OverUnity Research

New Controller is now built to explore the Explosions in the wire experiment.

It has 2 fet's and drivers and uses 2 digital mono sub boards for phase and pulse width control.

OK Board has been powered up and all checks out well.

Connected a coil as before and looked for the unstable spots that i was getting previously and found that this controller is the most immune controller i have built so far, i needed a supply voltage as high as 80Volts to get the cpu to start crashing as before.
The delay setting was about 120nS so i manually set this using the dip switches and started winding the supply to the coil up in volts.
I placed a magnet near the coil to indicate if i was going to get close to producing the Explosions and reached a point where there were loud crackles coming from the coil and before i knew it my FET went short circuit but only the FET that was on at the time which was the delayed FET.

I was scoping at the time and i certainly didn't see any pulses that went over the voltage rating of the fet so not sure why it has gone short S-D and the fet was running cold at the time.

Next i will swap the fet over and check to see how it's gate reads, maybe next time i need to try a bulb in series to try and limit peak current.

OK Failure mode for the delayed FET is as follows

Gate-Source = 230 Ohms
Gate-Drain = 233 Ohm
Drain-Source = 3.9 Ohm

Anyone any ideas what would cause the above failure.

I personally believe it to be peak current, even though i am using a 1amp psu but the pulses are so narrow that i would image the reservoir cap could potentially deliver a lot more.

Looks good!

You could put a choke between the power supply and the board and then a small filter cap on the board that will supply 1 amp

Sometimes back feeding HF into the power supply can cause them to loose their regulation and dump more than expected.

Only think i could think of off the top of my head..


What delay boards are you using?

Hi Ds
Yeah i could indeed try the choke to see if it stops, but i am trying to encourage the crashing CPU because it means i get re-triggering of my FETS which brings on the Explosions in the wire, and because this controller is better at handling the EM or RF bursts i have to crank up the psu to the coils to 80V to get the re-trigger and at this voltage my fet died as soon as the re-triggering started.

The Digital Monostable is my own design and build.
Schematic here
http://www.overunityresearch.com/index.php?topic=217.0

Place a small isolation transformer after each FET and before the bifilar coil.  This will reflect "it" back to the bifilar and you may get explosions at the iso-trans like I mentioned before.  ISo-trans need to be identical or compensate your delay to account for the difference.

Basically you reflect the ether back on itself at such a fast rate that it cavitates.  Much more too it but that is the starting point.  You crash two ether fields moving in opposite directions and at different velocities.  Ths is the only way that the delay bifilar effect can be achieved.  Without the ether, there can be no explosion - no RE.

EDIT:
moving the controller away from the bifilar will only save the FETs if the failure is cause by the effect.

The explosions only occur with a single wire coil(Monofilar), so each fet is connected together to drive the low side of the coil, the coil has a flyback diode across it also

I can get the pulse with both monfilar or bifilar coils, in monfilar the pulse always appears first then i hit various retrigger timing points as i scan up in nS delay settings.

Quote from: Peterae on 2010.04.23, 17:58:25
The explosions only occur with a single wire coil(Monofilar), so each fet is connected together to drive the low side of the coil, the coil has a flyback diode across it also

I can get the pulse with both monfilar or bifilar coils, in monfilar the pulse always appears first then i hit various retrigger timing points as i scan up in nS delay settings.

do you get explosions driving the high side?

I would still try isolating the FET's from the coil through a transformer

Hi G
Never tried driving High Side

This is the way i see things right now, a sharper brain should be able to refine it some.

The magnetic field that builds around a current carrying wire tends to resist changes in current(This happens on the first pulse)

but during the second pulse that follows shortly after, i can make the conductor sink enormous amounts of current without the normal restrictions or in other words without an induced back EMF to impede the speed at which the current can flow, hence on the 2nd pulse i can send a very fast magnetic pulse(Shockwave)

or Version2

The back EMF is helped by a collapsing new type of field that allows the back EMF to form much higher in amplitude and much faster than normal, this explains why the pulse is much higher in amplitude and much sharper in the pulse experiments i have done.
There seem to be multiple delay timings that create 2 distink effects, the first effect is always the pulse this shows in the form of a sharp pulse, then all the next delay timings show as bursts of EM that cause re triggering and crash the CPU in the controller or interrupt the chips.

Above i am talking about the pulse the pulse can be created using bifilar coils with each fet driving each coil in the bifilar, the pulse can also be created by connecting both fet's together and connecting to a monofilar.

The Explosion can only be created using the monofilar setup with both fet's driving the low side, the first delay setting during the delay sweep causes the pulse and then there are a few more delays further on that cause instability and retriggering, during this retriggering i sometimes get the explosions in the wire noise from the monfilar coil, so what ever is alowing the very fast creation of magnetic fields causes a condition where a collapsing pulse field intercepts another new expanding field, the consequences of this seem to be a snap of energy and an explosion inside the wire, by explosion i mean the sound this could be a mechanical slap of the individual multiple wire cores.


Here's the picture of the pulse to refresh what it looks like

Quote from: Peterae on 2010.04.23, 20:44:16
The magnetic field that builds around a current carrying wire tends to resist changes in current(This happens on the first pulse)

but during the second pulse that follows shortly after, i can make the conductor sink enormous amounts of current without the normal restrictions or in other words without an induced back EMF to impede the speed at which the current can flow, hence on the 2nd pulse i can send a very fast magnetic pulse(Shockwave)

You have my 'yes' vote for this explanation.

Quote
fast creation of magnetic fields causes a condition where a collapsing pulse field intercepts another new expanding field, the consequences of this seem to be a snap of energy

Sounds like you are working into something like... http://assets.cambridge.org/052148/1791/sample/0521481791WSN01.pdf
(This is a little easier on my brain...http://yesserver.space.swri.edu/yes2006/tomr.html)

Fun stuff if you go that way. If so, the only way to stop blowing FETs is to use them (in slower capacity) to drive a TL. The TL needs to soak up the energy before it hits the FET. Then you need to figure where to put your scope probe. Somewhere not at the FET.

I have the feeling your scope isn't showing the real amplitude.

>>Edit

Place a very small compass against the coil. It should point parallel to the axis of the core. Then cause the explosions. The compass should start pointing perpendicular to the axis of the coil.

Thanks Wave
It's great to know you are watching over my work.

You are totally correct about the scope probes, it's not just the speed of the scope either, as i have a 25Mhz and a 250Mhz and they both show about the same size pulse, there is a larger pulse using a x10 probe than using my x100 probe, and sometimes it's very apparent there clipping of the wave on top, also changing ranges on the scopes gives different amplitudes.

Some observed facts about the pulse.
Increasing the supply to the first non delayed fet alters the timing needed to get the pulse to appear.
Increasing the supply to the second delayed fet alters the amplitude of the pulse and will to a lessor degree alter it's timing slightly.
Also in this video it was interesting that the pulse alone crashed my scope, this was without re-triggering.


Using 2 coils the same dimensions wound at each end of a long perspex rod and driving each coil with each fet still creates the pulse even if the coils are over a foot apart at each end of the rod.


I have seen the pulse using an Avremenko plug, by connecting the diode common to a coil that has the pulse on, i am able to scope the pulse across each diode on the Avremenko plug

Using 3 delay channels i am able to create 2 pulses one after the other, but to achieve this i needed to use separate fets for each channel, using controller 3 which used 4 delay modules to try and recreate this but using only one fet stage being driven from all delay modules i did not see this, altering the timing on the 3rd channel altered the timing on the pulse being generated by the 2nd channel.


Using 3 saturable inductors in series with one of the bifilar coils and also in series with the flyback diode creates bigger pulses.

When I'm working on such pulse generation it helps me to think of it in terms of forensic ballistics.

When there is a need to obtain a bullet that shows only the markings of the weapon barrel, a round is fired into a very dense gelatin.
Imagine the wavefront created by that bullet and the displacement as the round travels (CEMF). Also realize the effect the gelatin has upon the speed of the round.

Now fire a second round, in the exact path of the first, immediately after the first. If the second round is fired while the gelatin is still expanded from the first, the speed of the second round is greater over the same distance. At some point, the second round will catch up to the first. At that point, calculate summation.

From there, add a third or more. All basic principles displayed with strings of saturable core inductors or TL pulse circuits.

Imagine the second round timed so it meets with the collapsing gelatin of the first round  :o

I was attempting to have a DC output from summed waves so sending multiple pulses down the same direction only gave me the summation (sometimes multiple, repeated summation of the same pulses(cracking and popping), mostly - lots of blown FETs with no apparent heat.

The DC output only came from sending multiple pulses to a common point from two or more separate points. With this, I could have the results of superpostion and a standing wave (the DC shift after the final massive superpostion -or- the long string of smaller wavelets after the big pulse).

I look forward to your experiments because I'm constantly living out of a suitcase these days and your work delves into the real matter of this field with presentations I can relate to and understand.

>>Edit

A quick note before I start my day....

I'm pretty sure your monster peak is also the result of a generated pulse meeting with the reflected pulse of the previous pulse (superposition). So, you have the critical adjustment of the voltage AND phase difference. Your coils must be long enough to house more than one complete pulse.

Cheers Wave

Work 1st banging a coil 2nd isn't that the way life goes, i am very lucky really being able to do this stuff at work during quiet periods.

The wire length can be very short or quiet long, i tried an experiment i wound a coil and found the pulse, i then proceeded to cut a couple of turns off at each test, each time locating the phase delay at which the pulse occurred and basically i was not able to predict the delay value at which it would show up, i could still get a pulse even with a foot of wire but the amplitude was getting less if i remember correctly

Another experiment i tried was to lay 1.5 meters or so of wire on the bench and set the phase to get the pulse, i then got a ferrite rod from an old LW aerial and started winding the loose wire onto it 1 turn at a time, and each time i put a turn on the ferrite again the timing of the phase altered, infact just moving the wire around also changed it, this was also apparent when i had the explosions, if i pushed really hard on the coil and wooden former i could deform the coil enough to change the timing at which the explosions occured, also moving the closeness of each turn altered it as well, from this effect i concluded that if SM used the explosion effect he would need to glue down each turn of the coil to stop even a mm movement otherwise he would have had to re cut his delay line to re tune.

I am not sure i know how to work out how fast a pulse travels in a piece of wire, it isn't the wavelength of the frequency for sure.

In the past i have pulsed a wire with 35khz at 32KV when i was doing Lifter experiments and could clearly see the pulses in subdued light which were from memory about 1cm apart and showed as corona discharges.
So from this i guess even 1khz i will have many pulses in a short wire.

One thing when i took the fundamental frequency above 460khz the explosions stopped, normally the master frequency makes no difference so i use either 10khz or 1khz and then adjust the delay for the second pulse.

Quote from: Peterae on 2010.04.24, 15:27:33
The wire length can be very short or quiet long, i tried an experiment i wound a coil and found the pulse, i then proceeded to cut a couple of turns off at each test, each time locating the phase delay at which the pulse occurred and basically i was not able to predict the delay value at which it would show up, i could still get a pulse even with a foot of wire but the amplitude was getting less if i remember correctly

Ah! Then yours has more to do with the rise time than the frequency or pulse width. In my experiments I determined similar results were do to the leading edge of the wave.

Quote
from this effect i concluded that if SM used the explosion effect he would need to glue down each turn of the coil to stop even a mm movement otherwise he would have had to re cut his delay line to re tune.

As I'm sure he did. The smallest TPU was covered in a resin. Both using the small toroid had (what I believe is ) Rob Roy Class I, Div. II sealant dripped over the small toroid, gluing caps down, holding loose wires in a fixed shape and position, holding the outer toroid to a support AND... Look at all that tape and heat-shrink. He didn't put that on to hide details. He had to prevent even the smallest movement.

I use sanded white tile grout. It is much cheaper and exactly the same thing except no epoxy base.  ;D

Quote
I am not sure i know how to work out how fast a pulse travels in a piece of wire, it isn't the wavelength of the frequency for sure.

VOP (Velocity of Propagation) depends on the wire, inductance and almost everything else. You move the wire or change the shape VOP can change. For a single conductor you can usually go with c x .85. Bifilar adding try .65. Bifilar opposing (cancel magnetic or 'bucking') try .95 or higher(some say 1.65).

Quote
One thing when i took the fundamental frequency above 460khz the explosions stopped, normally the master frequency makes no difference so i use either 10khz or 1khz and then adjust the delay for the second pulse.

Hmmm.... I never had any noise above 250kHz but my initiating pulse width was longer than yours. The wider the pulse the more results it will mask. Just so folks know I'm still quite crazy.... IMO, this interaction doesn't work above 250 because your pulse isn't interacting mainly with the mag field of the planet. ::)

QuoteHmmm.... I never had any noise above 250kHz but my initiating pulse width was longer than yours. The wider the pulse the more results it will mask. Just so folks know I'm still quite crazy.... IMO, this interaction doesn't work above 250 because your pulse isn't interacting mainly with the mag field of the planet. Roll Eyes

I will confirm this it's been a long time since i did that test.

Ok i did the video.



Interestingly over 300Khz the first pulse sharpens and in fact only 1 channel is needed without a delayed pulse to get a sharp pulse.
If i carry on up in frequency i reach a frequency of about 4.25Mhz and then i start getting the re-triggering and white noise on the coil with a magnet nearby and this is using only one fet channel with no delayed pulse again.No sign of explosions but my cpu is in constant reset and the touch pad on my lap top stops working.

Pulse width is 32nS throughout the tests

Well those results have confused me.

Here's a video of what happens above 250Khz, both fet channels are being driven by 0 delay phase so are in fact both on at the same time, so i am therefore only pulsing the coil once, but increasing the base frequency at which they are triggered.


So the way i see the pulse now.
Under 250khz the first pulse is normal, but the second pulse fools the coil into thinking it's being driven by a much higher base frequency, it's the second pulse that causes this effect, the same effect can be produced by driving with 1 pulse at much higher frequencies.

One advantage of using 2 sharp pulses to cause this effect is that because there are less pulses per time interval the net power being introduced into the system is a lot less.Instead of having 350,000 pulses in a wire per second, i can drive with 1000 pulses per second and have the same effect.

But then i suppose the question now is what parameters are changing within the coil above 250khz or below 250khz to cause this, and i think it's to do with energy saturation of the aether around the coil, but maybe a technical person maybe able to explain the effect using conventional theory.

Maybe if the Aether is pulsed below 250khz then it is able to create a more defined ripple pattern that can dissipate enough away from the coil to stop some sort of reflection from changing the way another set of pulses are seen.

speed of propagation is related to permittivity and permeability (see definition of speed of light)

http://hyperphysics.phy-astr.gsu.edu/hbase/electric/elefie.html

QuoteIn the equations describing electric and magnetic  fields and their propagation, three constants are normally used. One is the speed of light c, and the other two are the electric permittivity of free space ε0 and the magnetic permeability of free space, μ0.

alter either and SOP changes

you first pulse alters either

how does the speed of light change when these values change? 

Now I need to question my experiences.

The major differences between your tests and mine:
1. your signals are under much better control and more narrow
2. my tests were on toroidal cores and formers

My first memory of this 250k change point was part of a wave propagation class in the late 70's. The basic idea was above 250 the mag/elec vectors were much as explained in the common literature. Below, the magnetic was more pronounced and effective than the electric for communication and electronic warfare. The lower you went the more pronounced the difference.
This has been verified over the years in my HAM hobby using below noise floor magnetic loop antennae.

One point of your presentation struck a chord for me.

In the attached capture, which of the first three pulses are the two pulses you are injecting?
If the middle pulse is not generated by your electronics the group of the first three are an example of the 'three-sisters' effect. This effect only appears when your massive pulse is being generated. Most folks don't see the single result pulse because it goes off the screen or is too fast to display.

QuoteIn the attached capture, which of the first three pulses are the two pulses you are injecting?
If the middle pulse is not generated by your electronics the group of the first three are an example of the 'three-sisters' effect. This effect only appears when your massive pulse is being generated. Most folks don't see the single result pulse because it goes off the screen or is too fast to display.

I hadn't even seen that DOH

Those first 3 ripples are the result of the first 32nS pulse and maybe the 2nd peak and 3rd peak of it are the result of some sort of ring-down
EDIT: the first pulse is the injected pulse the slightly lower pk ripples that follow it.

This is for the above 250khz test where i was running both fets in parallel

OK 2 fet's in parallel and just 1 fet does make a difference see below and effects the ripple after the pulse.

So, if the first pulse alters the permittivity of space (or whatever medium) and then the second pulse is >C, it might charge a capacitor to a higher voltage than expected.  Like "voltage magnification".

We should be able to test this hypothesis relatively easily.  Oh yeah, we already did.

If we can figure out the necessary orientation for polarization to alter the permittivity for our pulse, we can alter it externally and then analyze the pusle to see if the previuous hypothesis may be correct.

In other words, the second pulse interacts with a field of altered space (increased permittivity or permeability via vacuum polarization? decreased speed of light...) which results in voltage magnification. Can this field be replaced with a static field? (we already know it can be)

Edit:
by the way - permittivity is a scalar

So guys, whats my next test.

I was hoping with this controller to track down the explosions, but from the tests i need 80+ volts on my coil to get retriggering and my fet died straight away, shall i try limiting the current or do you not feel it was excess current that caused the failure.
If it was a large dv/dt that killed the fet maybe i should try a fet with a higher volts rating.

Maybe to prove my theory with the white noise, i should build an amp and white noise generator and inject this into 1 coil of a bifilar and then inject sharp pulses in the other bifilat coil.

Do we know why the pulse appears with 1 pulse above 270khz but requires 1 pulse and then a delayed pulse below 270khz to get a sharp pulse.

Do i start trying to measure the power in the pulse verses power used to generate it.

I'm not very good with direction and go off track fast so any help would be much apreciated.

Cheers,
Peter

Peterae,

Perhaps replacing the fets with 12ax7 triodes at this point would be interesting for comparison.

Steven said that zener diodes on the grids helped a lot to reduce crossover distortion

Text only here so limited info. If you are using a bifilar put a zener as suggested and send the pulse directlyinto one and use a length of coax for a delay matching the gap between the pulses velocity factor for the coax used into the length calculation so there is a single pulse then a double pulse then another single seen by the coil.  A high voltage device will be neccessary for both switches.
I used irf 740 with 36 volts and they fry almost instantly. No heating at all dont forget what agent gates had as a problem his heated up so a compromise may be neccessary. I will soon be back so can see pics and video.

Quote from: Peterae on 2010.04.26, 21:05:58
I was hoping with this controller to track down the explosions, but from the tests i need 80+ volts on my coil to get re triggering and my fet died straight away, shall i try limiting the current or do you not feel it was excess current that caused the failure.
If it was a large dv/dt that killed the fet maybe i should try a fet with a higher volts rating.

You have added more questions.
You call it re-triggering. Maybe it isn't. Re-triggering at the FET may just be a result not a cause. Is it really related to voltage, current or both?
If voltage then add a small capacitance across the FET. If current then place a small inductance between the coil and the FET. Either will change the frequency but will they make the FET live through a higher coil drive voltage?
I suspect it is a voltage issue destroying the gate junction. Here is where I would try Lindsay's suggestion.

Quote
Maybe to prove my theory with the white noise, i should build an amp and white noise generator and inject this into 1 coil of a bifilar and then inject sharp pulses in the other bifilat coil.

I have doubts creating spatial effects in a single coil will result in much, but who knows?

Quote
Do we know why the pulse appears with 1 pulse above 270khz but requires 1 pulse and then a delayed pulse below 270khz to get a sharp pulse.

I think the second pulse simulates a higher frequency on a longer wire. I use an external pulse to pinch current on another wire so it will 'fit better'<?>... long story. I think I told that one to you and .99 already.

Reestablish the test with these results and trim the wire. If the cutoff moves higher then you'll know the point is related to travel time through the wire. It won't be related to wavelength, perhaps pulse rise time instead? Is 270 the change point for other coil lengths? Is there anything else that can vary that 270 number? Ambient temperature? Something crazy like time-of-day?

Quote
Do i start trying to measure the power in the pulse verses power used to generate

Your results are differing from mine so I'm not sure you are at a point where it will be easy enough to measure accurately. You do have a lot of 'hash' going on. Maybe you can wrap your noisy coil in a secondary coil for measurements?

http://74.125.45.132/search?q=cache:k7xHbMIFUGkJ:hi.baidu.com/5281566/blog/item/5a32b78633884f3d66096ef9.html+lightning+gate+oxide+damage+MOSFET&cd=9&hl=en&ct=clnk&gl=us

(Scroll down to "Spurious oscillation", The "Miller" effect, and "Radiated interference with controller" )

Have you checked if the pulse produces a charging effect of any sort?  I.E. does it increase the voltage on a capacitor (precharge to 50v) with coiled wire leads, or increase the voltage of a battery.

In any event, the energy of the explosions appears to be released into the ambiant environment and lost - try to capture it.

Thanks Guys
Not sure i have the resources to build Tube amps right now.

I will come up with some tests based on whats been suggested.

Cheers,
Peter

I am making good headway now.

I wound a coil using thinner wire and longer more turns and tails, this has allowed me to use lower volts 50 ish.

I can now produce a popcorn noise from the coil, with a magnet close i can see the tails jumping slightly so this is interwind movement making this noise i think, it's not easy to maintain as everything around keeps crashing, PSU's and pic but i will try to get some footage soon.

A 1 loop shorted turn really increases the sharpness of the pulses :)

This is getting very interesting

I am now getting the explosions consistently with this coil, and they appear at about 450kHz ish fundamental frequency.
I have one channel being driven by my pic and this is constantly reading garbage which would also mean the phase delay setting is random and constantly changing, i can see i need to build a white noise generator board now to plug in it's place.
Something really weird it appears i need to run the system for about 5 seconds before it appears, maybe the characteristics of the fet are changing to allow the effect

Quote from: Peterae on 2010.04.27, 14:34:24

Something really weird it appears i need to run the system for about 5 seconds before it appears, maybe the characteristics of the fet are changing to allow the effect

Sounds like you may be plowing through a static charge on the surface of the conductors insulation...

If the  interwinding movement is faster than 340.29 m / s you will hear a crack like a tiny whip as the wire ( wires ) break the sound barrier ( helical rail gun )

You need a SA shot of what is on the wire and an SA shot of the audio through a mic

sounds good!

check for the good stuff - this only takes a few minutes:

Charge a 100V (or higher - ceramic if you have one) cap to 50V (24v or 35v should work too). Cut two separate pieces of wire around 8 feet long that has good insulation. Connect one wire to one terminal and the other wire to the other terminal - no circuit!. Curl the wires up into two rough cylinders, diameter unimportant. Place near your coil with the explosions. Now pulse the coil whilst measuring the voltage on the cap, and scoping the cap. On the scope, the stuff looks like a very short burst of sustained oscillation. Voltage on cap will steadily rise if you got the stuff.

Cut wire down to get best pulse definition.

I will try G thanks when i get in tomorow.

Just uploading a video.

Yep got a SA shot in the vid

OK Here's the Video.

It's not very good, i am using a normal camera which has video function oin it and is normally pretty good but the sound is crap.

You need to wack your volume up to hear the crackling where as it's quiet loud in reality.
Maybe tomorow i can try and capture the sound using my laptop and an external mike, and try a SA plot from it.

The Spectrum video shot show dark blue bursts this is when the crackles occur you will also see it go flat line, thats where i flick the switch on the psu to reset it because it keeps crashing into Constant Current mode, you will see the SA go flat line and take a few seconds to re start the Crackle burst, even though the psu reset is instant the crackles take a short time to restart up.
The best crackling sound can be heard right towards the end of the video when i move right up close to the coil.

If I remember corectly the SA band is 0-50Mhz i think you can just see that in the video.

The coil is flat and rectangular, seems to be a better setup for the explosions.

Crackles1 picture = Crackle happening
Crackles2 picture = without crackle
Crackles3 picture = Crackle Happening

Plug your ears with your fingers - is the sound still loud?  (seriously)

Focus man focus!   ;D

What kind of wire? what is the breakdown voltage of the insulation?  Sounds like arc over from the vid..

It's automatic and consistently out of focus  :-\

the best cam i have is built into my laptop but since i got a laptop based scope i cant use it DOH

but i could do a screen dump on it which should be perfect if thats needed

Shall i try getting a sound sample tomorow with a mike, and record it as a wav, i think theres some free software somewhere never used it yet though

How does it look in the dark? like no light at all... give your eyes a while to adjust and look for flashes

I dont think it's electrostatic i can place my finger over it and it makes little difference as long as i dont deform the coil, it's always light when i'm at work LOL so it could be hard to get it dark

I think i need to try capturing the effect using a detector, i will try G's suggestion first

must try digging my compass out as well

Do the SA shots mean anything to anyone?

Dont forget the tug in the valve that SM mentioned is a mechanical tug or alignment with the earths mag field, no plasma there

This may sound stupid but stand the board up where the coil is in earths plane.. any better results?

Can you do a clearer screen capture of no effect / effect ?

I would really like to compare the two!!!

Yeah indeed i can do will be tomorow now.

QuoteDo the SA shots mean anything to anyone?

Yes they do!

They mean you need to SA above 50mhz or you need a much longer wire to bring the pulse down in frequency

Your many thin lines in your NON EVENT photo show two freq that are not harmonics of each other + intermodulations so the amplitude is not exactly additive

Your fewer thick lines in your EVENT photo show that the frequencies are much closer to being harmonics of each other and the products are closer as well so the amplitude is additive

This is the same thing i am getting with two freq that are just a few hz apart.

In the background there is a green pulse on a black background... what freq is the pulse ringing at?  ( frequency of the ringdown )

there is structure (order) to the crackles per the SA shot

Where was the SA probe located ?

Quote from: Grumpy on 2010.04.27, 22:06:54
there is structure (order) to the crackles per the SA shot

Where was the SA probe located ?

Agreed. There are two or more entrained signals. Sonic? Coil length vs. coil diameter? modulating the 450 carrier?

what is the spacing between the groups when the crackles are occuring?  (can't see - blurry picture)

EDIT:

OK looked again and it looks like the peaks are about 3mHz apart

what is the resonant freq of your coil?  I am wonder ing if you may be seeing the excitation of the coil itself

G - is correct it very well may be resonant rise at a very high frequency due to the short length of coil wire

looks like the board is 4" x .5" and the coil is about .75" ?  

so the coil would be 10" dia x .75"

28awg?

so say maybe 50 turns

1570" of wire

1180Uh self inductance and 26.40Pf self capacatance

so 900khz natural resonant frequency ??

that was just a wild guess... but  it fits nicely with the 450khz results   ;D

Thanks for all your help Guys i cant do this without you  ;)

I will attempt to answer all the above question during the day as i do tests ect.

My PSU seems to randomly switch to constant current mode during some of the EM bursts and it stops working because i had the Current in that mode turned to zero, so i thought i could save myself some trouble by winding up the current, bearing in mind it's only a 1A PSU, as soon as i did a FET blew, this time it was the non delayed FET, this is interesting because it means there was enough energy in the system to blow that FET, previously i had seen the delayed FET blew only it has blow slightly different to the last one.

G-S 1.5 K Ohm
G-D 1.510 K Ohm
D-S 3.5 Ohms

The Former dimensions are
81mm x 6 mm and i have used 4 layers of 1.5 mm card
Turns = 32
So total coil wire length = 5.568 meters or 219.212 Inches

L = 41.25uH @ 10 kHz
Q = 1.8508
R = 1.4006 Ohms
Z = 2.9464 Ohms

Driven Frequency where Explosions were heard was 378.79 kHz

Next Up Spectrum of Noise Sample

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.

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.

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  >:(

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

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.

try a small cap between each FET and the coil

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  ;D

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.

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

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

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.

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.

Quote from: Peterae on 2010.04.28, 13:38:55
Have i just reinvented the Radio  ;D

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.

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.

Quote from: szaxx on 2010.04.28, 22:39:17
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  ;D 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.

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

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

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?

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

Longitudinal stresses explain wire fracture under high current pulses.

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

Quote from: Peterae on 2010.04.30, 10:35:41


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...

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

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"

QuoteThe 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.

Thx for the info Mannix

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

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 ;D

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.

Hi
Not sure of the speed of transient suppressors ,varistors, they are worth a look at as they are a available in many voltages. Let us know if appropriate they do sink spikes very well.
Steve.

You can give the spark arrestor a try but this requires a wait period for charge to build up. This period may allow the damage point to be reached. It may be all in the sizing.
Same problem with most transient suppressors.

The cap won't be of much use unless you bleed it.

In the old days I would just connect a few NE1's or 2's in series across the switch. Add a pot in series for some adjustment. The problem with neon bulbs is just before lighting the gas they act like negistors. This may confuse things.

All this would make creating such short spikes difficult if not impossible.

Have you thought of just firing a part of the coil?
What I mean is - center tap your coil... fire fewer turns and let induction work on the whole coil. Less voltage would be required to create a much higher voltage across the complete coil.
I have doubt about getting the results you wish.

Maybe switch to FETs firing self-quenching spark gaps?(Pic shown is a sketch of a symbol I haven't seen in a while.)

Hi,
Theres a possible use for a zobel network here if correctly worked it should do. Ok on varistors speed, ww, Its knowing exactly what needs addressing to find the remedy. It could be a time related effect and this changes everything. The cure then would involve a different control device completely. Semi's have a relatively slow electron transittime.

I use the series circuit depicted in this paper:    http://cmosedu.com/jbaker/papers/RSI621991.pdf    and 100 pf caps across each transistor (except the triggered one) from collector to emitter to ensure avalanche (discussed on page 1).

I have run various versions of this for hours with a 1500v dc, 1ma supply pulsing various coils and resistors.  The trigger device must be isolated.

Peter, what voltage range to you have available? 

Hi G
I am using a variable PSU 0-125 VDC up to 1Amp for the coil supply.
The problem i have is that with this setup i need 70 volts on the coil to cause the instability and hence create the explosion noises to study, but at this high supply voltage my fet's are dying, when i first built a setup sometime ago now and i first discovered the explosions i only needed 45 volts to get the instability and explosions and my fet's lasted quiet a long time

So far i have blown 600v diodes across the fet and 900v Fet's.

I will have a read Thx

Hang on a minute, if it's over volting the fets, how come only 1 blows, after all both fet drains are connected to the same end of the coil, it's nearly always the delayed fet that blows. ???

EDIT: I have had the diode blow across drain source as well, so what does all this mean, a very large back emf in the coil when i get the explosions?
but then why only the delayed diode as well, surely both diodes would burn, but the non delayed fet is off to protect the fet from a current surge.

http://www.jensign.com/opto/speedoflight/index.html

2N2369 avalanches in the 50v to 70v range

about a pound a piece:  http://uk.farnell.com/jsp/search/browse.jsp;jsessionid=JU5VNVCUD1RU0CQLCIQZN4Q?N=0&Ntk=gensearch_001&Ntt=2n2369&Ntx=mode+matchallpartial&displaytext=&_requestid=171940

EDIT:

2N4401 av's at about 40v

When looking for the avalanche threshold, you want a clean avalanche and not a lot of noise or high resistance.  I use the circuit in the paper above with a small cap for setting the pulse width, and slowly tunr up the supply until is free-runs, then look at the signal for high spikes on the scope.  If spikes are low or stretched out, or nothing but noise, then set it aside and try another one.  Attrition rate is about 10% or less.  You need a load (choke?) to limit the current when it conducts.

Quote from: Peterae on 2010.05.03, 17:55:10
Hang on a minute, if it's over volting the fets, how come only 1 blows, after all both fet drains are connected to the same end of the coil, it's nearly always the delayed fet that blows. ???

EDIT: I have had the diode blow across drain source as well, so what does all this mean, a very large back emf in the coil when i get the explosions?
but then why only the delayed diode as well, surely both diodes would burn, but the non delayed fet is off to protect the fet from a current surge.

My best guess: Because the collision is happening IN the delayed FET and/or diode. Your source of coil noise is sneaking into the switching device.

Hi
The diode in the conducting device may be at a slightly higher temp so will conduct but this is an assumption based on thermal DD logic problems. As  they are in similar ambient  conditions its a slight possibility.
There is a piece of info on wiki on avalanche transistors  that may be of interest although brief.
Cheers.

maybe it's the internal diode in the fet thats melting and making the fet appear faulty, if they are both on the same silicon this could explain whats happening.

Could i place a fast diode in series with the coil to stop back emf to the diode but still allow the fet to do it's switching job.

wouldn't the diode just avalanche and fail if the current is too high? 

...and why only the delayed FET?

Or.. use basic devices that have caractoristic curves instead of hard specs.

Peterae,

Are you able to see the return EMF(E1) resulting from the delayed pulse? That big relaxation pulse(E2) and the negative oriented pulses(E3) after it should be the result of the return after the delay.

To better explain the terms used in experiments I refer to the attached mark-up of one of your pictures....

E1 does most of the damage to SS devices. It is normally not visible. I know it takes a 1gHz scope to show much of it, even when the firing sequence is as low as a 3 mHz PRF. E1 doesn't give enough time for protection circuits, even super fast fuses, to function.
E2 can cause mechanical movement or destruction in windings and cores. E3 causes metals to hold a static charge and can continue for long periods of time, even while loaded electrically.

Notice the similarity? Now you see why I like to follow your tests  :o

Place a brass or copper rod in your coils in the same fashion as a coil core. Measure voltage from one end of that core to the other - after you connect a load to it. The voltage builds up slowly only when a load is applied. Make that core a ring. The ring should become hot.

This all sounds like we should throw transformer theory out the window but it all fits current theory.

The only advice I can offer to protect your FETs is to place a very tight spark gap across the source and drain. The spark gap should be a flat, broad surface on the source side. It should be a very sharp point on the drain side. ( EDIT>> Sorry, reverse that idea. You are using N-channel FETs).

If your damage is caused by an E1, the spark gap won't help much either.

WW,
If its possible to calculate the E1s period from the pulse width/coil  then a UHF tank may eliminate it but at what expense on the pulse itself?
Thoughts?

The E1, if it is an E1, is not TEMW. I don't normally say that. I usually just say they probably can't see it on a scope.

What if you crank the current way down? (the "effect is supposed to be voltage induced so minimal current is required)

What if you add a choke (ballast) after the FET?

Facinating  Wavewatcher,

Im not sure that shunting the tiny little spikes would be useful

I'll agree with G. IMO, current and magnetic field change isn't the culprit. You can certainly create a reasonable EMP with a few micro-amps if the voltage is high enough and pulse width in the tens of ns.

If the culprit is an E1 then X_L won't matter. An E1 is almost all charge. In that time-frame there is no X_L, yet.

Is the targeted result avalanche of the FET or the anomalous patterns emanating from the coil?

Is the FET using silicon based junctions? Silicon junctions can change from a charge carrier to a charge hold and release. Avalanche and Snapback are two different things.

I think what is going on must be determined before working on a solution. If the damaging energy is an E1, primary result pulse of an EMP, then there isn't much to do except socket the devices. There are other things to try but not knowing them is probably why we aren't all carrying holstered EMP weapons  :-X

You may want to determine if the invisible reverse pulse is a result of a silicon junction. Tie the gate to S or D (depending upon the device type) and feed it like a detector diode. If you get powerful but incredibly short reverse conduction then the activity seen is probably more to do with release of stored charge than avalanche.(not talking about capacitive discharge - this silicon problem can create double digit ps spikes)

If the problem is EMP related then reducing voltage will make it harder to create the noisy coil. Decreasing pulse rate or increasing pulse width is also counterproductive. Not needing a delayed secondary pulse at higher frequencies is probably because the pulses are so close together each following pulse is delayed in reference to the leading pulse.

The primary of a true EMP is a sudden burst of charge not a sudden change in a magnetic field. The magnetic change part is secondary and tertiary.  

A diode in series? Maybe, if the anode is tied to the FET source and cathode to V-. you may want to try something besides silicon  :) If an E1 is the problem then that diode will produce an even charge across the FET and the new diode should fail instead of the FET.

Yes, the freewheel diode shunted in the FET is a waste of time for incredibly short pulses. Why? Diodes aren't on by default (forward current). They turn on when a current is applied. If you want a diode to always be on then maybe a zero-bias type is a better idea. Anything less may not conduct until after the spike is done. Another reason why the FET and diode with running current flow are the only ones to fry? They are already on. The collision happens inside them instead of the coil when timing is just right/wrong  ;D

Ok. The beers are starting to take hold. I'll stop pummeling y'all with my hot wind.

with my own explosions at my HV transformer terminals, I had a bridge of four nte517 diodes that seem unaffected.

is it possible that the FET is acting as a antinode of reflected energy?  (like the open end of a transmission line)

Would adding a bleed resitor across the FET reduce or eliminate this antinode?

Wouldn't this also provide a little bias to the coil?

Peter,

I have what might be a couple silly questions. Would the explosions be occurring at the nodes or the peaks? Does the wire looked melted? Or does it look like someone stretched it and pulled it apart? Like a mechanical stress verses a heat stress.

Ken

Hi Ken
There does not seem to be any real heat i can detect, it's just a noise, no sign of deformation or stretching as such infact i cannot see any movement of the wire.

I am currently not sure why it occurs or what it's effect is, i am hoping to answer some of these questions as experiments unfold.
My first task is to try and understand what conditions are needed to cause them to appear, once i am able to produce them for longer periods without blowing my equipment up i may be in a position to try detecting what ever it is.
I am clinging to the notion that there maybe free energy available  ::)

Peterae,
A good notion to have but don't let the details get lost.

I'm convinced if there is extra energy it is in the DC component that follows the result pulse. I'm also convinced that the HV arcs demonstrated in the TPU were the audible 5k or 6k PRF and reverse polarity of the suposed usable DC output.

Peter, are your heatsinks for the FET's grounded?

hard to read your blurry schematic, is the coil across the FET?   If it is, try in series with FET.

Hi Grumpy
Here's a block schematic
There are fet drivers driving the fet's as well

Edit Heatsinks are not connected to anything but the Drain does push up against them un insulated

In the first controller i used sometime ago when i first discovered the explosions i didnt have heat sinks

Quote from: Peterae on 2010.05.05, 19:36:29
Hi Grumpy
Here's a block schematic
There are fet drivers driving the fet's as well

Edit Heatsinks are not connected to anything but the Drain does push up against them un insulated

In the first controller i used sometime ago when i first discovered the explosions i didnt have heat sinks

I am wondering about the parasitic capacitance of the heatsink...

the first controller didn't smoke all of the FET's...

Are the connections between the FET's and the bifilar coils the same length physically?

Quote from: Peterae on 2010.05.05, 19:36:29
Edit Heatsinks are not connected to anything but the Drain does push up against them un insulated

Uh Oh! Having that much aluminum surface area electrically connected to the FET junction is a problem when such short duration spikes are generated. Having the heatsinks tied to the low side of the rail or 'grounded' is my suggestion before doing anything else.

I hadnt even thought about the insulation on the heat sink, they are black anodized so i guess this does insulate them, with regard to me generating the sharp pulse they are as good as not using a heatsink at all so dont seem to affect it.
I better get some insulation to try.

The 2 Drain wires are the same length from each fet to the bottom end of the coil.

Hi,
194-760 @ RS components 1700 volt 12 amp with no reverse diode. Quite cheap maybe worth a look. Hope they are not too slow for the project.

Hi szaxx

I have not yet tried IGBT, i always thought they were slow, the specs of this one seem pretty good, my only concern is the off speed @ 220nS, but they are really cheap and worthy of a try when i next place an order.

Thanks
Peter

Peterae,

Here is something for interesting results. You only apply power to the valve heater element. You must experiment with the resistors and capacitor. Attach your scope probe across the anode/cathode. Drive it with an inverter or invert your current drive pulse. You can connect two in the same way you are using your FETs. If you can't see anything then apply a small amount of voltage to the coil in your normal way without an inverted drive pulse.

All,

No jokes, please. I've already heard them. If you haven't tried this, with interesting results, or a null-voltage audion then you can say you are a tube person but really aren't.

This circuit will not work with an 811 or 811A. It works best with Nuvistors like the 8056, 2CW4 and 6CW4.

Now that i could build.

Not knowing much about valves, so what would be the best device thats a) cheap and b) provides a bit of power.

I assume i could use variable resistors for the 2 resistors what sort of values would you suggest and what power rating are they, i could then use a fixed value capacitor (What Value and Volts rating)

The driving pulse is that TTL? so i could use an inverter stage as a buffer.


Cheers,
Peter

You could experiment with the resistors by using a single potentiometer. Size the ohms so you have the correct load applied to the driving gate.
The capacitor can start as .01mfd while using low frequencies. This will allow you to see how it runs.

The passive devices run the same no matter what switch is used.

The function of most tubes are very similar to FETs with the exception that:
      1. tubes are not limited to best performance at 'full-on'
      2. the source of anode current can be only the energy radiated by the heater (for a very few tube types).
      3. the grid not only turns flow on or off but can also accelerate.

As far as higher pulse energy goes.... use shorter pulses and much higher voltage and current. Tubes don't break-down as fast as FETs.

The needed amplitude of the signal applied to the grid will change with the voltage being applied to the load. i.e.: If the load curve must change then the drive curve may need a change. These aren't normally used as digital switches (but they were the first gated digital switches).

I can't claim to be a master at using tubes. Simply, there was nothing else available for me to experiment with when my experimenting began.

WaveWatcher,

Sounds like a tube would be preferable in some instances. Could the semi-conductor devices stop certain things from occurring, which might hinder a successful OU device?

I know where there are a bunch of tubes sitting on a garage. I guess I should go see what they are going to do with them.

If I do acquire these, I will be asking you many questions as you are the tube guru.
And now for the jokes.


8)

Heh?...

If I were a tube guru I wouldn't have blown so many into little pieces  :D

I just started my bench shenanigans with tubes. I've destroyed too many to claim any expertise. Best to just read up on example circuits and schematics of others.

This weekend, I hope to rotate a charge field in a stripped-down magnetron. Guess I'll have to make sure it is pointed away from me  ;D

Never had OU from a semi-conductor device or tube. Can't claim OU from anything else, either. As far as SS preventing OU... If OU is possible, I doubt it will be easier to do with SS switching.

Don't spend much money on old tubes unless you get the 'tube bug'. Grab a few of each general type for games, maybe. Look for direct heater types (no separate cathode) or sub-minies and Nuvistors.

Nuvistors triodes and diodes were common in the tuners and video circuits of 60's - 70's color TVs. They are very small and require almost no plate voltage. You can see a picture on Wiki.

Having fried all but two fets the temptation is growing as  I have a few VHF tubes/valves a 4CX250B forced air cooled and a QQZ04... instant on heater and others but reluctant to use them as they are for 2m and 70cm amps. But .... The 6JS6 pair can be used as  they are retired from poor performance at 30MHZ probably too slow for this experiment.
If you choose to use tubes/valves apply the heater voltage allowing the cathode to heat up before applying the HT or it will eventually be stripped and dont heat the cathode leaving it with no HT for too long or it may end up poisoned.  Just a  couple of tips for anyone not too familiar with them. Ive not used them specifically for over 30 years but they may have a use to start with. Now secondary emission..... Sounds kind of familiar now. I wonder if that comment has a REAL meaning, it sure looks that way?
Steve.

I wasn't going to spend any money. I was just going to go ask them if they wanted to give them away.  ;D

I recall working as a helper on tv's when I was about 12. The only thing I ever did was check all the tubes with a tester. Half the time that was all it would take to fix it. I think all of these tubes are either radio or tv.

I recall the tuner tubes, they were always the smallest.

8)

Quote from: szaxx on 2010.05.07, 23:40:23
Now secondary emission..... Sounds kind of familiar now. I wonder if that comment has a REAL meaning, it sure looks that way?

No more than what I typed.

Secondary emission is why I suggested using a positive pulse applied to the grid.

In my small circuit suggestion there is no HT (HV). The amplitude of the output can be controlled by the heater voltage. The output will be extremely small until the right pulse timing is applied. Be aware, my tests were with air-core toroidal windings.

The negative connection of the coil can be broken and voltage applied at that point if so desired. If done then a capacitance should be connected across the power supply to complete the coil path for the pulses.

Peterae,

Any hardware reason why I can't duplicate the function of you PIC function generator using an 18F4550 & 20mHz crystal?

I have a couple left over from a recent project and would like to investigate your results from your perspective.

Thanks,

WW

You wont be able to get 1nS res using software.

Just build my digital mono, and if you do want to let me know i will forward the pcb files onto you.
Or if enough people were interested i could get some pcb's made up and mail them.
but you still need a pic to automate the dip settings.

Edit that's a real nice chip, i have done some dev on it myself, using a bootloader and a usb lead.

you also would not get the crackles.

I have quiet a complex system going to create the crackles which has to be a certain setup.

What happens is as follows.

I use a 8 pin pic to drive 3 led drivers, these drive 4 x 7seg displays and 1 8 bit latch, these are all set serially from the 8 pin pic.
The 7 seg displays 0-255 nS delay and the 8 bit latch plugs in place of the phase bank of dip switches using ribbon cable.

Now when i do a scan i get pockets of instability where my pic starts crashing, but in truth it's not the pic thats crashing, it's the led drivers receiving duff bit's serially, this means the 8 bit latch is also receiving random numbers between 0-255 which also means that the second delay unit is randomly setting the phase to 0-255 nS, the first digital mono is set using dip switches and this one is probably rock solid and fixed and it doesn't really matter as it's delay setting is only relative to the signal generator, once this first channel fires the fet it also triggers the 2nd phase delay which is the random channel.
Only once this cycle of instability and crashing do i get the occasional crackle/explosion.

To me what this means is there is a certain set of circumstances that created the crackles but only once in a while are the random events correct to do this.

So for you to produce this effect you will need to build my setup exactly, pic dip switch driver board, 2 delay mono's and my main fet driver and fet board, only then can i guarantee you will get them, and even then it's hard work as you know.


I am Cautious about the valve driver that people are suggesting to me to build as this may stop the circumstances to produce the effect, but maybe also it could make it easier to produce, worth a try as long as i don't have to throw buckets of cash at the valve build.

Mean time i wait for my tube of FET's  :) and pray i can get it stable again to capture some data, hopefully with ou guys helping we can iron out what ever is creating the crackles and the conditions needed to produce them.

One thing also occurred to me, when i ran out of irf840's i switched to a 900v fet, i got the white noise but no crackles, maybe this fet being higher voltage was not quiet able to get the sharp dv/dt that is required to produce the crackles, as it's like a snap or collision of 2 very sharp pulses.

The white noise from the coils is an indication i am close to getting crackles, looking at the spectrum plot it's looks more like pink noise but i know nothing ;D

When i say white noise i am referring to my untrained ear, i am hearing the sea at the beach, a sort of SHHHHHHH
This is present when the crackles are not happening, which is what you see in the spectrum shot without the crackles appearing.

Now my theory goes like this, SM was injecting pink noise into his bifilar delay loudspeakers testing for spacial and response when he noticed something on his Spectrum, ie bit like i could see on my Spectrum bands of energy grouped but equally spaced across the spectrum.

As i understand it an audiophile well indeed use pink noise :)

Do you etch your own PCB's?

I do unless the design is too critical.

So, part of the steps needed to create the effect is random misfiring of inline circuits. It is not totally avalanche on the FET.
I had the same problem some time ago. I found the only way to duplicate with control was to use a Brownian noise generator with full adjustment of the high and low frequencies and the slope between them. This one also required an additional white noise generator signal. The spectrographic result was sinusodal clumping(repeated grouping) of both signals. I wasn't hearing explosions, rather a noisy tone. 

I knew the PIC 18F wasn't fast enough for pulse building but I have the other chips you have posted on earlier schematics. I don't have a display but figured I would just move that function to my bench PC through USB. (It is personally hardened for my experiments and I build software interfaces to PIC firmware, when needed.)

If so many exact circumstances are required even exact duplication of your devices may not allow the same results. A change in PCB material, solder, power supply may also may make it impossible. I have seen this before.

I need to finish bench cleanup from the previous disaster before proceeding. (I wiped out my isolation transformer and I'm not sure about the performance of my good scope  ;)  Not from OU research. Just trying to make someone's bad design work.)

 it doesn't always pay to help  ::)

Yes i make my own pcb's

If you were to build my setup i would use my pcb for the pic, it only uses 1 8 pin pic which is pennys, but you do need this board design to get the correct instability, it's key.

I have used different fet and fet drivers and pcb's and still see the crackles.

The digital mono is very fiddly to build and is double sided with a lot of small soldering, you would need 2 of these.

Originally i used an inverter/ buffer on each digital mono board and then used an inverting fet driver to correct the logic levels, i used this method to help avoid ground noise, but have since taken the buffer/inerter off and used non inverting fet drivers, in both cases i had crackles.

Quote from: Peterae on 2010.05.08, 19:59:43
when i ran out of irf840's i switched to a 900v fet, i got the white noise but no crackles, maybe this fet being higher voltage was not quiet able to get the sharp dv/dt that is required to produce the crackles, as it's like a snap or collision of 2 very sharp pulses.

Is either FET not a true MOSFET? True, meaning "Metal Oxide Substrate". Not all use metal oxide. In my experience, only the true MOSFET's are closest to operation and performance of a valve with a direct heater.

Don't ask me what that performance detail is. I still haven't figured it out.
-------------------

Just read your last response. Perhaps there is a chance I can replicate your results.
The smallest PIC I have is a 18F2550. I think it is 22-pin.
I'll see what I can find for the display and the correct PIC.

Hopefully, I'll be asking for your PCB layout and I can see what you are seeing.

Lightning Warning Signs

Sometimes lightning will give you a few seconds of warning before it strikes:

• Your hair will stand on end
• Your skin will tingle
• Light metal objects vibrate or buzz
• Metallic taste in your mouth
• Palms get sweaty
• Crackling or "kee kee" sound
• Smell of ozone (swimming pool smell)
• At night, a blue glow may show from an object

Grumps,

I have a little problem with your list of warning signs...

Your hair will stand on end - Which one?
Your skin will tingle - Your's doesn't tingle all the time?
Light metal objects vibrate or buzz - Can't tell. The plate in my head is making too much noise.
Metallic taste in your mouth - Hmmm... maybe I've blown-up to many eletrolytics?
Palms get sweaty - Uh Oh! I think I'm always in trouble.
Crackling or "kee kee" sound - How can I tell this sound from my wife complaining?
Smell of ozone (swimming pool smell) - I never smell this anymore but my wife complains I smell of ozone after a few hours on the bench.
At night, a blue glow may show from an object - An object? You mean like, not me? At night only?

No clouds out tonight but I still feel like I'm screwed  :o

BTW: How did you have time to take all these notes before the lightning struck you?

Quote from: WaveWatcher on 2010.05.12, 23:35:37
Grumps,

I have a little problem with your list of warning signs...

Your hair will stand on end - Which one?
Your skin will tingle - Your's doesn't tingle all the time?
Light metal objects vibrate or buzz - Can't tell. The plate in my head is making too much noise.
Metallic taste in your mouth - Hmmm... maybe I've blown-up to many eletrolytics?
Palms get sweaty - Uh Oh! I think I'm always in trouble.
Crackling or "kee kee" sound - How can I tell this sound from my wife complaining?
Smell of ozone (swimming pool smell) - I never smell this anymore but my wife complains I smell of ozone after a few hours on the bench.
At night, a blue glow may show from an object - An object? You mean like, not me? At night only?

No clouds out tonight but I still feel like I'm screwed  :o

BTW: How did you have time to take all these notes before the lightning struck you?

I found this list on the web and noticed similarities with Peter's explosions. Like I was saying before, the first pulse appears to set the parameters of the medium and the second pulse extracts the goods.  You may recall that SM stated that tuning directly on the magical frequency (or delay?) would convert so much energy that the TPU woudl be destroyed - he may have been referring to this explosion effect.

I agree G

I have not been able to produce the crackles as loud as i first found them, probably because the conditions are different with this setup, the great news is that probably means last time i didn't have the conditions right either just a bit better, so i wonder what will happen if i can tame the conditions.

Now for the good news i had my FET's arrive today, so hopefully i can try again tomorrow to capture the Spectrum shot.

It's also interesting that i only get the crackles using copious amounts of feedback, but the feedback is caused by the release of the energy effect itself, when i send the first pulse and then the 2nd i get random impulses back to the fet drivers which causes a mash of rf bursts for a short period, in which sometimes the crackles appears, the noise eventually dies and then comes along another master pulse to start the sequence of events over again, this is clearly shown in the explosions video.

The random bursts of noise last different amounts of time each time they occur, and sometimes the crackle appears quiet soon after the second pulse and sometimes much longer, at this point i am not even sure if i need 2 pulses, it maybe just the correct sequence to create the energy release, bit like playing a sequence of notes to form a song.

or

maybe a sequence of very sharp fast pulses happening in a random fashion occasional have the same effect as mixing 3 correctly phased frequency's that are at presently unknown, but instead of using frequency's as such i am producing the correct mark/space impulses to simulate the same thing as 3 frequency's in the coil or harmonics in the coil.

It may just be i have found a random way of finding the correct frequency's to mix to cause the effect SM found.

Just a thought...

'accelerometer'

...use one to sense/display the acoustic spectrum.

One thing I will be doing when I duplicate your explosions - analyze them with my vibration analysis equipment and software. With that I can decompile the FFT into the individual waveforms.

(The SM videos don't have clear enough sound on the videos - I tried a couple of years ago)

Quote from: WaveWatcher on 2010.05.14, 03:09:00
Just a thought...

'accelerometer'

...use one to sense/display the acoustic spectrum.

One thing I will be doing when I duplicate your explosions - analyze them with my vibration analysis equipment and software. With that I can decompile the FFT into the individual waveforms.

(The SM videos don't have clear enough sound on the videos - I tried a couple of years ago)

What are you hacking the Fourier with?

Equipment I use in my real life.

Some of my duties include vibration analysis, ultrasonic assessments, thermography, MCA (motor circuit analysis) and power quality analysis.

I am a WaveWatcher  ;D

Hi WW
Let me know when you are close to building the pic board, there is an error on the pcb and you will need to put a link on the back, i will also email you the software to burn into the pic.

I did think about correcting the pcb error, but i think it maybe a good idea to build it exactly as i have to be a close rep just incase it makes a difference, and it's only one link anyway.

Peter

Ok guys
A few interesting things, i have replaced both fets with the new IRF840A's and i have lengthened the wires to the coil and also the fet sources.
Anyway it seems to have worked i can now get crackles at about 50volts which seems to be a pretty safe voltage to make the fet's survive.

I have been having trouble in my first tests this morning to find a spot where i could maintain a delay position and keep the crackles coming so i can grab some shots, but during this process i made a discovery, at my current setup settings, without using the pic board to drive the 2nd channel(I now have 2 mono boards all set using dip switches)
When i flick the 2nd channel between a delay setting of 63nS and 31nS i get a single crackle every time i flick the dip switch.

When the dip switch is set for 63nS i get 2 pulses appear across the coil and when i set the dip switch to 31nS i have only 1 pulse.
So what i am saying is the sudden change between a delayed pulse of 63nS to 31nS causes a single crackle from the coil and i can produce it every time i flick the 6th bit on the dip switches.

So what now, i set out to find a way to produce the crackles on demand and i seem now to be able to create a single one on
demand  :)

Although i am not using my pic board for this test to drive the dip switches, i still have it powered up and every time i flick the dip switch to create the crackle it crashes  ;D

OK the 2 values i have been switching between are not unique i have found others, it seems important to go from a short delayed pulse to a longer delayed pulse very fast, i have seen other effects so far, including a sort of mains hum come from the coil instead of a crackle, this lasted only a couple of seconds, because i was randomly flicking switches i have only managed to do this a couple of times.
One other thing i tried was increasing the pulse width of the 2nd pulse and this nearly caused something to melt, i could hear a sort of mains hum from the coil and smelt something getting very hot very fast so quickly turned of the coil supply, the current went upto about 200mA's to the coil during this, luckily everything seems ok still.

When i flick from 63nS to 31nS my coil draws 25mA and 35mA in each switch position.

I need to hold myself back from flicking too much i dont want to kill this system as i did previously until i have some data stored.

Note to self never switch the signal generator on while the rest of the system is powered up, it just killed both fets and a hell of a noise came from the coil, i guess my signal generator tends to wind up in frequency when switched on and therefore must have done a frequency sweep until it stabilized.

The tests above were carried out at around 450khz.

Having thought about this for a while maybe i have this wrong, it's a click more than a crackle.
Maybe it is this that creates the white noise i hear which is more frequent in occurrence where as the crackle or explosion only happens once in a while.
I think it's much more likely that what i am hearing would cause the white noise within the coil system during random triggering of the fets, it's still interesting that i can crash the pic cpu when i flick the switches so there must be a sudden release of energy either way, just may mean the conditions have to be correct to release a lot more of it.

Is there any way you can place a radiation barrier between the PIC and the coil?

For the type I suggest using something that could stop weak beta. A sheet of thick plastic - aluminum foil... whatever. It does not need grounding but should be electrically isolated from the circuit of the circuit common points.
I'm not suggesting you have beta radiation.

Here's a sound file recorded with a mike attached to a magnet sitting on top of the coil, note using my ear i can only hear a click when i switch from one delay setting to the other, but the microphone picks up a sort of humming noise in one delay setting but nothing on the other, very strange indeed, i was hoping you could pick up the click as i flicked between the two.

Ahh blown it up again this time i switched on without my laptop connected to the serial port of the scope, when i switched on there was a loud high pitched noise from the coils and the fet's went instantly.  :-\

16 fet's left and counting DOH

Peter,

Maybe you need to start a new thread. How to blow fet's 25 different ways.  ;D

I have been wanting to ask if you can tell whats happening by the KABOOM.

8)

Quote from: Ken on 2010.05.15, 15:17:44
Peter,

Maybe you need to start a new thread. How to blow fet's 25 different ways.  ;D

I have been wanting to ask if you can tell whats happening by the KABOOM.

8)

You can't tell what is happening by the KABOOM but you can tell the exact amplitude  ;D That is, unless you are such a good KABOOMER you can differentiate between the flavors of KABOOM!

WATZAMATTA? Uz guys don't speak KABOOM?

Quote from: Peterae on 2010.05.15, 13:53:45
Here's a sound file recorded with a mike attached to a magnet sitting on top of the coil, note using my ear i can only hear a click when i switch from one delay setting to the other, but the microphone picks up a sort of humming noise in one delay setting but nothing on the other, very strange indeed, i was hoping you could pick up the click as i flicked between the two.

It sounds like you are coupling with a 50Hz power source and hearing the second and third harmonic.

Causing misfire in your power supply? Try battery power.

You've invented a random energy sucking antenna? Try reorienting the magnet/coil combination. See if you can locate the source. It may not be in the coil.

Maybe your 5U4GB rectifier in your UK Tele is gassy and the filament voltage is superimposed upon the output voltage  ;D

LMAO@WW

Quote from: Ken on 2010.05.15, 18:25:21
LMAO@WW

Yea, I hope to find out just how funny Peterae's experience is. I have the missing parts on order. I'm hoping for some serious bench time next weekend  :)

Sounds great WaveW cant wait to have someone else join in, hopefully not too many people as the price of fet's could go through the roof  ::)

I can swallow the dying fets a lot easier if i knew they were dying because of free energy  ;D

Here's an audio spectrum video of the above audio file.

QuoteMaybe your 5U4GB rectifier in your UK Tele is gassy and the filament voltage is superimposed upon the output voltage  Grin

Now come on WW i have invented a a random energy sucking antenna

Not a time machine  ;D

Speaking of the cost of fets.... i have not blown the new ones yet! they have lasted the longest...


" London calling ... KABOOM "

;D

lol@kaboom. You guys make me laugh.

8)

Quote from: WaveWatcher on 2010.05.15, 16:43:03

Maybe your 5U4GB rectifier in your UK Tele is gassy and the filament voltage is superimposed upon the output voltage  ;D

That would more likely be  the filament current  if Stevens clue is worth while there .

Fantastic thread !

Filament current aiding plate current? That is how it works, kinda.

Lindsay, please check your PMs.

Hang on with one fet setting i get nothing on the mike and on the other delay setting i get hum, surely i have just built an isolated hum switch not to mention a KABOOM switch, i think the KABOOM switch is marketable but not sure what to do with the warranty policy  ;D

Peter,

The warranty is easy. Just put one of those stickers across the on/off switch that says, "removal of this sticker will void all warranties." ;D


8)

Well the KABOOM of my fet's is being caused by feedback which i think has been pretty obvious, i started isolating things.
First Test was with the mike against the coil and the scope connected as in previous tests, now when i flick the delay setting i would get a click in the mike and some of this hum noise when the mike was near the coil, first i removed the magnet from the coil and the hum was still there and also it still clicked when i flicked the dip switches.
I tried moving the mike away from the coil and i still had the click and varying amounts of hum, so i powered down and disconnected the scope probe from the coil and disconnected the scope totally from the laptop that was monitoring the mike output.
I then powered up and found it no longer clicked or had the hum on the mike.

But the coil still clicked when i switched the dips, this was all on a lowish voltage.
Once i started winding the voltage up again with the scope disconnected and just the mike connected to the lap top, i reached a voltage where everything became unstable, my laptop mouse pad stopped working and i had a really loud noise coming from the coil, so i am still getting feedback but this time it's from the coil which is feeding back to the fet drivers or delay boards, if the voltage was higher then i probably would have had a KABOOM event but instead i managed to record the noise on the mike without damaging the fets.

Up to now i have always run above 5khz master frequency, but i have just discovered running under 5khz is progressively deadly and causes the coil to take large amounts of current, this explains why i got KABOOM's when i switched on my signal generator, it slowly wound up from below 5khz and as you will see from the following videos is lethal.



For this test i disconnected the scope totally from the system.

I found with a coil voltage as low as 1.5 volts with a magnet near the coil i could get loud noises from the coil
If the magnet is placed near a tail wire to the coil the wire moves around dependent on the phase delay setting, during this video i had the delay phase on auto sweep from 0-255nS and back round constantly.

After this video test i started altering the master drive frequency, and this effect only starts occurring at 5khz and below, when i start at 5khz it only makes the noises occasionally with certain delay settings, as the master frequency gets wound down the coil starts making progressively more noise in more delay settings until when i get real low in master frequency the coil is making the noises all the time.

No wonder i get random KABOOM's if i can draw current at 1.5volts coil supply god knows what sort of peak current i get at 50 volts supply under these random feedback conditions.

Things are turning really weird now.

Any way for you to limit 'ON' time to a very short period, regardless of frequency?

Bursts of <5kHz can happen during connection/disconnection of nearby devices. This may also trigger your pulse & delay boards.

Yes WaveW my fet output is always only 32nS which is set using the dip switches, there some lags in the fet driver and fet and am probably really getting about 65nS.

I am not fully sure whats going on here because if i really have 65nS pulses it's hard to imagine such high current draw.

Just as another test i disconnected 1 channel and drove the coil with only 1 fet, i get similar results under about 3khz where the coil tail vibrates, i can only imagine for some reason at this low frequency something is stopping my fet from switching off.

Just in case i have a damaged fet Channel i tried the other and still get the same effect

Strange if i use over 5.5khz there is no longer this noise from the coil unless i start raising the coil volts dramatically until i reach a point where my pic starts crashing and i get re triggering.

Test Coil Specs

Uses a UF4001 diode across it.

Inductance = 41.25uH
Resistance = 1.40 Ohms
Q = 1.8508
Z= 2.946
Tested @ 10kHz

If i connect a 120pF cap across either Drain/Sources the noise get's more prominent and occurs more easily.

Here's another video i did varying under 5 Khz

As you probably know, the wire movement within a strong magnetic field is not surprising. The audible sound from the coil, while in a strong magnetic field is for the same reasons, I'm sure. Placing a magnet nearby simply puts the coil in a more dense magnetic field.

The frequency point of 5.5kHz doesn't surprise me either.

On my 'magnet with coil tests' the interesting frequency was around 4.9kHz and didn't change much even when the coil inductance varied a great deal.

When the 4.9 was hit multiple audio frequencies emanated from the coil. All lower than 4.9kHz. The resultant, lower frequencies did change with coil parameters.

I'm thinking about resurrecting the test to see if the odd hidden sounds can be heard during playback. There were no KABOOMs, anomalous DC output, or extra pulses from those experiments that couldn't be easily explained.

A quick sim of your cct:

Vsupply = 50V. PW = 65ns, T = 130ns. With and without flyback diode. Shown is VDrain.

.99

Certainly shows how close i was sailing when running at 70'ish volts

Something's really screwed with my current setup, i thought i would run the system up and capture the pulse using 50 volts supply and a drive frequency below 450khz and above to show the difference, but as i swept using only 1 channel it blew the fet.

It shouldn't have done this as i have done it many times before without any problems, it's either the longer wires on the coil or maybe the fet drivers are not working correctly.

Enough for today, next week i will plug some new driver chips in and see if that cures these strange goings on.
I wouldn't have expected this or the way the system behaved under 5khz so something is not right.

A couple ideas that may help save the lives of a few MOSFETs.

.99

That is a great idea to limit the on time fantastic thanks. ;D

And one final touch....450V worth of TVS protection on the Drain.

If they still pop with these 3 safeguards in place, then it's time to look deeper.

.99

OK guys
I have almost finished the new controller board, but am using bigger fet's TO-249

Now WW aired concern i was not using insulation washers on my previous builds, so if the heatsink is isolated how important is it for me to use a Thermal washer, because this is holding me up, they are not that easy to get here in the UK unless i place a separate order costing pounds.

Peterae,
If the heat sink is isolated, it is not important at all.  It's all right too if your parreleling mosfets, it just connects all drains together which you do anyway.

The TVS's are a better move than electric isolation between the FET and heat-sink. Don't let the washers hold you up. If you still have problems see below:

Sorry to disagree Room3327. In high energy pulse circuits it is best to electrically isolate the heat sink from the device. This is standard practice even with high-end switch mode power supplies and every HV generator I've used. (the HV wasn't at the fet. It wasn't a matter of arcing at the FET part of the circuit).

On sub-ns pulses the wavefront induces a very short spike in nearby metals. Aluminum is really bad. One solution for me was to mount a small flat square of Al sheet on the PCB then mount the FET flat on the sheet. As long as the source (source of the spike - coil, wire, etc.) was on the same plane as the heat sink there was little problem.

Of course, your heat sink can be one-half of an unintentional capacitor  ;)

Thanks Room

I realize the heatsink would be live also, but was not sure how the increased surface area of the drain would alter the sharp nS pulsing.

EDIT WW i missed your post as it appeared just as i was posting, i do have some TO220 washers which although are not large enough to insulate the whole metal area, i am wondering they may  insulate enough and move the rest of the metal far enough to isolate them totally, it's not as if i have large potentials between the heatsink and the drain anyway so there should be no chance of discharge or spark jumping across the gap, and i also will not be running these very hot hopefully so i don't really need 100% surface area in contact

WW, Peterae,
   it's OK to disagree with me, I guess I stand corrected.  I probably should not have said it is not important at all, I have just never found much difference myself in testing and I thought it would certainly be alright for Peterae to test without insulation washers.  But I can see that what you say is good info.

I must apologize.

Making spikes of odd variety has been a passion for a long time. I should have explained that I believe Peterae's situation is a bit unique.

Yes, in almost all cases there is little worry about the electrical connection. It is also very rare that anyone need worry about such short rise and fall times. This project is no different except when the 'noise' is happening. Regardless of the cause of the noise, noise can contain some incredibly short pulses.

If the failure cause is avalanche in the failing device there isn't much to do except limit current.
If the failure cause is the reverse potential created by the collapse of the magnetic field of the coil (due to FET avalanche activity) there is a strong possibility the rise and fall time of the BEMF pulses are short enough to create effects closer to radiation than Faraday induction.

Believe it or not... I realized I was being a windbag and just deleted a huge amount from this post  ;D

Room & WW
Thanks guys it's always good to hear views, i am really glad to have you both here reading what i am doing, as a collective we stand much more chance in achieving our goals.

Peter

Just leave the heatsinks off and keep the current low.

I tried that G trouble is during re triggering you have no control how long it's on for and can get very hot very fast without a sink to disperse the heat.

I use a delay line to control the on-time and force it to turn off when the delay line reflects.   The trigger just dumps the delay line (a capacitor made of coax) through the switch.  I think 10 feet is about 3ns (as a course rule of thumb).  You could use a cap if you don't need the square pulse shape.

Maybe i was wrong in saying i cant control the ON time as with the next controller i have implemented Poynt's mod as above, so hopefully this will control the ON time.

We will see how it goes, but anyway it has a heatsink right now  ;)