Safetly driving mosfet at higher voltages ?

[boguslaw]

Hi

Is there any electronic expert here ? I have composed schematic with optocoupler and mosfet driver and I very appreciate if somebody corrected it and posted if that makes any sense ?
I try to completly isolate oscillator section from mosfet because I want to dump large capacitor 8-30uF via 2-4 mosfets in parallel at 440V to low inductance primary and that would be quite dangerous
Attached is simplified schematic when oscillator is just replaced by a switch.
Please help.

[boguslaw]

Especially I don't know if 500V GND should/could be connected with mosfet driver ? How to pass signals sent to optocoupler to the mosfet driver side ? I just connected phototransistor inside optocoupler via resistor to +Vdd (12V) Is that correct way ?

[citfta]

Hi boguslaw,

I checked the data sheet for your mosfet driver and it looks like you have everything drawn up properly. I would make R1 and R2 about 1K to start with to limit the current and see how everything switches. If you do not get full on and off like you want you can try lowering R1 a little and then R2 if you need to. I would not go lower than about 100 ohms on each of them because you could burn out the opto coupler or the mosfet driver.

Good luck,

Carroll

[boguslaw]

Hi boguslaw,

I checked the data sheet for your mosfet driver and it looks like you have everything drawn up properly. I would make R1 and R2 about 1K to start with to limit the current and see how everything switches. If you do not get full on and off like you want you can try lowering R1 a little and then R2 if you need to. I would not go lower than about 100 ohms on each of them because you could burn out the opto coupler or the mosfet driver.

Good luck,

Carroll

Oh! Thank You Carroll ! You have given me hope that I could cope with this enterprise

[boguslaw]

Question arise : how can I check if mosfet is fully switching on/off in this schematic ? I have no experience in checking this, previously I have just built circuits from schematics but now I have old analog scope so maybe I can use it...

[Leviathan]

Yes looks very nice, what will your inductive load be? Also if your into
high voltage MOSFETS look up QFETs they are a special type of mosfet
designed to handle uber high voltage in the 1k volt range some also have lower
source drain resistance than normal FETs, unfortunately that's about all I know
about them might be worth a shot to test them if you have the free time though

[boguslaw]

Thank you. I will use IRFP450, maybe later IRFP460. I need to minimize costs unfortunately

Btw I need also oscillator schematic with interrupting.
Like the one I have built in the past with two NE555 chips when one is for example 30-50Kh carrier and second is switching /resseting the first one output with lower frequency (kind of circuit used in old times in Lawton HHO device). However while carrier wave can have variable duty cycle I need the interrupting signal perfectly 50% duty cycle so frequency would be stable.
I need it to drive flyback on ferrite core so carrier would be synchronized to resonant frequency of core ,while interrupting signal would turn flyback output on/off.
I only need some schematic or link which chip I can use to make it simple (KISS method ).

[Ufopolitics]

Hello Boguslaw

I see You are trying to reproduce my type of oscillation system with High Pulsed Spikes via MOSFET's...in order to get high Radiant Surges...Right?

I will try to help you, however I have not done this type of oscillation system before...

Every MOSFET's must have a Bias to opposite polarity from its Gate, whether an N or a P Channel, this is done through a resistor, like in my original oscillator it had a 47K Resistor from gate to ground on each FET, they were N-Channels, meaning Gate is Positive, therefore triggers (Close Drain-Source) at Low Oscillator signal, the Bias resistor will "call back" to reset position (Open Drain-Source).
I believe this Biasing of Gate is your way to keep the Low circuit from the HV at Drain-Source. You must have two ISOLATED supplies, one low V for oscillator and opto-coupler and the HV (Vcc-Vss). Therefore, the bias resistor to gate should feed from the Low Voltage ground. I believe that will keep Gate isolated from D-S circuit.

I have somewhere a schematic of this type of circuit...I will look for it and post it here...

Regards

Ufopolitics

[Ufopolitics]

Circuit...

Here is a circuit a friend send me...
The only thing here is that he is not driving MOSFET's...So, He is using the 48V supply to close the transistor side of opto-coupler...

I see it different though...I will use the same low voltage from oscillator (15V) to trigger Gate from Opto-coupler transistor side...must of FET'S gates max voltages are 20 to 30 Volts...and like I said before, using the bias resistor to low voltage ground side.

I am interested in also building this type of oscillation, but have not got the time to build it...

[citfta]

Hi again,

If you can get an SG3524 chip in your country you can use that to make a circuit that will give you an almost perfect 50/50 duty cycle. You could then use the output of that chip to control the 555 chip using pin 4 of the 555. This should let you have full control of the frequency of your pulses and control of the on and off times too.

You asked about your circuit if you needed to keep the ground from the 500 volt side connected to the ground of the mosfet driver chip. Yes you do need to keep them connected together so the mosfet driver chip can have good control to the gate of the mosfet.

If you have a probe for your scope that has a 10x position you can probe from the source to the drain of mosfet to see if it is turning all the way on and off. Put the ground of the probe on the source and the hot side of the probe on the drain. If you don't have a 10x probe you can check at the gate and see if you are getting almost the full 12 volts at the gate when the driver chip is turned on. If you are getting the 10 or so volts at the gate when turned on then the mosfet should also be turning fully on. Again just put the ground side of your probe on the ground of the circuit and the hot probe on the gate. Hope this helps some.

Carroll

[citfta]

Hi again,

In the "Use for the Tesla Switch" thread on page 19 post 551 there is schematic showing how we were using the SG3524 chip. For 50/50 output you would leave off the dotted lines labeled pulse width control.

Carroll

[Shanjaq]

Opto-couplers are orders of magnitude slower than BJT/FET/IGBT

For example:

Parameter    H11D1    IRF610
td(on)         5uS        8.2nS
tr               2.5uS      17nS
td(off)         6uS        14nS
tf               5.5uS      8.9nS

E= V^2*C/2
30uF @ 550v is 4.54 Joules, you could probably scale that up a bit


However, if your inductor is lacking impedance, your switch might be seeing 550 Amps at t=0 and ~10 Amps at t=120uS... Yeah.. most silicon switches affordable to hobbyists aren't well suited to capacitor discharge.. mmmmmm... Thyratrons .... :drool:

-Edit-
The following circuit could be modified to drive a low-side insulated-gate device, safely and very fast!
High-Side Gate Drivers

[boguslaw]

Opto-couplers are orders of magnitude slower than BJT/FET/IGBT

For example:

Parameter    H11D1    IRF610
td(on)         5uS        8.2nS
tr               2.5uS      17nS
td(off)         6uS        14nS
tf               5.5uS      8.9nS

E= V^2*C/2
30uF @ 550v is 4.54 Joules, you could probably scale that up a bit


However, if your inductor is lacking impedance, your switch might be seeing 550 Amps at t=0 and ~10 Amps at t=120uS... Yeah.. most silicon switches affordable to hobbyists aren't well suited to capacitor discharge.. mmmmmm... Thyratrons .... :drool:

-Edit-
The following circuit could be modified to drive a low-side insulated-gate device, safely and very fast!
High-Side Gate Drivers

Thank you.Yes this is what I'm afraid of. My coil inductance is low ,don't know how to avoid the initial current surge especially if switching frequency will be low ,probably in Hz range because I have also 11H choke coil which will recharge capacitor every cycle. I have no big hopes but I would be happy if I could make it working ....

[Shanjaq]

What about a triggered spark gap, or a gas-discharge tube?

Xenon Flasher

There are also big SCR's, but they have to go below a threshold current before shutting off, which would be a problem if charging were continuous.

IGBT "brick" modules are OK, a bit slow though.

That "high-side gate driver" page has been enormously useful, especially the final iteration shown.. try it out in the falstad simulator

Try it out with the biggest switches you've got (voltages must be within spec, power dissipation from amperage is probably cumulative, monitor for heating with an infrared thermometer and see how it goes!)