor, but only has
two terminals) near the heatsink for the LM317T regulator. It was OK in
my unit, but it seemed to me that whatever took out the regulator could
also take out that diode. It's worth checking.
So what caused the original failure? Maybe a surge from the mains,
though the unit was connected through a surge protector, and nothing
else in the house suffered.
Maybe just a chance failure of one of the FETs, the effects of which
cascaded to take out the other components.
But there was one further possibility. Some of the leads to the
transformer are held very tight against the transformer casing. Some
abrasion may have compromised the insulation. You might want to add some
more insulation in those places.
Good luck.
Sylvia.
Having spent considerable time and effort, and some wasted money, on
repairing it, I thought I'd post some notes about it here, so that
they'll be archived on Google Groups, and someone may benefit from them.
Others might want to add some comments.
The symptoms were that the UPS was abruptly almost completely dead. No
power was coming out of it, and the front panel was dark. The only
reaction was that a relay would click when the mains power was applied
or removed.
This seemed odd, in that even if its PSU part had failed, it has its own
battery backup. Conversly, if the battery backup part had failed, it had
a PSU. This conundrum was subsequently resolved by the realisation that
it is impossible to power up this unit from the mains unless it has a
battery connected with at least some charge in it. Without that, the
internal relays are configured so that the mains power is not connected
to anything useful.
I removed the cover. Before doing this, you should not only disconnect
the mains, but, less obviously, the internal battery. This is because
there is some possibility of shorting things out while removing the
cover. You'll need a Torx 9 (or possibly 10) bit to remove the screws.
Removing the circuit board involves unplugging a lot of leads. It's best
to note what goes where (or as I did, simply photograph it). Note that
one of the leads is plugged into the centre of the board. This lead,
together with one other plugged into the edge of the board, has a clip
that needs to be pushed away from the plug to release it. Most of the
other leads use spade connectors. You'll also need to thread the battery
leads through the slot that gives them access to the battery
compartment, since these leads are soldered to the board.
It was fairly quickly apparent that the 40A fuse has blown. (A 1000 Ah
version may have a different fuse, or two of them). This fuse is of a
type used in vehicles, and is easy to source, but is soldered directly
to the board. Removing it is reasonably straight forward if one first
cuts it in half (bit by bit, using wire cutters). When replacing the
fuse, make sure you're generous with the solder - these joints have to
carry a lot of current.
However, the fuse wasn't the only thing that had failed. I found that
two of the four power FETS mounted on large heat sinks were short
circuited, which certainly explained the failure of the fuse.
Removing the existing FETs is best achieved by cutting them off, so
their leads can be removed individually. Also I found that clearing the
holes so that new FETs could be fitted was best achieved by applying the
soldering iron to one side of the board, and a solder sucker to the
other. Knees are useful for holding the board.
Once the FETs have been removed (leaving the heat sinks in place), and
the fuse replaced, and all the leads reconnected, you should be able to
power up the unit on its batteries (by pressing the large button on the
front panel). If it's connected to the mains, it should start up in its
normal non-backup mode, otherwise you'll get a continuous beeb
indicating an error. Wear protective eye gear while doing this - you
don't want bits of liquified metal in your eyes if things go wrong.
With the mains connected, you may want to check that you have a
reasonable AC voltage between the two thick leads connected to the heat
sinks. Note that the peak to peak voltage is about 48V. This is
potentially dangerous. And, of course, you've got mains voltages anyway
on the board. So be careful! Note that these thick leads are so thick
that you cannot get a meaningful resistance reading from them with a
normal multimetre, so don't bother.
On the 750VA model it's about 32V AC RMS betwen the heat sinks. You can
also check between each heat sink and the third thick wire from the
transformer to the circuit bord, which on the 750VA model gives 16V AC
RMS. The primary purpose of these tests is to verify the integrity of
the transormer windings to which they're connected. If the transformer
has shortcircuited (which would probably trip the circuit breaker), or
open circuit, then it's certainly time for a visit to the tip.
The voltage across the larger two wire plug leading from the transformer
was about 27V AC RMS.
The voltage across the smaller two wire plug leading from the
transformer seems to be zero, despite the UPS functioning. It appears to
be some sort of sense circuit.
If the unit powers up once the FETs have been removed, then replacing
them should allow it to perform its inverter function again.
Now, this was where I made my biggest mistake. I replaced just the two
that appeared to have failed, powered the unit up, and was rewarded by a
flash of light, two new destroyed FETs and another fuse that I had to
remove and replace. Indeed, the two FETs that had appeared OK were also
now clearly damaged. Clearly, what I *should* have done was to replace
all four. Doing that subsequently made the inverter part work.
You may want to backtrack from the circuits that supply the FET gates
(through a resistor) to where there are two bipolar transistors that
drive the gates, and check them. In my unit they were OK.
At this point, I thought I'd finished. But after I'd run down the
battery while testing the inverter function until the unit turned itself
off because of battery depletion, I discovered that it wasn't charging
again. This turned out to be because the LM317T regulator attached to
the third large heat sink had also failed. Presumably it was taken out
either by the failure of the first two FETs, or by the more dramatic
event when I tried it after replacing the two FETs.
However, replacing the LM317T regulator restore the unit to full service.
There's a large diode (looks a bit like a power transistor, but only has
two terminals) near the heatsink for the LM317T regulator. It was OK in
my unit, but it seemed to me that whatever took out the regulator could
also take out that diode. It's worth checking.
So what caused the original failure? Maybe a surge from the mains,
though the unit was connected through a surge protector, and nothing
else in the house suffered.
Maybe just a chance failure of one of the FETs, the effects of which
cascaded to take out the other components.
But there was one further possibility. Some of the leads to the
transformer are held very tight against the transformer casing. Some
abrasion may have compromised the insulation. You might want to add some
more insulation in those places.
Good luck.
Sylvia.
)
