Hi all!
And please, forgive me for my English, it's not my native language. (and for my electronics skills too..)

I need to provide short (microseconds and less) pulses of high current (>50A) to a complex load, which resistance can change greatly from pulse to pulse. For example, first time it was 1Ohm, next pulse it is 0.1Ohm and so on. Traditional constant or adjustable voltage generator work bad on such load, any adjustments are "late" for at least one pulse.

I want to try another approach and use a modified buck-boost converter in discontinuous current mode with a deleted capacitor.

I'll rise some current in Coil L(50A desired) and then switch the MOSFET. I expect Coil to try to maintain it's current, rising or lowering voltage on it's end according to the load resistance. Energy of Coil will deplete, so the current in load will look like a high short peak up to 50A and then quick fall.
That's exactly what I need.

For the proof-of-concept scheme I plan to add current-limiting resistance to the coil part of circuit, so the current won't rise to short circuit state.
Sourcing coil circuit from a powerful 5V power supply, a ~0.1Ohm resistor will be enough. If concept works, will try to limit current using a comparator to detect MOSFET switching time.

Please, tell me, will this idea work?
And is MOSFET in this scheme in high side or low side configuration? Do I need to use high side driver with a bootstrap function?

Great thanks for any help!

 

You won't get a square pulse from this circuit. You will get a fast rise and then an exponential decay. Further the width of the pulse will depend on the resistance it is feeding following the time constant L/R.
http://hades.mech.northwestern.edu/index.php/RC_and_RL_Exponential_Responses

 

You will get a fast rise and then an exponential decay

That will be great, even better than square pulses.

 

Does the time varying in inverse proportion with the load resistance matter?

 

The desired value is AmpereSeconds (amount of current passed through the load, if I can say so), so I think no, pulse time doesn't matter, I'll just continue pulsing till I get desired A*s total.
And if pulses will be too short, maybe I can switch to bigger coil, storing more energy?

 

And if pulses will be too short, maybe I can switch to bigger coil, storing more energy?

Yes, that would work.

 

Yes, that would work.

Thank you for help!

Still can not understand high\low side of MOSFET in this circuit, now matter how long I read articles about it..
Theoretically, even if I use a simple low-side driver, such ad TC4420 with an external 15V source for switching the 5V load, +10V difference will be enough to drive MOSFET normally.
But, maybe, I'm seriously mistaken in this part.

 

Yes, if you use an N-channel MOSFET then the gate drive will need to be higher than your 5V supply by at least as much as the MOSFET needs to be fully on. If you use a P-channel MOSFET then a simple pull-down to 0V will do but the MOSFET must be chosen to be fully on with a 5V Vgs.

 

Thank you!
I will try now to draw a full circuit, will upload it here when ready.

 

Further to that, with N-channel MOSFET, during the inductive discharge, the source will go to a negative voltage (50A times load resistance) and the gate drive circuit must ensure that the gate source voltage remains zero during that time.

 

gate drive circuit must ensure that the gate source voltage remains zero during that time.

Thanks. Should it be a specific driver chip, or a specific topology? Could you please recommend something?
I have plenty of hi-current (or) high-voltage MOSFETS like IRF3703, so if possible, I prefer to use them.