P
Piotrne
- Jan 1, 1970
- 0
Hi,
I have a question about controlling a large current in an experimental
welder. The welder should generate a short pulse of current, about
50 ms long. Additionally, it has to be regulated with a simple PWM.
Now, I have following circuit: ATMega32 microcontroller which generates
the PWM signal (2 kHz square wave, turned on for 50 ms);
then an optocoupler; then a MOSFET driver IRS2001 driving
(through a 47 Ohm resistor 47 Ohm) the gate of IRFP4468PbF.
The transistor acts as a key turning on an off the welding current
(source to ground, drain to the burner, the burner to the positive pole
of a rectifier connected to a large transformer: 300A, 50V).
Additionally, there it is also a ionizer with a filter system (designed
by someone else). The microcontroller is powered from its own power
supply, the rest of the control system (driver) is galvanically
separated and also supplied with its own power supply.
The are (at least) two problems with this circuit:
(1) the microcontroller is resetting from time to time,
(2) the transistor is damaged after a single pulse...
The problem (1) is probably caused by the ionizer. I suppose, it can be
solved by closing the circuit in a metal housing.
The problem (2) is more serious. When I turn on the transistor for
the assumed period of 50 milliseconds, it does not turn off.
I had to quickly disconnect the transformer (as the trinsistor does not
stand 300A for a long time). However, the transistor has been damaged...
I suspect that the reason was a high voltage induced at the time of
disconnecting the high current (such moments occur 2000 times/second in PWM).
The sudden power interruption can induce high voltage, which probably
caused the destruction of the transistor.
Does anyone have experience how to avoid such interference? Varistors,
filters, etc.? Experiments are a bit expensive ... The high current part
(transformer, rectifier, ionizer) is ready, I would "only" like
to switch on/off a pulse of current, with adjustable parameters
(hence the microcontroller instead of something simpler).
Regards
P.
I have a question about controlling a large current in an experimental
welder. The welder should generate a short pulse of current, about
50 ms long. Additionally, it has to be regulated with a simple PWM.
Now, I have following circuit: ATMega32 microcontroller which generates
the PWM signal (2 kHz square wave, turned on for 50 ms);
then an optocoupler; then a MOSFET driver IRS2001 driving
(through a 47 Ohm resistor 47 Ohm) the gate of IRFP4468PbF.
The transistor acts as a key turning on an off the welding current
(source to ground, drain to the burner, the burner to the positive pole
of a rectifier connected to a large transformer: 300A, 50V).
Additionally, there it is also a ionizer with a filter system (designed
by someone else). The microcontroller is powered from its own power
supply, the rest of the control system (driver) is galvanically
separated and also supplied with its own power supply.
The are (at least) two problems with this circuit:
(1) the microcontroller is resetting from time to time,
(2) the transistor is damaged after a single pulse...
The problem (1) is probably caused by the ionizer. I suppose, it can be
solved by closing the circuit in a metal housing.
The problem (2) is more serious. When I turn on the transistor for
the assumed period of 50 milliseconds, it does not turn off.
I had to quickly disconnect the transformer (as the trinsistor does not
stand 300A for a long time). However, the transistor has been damaged...
I suspect that the reason was a high voltage induced at the time of
disconnecting the high current (such moments occur 2000 times/second in PWM).
The sudden power interruption can induce high voltage, which probably
caused the destruction of the transistor.
Does anyone have experience how to avoid such interference? Varistors,
filters, etc.? Experiments are a bit expensive ... The high current part
(transformer, rectifier, ionizer) is ready, I would "only" like
to switch on/off a pulse of current, with adjustable parameters
(hence the microcontroller instead of something simpler).
Regards
P.