Help troubleshooting an SMPS - I feel like I'm missing something simple.

[Vintagevalves]

Hello!  I'm attempting a repair on a switching power supply for a Pfaff computerized sewing machine.   I have the service manual for the machine, but it is scant on electronic details.  There is no schematic available or even any voltage callouts.    Someone attempted to 'fix' it before me.  And by that I mean, they tried to replace some caps, while ruining PCB traces and putting in too low voltage of caps.   

As it sits now the supply boots to hiccup / blink mode.   I was told by the previous owner than it was 'clicking' and wouldn't turn on.  I believe she was hearing the hiccup mode (the main transistor chirps a bit) 

This conditions happens whether the PSU is connected to the rest of the machine or not. 

What I've done:
-Repaired PCB traces as necessary.   
-Tested most resistors on line side out of circuit where necessary to verify value. 
-Tested and reinstalled the supplied old main bulk capacitors (in spec)
-Removed and tested all large output transistors and diodes (All appear to test good on my GM328 component tester)
-Tested all small diodes in circuit as best I could with DMM.  (No shorts or opens). 
-Replaced output filter caps (they were in spec but I had replacements) - removed and tested remainder of electrolytic caps.   
-I am finding some voltages on the output connector which I will list below.
-Drive IC is a 3844.    If I provide voltage to the LED side of the optoisolator the 3844 reacts by shutting down the supply (blipping stops and output voltages disappear).  If I apply voltage to the Vcc input the blipping stops, but again, the detected output voltages disappear. 

I'm not sure what to check next.  Is it possible that the drive IC is just not switching correctly?   I tried to scope the output, but the blips are so short, I'm not sure if I would be able to see the switching or not. 

main output voltages during blip mode:

1.) 46-47v (varies with pulsing) MA6D50
2.) Appears to go to Foot pedal/speed control area of circuit
3.) 2.8v - goes to Foot pedal control circuit
4.) 5v TL7705CP
5.) Ground
6.) Ground
7.) 4v MA644
8.) 4v MA644 (same one as above - maybe this one is the main output since it's on two pins?)
9.)11v MA749A
10.)5v TL7705ACP
11.)5v L7805CP
12.)Ground
13.) 0v 0122 BD577
14.) C3111
15.) Ground
16. Ground
17.) Foot Pedal
18.) Ground
19.) Ground
20.) Unused - goes to and unpopulated part of the board.

[Haenk]

Clicking might have been PSU going into protection mode, so maybe the fault is not within the PSU - did you check the rest of the machine for shorts?

[Vintagevalves]

I did meter the output supply cable going to the rest of the machine and the only thing I found dead shorted were pins which I traced back to the grounds on the PSU board. 

The supply goes into hiccup mode regardless of whether it is connect to the rest of the machine or not.  When connected to the rest of the machine I do not observe any additional current draw through the mains.     I have tried placing resistor loads on a couple of the supplies and that did not help.  I put about 300 ohms on the 48v and 8ohms on the 4 ohm - I'm not sure if those were adequate values or not to prove anything. 

[xavier60]

The UC3844's Vcc pin 7 needs to come up to 16V for the IC to become active, driving the MOSFET.
 It is usually supplied by a small startup current from high value resistors from the HVDC rail.
Vcc then needs to be sustained above 10V usually from a rectified transformer winding.
Trace the circuitry that supplies pin 7, hopefully finding a problem like shorted diode or open fusible resistor.
Does the PCB have cracks?

BTW: don't do anything to break the regulation loop.

[Vintagevalves]

Thanks xavier60,
When I view the incoming voltage on pin 7 on my scope, it appears to blip high enough to trigger turn on (around 17v).    I can also see the chopper pulse output on pin 6 as the PSU blips (momentarily of course).    I did try to connect a bench power supply in parallel with pin 7.       The blipping stops at around 12v on the bench supply, however the chopper output stops as well.   (In case anyone is about to freak out, I'm running the DUT on an isolation transformer).       

There are no cracks on the board that I have found.  I will trace back all the components related to pin 7 and re-verify. 

[xavier60]

The 12V should keep it active. When this family of IC is in their active state, the 5V reference should be present pin pin 8.
Maybe the IC is being externally disabled.
With the 12V applied measure pins 1, 2, 3, and 8. There should also be oscillator sawtooth on tin 4.

[Vintagevalves]

With external power feeding Pin 7 I get the following:
1: 6.7v
2. 0v
3. 0v
4. 0v (no sawtooth that I can find on my scope)
5. Gnd
6. 0v
7. 12v
8. 5v

[xavier60]

With external power feeding Pin 7 I get the following:
1: 6.7v
2. 0v
3. 0v
4. 0v (no sawtooth that I can find on my scope)
5. Gnd
6. 0v
7. 12v
8. 5v

There are various ways of shutting down these ICs. Stopping the oscillator seemed unusual at first.
The circuitry on pin 4 needs to be traced.
A clear photo of the area might allow us to help also.

[Vintagevalves]

BTW, no shorted diodes or open resistors related to Vcc. 

[eblc1388]

There is a capacitor from 3844 pin4 to common, check it for leakage.

You can also feed the 12V to the 3844 pin7 via a diode, thus allowing the pin voltage to rise above 17V to starts its proper operation.

[Vintagevalves]

Xavier,
Here's some photos of the PCB.  I've highlighted the relevant traces for pin 4.   A few notes:
I'm aware of the missing pads (thanks previous 'tech').  I had them jumped with wire and just took the wire off for this photo for the sake of clarity.

The trace terminates at the collector of Q2 a C3311  https://datasheetspdf.com/pdf/618093/Panasonic/C3311/1

Hard to get a photo of the topside of the board because of a heatsink Q2 is just behind R13

[Vintagevalves]

Based on the original looking solder on the bottom of the board, I looks like the missing components (C4 D4 R20 D5) were never populated.

[xavier60]

Check the Base voltage of Q2 to see if it's being driven.
Looks like either ZD1 or ZD2 are able to supply Base current to Q2.
But I can't see what is supposed to latch Q2 on and where the 2 Zeners are supplied from.
Don't disable anything, it could be OVP doing its job.

[xavier60]

Also, now is a good time to take note of the value of the Current Sense resistor/s in the MOSFET's Source circuit and Gate resistors in case a disaster occurs.

[Vintagevalves]

The Base of Q2 has 0.6v on it.   

While the supply is blipping ZD1 has about 12v on it's cathode, and once it bring up my external supply to VCC the voltage there increases to whatever my external supply is set at.   Voltage to ZD1 comes off a winding on the primary transformer and is dropped through two resistors.

While the supply is blipping, the voltage to ZD2 is blipping
With 12V applied to Vcc, ZD2 shows 6.1v on its cathode and 1.4v on it's anode which passes through R13.   

Power to ZD2 is coming from the optoisolator.

[Vintagevalves]

Good idea!  I will make a little spreadsheet of those values tomorrow before any more poking and prodding.  Thanks!

Also, now is a good time to take note of the value of the Current Sense resistor/s in the MOSFET's Source circuit and Gate resistors in case a disaster occurs.

[xavier60]

The Base of Q2 has 0.6v on it.   

While the supply is blipping ZD1 has about 12v on it's cathode, and once it bring up my external supply to VCC the voltage there increases to whatever my external supply is set at.   Voltage to ZD1 comes off a winding on the primary transformer and is dropped through two resistors.

While the supply is blipping, the voltage to ZD2 is blipping
With 12V applied to Vcc, ZD2 shows 6.1v on its cathode and 1.4v on it's anode which passes through R13.   

Power to ZD2 is coming from the optoisolator.

It can be difficult to work out what is the cause or a side effect. ZD2 could be part of an open loop protection,meaning that it's causing Q2 to be driven because the IC is disabled. Faults can be very dynamic, making it difficult to find the origin.
I would like to see a photo that includes the optocoupler.
One possible direct method of finding the original cause is to remove Q2, then externally powering the IC, being certain to get the voltage high enough for the IC to start.
Instead of using mains, supply the HVDC filter capacitor with another bench supply while monitoring the output rails. The rails will come up at a surprisingly low input voltage if there is no overload.
You will be looking for signs of regulation.
If so, also confirm 2.5V at the reference pin on the TL431 voltage sensing IC that should be controlling the LED side of the optocoupler.

Also: As eblc1388 suggested, apply voltage to the ICs Vcc pin via a diode. This will allow the external PSU to be switched off to see if the Vcc  voltage is being sustained normally.

[Vintagevalves]

I'm feeling you on the chicken vs egg conundrum. 

I've been powering the IC through a diode the whole time. 

The opto LED is being driven by another C3311 on the secondary side (no TL431's on this board that I can find).   It is receiving voltage from an MA644 rectifier.  While the supply is pulsing there is about 6v on the collector and base.   

[Vintagevalves]

FIXED! 

After a bunch of circuit tracing I was able to deduce which output and diode were the the primary.  I desoldered all of the other output diodes.   At that point I had no output on the primary voltage feed which previously had 4v (maybe bleed over through a zener diode or the optoled drive.  I didn't put much effort into figuring that out). 

I noted that there was no DC coming from the diode feeding the circuit.  I traced back and determined that there was no continuity from the transformer to the anodes!    Early on in my troubleshooting, I noted that the solder joints on this diode were broken - at the time I had high hopes that this would solve the problem.  I re-soldered them and it did not fix the issue so we dug in further. 

  Well, it turned out that not only were the solder joints broken, but the trace was broken too!   I bridged the broken trace, resoldered everything and now I have a live power supply and sewing machine!   I've got some new main filter caps coming to install today and then I will reassemble.

Thank you all for your patience, wisdom and guidance!  I learned a lot and am a lot more comfortable with SMPS troubleshooting now.