I think the duty cycle is a very important feature. Let me share my experiment. I'm just cap dumping into a primary coil of a Tesla transformer.
My coil is 31 uH, the capacitor is a 400 V 22 uF. That gives me a resonant frequency of 1.9 kHz. My SCR cap dump is using a neon lamp (100 V) for the trigger.
I set my function generator to 50% and at a high frequency (11 KHz) and the capacitor did not charge, because the spikes were not very high voltage. I lowered the frequency and the capacitor quickly charged, but once the dump was initiated I was not able to get the SCR to turn off, because current was always flowing in. I had to short the capacitor out so the SCR gets closed. If I did not, current was alwaysing flowing in from my DC power supply (that was set to 6 V, 20 mA limit and at that moment it was drawing 20 mA at 1.5 V, almost like a short circuit case). I don't know if it's the SCR (some 800 V one) or the frequency or the capacitor, but at 50% duty cycle I was not able to get it working at any frequency (tried up to 1 MHz).
I changed the capacitor with a 200 nF non-eelctrolytic one and I was able to make it work at 23 KHz at 50% duty cycle, but the cap dump was very slow - 11 Hz and a lot of power has been drawn - 1.3 V at 20 mA (26 mW). I changed the duty cycle to 8%, the cap dumps' frequency was the same (11 times a second), but the power consumption was 6 V at 2 mA (12 mW). Mostly voltage, as it should. The power consumption is just from the DC power supply. I'm not including the frequency generator in the calculation.
With that I just wanted to note the importance of the duty cycle. I guess that a 50% duty cycle would work fine with smaller caps, but not smaller coils because they won't give you enough voltage spikes to charge a capacitor up to 100 V. This means you'll probably need to trigger it with a Zenner diode. I tried to connect it to the base of the SCR (just like the neon lamp) and it did not seem to work as expected though (I only have a 5 V one). I don't have much time to experiment right now in order to find out the reason.
My coil is 31 uH, the capacitor is a 400 V 22 uF. That gives me a resonant frequency of 1.9 kHz. My SCR cap dump is using a neon lamp (100 V) for the trigger.
I set my function generator to 50% and at a high frequency (11 KHz) and the capacitor did not charge, because the spikes were not very high voltage. I lowered the frequency and the capacitor quickly charged, but once the dump was initiated I was not able to get the SCR to turn off, because current was always flowing in. I had to short the capacitor out so the SCR gets closed. If I did not, current was alwaysing flowing in from my DC power supply (that was set to 6 V, 20 mA limit and at that moment it was drawing 20 mA at 1.5 V, almost like a short circuit case). I don't know if it's the SCR (some 800 V one) or the frequency or the capacitor, but at 50% duty cycle I was not able to get it working at any frequency (tried up to 1 MHz).
I changed the capacitor with a 200 nF non-eelctrolytic one and I was able to make it work at 23 KHz at 50% duty cycle, but the cap dump was very slow - 11 Hz and a lot of power has been drawn - 1.3 V at 20 mA (26 mW). I changed the duty cycle to 8%, the cap dumps' frequency was the same (11 times a second), but the power consumption was 6 V at 2 mA (12 mW). Mostly voltage, as it should. The power consumption is just from the DC power supply. I'm not including the frequency generator in the calculation.
With that I just wanted to note the importance of the duty cycle. I guess that a 50% duty cycle would work fine with smaller caps, but not smaller coils because they won't give you enough voltage spikes to charge a capacitor up to 100 V. This means you'll probably need to trigger it with a Zenner diode. I tried to connect it to the base of the SCR (just like the neon lamp) and it did not seem to work as expected though (I only have a 5 V one). I don't have much time to experiment right now in order to find out the reason.