Building Low Wattage ±330V Signal/Square Generator Help...

[Mechatrommer]

objective: to measure/verify my 1Mohm input differential probe output accuracy and CMRR at greater resolution, my ±10V FG is not enough.
HV-FG V output: Mains voltage ±330V (240Vac) capable of driving 1K, 10Kohm or 100Kohm load. whichever is possible, lower ohm (hi wattage) is better.
HV-FG F output: 50Hz - max Hz. max Hz will depends on simplicity of the project. if its up to me i'll put 100MHz there, but for the start i believe 300KHz up to 2MHz or 10MHz should be feasible?
distortion spec: close to none. for the start, if i can get a square pulses instead of pure sine, then thats workable or good enough i think. as always, pure sine is better, if the circuit's complexity is managable.

idea: i have two ideas in mind (attached picture).

idea 1: is using mosfet or bjt to switch the mains voltage. the control is from normal ±10V FG to optocoupler controlling the mosfet gate. the problem... my PC817 opto only rated at 35Vce, how do i make a circuit to accomodate the opto on the hot 330Vdc side?

idea 2: is using step up transformer. this requires more solid 4 quadrant PSU to drive the Tr primary rather than the normal 50ohm FG which luckily i have built on smaller scale (around 10W, 2MHz BW) sometime ago. the problem is, as i connected the 4Q-PSU to the primary, i got very nasty ringing on both side probed with DSO. do i need extra circuit to avoid this ringing? a snubber maybe? can i use the same snubber structure as usually found in SMPS flyback designs? any example more proper for my application?

hope you people can help thanks.

[flolic]

You should consider using dedicated low side opto mosfer driver instead 817...

Just search for app notes
http://www.vishay.com/docs/81227/81227.pdf

[robrenz]

You have the BOSS schematics I sent you. That should be a very good road map to build exactly what you are after. Here is the user manual of the 500V version of the Kepko BOP.

Edit:  For sine wave only it may be easier to use a step up transformer on the output of a lower voltage 4 quadrant supply.  A 17:1 transformer would give 340V and .58A from a 20V-10A BOP

[Marco]

Silly question ... what's wrong with a class B BJT push pull amplifier?

[blackdog]

Hi Mechatrommer

The higher you go with a transformer (the end of its frequentie range) the loser the coupling between  the primary and secondary windings.
The amp who is feeding the transformer cant control the secondary anymore (less damping)

Kind regarts,
Blackdog

[Mechatrommer]

what's wrong with a class B BJT push pull amplifier?

i'm not sure or am i reading it upside down... http://www.circuitstoday.com/push-pull-amplifier using 2 NPN? the one in my head is 1 NPN at the top and one PNP at the bottom, the problem is can i get 240Vac out driving from 10Vac base input? isnt it safe for my 10Vac FG is it? no isolation? any circuit example?

(less damping)

you mean the driving opamp/push pull bjt will oscillate as what i've experienced?

thanks Rob i didnt know that PSU family can produce 200Vac, schematics printed i'll study tomorrow, that one looks complicated.
thanks flolic i still cant open the link, my net speed is very bad now, will try tomorrow again.

[Marco]

Nevermind, I wasn't thinking right, transformer coupling just doesn't work very well over wide frequency ranges for square waves ... I did remember a post by Winfield Hill concerning high voltage amplifiers :

http://groups.google.com/group/sci.electronics.design/msg/9376a9d15aa6d062?hl=en

(need a fixed width font to view the schematic.)

Getting MHz+ bandwidth at these voltages won't be easy.