My crusher uses 1/2" square copper bar to connect the capacitor to the coil and spark gap. The coil is four turns of 1/2" copper tubing wound around a twelve inch length of 3" OD X 2 3/4" ID polycarbonate tubing (the coil was formed with a 1/2" lever type tubing bender) to provide insulation between the coil and can. The plastic tube is supported on 1/2"-13 threaded fiberglass rods, which are mounted in a drilled and tapped 12" X 6" X 1/2" sheet of polycarbonate. Nylon nuts hold the coil form in position. The gap is made from two 1 7/8" chrome steel ball knobs, screwed onto lengths of 3/8"-16 threaded 18-8 stainless steel rod. The square copper bars are tapped on the end to accept the SS rod and both contain 3/8" holes for connection to the capacitor and 5/32" holes to accept the banana plug terminated cables from the power supply. The ends of the coil are just slip fit over the threaded rod.
I started with the gap set to a small distance (~1 mm) and turned on the power. After a few seconds there was a snap and I turned off the power, discharged the capacitor through a limiting resistor and saw no visible effect on the can. I opened the gap up to ~5 mm and tried again. About a minute after turning on the power, there was a loud bang and bright flash. Again no effect was seen on the can. At a gap of ~8 mm, five minutes of charging led to a very loud bang and a shower of vaporized steel from the gap. Still no effect on the can, and the C-W power supply no longer produced any output.
A post mortem on the supply showed that the capacitors were OK. However, the diodes now conduct freely in both directions (this is bad). This took me by surprise, since, to reverse bias all of the diodes at once (12 in each leg, each rated for 12,000 volts) should require more than 144,000 volts. Since the capacitors (only rated for 90,000 volts in this configuration) were unaffected, I can only assume that high frequency feed-back from the crusher caused this problem. The coil has an inductance of ~0.85 microHenrys, meaning that the coil capacitor combination should resonate at ~100 kHz. I will rebuild the C-W supply and in the future will place my Tesla coil RFI filter between it and any likely source of RF feed-back.