Very odd PS issue for a washing machine motor controller ... HELP !

[Journey]

Hi All,
I have a interesting problem with a Motor controller brd from  Kenmore 417.44042400 front loader washing machine ... motor control part# 134393900.
Initial problem was obvious but odd ... The inrush current limiter decided to desolder itself from the brd and fell out. Owner saw it and resoldered it back in and it ran for 12 loads and quit again. Some other repair idiot decided it wasn't needed and jumped where the thermistor sat ... moron. Of course it still didn't work. Then it came to me and found the thermistor was blown and a cap had lost it's guts ... Replaced both and it still wouldn't work ... I took it back and decided to draw up a schematic and during that process discovered a broken solder joint ... hooked it up to AC and took some voltage measurements.Everything seemed fine and then wanted one more reading for curiosity sake and spark/pop and a transistor blew. Replaced tranny and tested no sparks but no voltages for 12v and 5v. Found that both zeners has gone short ... replaced those and tested again and all hell broke loose. All 3 trannys popped and both zeners again and 2 resistors and another diode.
So I replaced it all again and tested it blew up again.
I started looking at the voltage that feeds the regulator part and thought the measured 160vdc was very high for a 12v and 5v simple zener regulator. Also the measured output from the rectifier was at 325v ridiculously high.
When I drew the schematic part of the mains in and rectifier et al ... I saw something very odd never seen it done quite this way before.
The second time it blew up ... I discovered the 1st tranny replaced went in backwards ... don't ask how/why long stupid story. SO then I built a mock up on bread brd and got everything straight and worked fine. Replace all blown units again and poof

So with that history so far ... here's what we are looking at. First drawing is the mains/rectifier/caps section with initial voltage reading with ref to GND.
What the hell is this with both AC inputs to bridge tied to hot and neutral between the caps ? And again how could the bridge positive output to GND be 325VDC ?
The 2nd image is the regulator section that is fed from between those 2 filter caps which initially measured 161VDC
The voltages shown here are from the breadbrd mock up and I was only using 40v so I wouldn't blow anything up and waste parts and it seemed to work fine.
I am pretty much all out of clues as to why after replacing every blown component it keeps blowing up ?
At this time I am not knowing how to insert images so hopefully the attachments are available after I post this.

[Journey]

For some reason my 1st attached image didn't post here it is ... this the mains section.

[Journey]

Quick Update,
I decided to use the neutral line between the 2 filter caps a reference ... I got 157v from positive bridge terminal and -157v to the negative terminal.
So now that design is finally making sense but using -157v to feed the 12v and 5v regulator doesn't make sense given the that all 3 transistors are NPN's.

[Journey]

Minor revision to regulator section ... one more connection I missed.

[Journey]

Just a speculation but according to Q10 datasheet it IS a NPN ... is it possible that it should be  a PNP ? just spit ballin here ...

[abdulbadii]

shouldn't U8 node 3  straight down to N ?

is source own REed put on Kicad?

[Journey]

Hi There,
Yes ... in a normal design pin 3 would be connected to neutral ... but this design is an oddball.
It seems they did it this way to have a split PS pos/neg voltages without the need of a center tap transformer.

And yes this is a KiCad drawn schematic that I did to help troubleshoot.

[fzabkar]

These dual capacitor designs are intended to service both 120VAC and 240VAC markets by simply moving a jumper wire. Your current configuration is a voltage doubler.

See the AC input section of this dual voltage ATX PSU design:

http://www.pavouk.org/hw/en_atxps.html

Note the 230V/115V selector switch.

In short, you should measure 160VDC across each capacitor, and a combined voltage of 320VDC. That's normal.

[Journey]

Hi There,
Well that's an interesting take but that circuit clearly shows the AC inputs of the bridge in a normal configuration ... L or Hot to one AC input and N Neutral to the other AC line in. And you are correct that it seems to be a doubler kind of setup because if I measure the positive out from bridge using ground as reference I get 325V. If I use Neutral as reference I get the usual RMS x Square root of 2 for DC output on positive out and and equal negative voltage on the minus pin of bridge so about +160 and minus160V.
I had an idea though to bring that 160V down that feeds my zener regulator. There is a wire jumper that connects and brings the 160V into the reg circuit ... SO I decided to put a 150K resistor to drop it down to about 40V which seems to work fine on my mock up on the bread board.
However it might limit the current way too much.

[fzabkar]

You should leave the voltage at 160VDC, in normal operation.

The regulator is a switchmode type, not linear. Some other part of the circuit, perhaps the microcontroller, must be sensing the 12V supply and then controlling the regulator by pulsing the base drive of Q9.

Q9 and Q10 constitute a Darlington pair. Q10 has a current rating of 300mA. Q9, Q10 L2, C43 and D12 constitute a buck regulator (step-down converter), and R78 (12 ohms) is the current sense resistor. If the load current increases beyond 50mA, the voltage drop across the resistor becomes 0.6V, and this then turns on Q8, which then diverts the base drive away from Q9.

ZTX458, Diodes Inc, NPN transistor, 400V, 300mA, 1W, TO92/E-line:
https://www.mouser.com/datasheet/2/115/ZTX458-92715.pdf

US1J, Vishay, ultrafast rectifier diode, 600V, 1A, 75ns, marking UJxx, DO-214AC:
https://www.vishay.com/docs/88768/us1.pdf

In short, the high supply voltage is OK because Q9/Q10 are being pulsed on and off, and Q8 is limiting the current ... when everything is working as it should.

[Journey]

wow cool that you explained that ... but doesn't really explain why after replacing all faulty components why it keeps blowing up ?
I have included a bigger picture of schematic show micro controller.
The only access micro had to the regulator is seen at U4b pin 15 so how can it pulse Q8 base ? I took scope readings of all 3 transistor bases on my mock up on the bread board to to look for pulsing and found none. So I await your reply ...

[Journey]

Just had a thought ... if that VAUX on U4 is cooked then we may need a new controller. Only other idea I had was to make whole separate PS using transformer bridge and cap and two linear regs to supply the 12 or 18V and the 5V and then McGuiver it to the main chassis of the unit.

[fzabkar]

I'm stumped. I can't see any source for the pulses that I'm expecting, nor can I see how the output voltage is sensed. :-?

Maybe I've got it wrong. :-?

[fzabkar]

Just had a thought ... if that VAUX on U4 is cooked then we may need a new controller. Only other idea I had was to make whole separate PS using transformer bridge and cap and two linear regs to supply the 12 or 18V and the 5V and then McGuiver it to the main chassis of the unit.

VAUX0 is sensing the 160VDC supply by way of a 100:1 potential divider (470K / 4.7K). It is not involved in sensing the 12V supply. Normally you should see 1.6VDC on this pin. However, because you have reduced the supply voltage to 40VDC, the microcontroller sees only 0.4VDC, thinks that there is a brown-out, and then goes into failsafe mode. At least that's what I would do if I were the designer.

[Journey]

What do you think about just adding a second PS via Xfrmer bridge cap and linear regs and fasten to main chassis and jumper it ?

[Journey]

And as far as the 160V ... I can jumper that to the Vaux sense and the micro won't have a hissy fit

[Journey]

Just one other note that may be a hint ... Q10 is the transistor that actually loses pieces off it ... the emitter leg usually gets blown right off and the front side gets blown off. The other 2 Q8 & Q9 just get cracked on surface of case and internals messed up.

[Journey]

Another idea ... I could use a voltage divider to bring reg input down to say 47V and again fool the micro with a jumper to still see the 160v.

[Journey]

Do you have KICad ? I will post the whole schematic for ya ... I tried a screen cap of it all and it's not readable.

[fzabkar]

I guess you could add your own 12V PS, after removing the switchmode regulator. You wouldn't need to fiddle with the brown-out detection divider.

[fzabkar]

Do you have KICad ? I will post the whole schematic for ya ... I tried a screen cap of it all and it's not readable.

I don't have any CAD tool. I can understand your circuit drawings, it's just the design that I don't understand.

[Journey]

I just thought you might like to have the whole schematic ... no worries. KiCad is free if you want a fairly decent electronics schematic design tool.
It also does pcb layout as well ... should check it out. I used to use OrCad 9.2 but this one is newer and does pcb's too.