There was smoke, but no fire...

The first half of the tale of Halloween. 

In the course of my evening, I got a little work done on a few projects, notably I made an adjustment to the Elco No 7 capacitor situation. I found some 630V .1uf mylar film caps, put them on the breadboard and wired them in to see if they offered better performance over the ceramic film caps I've been using.

I started the experiment with 6, then 5, then 4, then 3 to observe the difference in output. As I started to play around a bit with 3, the overall output was much, much stronger, and I wasn't really paying so much attention to the unit itself, but the spark length from the electrode to the tweezers in my hand. A couple of minutes had passed by this point, and while that is really nothing in terms of what the caps should be able to handle, fine white smoke suddenly started to waft upwards at me. I killed the power, and looked down to see the corners of the mylar caps blackened and bubbling which subsided after about 5 seconds of the power being off. So that clearly won't work.

After disposing of my electronic popcorn, I repeated the same experiment with 3 ceramic caps, which are 1000V rated, which worked like a champ. I had previously replaced the guts of the "condenser" shell with 6 ceramic caps, but the output was way too weak. My theory was flawed, in that by lengthening the pulse out, that there would be greater peak output. What I failed to appreciate was that while it might have increased peak voltage, the duty cycle actually would be shorter due to the additional charge/discharge time. New knowledge in hand, I removed half the caps, and it's been running much more like its former self. I will eventually try dropping down to .2uf, but I also need to make sure that I'm not otherwise overheating the coils or anything else that will be a nightmare to repair/fix. Burning out caps isn't ideal, but is a hell of a lot easier to fix than trying to have a 100 year old coil rebuilt.

At any rate, I did manage to take a picture of the 2 visibly burned up caps, after thinking I should probably document the failure. I stuck them back into the breadboard haphazardly for the picture.

So after a little bit of conferring with the Tinkerer, more measurements will be needed to validate that the voltage at the caps remains <550V when running with reduced capacitance compared to the original cap that was in it. Regardless, the voltage should still not be exceeding 1kv.