They are formed in the way I described — sudden discharge of a capacitor through primary winding (one turn) of a transformer, secondary has several turns, and is in series with the welding output. Hence name: coupling transformer, to couple the high voltage to the welding supply. The actual generation of the sudden pulse has similarities to the same in petrol engine spark ignition arrangements — step-up transformer (coil). Sudden change in current in primary produces high voltage in secondary, or, if you prefer, voltage induced across primary by sudden current change (think inductor) multiplied by turns ratio gives voltage pulse across secondary. Yes, I see the similarity with fluorescent lighting. High voltage ionises the gas, then mains voltage or much less maintains current. But in the present welder problem I think that the high voltage pulses are being terminated prematurely because the MCU thinks that the welding current has started, which it has not.
This thread has moved some way from repair obstacles created by mystery code in MCU. Just to go back to that for a moment, campaigners for manufacturers to be forced to release information and parts helpful to people attempting to repair their products should perhaps campaign also for release of circuit diagrams AND code (programme) in any microcontroller or microprocessor.
This thread has moved some way from repair obstacles created by mystery code in MCU. Just to go back to that for a moment, campaigners for manufacturers to be forced to release information and parts helpful to people attempting to repair their products should perhaps campaign also for release of circuit diagrams AND code (programme) in any microcontroller or microprocessor.
Well, all troubleshooting should proceed with a full check of the power supply, before you start looking for more complex faults. Sensors get dirty, hot areas get dry joints.
They never would release their codes to cpu and some other data ,because china or simply other manufacturers would copy paste that after few minutes , copyrights and design techniques are cost money to them and they would not share that .
So ok , if cpu is thinking that welding has started , need to determine which input is responsible for that .You may disconnect inverter control and play a little with that . Maybe try to paint schematic how cpu is connected , like pins and what they are doing , where sensor is connected , who controls inverter on off , which output starts hv pulses and so on .Try to measure voltages with dmm ,maybe you will find abnormalities ,like some voltage is present from current sensor ,even no current ,but not stable ,which would need to backtraced where it comes from .Sometimes this is very interesting to find that unbelievable failed part ,like broken resistor ! Or faulty opamp ... Few months ago had chance to repair car amplifier ,where all tl084 (if i remember correctly 4 pcs) in preamplifier have been failed ,most of their outputs gives dc voltage ,each output different voltage ,which was unstable .Supply voltage was zener regulated and ok . After changing all ic's amplifier started to working properly .
So ok , if cpu is thinking that welding has started , need to determine which input is responsible for that .You may disconnect inverter control and play a little with that . Maybe try to paint schematic how cpu is connected , like pins and what they are doing , where sensor is connected , who controls inverter on off , which output starts hv pulses and so on .Try to measure voltages with dmm ,maybe you will find abnormalities ,like some voltage is present from current sensor ,even no current ,but not stable ,which would need to backtraced where it comes from .Sometimes this is very interesting to find that unbelievable failed part ,like broken resistor ! Or faulty opamp ... Few months ago had chance to repair car amplifier ,where all tl084 (if i remember correctly 4 pcs) in preamplifier have been failed ,most of their outputs gives dc voltage ,each output different voltage ,which was unstable .Supply voltage was zener regulated and ok . After changing all ic's amplifier started to working properly .
Bad soldering at first have not stable fault ,when something is working for some time ,when cold , or when hot ,or when mechanically touched ,but in this case as i understand ,fault is stable .But i think op has already resoldered most of such things ,that may look suspect .
About cpu operation - at moment i think so , there are three possibilities ,why cpu terminates ignition pulse :
1.It receives something ,that should not ,like overcurrent signal or protection signal ,maybe noise from sensor input .
2. It didn't receives something, like hv feedback ,or DC feedback ok signal .
3. Its working as it should ,and pulse length is programmed to be such, and should be enough for ignition ,but something is weared ,or capacitor lost capacity ,or hv transformer winding insulation broken ,or leakage somewhere ,so that pulse don't make expected result .
Yes ,again thinking about cpus ,some can be easily programmed ,like arduino with board, for testing .But again ,need to know signals ,and what cpu should do with them ,even for temporary replacement .Need to study feedbacks then and controls of your welder .Code extraction can still be usable ,if extracted code will contain bootloader and eeprom ,and you can reprogram the same into new identical cpu (in case some input of cpu is broken,shorted protection diode in example , you can check that easily with dmm ) .If bootloader will be different ,probably locations in code need to be adjusted ,or it will not work at all i think .
1.It receives something ,that should not ,like overcurrent signal or protection signal ,maybe noise from sensor input .
2. It didn't receives something, like hv feedback ,or DC feedback ok signal .
3. Its working as it should ,and pulse length is programmed to be such, and should be enough for ignition ,but something is weared ,or capacitor lost capacity ,or hv transformer winding insulation broken ,or leakage somewhere ,so that pulse don't make expected result .
Yes ,again thinking about cpus ,some can be easily programmed ,like arduino with board, for testing .But again ,need to know signals ,and what cpu should do with them ,even for temporary replacement .Need to study feedbacks then and controls of your welder .Code extraction can still be usable ,if extracted code will contain bootloader and eeprom ,and you can reprogram the same into new identical cpu (in case some input of cpu is broken,shorted protection diode in example , you can check that easily with dmm ) .If bootloader will be different ,probably locations in code need to be adjusted ,or it will not work at all i think .