After a long period of being left sitting on the dining room table, I decided to warm the beast up again ! I wound a new primary set on a plastic ice bucket to increase the diameter a bit and prevent arc-over.
The new primary changed the output from long and thin to shorter, fatter sparks, probably due to different coupling values. It was now VERY loud and kept melting my breakout point on the topload. All was going well until…
You can’t ignore the acrid stench of burning insulation ! Oh dear !
There goes the ice bucket theory. I think what had happened is that, during it’s idle break, the coil fell over and was stood up again but I did not check the positions of the coil windings and connections before juicing it up with 2000vdc. Just goes to show that a 5mm gap is not good enough between the feedback coil and the primary coil. It probably was not being helped by the fact that the primary had excess turns on it which were acting as an open auto-transformer and stepping up the tank circuit voltage even more.