High voltage mosfet driving

[James Rollins]

I'm a bit confused about how to drive a mosfet as a switch when
dealing with high V_DS voltages. I have several nch mosfets with 100V+
up to about 1000V for max source to drain voltages.

I want to create a few buck converter circuits and others to test the
quality of the output but I cannot for the life of me figure out how
to drive the gates without requiring other high voltage components and
without the circuit becoming overly complicated.

If you take a simple voltage diving circuit with the nch mosfet
replacing the high-side resistor and try to drive the gate then the
source terminal is "floating" in the sense that it depends on the
load. If I tie the gate to the source through a resistor then this
solves the problem by turning off the mosfet. Unfortunately it then
causes major problems when trying to drive it. Not only that the gate
then has to discharge through that resistor plus the low-side
resistor, which here is the load, which reduces the fall time
significantly for low loads.

One can't simply use another nch HV switch to connect the first
mosfet's gate to ground because it brings the source of the first
mosfet way below spec.

The issue is driving the gate so that it is only within +-20V of the
source which can be up to several hundred voltages above ground almost
up to Vcc.

Is there something I'm missing here? I imagine using a p-ch mosfet
would be much easier yet I can't find ones with spec for my
application yet N-ch's are abundant.

 

[MooseFET]

I'm a bit confused about how to drive a mosfet as a switch when
dealing with high V_DS voltages. I have several nch mosfets with 100V+
up to about 1000V for max source to drain voltages.

I want to create a few buck converter circuits and others to test the
quality of the output but I cannot for the life of me figure out how
to drive the gates without requiring other high voltage components and
without the circuit becoming overly complicated.

If you take a simple voltage diving circuit with the nch mosfet
replacing the high-side resistor and try to drive the gate then the
source terminal is "floating" in the sense that it depends on the
load. If I tie the gate to the source through a resistor then this
solves the problem by  turning off the mosfet. Unfortunately it then
causes major problems when trying to drive it. Not only that the gate
then has to discharge through that resistor plus the low-side
resistor, which here is the load, which reduces the fall time
significantly for low loads.

One can't simply use another nch HV switch to connect the first
mosfet's gate to ground because it brings the source of the first
mosfet way below spec.

The issue is driving the gate so that it is only within +-20V of the
source which can be up to several hundred voltages above ground almost
up to Vcc.

Is there something I'm missing here? I imagine using a p-ch mosfet
would be much easier yet I can't find ones with spec for my
application yet N-ch's are abundant.

You need the actual driver to fly up and down with the source of the
MOSFET. It needs its own little power supply.

You could also use a gate driving transformer.

Smaller
MOSFET
--- ------- -------- Gate of big MOSFET
)!!( s! ! !d
)!!( -----
)!!( !
--- ---------+---------- Source of big MOSFET

 

[James Rollins]

You need the actual driver to fly up and down with the source of the

MOSFET.  It needs its own little power supply.

Yes but I'm not sure how to do this without running into similar
problems and making it too complex. I believe the problem of the the
gate attached to the load is still an issue. The load is 100kohms
which will effect the fall times.

You could also use a gate driving transformer.

             Smaller
             MOSFET
---    -------     -------- Gate of big MOSFET
   )!!(      s! ! !d
   )!!(       -----
   )!!(         !
---    ---------+---------- Source of big MOSFET

Are there transformers created just for this sort of thing? I suppose
it would work but I'm a little confused how I would drive the smaller
mosfet. I assume the transformer simply floats from the source and
one applies the primary side voltage to the gate. So if 10V input and
it's a 1:1 transformer then it would lift the gate 10V? The small
mosfet is to remove the transformer from the gate? I still have the
problem then of bringing the gate low quickly. I'm thinking that some
voltage follower might work where it would follow the source voltage
and sink the gate charge but this introduces similar problems with the
powering of the opamp and keeping it within spec.

Is this correct? Or do I not need to worry about the gate discharge
when floating since it is at a higher voltage and hence will be
quicker and might act as a virtual ground?

 

[Jamie]

I'm a bit confused about how to drive a mosfet as a switch when
dealing with high V_DS voltages. I have several nch mosfets with 100V+
up to about 1000V for max source to drain voltages.

I want to create a few buck converter circuits and others to test the
quality of the output but I cannot for the life of me figure out how
to drive the gates without requiring other high voltage components and
without the circuit becoming overly complicated.

If you take a simple voltage diving circuit with the nch mosfet
replacing the high-side resistor and try to drive the gate then the
source terminal is "floating" in the sense that it depends on the
load. If I tie the gate to the source through a resistor then this
solves the problem by turning off the mosfet. Unfortunately it then
causes major problems when trying to drive it. Not only that the gate
then has to discharge through that resistor plus the low-side
resistor, which here is the load, which reduces the fall time
significantly for low loads.

One can't simply use another nch HV switch to connect the first
mosfet's gate to ground because it brings the source of the first
mosfet way below spec.

The issue is driving the gate so that it is only within +-20V of the
source which can be up to several hundred voltages above ground almost
up to Vcc.

Is there something I'm missing here? I imagine using a p-ch mosfet
would be much easier yet I can't find ones with spec for my
application yet N-ch's are abundant.

You have some options where, use a Gate driver with a self generated
inverter powered from the high potential or, do what I did once and
that was to use a 555 timer into a toroid xformer where on the other
side I used a high speed switching full way rectifier config (2 diodes
CT), no caps were needed due to the gate capacitance but you may want
to have that your self. I had fets that already had high value cap on
the gate. A shunt R is needed of course.

By connecting the (-) output to the source and (+) to the gate, I was
able to increase the gate voltage above the source voltage.

something to think about I guess.

http://webpages.charter.net/jamie_5"

 

[MooseFET]

Yes but I'm not sure how to do this without running into similar
problems and making it too complex. I believe the problem of the the
gate attached to the load is still an issue. The load is 100kohms
which will effect the fall times.



Are there transformers created just for this sort of thing? I suppose
it would work but I'm a little confused how I would drive the smaller
mosfet.  

Notice how I hooked it directly to the transformer. Try spicing it up
with a squarewave driving the transformer primary through a resistor.
You get blips in the up and down direction on the transformer instead
of a squarewave. The small MOSFET turns this back into a squarewave.

I assume the transformer simply floats from the source and
one applies the primary side voltage to the gate. So if 10V input and
it's a 1:1 transformer then it would lift the gate 10V?  The small
mosfet is to remove the transformer from the gate? I still have the
problem then of bringing the gate low quickly.

When you yank the transformer the other way, the small MOSFET is
turned on and snaps the ate down.

Try spice and see what happens.

 

[Tim Williams]

When you yank the transformer the other way, the small MOSFET is
turned on and snaps the ate down.

I once tried that (or rather, a somewhat more complicated approach,
but same idea).
http://webpages.charter.net/dawill/tmoranwms/Circuits_2008/Pulse_Coupled_Gate_Drive.gif

Had erratic behavior though. Nasty business when you're humming along
at 20kHz like a good half bridge, then suddenly a transient (I never
decided what it was, but I'm guessing related to dV/dt coupled in the
input transformer there) ticks it "ON" and gulps three hundred amperes
short circuit through your bridge. Oops. So the desat kicks it off,
but it takes a cycle before either gate drive's state is determinate.

That's what I don't like about these sorts of "I trust you'll stay
put" methods. To mix quotations: trust but verify, while carrying a
big stick. ;-)

Tim