Controller No 4 - Dual Fet Driver - Looking for Explosions in Wires

[Peterae]

While trying to capture the sound there was a much louder explosion than i have heard so far and it's taken out both fet's.

Remind me to socket them in the future.  :

EDIT
OK this is no longer fun, i have blown 6 Fet's now in a row, i cannot get it to be stable as it was yesterday, i had to switch to STP5NK90 FETs these are 900V and the effect has changed, they seem to produce an even sharper pulse but the device no longer goes unstable at the previous frequency, now i have to go up to over 1Mhz and the sound from the coil is more like a loud rushing noise without random explosions.

I need some more IRF840's and maybe try Ds's cap idea across D-S.

[Grumpy]

rather than FET's - try BJT's switched or in avalanche mode

I have had no failures with 2N5551's (1ma supply at 1500vdc)  At higher voltages (over 1000v), you can use an avalanche diode to reduce the number of transistors, but only one diode as the avalanche threshold is not cummulative with multiple diodes (which is interesting and leads me to think that a diodes only blocks one side of the aether flow and not the other).

another idea is to "isolate the FET's" - opto devices will trigger randomly as well, so try small isolation transformers - if something is propagating along the line it will reflect at the iso transformer (they work by field reflection), but it will pass right through a diode or bridge (or other SS device)  In my test with 10kv, the stuff never got past the transformer.  You may recall that Tesla destroyed the generator at a power plant in Colorado Springs with energy reflected back to the plant.

By the way, in avalanche mode with a delay line (open coax about a couple of feet long) for pulse width, the reflected pulse shuts the transistor off - you can apply this in other arrangments and it is easy to setup for any switch.  It doesn't allow constant current even if the switch shorts as the delay line HAS to charge and discharge.

EDIT:

Here is a 2N2369 in avalanche mode for driving a laser diode - pretty simple: http://www.elexp.com/t_SpeedofLight.htm

Replace C1 with your coax to set the pulse delay - try about 3 feet of coax

(see this page for a good example: http://www.holmea.demon.co.uk/Avalanche/Avalanche.htm - this one uses a 2N3904 but does not have a huge cap to keep the supply volts up)

[by the way, hitting your coil with that much energy has to generate reflections]

Trigger them, do not let them free-run as the pulses will be all over the place.

[Peterae]

OK i can confirm it's a voltage death for my fet's when i switched to 900v FET's the fet blew but also the MUR160 blew across the D-S i just looked up the spec and it's a 600V diode.

I have run out of IRF840's need to look at home.
I am just about to try some different 500V 8A ones i had tucked away.

If i have no luck with these i've had it for the day, need to get some real work done, and i've spent all morning swapping fet's 

[Grumpy]

transistor are friends!  Overvoltage and they avalanche, forever if you keep the current low

[Peterae]

Well i guess i could try Transistors, i am not sure if the effect is possible with them or not right now, it's been hard enough trying to replicate the explosions with fets, and even different rated fets alter the effect by the looks of things.
I would prefer to try and get things stable again as i had yesterday.

[Grumpy]

try a small cap between each FET and the coil

[Peterae]

OK this is getting weird, it sounds like voices , i cant even hear the crackles that i hear from the coil

Funny thing, in loads of the video's i have filmed over the last year or so, these noises have been in the background of many of them, i always wondered why because they were filmed in silence or m speaking.

I will Try recording another just to verify.

www.overunityresearch.com/uploaded/CrackleSound1.zip

Have i just reinvented the Radio  

EDIT
Oh Dam another FET gone, that's me finished for now

Tomorrow i will start work on Ds's FET Board, need a day or so break from this.

I couldnt find 180pf without ordering a shed load, so ordered 120pf 500V from ebay which is the nearest they had.

[darkspeed]

Peterae, i use a silver dipped mica 180pf across the S/D 500v for 500v mosfets 1000v for 1000v mosfets etc...

[Grumpy]

I'll bet that a small air-cap near your crackling coil will indicate that the permittivity of space is going crazy

[Peterae]

I suspect you maybe right G, the mike is picking some weird things up in the above recorded wav file, it's an electret mike, i wonder whats in my camera as that seemed to record correctly, maybe because it's screened in the camera or maybe it's a tiny magnetic pickup instead of capacitive, or maybe because it wasnt as close.

Many of my recorded videos have these strange noises in the background and they were recorded on my laptop webcam using the built in mike which again is likely to be electret.

[Grumpy]

if you have an extra electret mic - try removing the electret (usually a piece of plastic) and see what you here via changes in the capacitance of the mic and not physical movement of the electret.

[BEP]

Have i just reinvented the Radio  

SubSpace Radio!

Seriously, don't be surprised if you've built a very funky form of superregen radio receiver with capacitive coupling to the electret mike. You probably wouldn't here it until you played the video back.

I've heard very similar. If I remember correctly I was trying to tune in a phase modulated wide-band RTTY signal. There were several international AM broadcast stations on nearby frequencies. The damned receiver kept trying to lock in on the broadcast stations.

[szaxx]

Hi all
An oscillator with non linear elements, a direct conversion  receiver of sorts. These can function on multiple  frequencies ie harmonics. The signal may be from effective energy sucking  antenna properties. This isnt in the arrl manual lol.

[BEP]

Hi all
An oscillator with non linear elements, a direct conversion  receiver of sorts. These can function on multiple  frequencies ie harmonics. The signal may be from effective energy sucking  antenna properties. This isnt in the arrl manual lol.

It isn't in the ARRL handbook but HAMs use it frequently   You are right about tuning multiple fundamentals AND most of their harmonics. In many cases an antenna isn't even needed.

Basically, the stronger you cancel out a signal the stronger it is received. Folks laugh at 'energy sucking antennas' but I'll stop talking about them when they are taken out of some GPS, cell phone and weather radios.

The zipped wav file sounds like the coil has also gone microphonic.

[Grumpy]

I don't think this is related, but...

Crackling Wires
http://www.sciencemag.org/cgi/reprint/318/5848/207.pdf

[Grumpy]

Peter,

your crackling sound reminds me of an overcharge piece of coax

same sound in high tension lines (overhead) on a summer day

EDIT;

next time you get this going, try a coil of bare wire with space between the turns

EDIT 2:
http://www.procosound.com/downloads/whitepapers/Understanding%20Instrument%20Cables.pdf

Start at "What does the electrostatic shield do?" and read down the "Why are some cables microphonic?" and read about "triboelectric noise".

hmm - can you varnish or otherwise bond the wire to the coil form?

[Peterae]

Well i had another stab at it this morning with the new 120pf Silver mica caps across each fet.

I only had 1 fet left which wasn't a IRF840, i haven't been seeing the crackling with this fet, but at 1mhz i had a rushing sound.
The fet's seem to survive better with the caps, but as soon as i put the frequency back to 350kHz and did a scan one fet popped immediately while i was looking for the explosions.

Unfortunately i had the clever idea of ordering IRF fet's from Hong Kong in tubes of 25 which is good the way i'm blowing them but it's probably going to take a good week or so for them to arrive.

I think the plan now is to think about new drive systems.

I need a fast output stage that's resilient to voltage spikes, i know valves are the way to go but they cannot drive a low ohm coil in pulse mode as i understand, anyone know how to do this

Either that or i just persevere as i know i can get stable running and the explosions.

I could maybe mean time try G's idea i have a couple of BU2508DF line transistors at home somewhere i guess all i need is a base resistor and can then drive them from my fet driver, any idea what sort of value base resistor would be best suited.

Grumpy if i can get this stable indeed it would be interesting to dope the coil in varnish and see how the noise changes.

Peter

[Peterae]

Longitudinal stresses explain wire fracture under high current pulses.

http://www.df.lth.se/~snorkelf/Longitudinal/node3.html

[darkspeed]

i know valves are the way to go but they cannot drive a low ohm coil in pulse mode as i understand, anyone know how to do this

Yes they can - http://www.google.com/search?client=opera&rls=en&q=vttc

- http://www.stevehv.4hv.org/VTTCfaq.htm

That is what my 811 triode amp is for...

[Peterae]

Thats great news Ds thanks for the info, now i know what i got to do.

[Lindsay Mannix]

Yep, DS good info !

Wen considering the 10 year gap between tubes and fets as described
its hard to to try and work with basic differences even though some parts are similar

http://www.w8ji.com/vhf_stability.htm

It seems likely that we dont want  to supress the instability but capatilse  around it .
Afte all we WANT noise  (apparently)

The anode/grid  circuit might oscillate in its instability mode and the grid would be  pulsed about in  this state.

The transition from stability to insatbility might...just might  just be what brings in something other that what we think we  are putting in

so ,3 unstable oscillators of different impedances pulsed at 3 different stable rates . All around a conductor that somehow harvests the new interaction

An area worthy of  investigating..at least  in my bewildered mind.

I and others would love to see the many possible bottle methods get some good quality air here

BTW Peterae, that wave file sound is  rather spooky


Another interesting point is that these circuits do not "work" in multisim at all ..after all thay are not really working "properly"

[darkspeed]

The gradient coil is one of several internal parts of the MR system that you cannot see. The gradient coil is made up of loops of wire which are embedded in a hard plastic tube. During the scanning process an electric current is switched on and off through the gradient coil approximately every few milliseconds. Because the switching is so rapid, the wires vibrate within the hard plastic and cause the knocking sound. This knocking is not harmful but the sound can be irritating to some patients. You will here different knocking sounds during the MR exam, this means that different types of "MR sequences" are being run to acquire different views and images of your body.

[Peterae]

Thx for the info Mannix

Yep it would be nice to have a working Valve pulser.

[BEP]

Peterae,

If you choose to try it, be aware there is a definite delay between the trigger signal applied to the grid and the output. This would not matter for most but you are on the bleeding edge of home-brew pulsing.

Now the good news: The instantaneous voltage and current can be magnitudes higher than the device rating, provided the pulse time is kept very short. This is something you can't have with FETs.

My old BT79's lasted a couple of years (maybe a dozen capacitor bank discharges) before they cracked. Some firings were approaching 400V with peak current estimated at 20kA. (they don't like to turn-off once triggered)

I trust you will stick with making noisy coils rather than coils jumping off the bench or causing destruction

[Peterae]

Cheers for the help and info guys.

I have thought about it and for now i am going to give the valve build a miss.

Is there anything i can do to help protect my FET's the 120pf cap across it didn't work.

How about a 300v spark arrestor.

I also still have the transistor idea to give a try.