How to enlarge voltage in Maxim boost converter?

[Searching]

Hi

I made a geiger muller detector from Maxim application note from this site:
https://www.maximintegrated.com/app-notes/index.mvp/id/3757
https://www.maximintegrated.com/images/appnotes/3757/3757Fig01.gif

Detector works perfectly. I made also Atmega powered counter and connected it with detector

Few months ago when I was testing boost converter block while building device prototype I turned on converter and drive it to maximum power what means - minimum transistor-current off-time and then there appeard huge spark at least one centimeter long at the end of multiplier ladder. I dont remeber if some parts like resistors or capacitors was used with another value.

Today I tried to achieve that effect with another prototype version of my geiger muller counter turning on only converted block but nothing happens.

I switched transistor-current off-time to minimum connecting 100k resistor upper pin directly to Vcc but nothing happen. The voltage at the end of multiplier was approximately 800V.

Then I replace 10k resistor with potentiometer and change transistor-current duty cycle to aproximetly 50% and nothing happen - there wasnt any sparks, even ~0.5 milimeter long between two very close wires. Only transistor was heating very very fast and i had to turn duty cycle lower and turn off all device avoiding transistor dead.

Then I replaced all multiplier capacitors from 100nF to 4,7nF thinking that higher impedance will drive coil to produce higher voltage to remain given coil current when transistor turn off. Nothing happens.

I am not surprised that crazy blind-eye non-science tries with making working converter have to fizzle out but maybe You can tell me which part shuld I replace to achieve spark? I will be very happy if You can help me.

 

[BradtheRad]

When the first spark occurred, it may have altered one or more component's characteristics.
Example, it may have lowered its internal resistance, etc.

Afterward that component(s) could be hindering the spark.

You probably will have to test each component individually, to find the rogue component.

Or test the entire circuit at a reduced volt level. Put the voltmeter across each component. The odd one (or ones) should reveal themselves.

 

[Searching]

Maybe I write previous text with some ideas shortcuts because I don't think that last version of converter is made with lower quality parts or some parts are broken. The parts are fine - voltage and possible output current are ok for geiger muller tubes.

I am courious if in previous version, where there was a spark, didnt I made any spontaneous part changes.
I remember there was some problems with previous version because the comparator output was open collector type and I didnt use pull-up resistor. I didnt know where was a problem. I was sad and this made me do some bizarre tries to repair and turn on converter. Maybe I change resistor values or capacitor values.

I did not expect answer supported by difficult calculations. I know that making converters is very difficult thing and there are many tough books and I should not wonder why there isnt any spark. I rather hope that You can show me which part, parameters, values are responsible of converter output voltage.

 

[BradtheRad]

From what I understand, a one centimeter spark requires greater than 1,000 V. Maybe 10,000V?

You constructed the voltage multiplier using several components. For a fraction of a second, each contributed some volts to the total. Some components were probably exposed to several kV. It will be surprising if they all are still the same afterward.