Using mosfets to switch capacitors into series

Thread Starter

Musakck

Joined Jan 11, 2023
5
Me and my friends have been building a project for a couple of months now and we got stuck on on something we dont have much experience with. We need to connect 4 charged capacitors into series in order to have a higher voltage than the supply, we have been using relays and a arduino nano to do this but recently we have been experiencing a lot of issues like contacts getting stuck or the relay stops responding to the Arduino.We already selected a mosfet(IRF840) that can be switched by arduino and can handle the voltage and current of our circuit. Now the problem we got stuck on is how we need to connect the mosfets and capacitors to get the results we want. We have tried a bunch of ways using simulators and couldn't find a proper way
 

ericgibbs

Joined Jan 29, 2010
18,766
Hi Musak,
Welcome to AAC.
May I ask what voltages. Capacitor sizes etc, you are using, also how do you charge up the caps.?
E
 

BobTPH

Joined Jun 5, 2013
8,808
Why not tell us what power source you have and what voltage / current / duty cycle you need. Then we can suggest the best solution instead of fixing your guess at a solution.
 

Thread Starter

Musakck

Joined Jan 11, 2023
5
Each capacitor is 200v and 2200 microfarads. They are charged by a 12 vdc source which powers an inverter that gives 220vac output, this then goes to a full bridge rectifier to become around 200vdc which gets controlled by mosfets and an arduino nano. The charged capacitors then get connected in series to achive 800v. we tried putting the capacitors in series beforehand but couldn't charge them properly that way.
 

Ian0

Joined Aug 7, 2020
9,667
Each capacitor is 200v and 2200 microfarads. They are charged by a 12 vdc source which powers an inverter that gives 220vac output, this then goes to a full bridge rectifier to become around 200vdc which gets controlled by mosfets and an arduino nano. The charged capacitors then get connected in series to achive 800v. we tried putting the capacitors in series beforehand but couldn't charge them properly that way.
Exactly the job for a Cockcroft-Walton multiplier (But a capacitor charged from the output of a bridge rectifier with 230V input, would be 325V DC)
 

BobTPH

Joined Jun 5, 2013
8,808
Or, use a different transformer to boost to 800V directly, which would be more efficient.

But, I fear that the use of this is likely a banned topic which we should not be helping with.
 

KeithWalker

Joined Jul 10, 2017
3,063
The problem you will encounter if you try to use MOS FETs is that each one will have a different source voltage when connecting the capacitors in series. That voltage will vary depending on the state of the charges on the capacitors. This is important because the voltage used to switch each gate must be referenced to its appropriate source. That makes the switching circuit very complex.
The problems you experiencing with using relays are caused by high currents that flow between the the charged capacitors when the relay contacts are switched.
One simple solution is to use two sets of switching relays. The first connects an adjacent pair of capacitor terminals through a resistor to limit the initial currents. After a suitable delay, the second set of relay contacts close, shorting out the resistors. To disconnect the capacitors, the shorting relays are opened first and after a delay, the second set are opened.
 

LesJones

Joined Jan 8, 2017
4,174
Have you considered using Photovoltaic optocouplers to drive the gates of the mosfets ? This would drive the mosfet gates without requiring multiple isolated power supplies. I have never tried these but read about them and thought that they would be useful for this sort of application.

Les.
 

BobTPH

Joined Jun 5, 2013
8,808
Have you considered using Photovoltaic optocouplers to drive the gates of the mosfets ? This would drive the mosfet gates without requiring multiple isolated power supplies. I have never tried these but read about them and thought that they would be useful for this sort of application.

Les.
That does not solve the problem of hundreds of volts difference between the different switches.

A while ago I remember reading about an opto-coupler that drive a solar cell which would develop it’s own isolated voltage. That would solve the problem.

Edited to add: Look up photovoltaic optocoupler.
 

Thread Starter

Musakck

Joined Jan 11, 2023
5
wow! i was not expecting replies this fast, thanks for all the suggestions.

#6:We have looked into the Cockcroft-Walton multiplier but saw that the output current was too low which gave us a very long charge time
#7:We would but a transformer of that kind seems too dangerous to work with both for us and the circuit. Why would working with hv switching be a banned topic? isn't this what this is category about?
#9:This seems like a simple and effective solution, we will give it a try,thanks for the suggestion.
#10:No, the switches only need to connect the capacitors in series.
#11:Thanks for drawing a circuit and providing a graph. We will try this circuit and see if we can get similar results, thank you very much.
#12,#13:We have looked into them but didn't see them fit for this purpose, although we will look into the opto-coupler mr.bob mentioned.
#14:We did want to make one for some time but we do not have the resources or knowledge to make a railgun.
 

Ian0

Joined Aug 7, 2020
9,667
The problem you will encounter if you try to use MOS FETs is that each one will have a different source voltage when connecting the capacitors in series. That voltage will vary depending on the state of the charges on the capacitors. This is important because the voltage used to switch each gate must be referenced to its appropriate source. That makes the switching circuit very complex.
The problems you experiencing with using relays are caused by high currents that flow between the the charged capacitors when the relay contacts are switched.
One simple solution is to use two sets of switching relays. The first connects an adjacent pair of capacitor terminals through a resistor to limit the initial currents. After a suitable delay, the second set of relay contacts close, shorting out the resistors. To disconnect the capacitors, the shorting relays are opened first and after a delay, the second set are opened.
Doesn’t that do exactly the same as the diodes in the Cockcroft-Walton multiplier?
 

Ian0

Joined Aug 7, 2020
9,667
wow! i was not expecting replies this fast, thanks for all the suggestions.

#7:We would but a transformer of that kind seems too dangerous to work with both for us and the circuit.
What makes you think that 800V from a transformer is any more dangerous that 800V from any other circuit?
#11:Thanks for drawing a circuit and providing a graph. We will try this circuit and see if we can get similar results, thank you very much.
@Danko has drawn a Cockcroft-Walton muliplier.
 

Thread Starter

Musakck

Joined Jan 11, 2023
5
What makes you think that 800V from a transformer is any more dangerous that 800V from any other circuit?

@Danko has drawn a Cockcroft-Walton muliplier.
well even if we skip that a transformer that does exactly what we want would have to be wound by hand.

We have looked a bit more into the cockroft walton multiplier and got it working properly, Thank you so much for suggesting it in the first place.
 

LesJones

Joined Jan 8, 2017
4,174
What charging current do you require to the 4 capacitors in parallel and what will be the maximum discharge current from the 4 capacitors in series ?

Les.
 

Ian0

Joined Aug 7, 2020
9,667
well even if we skip that a transformer that does exactly what we want would have to be wound by hand.

We have looked a bit more into the cockroft walton multiplier and got it working properly, Thank you so much for suggesting it in the first place.
As an alternative to the 800V transformer, you could put three 230V transformers in series, or two 400V transformers, which are probably available off the shelf for phase-to-phase connection in 3-pbase systems.
 
Top