XANTEX XPH75-2D Repair FIXED, still baffled

[Seekonk]

Not many postings for repair of Xantrex power supplies. Thought this might be helpful
to someone. There are some partially drawn schematics that show major power flow that
may not be totally accurate, but enough detail to provide circuit understanding. These
are the steps I use in approaching a repair. Understanding and luck are the two factors
in a sucessful repair. A little understanding makes you a lot luckier.

This nice 2006 Xantrex supply has a switch on the front panel it can be switched from
dual 75V 2A to single 150V 2A. It can also be "paralleled" for 4A. The two supplies
don't really connect together, one of the supplies is just beefed up for 4A. This supply
happened to be made in England.

The supply would only tick about every seconds with a short LED flash.  The switcher was
trying to start up, but the inverter was not sustaining it. Scope trace shows the inverter
output top and the shutdown signal from the over voltage opto isolator bottom. For better
or worse, supplies have become a lot more complicated. This one has power factor control.
The main supply cap will not even charge up unless the control electronics work. That cap
then powerers the inverter transformer which powers the two nearly identical supplies.

That inverter transformer also powers the control electronics. Everything has to work in
order to remain working. One nice aspect of this is that any failure immediately shuts
down the supply. It does make it hard to service when the initial power bump up circuit
only lasts a fraction of a second.

Interesting to note that the AC switch and fuse are after the AC line filters. No protection
in the event of an AC filter component failing.

[Seekonk]

I find it helpful to do a simple sketch of the board and major power circuits. This
ferments in my mind where everything is and where I want to look. I always use a dim bulb
tester in the AC power to protect against severe damage when testing.  However, these do
not work well with PFC supplies. At lower voltage they tend to just draw more current.
Here are some diagrams that might help someone with the same supply, not guaranteed to be
accurate. Initial ohm checking of major power components came up with nothing.I then powered
the PFC chip with an external 16V supply. That showed some life and indicated up to and
including the inverter was working. The problem had to be in one of the isolated output
supplies.

From each of these supplies came an opto isolator monitoring the 100V power supply capacitor.
The output of these were paralleled for a single voltage feedback. I assumed that these
just regulated the inverter output. The output supplies had an inductor, diode, capacitor
and FET. What you would expect in a buck converter. What I had drawn out didn't look right,
but I hadn't been too careful with the details.

[Seekonk]

I was getting nowhere fast and became impatient. Time for a little destructive testing.
Nothing on the board was really that sacred. I decided to momentarily short out the opto
LED. That couldn't cause any problem! Immediately after shorting with the scope probe the
front panel came to life and it kept workin after the probe had been removed. Must have
zapped some intermitant. Then I started to see smoke. Before I could hit the power switch
the capacitor exploded.  At least I'm getting somewhere.

An ohm check now revieled that one zener had shorted and the other opened on the opto
monitoring circuit. Besides that two 200 ohm 1/4W resistors had opened up. Not a sign of
heating or cracking. I found that unusual.  can usually get 1/4W resistors to glow red
minutes with hardly any change in resistance. Here two of them failed and yet the opto
isolator LED was not damaged. Something was really crappy about these resistors! Now that
the bad channel had been isolated I dug a little deeper and found another 100 ohm resistor
of the same type had failed open with no visible damage.

The simple buck converter was not a buck converter at all. It looks like they are using
the DC saturation of an inductor to control the output and the FET was operating in the
linear region. Sure would like to see that circuit in detail. So it appears that anyone
with a free Harbor Freight meter could have fixed this by the drudgery of measuring each
resistor.  Xantrex usually takes a lot of care in selecting parts.

[JFJ]

Sure would like to see that circuit in detail.

The Xantrex XPH75-2D is also marketed as the TTi EX752M. An EEVblog member obtained a service manual or circuit diagram (it's not clear which, from their post) for that power supply, free of charge from TTi:
https://www.eevblog.com/forum/testgear/does-anyone-have-a-tti-ex752m-bench-power-supply/

... two 200 ohm 1/4W resistors had opened up. Not a sign of heating or cracking. I found that unusual.

They may have been fusible film resistors, as used in other TTi power supplies. In the attached diagram, for example, the "+" sign after resistors R32, R100, R103 and R110, indicates that they are fusible:

(Taken from a TTi EX354D service manual).

It looks like they are using the DC saturation of an inductor to control the output and the FET was operating in the linear region.

The voltage drop across the MOSFET linear regulator is used to forward bias the Base-Emitter junction of the KSE350 transistor (in your last circuit drawing). That transistor provides a feedback reset current to the magamp (MA2), which keeps the MOSFET voltage drop to under a volt. Thereby, minimising the linear regulators power dissipation.

[Seekonk]

Thanks for that schematic, it is very similar and lets me understand the circuit.  All those resistors were certainly fuseable.  The two on the opto isolator though seem to defeat the purpose of limiting damage to the supply.