I picked up another power supply on ebay that I needed like a hole in the head. I just
can't resist buying a Xantrex supply for under $50 shipped. Physical condition was
gorgeous. Thought I would document the steps in the diagnosis and repair for those
starting out in repair.
I always test equipment with a lamp in series. I used a 100W for this testing as I
couldn't find one with a little less wattage. Lamps have a positive temperature
coefficient and have minimal resistance at low currents. Should a fault be present,
currents are limited by an increase in resistance of about 10 times. This allows you
to test with power and at least gives you a chance to turn it off before major damage
happens.
Unit as received will not power up, no display. Checked the large replaceable internal
fuses, these were ok.There is an on board soldered in tan rectangular block 3A fuse F3
that I can not check located just below. The underside of the board is blocked by heat
sink. From appearance of board and printing, the BUZ80 FET that powers control electronics
appears to have been warmer than I like. This FET will at least get an additional heat
sink. FET gate to drain resistance is 138 ohm, D-S is not shorted. There is a small
fuse F4, located just below the screw, connected to the FET and it is OK. Will check
this section further later.
Powered up unit and then disconnected AC. Plus and minus of bridge rectifier measured
300ish volts before decay. Capacitors are powered, but that power does not get to the
inverter FETs or control power FET. Control power FET does connect to the converter
FETs. It appears F3 is open. The converter FET gates are not shorted or D to S. I love
Xantrex supplies, but getting to parts of them is a mechanical nightmare. This unit has
the additional M5A remote board which would complicate disassembly.
It became obvious that F3 was bad. I always try to exhaust every possibility before
commiting to a tough disassembly. It was tight, but I believed I could reach the board
with the soldering iron to add some wires for a remote fuse. I went in with a needle
nose and bent the fuse till it broke off the board. Checking it, found it was open.
Soldered in a length of wire to be used for a more convenient fuse.
I powered up the unit with a 100W lamp in series with the AC. The surge limiting resistor
(shorted by a relay after delay) began to heat up and smoke. No damage done. Q4 the
control power FET was a little warm. Checked the diodes of the control supply output
and no shorts were found.
Checked the Q4 gate resistance again. This time it was 77 ohms instead of the 138 it
had been. One of the resistors looked a little overheated near it. The four to the right
are of interest Pulled U6 the UC3842 switcher out to see if the gate resistance changed,
it didn't. Things are pointing to Q4 BUZ80 that I had suspected earlier. Pulled up the
data sheet for the 3842 to see what went where.
When I checked the resistors I found something interesting. R55, the current sense
resistor, red black black silver (2 ohm) was measuring in the thousands of ohms. Lifting
it up I found a small chunk had blown off the bottom. This was replaced with two 4.7 ohm
resistors and another 15 ohm all in parallel. That combination measured 2.033 ohms. The
BUZ80 Q4 located top center of the board, which I stock, was just replaced without further
testing and I added a small heat sink to it. R56 which had been heated measured 28 ohms,
seemed about right for a gate drive resistor and it wasn't changed. Pretty amazed that
the 3842 switcher chip didn't blow. It is nice that they socket this and other chips.
Powered it up with the lamp still in series with the power and it functioned perfectly.
All repairs were made to the top side of the board and no disassembly was necessary.
All non functional dust was removed. So that was it, a fuse, 2ohm and a FET. As this is
a little smaller footprint it will take the place of the Lambda on the bench.
Home
Help
Search
Login
Register
