UK Vintage Radio Repair and Restoration Discussion Forum - View Single Post - Design-and-building a large valve Tesla coil (class C, Armstrong oscillator).

Al (astral highway) · 8th Nov 2017, 3:32 pm

Thank you, Dave!

Progress. I've been using my modulator to analyse waveforms in the grid feedback coil. This is usually a huge guessing game for people who power up at EHT without knowing what's really going on in the feedback or 'tickler' winding. It can put the power oscillator valve under unusual and indeterminate conditions.

It seems crazy to me, now that I have made the same mistake in the past.

I have clearly been able to demonstrate that loading the grid feedback coil also substantially loads the primary. Current consumption in my MOSFET modulator increases vastly at some frequencies, showing that coupling is rather close.

The photos keep attaching in a random order, how ever many times I try! So please excuse the confusing numbering, this is how they uploaded!

2) Blue trace x 10V shows the Grid feedback coil waveform with RC components (variable 20K resistor and 680pF 2KV capacitor) at 205KHz. Note the extreme symmetry of the kink in the blue trace and the kink in the yellow trace. The waveform is nearly sinusoidal at resonance.

4) Same as above at 376KHz, RC components only, no damper or 'recovery' diode. Current in the modulator rose to 140mA from 40mA as I swept up frequency.

1) Using a flyback damper diode. My idea, in this application! I found this paper by Dr Hugo Holden, who happens to be forum member Argus25! It is relevant and extremely detailed. The diode is in series with the RC components and is a Schottky diode. It does seem to have a pronounced effect on the waveform here.

(If I reduce the output voltage of the modulator, the ringing in the waveform disappears) The top of the wave becomes almost flat, like a square wave, and its shoulders are also squarer.

The waveform goes negative with a slope that much more closely matches the falling edge of the MOSFET drain square wave - way faster than the sinusoidal decay without it. Because of the RC components, it starts oscillating again when the MOSFET is still switched off, and continues to resonate until the end of the next cycle. That ringing is really very pronounced, right?

3) Same as above, flyback damper diode, but at 376KHz.

In other news, I've just heard back from Richard (Mole42UK) who has done a superb job of lathe-winding the secondary inductor. I can't wait to see it in person. Thank you, Richard!

I haven't heard back from the other Richard yet. Ed Dinning, our resident transformer expert, is also on the case to price up a bespoke wound transformer with various taps. Thanks also to Ed!

I'd be interested to hear comments on these waveforms and my observations.

Thanks folks!

8th Nov 2017, 3:32 pm	#110
Al (astral highway) Dekatron Join Date: Nov 2006 Location: London, UK. Posts: 3,496	Re: Design-and-building a large valve Tesla coil (class C, Armstrong oscillator). Thank you, Dave! Progress. I've been using my modulator to analyse waveforms in the grid feedback coil. This is usually a huge guessing game for people who power up at EHT without knowing what's really going on in the feedback or 'tickler' winding. It can put the power oscillator valve under unusual and indeterminate conditions. It seems crazy to me, now that I have made the same mistake in the past. I have clearly been able to demonstrate that loading the grid feedback coil also substantially loads the primary. Current consumption in my MOSFET modulator increases vastly at some frequencies, showing that coupling is rather close. The photos keep attaching in a random order, how ever many times I try! So please excuse the confusing numbering, this is how they uploaded! 2) Blue trace x 10V shows the Grid feedback coil waveform with RC components (variable 20K resistor and 680pF 2KV capacitor) at 205KHz. Note the extreme symmetry of the kink in the blue trace and the kink in the yellow trace. The waveform is nearly sinusoidal at resonance. 4) Same as above at 376KHz, RC components only, no damper or 'recovery' diode. Current in the modulator rose to 140mA from 40mA as I swept up frequency. 1) Using a flyback damper diode. My idea, in this application! I found this paper by Dr Hugo Holden, who happens to be forum member Argus25! It is relevant and extremely detailed. The diode is in series with the RC components and is a Schottky diode. It does seem to have a pronounced effect on the waveform here. (If I reduce the output voltage of the modulator, the ringing in the waveform disappears) The top of the wave becomes almost flat, like a square wave, and its shoulders are also squarer. The waveform goes negative with a slope that much more closely matches the falling edge of the MOSFET drain square wave - way faster than the sinusoidal decay without it. Because of the RC components, it starts oscillating again when the MOSFET is still switched off, and continues to resonate until the end of the next cycle. That ringing is really very pronounced, right? 3) Same as above, flyback damper diode, but at 376KHz. In other news, I've just heard back from Richard (Mole42UK) who has done a superb job of lathe-winding the secondary inductor. I can't wait to see it in person. Thank you, Richard! I haven't heard back from the other Richard yet. Ed Dinning, our resident transformer expert, is also on the case to price up a bespoke wound transformer with various taps. Thanks also to Ed! I'd be interested to hear comments on these waveforms and my observations. Thanks folks! Attached Thumbnails __________________ Al Last edited by Al (astral highway); 8th Nov 2017 at 3:50 pm. Reason: Confusing order of photos!