Dynasty 200SD low buss voltage

[jjohn76]

I had to register on this forum to do one thing. To say THANK YOU!
I was in the process of reverse engineering that PCB myself and was wondering why the flyback was only working occasionally.
You mentioning that it needs Vbus operational was very helpful.

It lead me to check where the 810V goes from RC1 on PC10. Found the first 750k resistor, part of the voltage divider of 3x750k, to be open circuit.

You just saved me a couple of hours of troubleshooting. Thanks a lot!

Glad it worked out!

[thomasbowen]

Hello Jon,
I too want to thank you for all the investigation and documentation you have supplied. I have an older Dynasty 200DX that worked perfectly for years until it was returned to me non-operative.
After sitting on a shelf for a long time I decided to try and fix it because I have a car restoration I am working on and my mig is not providing the results i need.

Upon inspection I found the following damages on PC2:

C10 - large capacitor, 5uf @ 1000V blown almost off the pc board.
hole in pc board under what used to be C10
C4 and C8 - 2 large buss capacitors show same blistering at the top you photographed
C7 - small electrolytic shows some bulging.
All diode tests pass EXCEPT the one you mention in post 11: RC4-6/8 and the C12/C13 point directly on the module. This fails (open).

My first question is: have you put together any sort of schematic for the PC2 board?
second: what would cause the roasting of C10 in such a manner?

I am a EE, but not much experience with circuits like this, and I appreciate any help you can provide!

Thanks,
Thomas

[alvideo]

Hi everyone,
I need help.
My Dynasty not powering up at all.
After buck drive repair, i powered it up, and found, that DC BUS volgate is 560V instead of 810V, panel wouldn't light up.
So, PC10 board not functioning.
In red color are my real measurements. It seems, all power signals are ok: 30V, +15V -12V.
I found, that boost IGBT signal comes from U3 CD4093 4 pin (+15 V) and goes via transitors to RC1-4 pin. A this point is -12V, no generating any PWM signal. Do You have any ideas?

I don't understand, what purpose is UC3845 PFC controller? U5 UC3854 14 pin shorted to ground. This pin is CT (Oscillator timing capacitor). A capacitor from CT to GND sets the PWM oscillator frequency. If 14 pin is grounded, this IC can't wok, i think. Also i can't find where 16 pin (Gate drive) is connected?

[jjohn76]

John,
PM sent. I never got around to the boost IGBT driver/controller section before giving the welders back. That said, it looks like the BJT Q29/Q30 is the totem pole driver pair, signaled from pin 4 of the CD4093. Pin 4 is the NAND output of pins 5 and 6, which is worth tracing back to their sources. My guess is one will originate at UC3854 and the other one will likely originate from an opamp or comparator used as a zero crossing detector, or at least low threshold detector for the inductor current (LEM feedback). Here's why:
The UC3854 is a fixed PWM controller that works well in continuous conduction mode through discontinuous conduction mode. However, I don't remember seeing turn on snubbers (inductors present) for the boost IGBT, which likely rules out continuous conduction mode because of the high turn on loss. Discontinuous conduction mode works, but will mostly likely have higher peak currents than a Critical/Transition Conduction mode controller, which charges up the current in the boost inductor to the reference current (duty cycle will always be below 50% because the bus voltage is so high), then shuts off and waits until exactly when the inductor current returns to zero before it turns back on. This creates a variable frequency PWM, which the UC3854 isn't designed to do.

The way around this is to use the UC3854 only to determine the correct turn on time based on the output voltage error signal (difference from 810V reference), the input voltage waveform (input current needs to follow the input voltage sinewave shape for greatest power factor), and compares this to the current across the boost inductor. To figure out the off time, use a circuit that measures the inductor current, and when it drops below a certain value, turn on the IGBT.
Hopefully that makes sense. Either way, if you trace back pins 5 and 6 on CD 4093, we can figure out if that is what's happening.

Jon

[jjohn76]

John,

Last thing, check out this guy's work on that board. It's much easier to trace out the schematic with the board unpopulated in these photos. It doesn't show the ground or power planes, but you can get those pretty easily by holding your board up to the light and seeing where the cross-hatched connections are.

https://www.element14.com/community/message/273474/l/miller-boost-circuit-board-assembly-no-200841#273474

[jjohn76]

For the group, John (alvideo above) converted all of my chicken scratch schematics (and then some) for the boost card below. This is awesome work! He has also drawn out the PFC boost circuit, though which seems to me like it's using a fixed frequency PFC chip (UC3854) in a variable frequency control system. As far as I understand it, it looks like average mode control to set the on pulse width, then uses the rectified voltage to determine the variable off time. Please let me know if I am off. Also, if you're working on the board, please confirm the schematic, particularly in the lower right and upper right. Something seems off about the wiring for that quad NAND IC.
I am about to start working through an XMT 350 (the micro-controller and peripherals look ok right now), and will do what I can for a schematic. It probably won't be as clean as John's work though...

Jon

[Bushytails]

Nice!

Miller really should be providing these...

[jjohn76]

Bushy, I don't disagree, but would imagine it's more liability than anything. I am making a big assumption that folks who would spend the time going through that schematic would spend the time drawing the schematic for the board when the welder comes up in the queue - it's just a time saver and bumps the welder up the queue a bit. I throw out the caveat that I haven't been able to compare the PFC circuit (bottom) and boost driver circuit to the board. It doesn't make sense to me how the output signal from the boost inductor current comparator (bottom right, traces along the right edge of the paper) connects to the SR latch on the Quad NAND gate in the upper right. Also it looks like that opamp circuit in the upper right sets the off time based on the input rectified voltage, but I don't quite see how it triggers the SR latch and the NAND gate that signals the output driver. Hopefully that made sense...

Anyways, I am starting on the XMT 350 shortly - I have one with a fried auxiliary power circuit.

Jon

[kvans32]

Thanks for sharing jjohn. This thread has nugged me in the right direction.

I have an sd200 that I am looking at for a neighbor. It looks like the -Bus rectifier diode shorted and put AC on -Bus. Or something happened on PC2 that took out that diode. I believe I have all of the bad components identified, plus many that still test good but were near ground 0 on board 207818. D2 and Q11 are shorted, but not blown up like I have seen in photos. R6 is toast, and R4 is roasted. I think I'll also replace R5 due to it being blackened by R4, and R8&9 because they look bloated. I'm not sure if that is just the way they are manufactured or due to excessive heat.

I contacted Miller support, asking if they could provide me with values for R4-R8. The first time they said "R4-R8 should each read less than 1 ohm". I knew that was BS, but proceeded with disassembly, hoping that I could find those answers from another source.

The two components that I have the most trouble identifying are R4 & R6. R4 is toast, but still reads 127.27ohm, so I assume it should be 130ohm (Br O Bk Bk Gd). R6 was burnt in two and gives me a value in the Mohm if I try to probe the larger half. Someone on the E14 sites says it should be 0.18ohm. I was initially gun shy on that when I saw references to +30V on the one side (RC2,4) with the other side connected to C (RC2,2). But it's feedback "for 30V", so I feel better about that number than I did, but have not found a super clear picture that gives me confidence in 0.18ohm.

After questioning Miller on their less than accurate info, and hoping to get just the value of R6, they responded "1000 ohms at room temperature". Which sounds like BS also.

Can anyone confirm any of this?

[jjohn76]

So you provided the Miller Techs the board number and component number, and they wouldn't give you the nominal values for the resistors? Things may have changed in their tech support since I called last - I haven't worked on one of these in a long while.

Are you checking across the power module directly, or going off measurements from the pre-power checks in the tech manual? It's odd only one diode from the bridge rectifier is bad. Of the four diodes in the rectifier (can't remember if these units took 3-phase, so am assuming single phase), a diagonal pair would need to be shorted to feed AC. Your buss capacitors would certainly be bulged as well.

Check the board picture above and look for the buck drive section of the schematic for those resistor values. I don't have one to check against anymore.

I'll see if I can scrape away some time to look through the schematics and your issues again.

[kvans32]

So you provided the Miller Techs the board number and component number, and they wouldn't give you the nominal values for the resistors? Things may have changed in their tech support since I called last - I haven't worked on one of these in a long while.

Worse than that, they are giving me bogus numbers. Clearly, liability is not the reason they do not provide a schematic. Giving bogus values carries far more danger than proper values.

Are you checking across the power module directly, or going off measurements from the pre-power checks in the tech manual? It's odd only one diode from the bridge rectifier is bad. Of the four diodes in the rectifier (can't remember if these units took 3-phase, so am assuming single phase), a diagonal pair would need to be shorted to feed AC. Your buss capacitors would certainly be bulged as well.

I did both. Good point. It would not have actually been AC on the bus. I'm not sure what alternating voltage would do on -bus. The buss caps have an air gap under the top, which are ballooned, but I have seen where the aluminum can is bulged on the top and that is not this. The hard dimensions appear to be where they should be. It is a 3-phase possible unit.

Check the board picture above and look for the buck drive section of the schematic for those resistor values. I don't have one to check against anymore.

If that schematic has PC2 stuff on it, I do not see it. Unfortunately the schematic is not searchable. I wonder if alvideo could provide his original for editing / making a proper pdf. I do see refs to some of that stuff on the one you drew in post #24. I never actually inspected that until now because I thought it was all in alvideo's schematic. R4 I can not make out. R6 looks like 150mOhm, which would make sense if the guy on E14 measured it @ 180mOhm, as the leads might add some. I debated on ordering a 150mOhm resistor on my original parts order for that exact reason, and because 180mOhm is an odd value. 150 & 200 are common, especially from Ohmite, which is what the original is I believe. I, however, went on the safe side and ordered a 0.2Ohm. For all I know, R6 was the first thing to go.

Do you think I should reorder a 150mOhm one?

I have a 130Ohm for R4. If that does not match what you wrote, I may reorder both.

[alvideo]

I contacted Miller support, asking if they could provide me with values for R4-R8.

Here are R4-R9 values:

[kvans32]

Cool, thanks alvideo. I'll buy a 10ohm. That's what j's chicken scratch looks like indeed. I was biased toward 100's from the measurement that I got after pulling it.

When I look back at the pictures that validates. (Br Bk Bk Gd Br) - 100 x.1 1%.

Would you be willing to share your schematic so that I may add PC2 stuff to it?

[alvideo]

I don't have PC2 schematic, only PC10, posted No 81