SSTC build as per Steve Ward's Mini SSTC - Final Design

[Thunderstruck]

Hello everybody,

First I would like to mention that I am aware that I am starting a new thread straight after a thread based on the same design, which could be considered as Double Posting. So I leave it up to the admin to allow it or not.

Anyway, after successfully building a SRSG Coil a few years ago, I decided to have a go at building a SSTC because it will allow me to experiment without a noisy spark gap, also everything is smaller and easier to handle.

For my first attempt I chose Steve Ward's Mini SSTC Design, it looks simple enough and easy to understand.
My understanding of electronics is OK, should be sufficient to build this circuit, but please forgive me if I ask a "dumb" question ( I'm only dumb till I get an answer  )
Anyway, I built the circuit and powered up the low voltage side to see if anything will go pop, aaand to my delight, there was no pops !
I recently got myself a new digital scope to replace a very basic CRO which is at least 25 years old, so this is a perfect opportunity to put it to a good use so I had a poke around to see how things are buzzing.
There are a few things that I do not make sense, perhaps of my lack of understanding, perhaps there is a fault somewhere, so I decided not to give it a full test until I get some clarification.
I hope somebody will be able to help me out with this. I used this fantastic feature on my new scope to capture a few screen shots ( what a great feature !!! )

Image 1 shows output of the 555 timer, looks ok, but it gets bit wonky at higher frequency
Image 2 shows UCC37322 output, so far so good
Image 3 shows UCC37321 output - that looks weird, tried another ic, same result. I would like your opinion on this one
Image 4 shows input signal on Schmitt trigger - from my ancient function generator
Image 5 shows output signal from the Schmitt - it is Schmitting quite nicely I think
Image 6 shows output from the UCC37322 with Schmitt only connected 555 has been disconected
Image 7 shows output from the UCC37321 with Schmitt only connected 555 has been disconected - Why is this not inverted signal ? I was
               expecting it to be...
Image 8 shows output on the UCC37321 / 22 with the 555 and Schmitt connected - What a mess... What is going on here ??

GDT is built and it works quite well, I tested it as per the instructions on another page ( with a function generator and a scope ).
I dare not plugging this thing in, because I reckon something will burn unless I clear this up.
Any help will be appreciated.


 

[Fumeaux]

(*) I don't think your 555 timer output is working as intended. It should output something in the audible frequencies, not higher than 500Hz. Are you sure you have a cap where it states cap sub point? I'm not sure what it mean but look at this reference (http://kaizerpowerelectronics.dk/wp-content/gallery/2009_01_22_-_kaizer_SSTC_I/KaizerSSTCIdriverschematic.gif)
(*) The other thing is that the UCC's don't really like to be operated without any load. You don't have to use the GDT, as a high power resistor would be sufficient to provide a load (pls check with ohms law). If you use your GDT make sure to feed a "fake" feedback signal as the inductor would act like an open curcuit with DC (the UCC's don't like this too).

The first point should be able to fix your "UCC with the 555 and Schmitt connected" waveform.

I hope this helped, if you have more questions for me send a PM so I see it.

[Thunderstruck]

Hi Fumeaux,

Ok, I will look at my 555 circuit, yes there is a cap on that sub point and a resistor on the resistor sub point. I connected those two with screw down terminals, because I want to be able to take them out easy in order to try different options - apparently spark characteristics change depending on values of that cap and resistor.
Honestly I can’t se how becuse 555 is there just to start the oscillations, after that the coil does it’s own thing so it will produce a spark depending on it’s own characteristics, regardless of a cap and a res in 555 circuit.
GDT is connected to the UCC’s so they are not running completely load free. I read somewhere that UCC’s can burn out without a load, so I had a GDT connected before I powered up the circuit.
But, I did not know that there should be a feedback from GDT. Should I just connect my signal generator to one of the windings and give it a square wave signal ?

[Mads Barnkob]

Hi Thunderstruck and welcome to HVF!

It is good that you make a new thread, instead of high jacking another thread. While its the same build, its most likely not the same problem, also who wants 100 pages long threads to look for answers in when searching from google

There is nothing wrong with the 555 running at up to the resonant frequency, that is just running it in CW mode, which would be better to implement by tying UCCs ENBL to on all the time instead, as the 555 properly cant handle the resonant frequency well enough over such a wide span. The 555 interrupter is only turning ENBL on/off on the UCCs to interrupt the CW mode and spare the MOSFETs from heavy heat dissipation due to conduction losses.

Fumeaux raises a few good points about UCCs destroying themselves if used without a load, they can simply oscillate into MHz region with their own output capacitance/inductance and burn to the ground.

You can just hook it all up driver -> GDT -> MOSFET gates, without mains power, then you have the UCCs running under right conditions and use a signal generator to simulate the antenna feedback. If you hook it all up with secondary coil in place and such, the driving energy from the GDT to MOSFETs can actually supply enough current to "leakage" drive the resonator as I demonstrate here:

/>

[Fumeaux]

My point was that if you have the working frequency too close to the interrupter frequency you won't get a clear viewing on the osci, as the output is too unpredictable and not periodically.

[Thunderstruck]

Hi Mads,

Thanks for your comment, what would be an ideal frequency for the 555 ?
I noticed that 555 is connected to the enable pin of the UCC’s and it did cross my mind that it will switch them on and off at a given frequency but I could not understand why ( I mentioned that im OK with electronics, there is lot to learn )
From your comment I gather that is to allow the transistors to “take a break” so they do not overheat.
Thinking along those lines, there must be a frequency that is optimal for that job ?

[Fumeaux]

There is no ideal frequency, but the problem is, that when the signal frequency (signal generator) and the interrupter frequency (555) are too close together you won't be able to see either on an oscilloscope. The signal frequency is typically between 200kHz to 600kHz (for sstc) and the interrupter/enable pin is there to well interrupt the signal. Typically the interrupter goes only up to 300Hz.

And the funky waveform you see comes from the fact, that the signal and interrupter frequencies are not matched and the interrupter can turn on/off mid pulse, resulting in the chopped up squarewave. This not the big of a deal if the interrupter does that every now and then (300Hz), but at 24kHz it does the copping pretty much every cycle => funky waveform | slower interrupter => better/normal waveform

In my video in the background you can see the waveform on my oscilloscope. There is the ontime with very fast switching (for me 130kHz) and an off phase. This on off is around 40Hz.

PS: I swear I didn't find a scope shot of an interrupter. This is not just self promotion.

[Thunderstruck]

Ok, I’m starting to understand what the problem is, my 555 is going too fast.
So, jobs to do :

- Make 555 output about 300 Hz
- Feed between 200 to 600 kHz into the antenna
- Check output on UCC’s

I thought that the 555 becomes irrelevant once the coil starts, but now it all makes sense.
Things you learn... 

[Thunderstruck]

As planned I carried out some tests, and I think I made some progress.

Image 1   I slowed down the 555 to around 300 Hz
Image 2   Connected 350 kHz signal to the antenna
Image 3   Schmitt output at 350 kHz - looks a bit messy
Image 4  UCCxxx22 output - also messy
Image 5  UCCxxx21 output - messy, but a LOT cleaner than before
Image 6  Both UCC's together ( xxx22 CH1, xxx21 CH2 ) apart from that weird waveform, I am really pleased to see that now I can see inverted signal on UCC xxx21

Next job is to figure out the reason for that bad signal form on the  UCC's.
I disconnected the 555 and fed exactly the same signal from a generator, UCC output did not improve.

[Mads Barnkob]

How sure are you about your GDT being suitable? To me it looks most like the first case here in Richies GDT troubleshooting: http://www.richieburnett.co.uk/temp/gdt/gdt2.html

[Fumeaux]

Looks good. Currently you dont have a 50/50 pwm but if you have the antenna connected and the coil running it should solve itself. For now it could help to increase the voltage output of your signal generator if possible, if not it's not a big deal either.

The second thing is the bad waveform on your output. It may be due to a unsuitalbe core, as Mads pointed out, or, and something similar happend to me, it could be due to too little capacitance on the 12V rail, which causes the voltage to dip too low during switching of the UCC's. Try to find something between 100-1000uF and put it between 12V and GND. [Scope shots of the 12V rail before and after pls]

PS: Can you post a photo of your gdt for "remote diagnosis"?

[Thunderstruck]

Mads, you had exactly the same link posted before on another thread, it looked interesting so I connected my signal generator to one coil of my GDT and scoped the other two, they looked quote good which makes me think that my GDT probably works ok. My opinion might be wrong.

Waveforms you see are directly from the outputs of the UCC’s, so signal going into the GDT is wrong, I did not even bother checking the output - it is probably tragic.
I would like to get a nice clean signal from the UCC’s before I move onto the next component, otherwise I will get lost...

[Thunderstruck]

I used this core:
https://www.jaycar.com.au/l15-35x21x13mm-toroid-or-ring-cores-pk-2/p/LO1238

[Mads Barnkob]

That is a L8 material ferrite/iron powder ring core, that material is not suitable above 100 kHz, as you can see here its initial permeability start dropping hard after 300 kHz: https://www.forward.com.au/pfod/HomeAutomation/BLEHighPowerLightSwitch/M7_HY2_ferrites.pdf

Get one with a Ui/AL value around 4000-5000, follow this excellent guide: http://thedatastream.4hv.org/gdt_parts.htm

 

[Thunderstruck]

I guess I was wrong regarding my GDT. I just ordered a couple of these as per your advice:
https://au.rs-online.com/web/p/ferrite-rings/2120910/?relevancy-data=636F3D3126696E3D4931384E525353746F636B4E756D626572266C753D656E266D6D3D6D61746368616C6C26706D3D5E2828282872737C5253295B205D3F293F285C647B337D5B5C2D5C735D3F5C647B332C347D5B705061415D3F29297C283235285C647B387D7C5C647B317D5C2D5C647B377D2929292426706F3D3126736E3D592673723D2673743D52535F53544F434B5F4E554D4245522677633D4E4F4E45267573743D3231322D30393130267374613D3231323039313026&searchHistory=%7B%22enabled%22%3Atrue%7D

[Thunderstruck]

Drying Secondary is taking up space on my bench, so I can't work on the circuit until it completely dries...

[Mads Barnkob]

That is a fine looking secondary, how many layers of varnish will you end up giving it? And what type did you use?

[Thunderstruck]

For this coil I used Polyurethane varnish, I’m already on my 4th layer.
Usually I use Ultimeg varnish - designed for the electrical industry to resist high voltages.
I do not expect a huge output on this coil so I did not want to use expensive material.
Secondary on my SRSG TC is done with Ultimeg and it works really well.

[Thunderstruck]

As Fumeaux suggested, I took some measurements of voltages at different points.
Just remember that I am still learning to use my new digital scope, apologies if something does not make sense.
From what I see and understand there is a bad noise problem, comments are appreciated.

Image 1 - Output of LM7812
Image 2 - Output of LM7805
Image 3 - Voltage at 555 ( trace is a bit dim, I later realised that I can make it brighter )
Image 4 - Voltage at UCCXXX22
Image 5 - Voltage at UCCXXX21
Image 6 - Voltage at SCH Inverter

[Fumeaux]

That looks quite good. Is it while having a gdt connected and a generated signal for the feedback? If so, there shouldn't be a problem. Maybe add three 100nF ceramic disc capacitor on the lm7812 input, output and the lm7805 output for filtering the sharp spikes.

The reason I said it was because I had 2 Volts of ripple on my 12V line. That was at ~130kHz, and the higher the frequency the more often the ucc's switch -> more drop. But I guess you are fine.