Here's a very strange problem I've come across. I need to shift a high-frequency (50 kHz - 10 MHz) 5V amplitude square wave signal down to -100V, in order to drive the gave of a MOSFET. I did this with a high-pass filter:

This works fine at high frequencies, but for some reason not with low frequencies. In this case, 200 kHz.

I measured the voltage at IN+, and as you might expect, it is a clean 200 kHz square wave.

HOWEVER, when I measure the voltage at V3.pin1 (between C5 and R8), I get this horrible thing (200 kHz):

This doesn't make any sense to me. It's a high pass filter with a cut-off frequency of 1.5 kHz, why is it doing that?

Does it have something to do with the gate-capacity? Or Miller capacity?

Here's the entire circuit:
http://imgwiz.com/images/2013/05/06/BJgDu.png

The signals IN+ and IN- are as follows (5V digital):

Thanks for any help you can give me. If you need more information, I'll be happy to supply it.

 

Can't see any reason for the strange signal between C5 and R8.

Why are you running the MOSFET source to -100V?

 

I can't explain it either.
If you are expecting to get ±100v out, prepare to be disappointed. Since V2 is operating as a source follower, the high output level at TP1 will only be a volt or two.

 

Why are you running the MOSFET source to -100V?

So D3 isn't supplied a negative voltage when the output is -100V.

If you are expecting to get ±100v out, prepare to be disappointed. Since V2 is operating as a source follower, the high output level at TP1 will only be a volt or two.

The circuit works, and is outputting ±100V at 10 MHz with a rise/fall time of <5ns. The reason it works is because V2 is bootstrapped.

@ all

From my observations, I'm assuming the culprit is the Miller capacity of V3. When the Gate is switched on, the voltage on the drain terminal is ripped from +100V down to -100V. I'm assuming this has a capacitive influence on the Gate voltage, ripping it down with it.

TheComet

 

So D3 isn't supplied a negative voltage when the output is -100V.



The circuit works, and is outputting ±100V at 10 MHz with a rise/fall time of <5ns. The reason it works is because V2 is bootstrapped.

I missed the bootstrapping.