Development of an Avalanche Output Stage

[Peterae]

Grumpy is already using an Avalanche output stage and so i would like to look at building one also.

Reading from Spherics instructions on pulsing a bifilar coil the specs need to be as follows.

1kHz-3.5 MHz
Pulse minimum 150 Volts.

Must be able to drive a very low impedance coil.

I will use FMMT427 which is an Avalanche transistor capable 320V pulses and can run at over the required 3.5MHz, peak current is 60A
http://www.diodes.com/datasheets/FMMT415.pdf

Only 1 device is needed and with the voltage spec of this device i will chose to run it at between 200-250 volts.

The board will have an on board psu to supply the HV and allow TTL input to drive it.

[Grumpy]

preferably over 1kv - this agrees with that Italian device as well (Energia Celeste ?)

isolate the trigger or learn the hard way - LOL!

[Peterae]

Pigeon steps G 300V first as it's easier to work with.

Here's the frist draft, need to round some of the component values, i decided to go for 300V first try.

[Peterae]

isolate the trigger or learn the hard way - LOL!

Good advice thx G is this what blew your sig Gen

[Grumpy]

No, I always isolated the sig gens.  I think it was either close proximity to pulsed coils or just age that took out my gens.

555 is running and adjustable.  Looks real stable but has a slight overshoot at the rise of the pulse.

Pulse looks a little wide too, no that it matters, just wider than it needs to be.  I'll probably differentiate anyway so it doesn't matter.

555 should have no trouble triggering the avalanche stack.

It will be a real kick in the ass of the nay-sayers to have an ou device running on a little 555 timer.  I honestly think that SM used a multivibrator (they work with 100% feedback to the transistors) and he could easily crystal control this - pretty solid clock.  He could have used hv transistors too, or placed them in series for higher voltage.

[Peterae]

Hi G
thats great news with the 555, if a viable device was made using them we would be a laughing stock 

With your sig gens have you looked in to seeing if they have nov rams, as Ds found out it's the first thing to get corrupted, with my old scope it was constantly getting corrupted and i had to reset back to factory and a cal on it regularly.

Now over dinner i have been contemplating your specs for the output stage, maybe i should just go for it at 1200v

[Peterae]

G you have made my life more complicated  

The good news is i am going to build a 1.2KV Psu that uses a 555  

Not my circuit
http://www.cosmicrays.org/muon-power-supply.php#captain

Edit
G you can alter the Mark/Space if you need to

http://en.wikibooks.org/w/index.php?title=Practical_Electronics/Astables/555#Design_equations_2

[Grumpy]

go higher than 1200v if you can as you need to drop at least 1000 volts to get cavitation

(I actually found an image of the cavitation oscillation in reference to the density - think I save it at work as I can't find it at home

cavitation is not the freeing of energy or anything like that it just moves the medium

remember you can tap into the stack at any point to see what is going on with a lower votlage 

for example, you ground the first trans and watch everything on the scope at a few hundred volts

thanks for the 555 tips.  I am pretty  sure that the unit I install in my car will have a more stable clock, but this is good for now.

[Grumpy]

555 straight to the av stack sucks - pulse is too wide

make the trigger shorter than the pulse

[Grumpy]

Keep in mind that you need a short sharp trigger pulse to the base of the transistor and that HV will be at the base when it fires.  So, you can block with a diode or isolate the trigger with a transformer (can be 1:1 or an actual trigger transformer).  Make sure your transformer can handle the HV and will not short through the core.

Methods to fire SCR's (thyristers) are suitable - hint hint...

You can differentiate your pulse to the transformer and clip off the neg spike (minimizes ringing and saturation issues) and then the trans former will differentiate again and you have to clip that neg off too - see attached for some good examples:

EDIT:

These are good to 1600v isolation: (just an example)
http://www.datasheetcatalog.org/datasheet/vishay/pt50.pdf

[Grumpy]

diodes - he he - 1N914 at 4ns recovery isn't fast enough

555 will drift like an SOB if R1 is 50 ohms - hot as hell to!

So, R1 (sets width) at 1k on 555, differentiate and restore dc with 1N914 to a trigger transformer of something like 1:35 ratio - decent pulse to av stack.

Getting noise from the overshoot on the output of the 555 - looking into that.

Ready to sweep with the 555 now.

[Grumpy]

for the trigger coil, I have a TC-21 and a TC-31:

http://www.qimeioe.com/en_product1.5.asp

These easily withstand the pulse voltage of 1500v and require no modification.

output pulse is low as little current is applied from the 555 or other controller but higher than a 1:1 would be

was hoping to have the sweep gen by now, but not here yet

[Peterae]

Am i missing something if you look at this pdf they use 5v to trigger the stack

http://www.ccsenet.org/journal/index.php/apr/article/viewFile/3287/3601

OK i know this is not the config i need, but other documents dont seem to mention trigger problems either.

Have you tried a pull down resistor on the base to stop it floating.

Without a diagram to see your config i dont know really what you have their.

I need to drive my coil from the emitter though via a low ohm resistor i guess, i have not played with this yet, still trying to get the psu built.

[Grumpy]

Am i missing something if you look at this pdf they use 5v to trigger the stack

http://www.ccsenet.org/journal/index.php/apr/article/viewFile/3287/3601

OK i know this is not the config i need, but other documents dont seem to mention trigger problems either.

Have you tried a pull down resistor on the base to stop it floating.

Without a diagram to see your config i dont know really what you have their.

I need to drive my coil from the emitter though via a low ohm resistor i guess, i have not played with this yet, still trying to get the psu built.

That is correct.  Only a few volts to trigger and a trigger transformer is not required.  I am having trouble getting a clean trigger and trying some different things.  Playing with the differentiator components and all that.  Think I need a decent cap across the output of the 555 as it doesn't like the trigger coil or the isolation transformer - rings or low output - hence I think I need a cap across it.

in your pdf link - I have not tried this method of triggering and just found this same file yesterday.  Would be nice to not need the isolation!  However, keep in mind that collector overvoltage switching results in a rise of about 2ns, but takes more current and exposes the trigger to full voltage.

[Grumpy]

here is an other stack triggered at the last transistor and not isolated - this config is for pos pulse

EDIT: wow! they triggered it with a 555 - LOL!!!

back to the bench!!!

[Grumpy]

OK

555 - differentiated - to base of 2N3906 with 12v supply = nice pulse across 50 ohms

Nice pulse to 1:1 transformer looks pretty good too

ready to pulse av stack with it

[Grumpy]

Well..

that sucked

tried one av trans at 240v - no 100 ohm resistor at the trigger - pop! shot to about 500v and grounded through the 555 circuit - not sure what fried yet - got a short and no oscillation

now, about that iso-trans thing...

[Peterae]

Well..

that sucked

tried one av trans at 240v - no 100 ohm resistor at the trigger - pop! shot to about 500v and grounded through the 555 circuit - not sure what fried yet - got a short and no oscillation

now, about that iso-trans thing...

Sorry to hear that G, hopefully just the 555 has gone, i wonder if we were to use a 6v TVS or something to protect the driving stage, 555 are not so bad to blow but i am using my delay boards so will need to be sure all is ok and then maybe use a logic gate as a buffer, having said that just in case, i did order some isolation pulse transformers in my psu build component list.

https://www.rapidonline.com/netalogue/specs/88-2423.pdf

They are only good for 500v isolation, but that's all that supplier had so maybe next order from farnell i can get something better, i am still hoping to avoid using one, but i will need to experiment to see.

[Peterae]

I have finished building 2 1.2KV switched psu's, so far they seem to work well, although in the circuit they published, the 555 was free running at 85kHz using his published values, the circuit specified 2kHz LOL so i changed the 330pF to 22nF and am running at about 2.6kHz and all tests well so far.

Still need to order the reservoir caps for the output stage but have loaded with 1.6kv 1Uf caps and the circuit seems to regulate and remain stable.

More load tests need to be done once i have the final caps.

I have used 12VA Transformer so i presume the max i would get at 1.2kv is about 10mA

I seem able to vary the voltage from 4 to 1200v using the pot which is great.

[darkspeed]

Looks good!

Is this based on your  1KVPSU104.jpg diag.?

[Peterae]

Hi Ds Thx
Yes thats right, there is a link to the original design just above that diagram, it looks like theres a lot missing but it's all SM and each heatsink carries 2 fet's.

the difference is that i am using a 0-250vac / 6-0-6 transformer

and i am using 3 series 2M2 resistors instead of one 5M1

[Peterae]

Not to size

[Grumpy]

sweet!

[Peterae]

Cheers G

A lot of testing to do on it yet.

I am starting to think about the Avalanche stack now, and i am very worried about using the output drive from the emitter because of the chance of serious damage to the drive circuit even using a pulse transformer at such high voltages.

It looks to me as i am stuck with pulsing low if i understand this correctly it means i need to have one side of the output coil to +HT and then pulse the other end low with the stack, i guess from the coil perspective it doesn't matter?

EDIT which actually is what i am doing anyway but with lower voltage on my coils.

[Grumpy]

Cheers G

A lot of testing to do on it yet.

I am starting to think about the Avalanche stack now, and i am very worried about using the output drive from the emitter because of the chance of serious damage to the drive circuit even using a pulse transformer at such high voltages.

It looks to me as i am stuck with pulsing low if i understand this correctly it means i need to have one side of the output coil to +HT and then pulse the other end low with the stack, i guess from the coil perspective it doesn't matter?

EDIT which actually is what i am doing anyway but with lower voltage on my coils.

been at the bench most of the day:

wasted a few hours with a 2N3906 that I thought was a 2N3904 - wtf?

my isolation transformer is now a tiny toroid (choke from radio shack) with 5 turns on it

circuit is a 555 to a half-shot multivibrator with the 2N3904 in it - no differentiator now - makes pulse shorter than 555 is capable of

almost there...