PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2022-01-24, 16:34:08
News: Forum TIP:
The SHOUT BOX deletes messages after 3 hours. It is NOT meant to have lengthy conversations in. Use the Chat feature instead.

Pages: 1 2 3 4 [5]
Author Topic: Development of an Avalanche Output Stage  (Read 61788 times)

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
they only squeal when pulsing the load

Is this from low impedance pulse into high impedance load?

EDIT:

replaced transistor (was a 3055 not 2055) and pulled a new wire from electrical panel to fix outlet.

Now, where was I?

EDIT:

pulsing again at 3.5kv ...
« Last Edit: 2010-07-06, 22:31:22 by Grumpy »
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
I'm ready to start testing now with 2 stages of Transistors @ 600v

First thing there's no way i am going to scope my coil because both sides are @ 600V and i dont fancy the earth lead of my scope sitting at this higher voltage(Scope is connected to usb of my laptop also), to start i will use my bifilar pancake coil, i can scope the 2nd bifilar coil which will give isolation and hopefully show me whats happening.
   

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
I use a resistive divider and other means of detection.
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
My probes are x100 so have the divider built in but because both sides of my coil sit at 600V i dont want to connect my probe earth to it.


Anyway seems i have trouble, the 2 transistors drain my psu and shut it down.
When i measure the base+emitter(which are shorted) to the collector of that transistor on a diode meter i get a reading of 0.3 which a normal diode would give me 0.66, on ohms i get about 730 Ohms, well there's no way i could poke 600volts on the collector and hope it stays up, so what did i miss.
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
Well i guess it's obvious that the transistors died, maybe 600volts was too much for only 2 transistors in the chain, anyway looks like i'm going to be joining G, i have ordered a pack of 50 2n5551 transistors as these smt417 transistors are too expensive to burn.
   

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
My probes are x100 so have the divider built in but because both sides of my coil sit at 600V i dont want to connect my probe earth to it.

Anyway seems i have trouble, the 2 transistors drain my psu and shut it down.
When i measure the base+emitter(which are shorted) to the collector of that transistor on a diode meter i get a reading of 0.3 which a normal diode would give me 0.66, on ohms i get about 730 Ohms, well there's no way i could poke 600volts on the collector and hope it stays up, so what did i miss.

They blow when you have to much current and those only need something like 100uA to avalanche.  600v doesn't matter as it can take thousands of volts if the current is kept low.

Well i guess it's obvious that the transistors died, maybe 600volts was too much for only 2 transistors in the chain, anyway looks like i'm going to be joining G, i have ordered a pack of 50 2N5551 transistors as these smt417 transistors are too expensive to burn.

2N5551's require 2mA to avalanche.
 
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
Something i dont understand with the FMMT417 Vcbo is 320V but the Vceo is only 100V, now if i have 300 volts between my base and collector that should be ok but the emitter is also connected to the base on each stage bar the first and therefore surely i am going over the Vceo by 200V  ::)

OUCH i also didn't realize you have to watch the base current that much.

http://www.diodes.com/datasheets/FMMT415.pdf
   

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
Something i dont understand with the FMMT417 Vcbo is 320V but the Vceo is only 100V, now if i have 300 volts between my base and collector that should be ok but the emitter is also connected to the base on each stage bar the first and therefore surely i am going over the Vceo by 200V  ::)

OUCH i also didn't realize you have to watch the base current that much.

http://www.diodes.com/datasheets/FMMT415.pdf

avalanche is determined by the Vcbo: see attached
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
OK G i was getting confused.
   

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
Also, notice that the avalanche voltage depends on several things.  (look at the charts)
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
I decided to start a fresh with the Avalanche Pulser after i determined that the efficiency of a Saturable pulse shaper was only a few % efficient.

Using the details in this doc posted by Grumpy
http://icecube.wisc.edu/~kitamura/NK/Flasher_Board/Useful/research/RSI00308.pdf

I built the transistor tester circuit Figure 1.
but used my digital mono and a hex inverter to drive the stage.

Here's a proto test of the stage, in the final pcb i will build the marx device with plug in 900v modules

The circuit should allow me to evaluate the performance of the bag of 2N5551 transistors i bought from china.

I am still testing things right now, here's 2 pulse shots,

One with a base capacitor of 10nF and one without
The one without was a lower voltage Peak but enabled me to drive it with a shorter pulse
The one with had a higher peak switched voltage.

Hopefully when i switch away from standard caps to a enamel wire wound cap i should get a more square pulse.
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
A couple of pictures of a 3 2N5551 transistor stack, 300pf cap,340K feed resistor. 47 Ohm load resistor, supply was 900V

Green Trace is across 47R load resistor
Yellow trace is on collector of the top transistor
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
I am almost in business

I tried my cap idea by winding enamelled wire axially along a length of wire to form a capacitor, although my cap meter measured 100pf i could not get it to work properly so have given up on the cap idea for now.
This means i need to buy some 200pf caps    :(

Anyway i have etched 7 Marx modules , 2 driver boards and 2 output boards, i already built sometime ago the 2 900V switched psu's

At the moment i have used 100pf caps and with a 47R resistor load the Marx boards do not work(The pulse doesnt get bigger with more modules GULP), as a temporary test i loaded with a 680K resistor and this verified they do work, once i get the caps i can try a coil.

A photo of the finished boards some still need making up

3 Scope shots of the pulse across a 680K load resistor (Yellow Trace) with 1,2 and 3 modules stacked as you can see each Marx module does indeed add to the final output pulse amplitude.
   

Group: Administrator
Hero Member
*****

Posts: 3173
It's not as complicated as it may seem...
Good stuff Peter!

You're becoming quite the "marxman" ;)

What's next in your "sights"?

.99
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
Quote
You're becoming quite the "marxman"

 ;D Thanks Darren

Well once i have 2 of these working i will try nS phase shifting the 2 pulses connected to a coil, but primarily i want to use the 2 channels for injecting pulses into a closed toroidal ring and see what sort of effects i can see.

Here's the circuit, incredibly easy to build and i am yet to find a failed transistor and have tested over 20 of the 2n5551 from the 50 i ordered from china.

   

Group: Tinkerer
Hero Member
*****

Posts: 3482
tExB=qr
I tested my 2N5551's for avalanche point and at Zener breakdown and found that any that were noisy at Zener had higher resistance at avalanche (as evidenced by lower pulse), so I put those aside.  May not really matter though.
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
Hi G
I have not seen much variation in Avalanche pulse size really so far, i did initially run some tests looking for the avalanche voltage and this did seem to vary quiet a bit.
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
I still do not know how a low impedance coil will affect the pulse loading yet, but with a resistor even a 1K starts loading up a 3 stage setup.

When i upgrade to 200pf caps hopefully i will be able to pump twice the impulse energy and increase the loaded peak amplitude.

Here's 2 pictures, 1 with a 1K and another with a 50K load using 3 modules stages.

Looks like i am getting a fall rate of about 18nS for the 1K load @ 740V pulse

and for 50K the pulse is about 1100V
   
Group: Guest
Grumpy,
The .PDF in posted Reply #14 has no "root" or it's invalid.  I get a blank black screen for an output of the .PDF

--Lee
   

Group: Administrator
Hero Member
*****

Posts: 3848


Buy me some coffee
Looks like it was one of those files that got damaged about a year ago, It will need to be re uploaded by Grumpy if he still has it.
   
Group: Guest
Okay, thanks.  He might have another file, or not?, that has a circuit which does something similar to the original?  In any event, I like to see it.

--Lee
   
Pages: 1 2 3 4 [5]
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2022-01-24, 16:34:08