I need an HV fast sqaure wave generator. / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

I need an HV fast sqaure wave generator.

first previous 3 4 5 6 next

Move Thread

LAN_403

Patrick

Sun Nov 06 2011, 03:53AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Is the transsimision path of the cable what makes the pulse width vary? IE. the speed of propagation?

Pinky's Brain

Sun Nov 06 2011, 05:06AM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

Patrick wrote ...
is your earlier schematic still what you suggest?

On second thought as I mentioned above I don't think that will work because for a transformer with a 100 ns rise time you are looking at only a couple of turns, which isn't going to work well at 100 kHz. You need separate transformers for turn on and turn off :/

Like this :

The MOSFETs look fine.

teravolt wrote ...

hi Patrick I have made a cuple of changes to your 2n5551 shematic what do you think? the pulse width is detirmined by the lenth of cable.

Which at the original requirement means 2e8/100e3/2=1000 meter of cable, nevermind the amount of energy you burn each cycle in the 50 ohm termination.

Patrick

Sun Nov 06 2011, 05:31AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I want to try a single MOSFET square wave generator using a 555+MOT just to see what waveform i get. Beforei go making this too complicated.

teravolt

Sun Nov 06 2011, 06:00AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

yes that's part of it. the pulse width is the length of cable / (speed of light X velocity factor). the velocity factor is given by the manufacturer like rg58 is 70%, foam is less maby .8 ie 75ohm video cable. rg58 is around 1.2 ns per foot. your first edge (rise or fall) will be determind by the transistors and there inductance and resistance. the transistors have lots of ground plane around them and they are layed out in a zigzag patern to reduce inductance. the second edge is dependant on the 50ohm resistor. you could build the pcb as perscribed in the paper and modifie it to my version later. the advantage with cable or a PFN is that a square wave can be achieved

Herr Zapp

Sun Nov 06 2011, 06:30AM

Registered Member #480 Joined: Thu Jul 06 2006, 07:08PM
Location: North America
Posts: 644

Patrick -

For a really low inductance layout, you might consider using MMBT5551 devices, the SMT version of the 2N5551. The lead length of the TO-92 devices used on the board in the RSI writeup certainly doesn't look like it was chosen for low inductance. I agree with you about the quality of the article; it mentions that the 50pf coupling capacitors were integrated in the foil of the double-sided PCBA, but if you expand the image of the PCBA you can see that they are actually ceramic disk capacitors.

You did see the comment that they "burned-in" a large number of transistors for MONTHS, and then selected those that had the most closely-matched switching characteristics? Sounds like a real PITA.

Herr Zapp

Patrick

Sun Nov 06 2011, 06:57AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Herr Zapp wrote ...

....but if you expand the image of the PCBA you can see that they are actually ceramic disk capacitors.

You did see the comment that they "burned-in" a large number of transistors for MONTHS, and then selected those that had the most closely-matched switching characteristics? Sounds like a real PITA.

Herr Zapp

yeah the cap descrepency i noticed, and i was going to skip the "burn in " they describe, im gonna start with a simple MOSFET + 555 timer + 900 Volt source.

Mattski

Sun Nov 06 2011, 07:44AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

Herr Zapp wrote ...

For a really low inductance layout, you might consider using MMBT5551 devices, the SMT version of the 2N5551. The lead length of the TO-92 devices used on the board in the RSI writeup certainly doesn't look like it was chosen for low inductance.

They mention in the paper that the series inductance of the transistors is used in conjunction with the pad capacitance to form a synthetic 50ohm transmission line in the signal path (Z=sqrt(L/C)). So as long as you keep the lead length consistent the inductance is not necessarily a problem.

Pinky's Brain

Sun Nov 06 2011, 11:40AM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

teravolt wrote ...
the advantage with cable or a PFN is that a square wave can be achieved

At 2 kV? That's 40 amps of current during the top, well beyond the 2n5551's capabilities, with an average power being burned in the termination of 2e3^2/50/2=40 kWatt! On top of the fact that you need 1000 meters of cable for 100 kHz the whole concept is just plain weird.

The only realistic way to use a Marx for a 100 kHz square wave is to use the dual pulser setup, dumping energy into a capacitive load and using regulation (avalanche diodes) to convert the capacitor voltage into your test signal, and then taking the energy out of the capacitor with a pulser of opposed polarity.

Sure it's hopelessly complex, but conceptual simplicity is completely useless if what you end up with in practice is 1000 meters of simple cable and 40 kWatt of simple power.

Patrick

Sun Nov 06 2011, 12:42PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I have the HV probe nearly done, so we'll see soon what the callibration situation really is.

Sulaiman

Sun Nov 06 2011, 01:01PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

It would be much easier to make a 'low' voltage squarewave calibrator (74ACxx logic for example) and build a high-bandwidth pre-amplifier for the hv 'scope probe, which would also be a useful bit of kit.

first previous 3 4 5 6 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob