SSTC won't self-start

[ResonantEMF]

I've built an SSTC using the circuit from Labcoatz's video. The way I built it is pretty much exactly the same but with irfp460 MOSFETs. And it works perfectly, I'm getting 6 inch arcs at 170v. But I need to hold my little slayer exciter wand near it to get it to start oscillating... I don't have much room in the base for a baby SSTC to get this running, so what can I do to troubleshoot this?
The schmitt triggers, driver, and interrupter are all healthy.

Another main difference is that I got impatient with the whole ground plane thing and just connected the base of the secondary directly to the antenna. I theorize that this might be the issue. I wonder if inverting the schmitt trigger might solve this and let me keep getting away with the direct secondary feedback. What do you think?

[davekni]

The way I built it is pretty much exactly the same but with irfp460 MOSFETs.

With FETs instead of IGBTs, I'd recommend using a 1:1 GDT.  Can be made with much lower leakage inductance (by winding with two twisted pairs and paralleling one wire of each pair for primary).  Also reduces load on driver chip.

Another main difference is that I got impatient with the whole ground plane thing and just connected the base of the secondary directly to the antenna. I theorize that this might be the issue. I wonder if inverting the schmitt trigger might solve this and let me keep getting away with the direct secondary feedback. What do you think?

Feedback from secondary bottom is used by other coils (including my SSTC).  Secondary peak current is likely to be 1-2A.  A bit high for 1N4148 diodes even at low duty cycles.  Also can cause an issue of over-powering +5V supply.  Diodes feed current to +5V, more than circuitry consumes.  All designs I've seen using secondary feedback either use a current transformer (CT) or use high-power zener or TVS diodes to clamp voltage rather than clamping to +5V.  (Additional lower-power clamp to +5V may exist after series resistor limits current.)  If you decide to use a current transformer, clamping needs to be symmetric around 0V and/or AC-coupled before clamp.  Otherwise DC component of voltage will saturate CT core.

Good grounding is still necessary.  With secondary bottom feedback, grounding of logic supply negative rail becomes the most important.  At least connected to line ground.  Counterpoise is better.

Yes, polarity may need inverting when changing from antenna to secondary feedback.  This can be done by swapping GDT primary wires or SSTC primary wires or by adding an inverter stage.  You'll just need to try both polarities to see which works.

Adding bleed resistors across 0.82uF caps or across IGBTs can help with startup.  Makes sure that initial state of half-bridge output is centered so that initial half-cycle will generate some feedback.  A resistor across only one side can be better or worse depending on which IGBT turns on first, which is nondeterministic with this circuit.  My favorite fix is to add a resistor from HC14 pin 2 to far side of 1k resistor (junction of 1k and 0.1uF).  That keeps the HC14 toggling even when coil is off (between interrupter pulses).  There are some other threads here discussing this change.  Details for your case depend on what other changes you make first.

Good luck!

[ResonantEMF]

The way I built it is pretty much exactly the same but with irfp460 MOSFETs.

With FETs instead of IGBTs, I'd recommend using a 1:1 GDT.  Can be made with much lower leakage inductance (by winding with two twisted pairs and paralleling one wire of each pair for primary).  Also reduces load on driver chip.

Another main difference is that I got impatient with the whole ground plane thing and just connected the base of the secondary directly to the antenna. I theorize that this might be the issue. I wonder if inverting the schmitt trigger might solve this and let me keep getting away with the direct secondary feedback. What do you think?

Feedback from secondary bottom is used by other coils (including my SSTC).  Secondary peak current is likely to be 1-2A.  A bit high for 1N4148 diodes even at low duty cycles.  Also can cause an issue of over-powering +5V supply.  Diodes feed current to +5V, more than circuitry consumes.  All designs I've seen using secondary feedback either use a current transformer (CT) or use high-power zener or TVS diodes to clamp voltage rather than clamping to +5V.  (Additional lower-power clamp to +5V may exist after series resistor limits current.)  If you decide to use a current transformer, clamping needs to be symmetric around 0V and/or AC-coupled before clamp.  Otherwise DC component of voltage will saturate CT core.

Good grounding is still necessary.  With secondary bottom feedback, grounding of logic supply negative rail becomes the most important.  At least connected to line ground.  Counterpoise is better.

Yes, polarity may need inverting when changing from antenna to secondary feedback.  This can be done by swapping GDT primary wires or SSTC primary wires or by adding an inverter stage.  You'll just need to try both polarities to see which works.

Adding bleed resistors across 0.82uF caps or across IGBTs can help with startup.  Makes sure that initial state of half-bridge output is centered so that initial half-cycle will generate some feedback.  A resistor across only one side can be better or worse depending on which IGBT turns on first, which is nondeterministic with this circuit.  My favorite fix is to add a resistor from HC14 pin 2 to far side of 1k resistor (junction of 1k and 0.1uF).  That keeps the HC14 toggling even when coil is off (between interrupter pulses).  There are some other threads here discussing this change.  Details for your case depend on what other changes you make first.

Good luck!

Wow, this all sounds like great advice. I can see why adding that resistor to the HC14's pin 2 is your favorite, it sounds brilliant and I don't know why I didn't think of that myself. I'll give all that a try, thanks.

[MarcusFunt]

The way I built it is pretty much exactly the same but with irfp460 MOSFETs.

With FETs instead of IGBTs, I'd recommend using a 1:1 GDT.  Can be made with much lower leakage inductance (by winding with two twisted pairs and paralleling one wire of each pair for primary).  Also reduces load on driver chip.

Another main difference is that I got impatient with the whole ground plane thing and just connected the base of the secondary directly to the antenna. I theorize that this might be the issue. I wonder if inverting the schmitt trigger might solve this and let me keep getting away with the direct secondary feedback. What do you think?

Feedback from secondary bottom is used by other coils (including my SSTC).  Secondary peak current is likely to be 1-2A.  A bit high for 1N4148 diodes even at low duty cycles.  Also can cause an issue of over-powering +5V supply.  Diodes feed current to +5V, more than circuitry consumes.  All designs I've seen using secondary feedback either use a current transformer (CT) or use high-power zener or TVS diodes to clamp voltage rather than clamping to +5V.  (Additional lower-power clamp to +5V may exist after series resistor limits current.)  If you decide to use a current transformer, clamping needs to be symmetric around 0V and/or AC-coupled before clamp.  Otherwise DC component of voltage will saturate CT core.

Good grounding is still necessary.  With secondary bottom feedback, grounding of logic supply negative rail becomes the most important.  At least connected to line ground.  Counterpoise is better.

Yes, polarity may need inverting when changing from antenna to secondary feedback.  This can be done by swapping GDT primary wires or SSTC primary wires or by adding an inverter stage.  You'll just need to try both polarities to see which works.

Adding bleed resistors across 0.82uF caps or across IGBTs can help with startup.  Makes sure that initial state of half-bridge output is centered so that initial half-cycle will generate some feedback.  A resistor across only one side can be better or worse depending on which IGBT turns on first, which is nondeterministic with this circuit.  My favorite fix is to add a resistor from HC14 pin 2 to far side of 1k resistor (junction of 1k and 0.1uF).  That keeps the HC14 toggling even when coil is off (between interrupter pulses).  There are some other threads here discussing this change.  Details for your case depend on what other changes you make first.

Good luck!

Wow, this all sounds like great advice. I can see why adding that resistor to the HC14's pin 2 is your favorite, it sounds brilliant and I don't know why I didn't think of that myself. I'll give all that a try, thanks.

use an antenna. Im pretty sure that will solve your problem