Products > Other Equipment & Products
Gordak 952-v hot air solder station - high failure rate - what cause?
T4P:
Check the triacs
amyk:
"6mm" is the length of the jumper wire below the cap. Go ahead and annotate it, that would be good.
You don't need a full parallel programmer if all you need to do is unlock and erase chips. I've written about it before on the MCU board here. (Hint: You don't even need a PC if you've printed out the relevant pages from the datasheet! 8))
Here's the power board schematic. Took a little longer since there were more parts.... and I have no idea what a fen nor a han is. Perhaps they meant a fan and a hen? The basic operation looks straightforward, a bunch of isolated triac controls via active-low signals, although I'm not good enough at analog electronics to decipher what they're doing with the comparator on the HV side and the optocoupler driving the pin I've decided to call SENSE.
The PSU section is a half-wave for +5 and -5, and full-wave for the negative VFD drive voltage. 3VAC from the transformer is likely for the VFD filament. Besides the half-wave rectification I don't see anything else suspicious, but certainly checking the electrolytics is recommended. Half-wave produces a lot more ripple than full bridge, and non-branded caps aren't known for their longevity. Maybe they were already marginal when they left the factory.
--- Quote from: DaveXRT on July 26, 2012, 07:36:12 am ---Check the triacs
--- End quote ---
I doubt it. The display would still work, but without heat, if it was their fault.
asbokid:
Hi amyk!
--- Quote from: amyk on July 26, 2012, 08:33:37 am ---"6mm" is the length of the jumper wire below the cap. Go ahead and annotate it, that would be good.
--- End quote ---
Aha! I see. They are PM'ed to you. :)
--- Quote ---You don't need a full parallel programmer if all you need to do is unlock and erase chips. I've written about it before on the MCU board here. (Hint: You don't even need a PC if you've printed out the relevant pages from the datasheet! 8))
--- End quote ---
Hopefully the MCUs won't be locked. Otherwise we are a bit stuffed with no code to flash into them again :(
--- Quote ---Here's the power board schematic. Took a little longer since there were more parts.... and I have no idea what a fen nor a han is. Perhaps they meant a fan and a hen? The basic operation looks straightforward, a bunch of isolated triac controls via active-low signals, although I'm not good enough at analog electronics to decipher what they're doing with the comparator on the HV side and the optocoupler driving the pin I've decided to call SENSE.
The PSU section is a half-wave for +5 and -5, and full-wave for the negative VFD drive voltage. 3VAC from the transformer is likely for the VFD filament. Besides the half-wave rectification I don't see anything else suspicious, but certainly checking the electrolytics is recommended. Half-wave produces a lot more ripple than full bridge, and non-branded caps aren't known for their longevity. Maybe they were already marginal when they left the factory.
--- End quote ---
Amazing work! Though you lost me at the beginning of the first paragraph! So far as I can work out, FEN = hot air gun pump motor and HAN = soldering iron.
What would be a good repair plan? Maybe replace the MCU, and swap in some known good electrolytics, perhaps uprating them? On both the power board and the control board?
The MCU must be borked in some way, yet everything seems in order. There are some replacement MCUs on their way. Maybe it would be good to install a DIP socket for them first. Before they arrive, the VFD output pins and those "CPU DRIVER" lines can be scoped. Though not sure what that will show. With some Chinese good luck, the replacement MCUs might not be locked, so flash/EEPROM dumps could be extracted from them for the greater good ;)
The integrated EEPROM on the MCU must hold the 'dynamic user configuration' since the station settings are preserved across power cycles.
Those settings are for the pump motor, and for the two temperatures (gun and iron). If the EEPROM contents became corrupted during a write operation, it could in theory could cause badly-written firmware to crash. And with no EEPROM reset function, the f/w crash could be irreversible (except for re-flashing).
EEPROM corruption could be caused, perhaps, by pulling the plug on the station, rather than gracefully shutting it down via the power switch on the front panel. A graceful shut-down asserts the MCU's ON/OFF line (PD4 pin). Whereas pulling the wall plug does not. Only a theory, though ???
Thanks once again for all your effort in working out these schematics, Amyk, and to everyone else who has offered advice, too.
cheers, a
amyk:
I've updated the schematics posted above with the new information.
I wouldn't be too optimistic about the MCU being unlocked. As I mentioned before, if you erase it you can upload a simple program that just blinks the display so you can see if there are glitches/unreliability causing it to lock up.
But understanding the design, it doesn't seem too hard to write your own firmware. I think it just uses PWM on the triacs, synchronised to the zero crossings of the line frequency. You can find one open-source design here: http://dangerousprototypes.com/forum/viewtopic.php?f=56&t=2457&start=15
The signals on "CPU DRIVER" socket are active-low and used to switch on the heater, iron, and pump via the triacs. There's one SENSE signal which is an input to the MCU and probably used for zero-crossing detection. You can scope that one (on the isolated side, not the mains side) to confirm this.
Your hypothesis about EEPROM corruption does make sense, although the EEPROM is usually most vulnerable to corruption during a write; a hard shutdown doesn't trigger a write, unless it's some bad design that continuously writes to the EEPROM every second or something like that. EEPROM endurance would also be a problem if it was doing that. (A graceful shutdown should trigger an EEPROM write. I'm not sure of the power connections so I don't know whether it's the relay that turns on a main transformer and there's a standby one, or it's used for something else.)
asbokid:
Thank you once again Amyk for going to all this trouble. :)
Using your schematics, I was up until 5am studying the board(s). ::)
Just as you expected, the "SENSE" line delivers a 50Hz square wave to the MCU. But, surprisingly, the MCU is still running, at least initially. It is continuously generating an arbitrary PWM signal on one of the "DRIVER" lines. And on power-up, it initially sends 'sane' signals on the three wire serial interface to the VFD controller (strobe, DIN and clock).
This does look like a firmware problem, rather than faulty hardware.
I will post some scope traces once the cable/driver has come to light.
cheers, a
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version