I would like to switch a mosfet at 10khz (not that crazy of a thing to do from my understanding) then have it recreate a second square wave over top if that makes sense.
so lets say the square second wave is 4hz (probably would be faster but for simplicity's sake) so we would have the initial square wave switch from 0-20 at 10kHz for 250mS then switch from 0-10 at 10kHz for 250mS and so on
Hopefully I did an okay job explaining this. Is this feasible to do with one mosfet or would I need 2, one switching at the 4hz frequency then one switching at the 10khz frequency?
Is this where Dutycycle begins to play a roll to get a simulated voltage like this?
better yet would be if i could accomplish the same task except setting current instead of voltage. So fluctuate between 20amp and 10 amp in a 4hz 50% duty cycle square wave over top the 10 khz wave.
Im semi new to mosfets and and arduino all together. any help is greatly appreciated!
Such a signal can be created from two digital outputs and a resistor network (2 bit DAC). Question is the nature of the load, that determines the amplifier circuit between the signal source and load.
The idea is for a welder power source. So i plan to use an igbt module(which seems to behave like a mosfet) I can get away with a constant voltage, but being able to do constant current would be even better. Both would be best(select between one or the other) but current can be up to 300amps, voltage is probably 15-120v
DrDiettrich:
Control of so high currents is best done one the primary side of the transformer, by inserting variable current limiting resistors.
That was the plan, and why it was switching at 10 kHz (for the transformer) which could still be seeing as much as 50 amps, variable current limiting resistors? I'll have to look into that.
DrDiettrich:
I don't see a way to make a welder work with PWM.
Why is that? This is the standard for modern Inverter welders. They use an IGBT Module before the transformer switching at very high speeds (such as 10KHZ) this is why they are so small. Obviously there are many more electronics involved, but i am more looking for a proof of concept right now. If I'm not mistaken after delving more into the topic i should be able to limit the current with pulse width? and the voltage with Duty Cycle?
Right now I will Have a High Current rectifier, some Large Capacitors and probably 2 inductors to smooth the ac out, to a nice pure DC Wave, then using the arduino, a driver and probably a few other things i will output a varying waveform across the transformer, as long as it is switching fast enough there should be almost zero noticeable difference to the welder. Then if i can create a simulated slower wave form I can have pulse tig, mig, and even stick (not popular in the US yet)
No, they're not the same thing. But they're close enough that you could do most of your design with one type and then switch to the other when you find out you can't get a MOSFET to do what an IGBT does.
antleo:
i should be able to limit the current with pulse width? and the voltage with Duty Cycle?
You should limit the current with current-limiting feedback. Duty cycle, pulse width and frequency are linked together. You can't change one without changing one (or both) of the others. Most of the time we use constant frequency and we change duty cycle. Pulse width does change but we don't care what that width is. We just look at the % duty cycle to work out how much power is going to the output.
We have had people trying to do welding control projects here in the past and the stumbling block has always been trying to cope with the interference switching such large currents generates. I have not seen anyone report a successful project.
MorganS:
No, they're not the same thing. But they're close enough that you could do most of your design with one type and then switch to the other when you find out you can't get a MOSFET to do what an IGBT does.
Correct. I planned to do a small scale version on a test bench through a little transformer and mosfet to verify functionality.
MorganS:
You should limit the current with current-limiting feedback. Duty cycle, pulse width and frequency are linked together. You can't change one without changing one (or both) of the others. Most of the time we use constant frequency and we change duty cycle. Pulse width does change but we don't care what that width is. We just look at the % duty cycle to work out how much power is going to the output.
any suggestions on a specific one for my purpose? would a shunt fed through an amp and to the arduino be a reliable method since it will be responsible for setting the on state of the IGBT module?
Would I be better off rectifying after the transformer and doing just low frequency switching? This seems to be a somewhat easier route and would allow manual control of the current through the original transformer. Ill be getting an igbt module, rectifier, and capacitors capable of either, but would need to create a different inductor(if If not mistaken the more important of the two for steady current)
Grumpy_Mike:
I suspect that they don’t use a Micro controller.
What would you guess they use then out of curiosity?
If you take a commercial unit apart ( I have, out of curiosity) you'll find it's chock full of interference suppressing devices and very clever magnetics as well as a lot of juicy semiconductors.
Tricky design - not one for even a professional engineer unless experienced in that industry.
I'm not, and wouldn't attempt it. There'd be a lot of magic smoke.
Like high power RF ( which I have done),as in the old movies, 'one false move and you're dead!'
