Well.... that certainly is confusing. Maybe the coil should be called the "can of worms" coil instead.

It's no wonder that people are having trouble with the various techniques for finding resonance of that coil. Particularly if some of the coil is wound in the opposite direction from other parts of the coil.

Well.... that certainly is confusing. Maybe the coil should be called the "can of worms" coil instead.

It's no wonder that people are having trouble with the various techniques for finding resonance of that coil. Particularly if some of the coil is wound in the opposite direction from other parts of the coil.

Multiple series connected windings having partial magnetic flux overlap with interdependent transmission line bandwidth stability.  Not your typical Electronics 101 project by any means.  Herding cats would be far easier.

Seriously TK, you should jump into the fray.  In the end it's only time and money, something everyone will have give up.

... like this???    https://www.youtube.com/watch?v=qC2fKEl12i0

Except, we could do without:

... a project that may or may not work and if it doesn't work, you might get to see this video

I kind of completed the PLL part on the breadboard and made some measurements.

As i have not installed the Kacher MOSFET (ended at the UCC27321 driver) and the IR2184 driver, i simulated the feedback from the yoke
(via the current transformer) with my FG which i put at 15Khz (that seems to be the new inductor (bifilar coil) resonance frequency).

I setup the various pots/caps for a frequency locking range between 9 and 21 KHz so with the 15KHz in the middle.
C5 is 320pF (220 + 100), C2 is 6nF

The screenshot shows:

Yellow trace is output pin 4 (74HC4046)
Blue trace is   comparator pin 3 (74HC4046)
purple trace is feedback pin 14 (74HC4046)
green trace is the output pins 6/7 of the UCC27321
All referenced to ground

Video here: https://www.youtube.com/watch?v=pnEXk7qnRKE&feature=youtu.be


Regards Itsu

Video here: https://www.youtube.com/watch?v=pnEXk7qnRKE&feature=youtu.be

Excellent work and video demonstration Itsu.  I only wish you were the one to develop this device originally instead of Ruslan.  If that were the case we would all have this thing replicated and probably highly optimized by now.


Something I mentioned quite a while back on this thread.  I purchased a pair of 40Hz to 200kHz inductive pickup transformers to use specifically for monitoring the current in the induction heater coil.  These devices work extremely well and would interface with the LM339 just fine without any fear of overvolting the comparator chip.


On the behavior of your bread-boarded circuit...   This particular circuit does not burst the Tesla coil.  It only outputs a single impulse at the proper position within the lower frequency signal.  It would seem to me, this is a relatively decent starting point.  However, I suspect with a burst output of controlled width having the exact frequency resonant to the Tesla coil, superior performance could be achieved.  Someone mentioned Ruslan had improved his circuit to do this.  I have also done it using a micro-controller implementation.  After getting my grenade coil re-wound for the fourth time, I will make an attempt to demonstrate this as a comparison.

Excellent work and video demonstration Itsu.  I only wish you were the one to develop this device originally instead of Ruslan.  If that were the case we would all have this thing replicated and probably highly optimized by now.


Something I mentioned quite a while back on this thread.  I purchased a pair of 40Hz to 200kHz inductive pickup transformers to use specifically for monitoring the current in the induction heater coil.  These devices work extremely well and would interface with the LM339 just fine without any fear of overvolting the comparator chip.


On the behavior of your bread-boarded circuit...   This particular circuit does not burst the Tesla coil.  It only outputs a single impulse at the proper position within the lower frequency signal.  It would seem to me, this is a relatively decent starting point.  However, I suspect with a burst output of controlled width having the exact frequency resonant to the Tesla coil, superior performance could be achieved.  Someone mentioned Ruslan had improved his circuit to do this.  I have also done it using a micro-controller implementation.  After getting my grenade coil re-wound for the fourth time, I will make an attempt to demonstrate this as a comparison.

Good day Dog-One

OK, I been pondering this bit for a while now............. 
Yes, please note that the VCO is running @ 15kHz in the O-scope shot from Itsu.
And, the *nano-second* pulse (Green waveform) is ALSO firing @ 15kHz.

We know that the Karcher is *tuned* to approx. 1.2MHz (2.4MHz?), right?

So, when the FET fires @ 15kHz (or what ever the VCO/PLL is 'Locked' to) the Karcher will *RING* @ it's Fres (Resonant Frequency). 

Please note that the mosfet only has to fire once/cycle to *RING* the secondary of the Karcher and thereby the extra coil (antenna) concurrently.  See TK's post above regarding 'Single-Shot sweep mode'.  In principle this is exactly the same animal.

This is analogous to a Bell being *hit* with a piece of pipe.................. now matter how fast you can hit the bell, you will never hit it in tune with it's Fres.
You will only be repeatedly striking it *below* it's Fres (unless you are capable of striking it many hundreds or thousands of times/second). 

Of course the sound you hear is the sound emanating from the bell which reverberates @ it's Fres. This is the natural resonant frequency of the bell, which give each bell it's unique tone/sound.

I believe this is what happenings with the Karcher, because we see and know that the nano-pulse circuit from Ruslan does *not* drive the mosfet @ 1.2MHz.
It only repeatedly fires the mosfet at the frequency of the PLL locked signal which happens to be the same signal which drives the Yoke core push-pull XFRMR (24kHz??).

So what happens in the space/time of that nano-second pulse width?

If we know the marktime of the pulse width of the Green wave form we can then calculate how many 1.2MHz pulses can occupy this space in time.

This will be your MHz pulse train, not a single pulse, but a train of pulses at the Fres of the Karcher which fire the packet of MHz pulses in time (once/cycle) with the VCO signal of the 4046CD.

This is precisely why the nano-pulse part of the circuit from Ruslan is *adjustable*, the pulse width and phase are variable.

By adjusting the pulse width on this part of the nano-pulse circuit, one can *choose* the number of 1.2MHz pulses that are in the pulse pack (number limited/influenced by ringing decay/damping rate, ie; coil Q, coupling, etc.).

By adjusting the phase control, the 'packet of MHz pulses' can be positioned anywhere along rising/falling/apex of the input signal waveform.

Please let me know your thoughts on this.

take care, peace
lost_bro

Ringing like a bell, a spring or a tuning fork.  It is my belief the Tesla coil rings down with a single impulse as shown on many scope shots.  However, if we strike the Tesla coil with a tuned burst it will instead ring up, increasing in amplitude at each high frequency cycle.  So we start the burst a bit early and continue it right through the peak of the low frequency signal.  This should give us the maximum amplitude at the antenna achieving the maximum effect we are looking for.  Back earlier in this thread I posted a scope shot of what I named the "sperm wave", which I'll post again for reference.   Though the frequency is way low, you'll get the idea--the peak amplitude it considerably larger than the initial one.  The peak can be phase shifted and the number of impulses per burst can be configured to maximize the performance of the Tesla coil for this application.  Just a note about this scope shot:  It was derived with relatively primitive ICs.  I have improved it greatly using a micro-controller, ensuring each impulse is precisely timed (unlike in the scope shot where the first impulse is obviously too short, creating unwanted harmonics).

Good morning Dog-One

I do agree that a properly implemented micro controller would sure be convenient way to go if it could be made to work.
I can get into the MHz range using an AD9850 dds controlled by an Arduino Uno.  It's a pretty clean signal and pretty accurate also.
Haven't really put much thought into the actual interfacing with this particular circuit.
I remember now that you are not using Arduino but some other type of micro-controller.  I am interested in your solution for this application.
Are you thinking of  interfacing your micro with a discrete PLL chip? Or going to run complete logic control/signal generation via the micro?

take care, peace
lost_bro