Rotary Spark Gaps

Click on the photos for bigger picture. 
My synchronous rotary spark gap is my own 3000RPM-version of Finn Hammer's very successful 1500RPM "Millenium-Gap" design. It has been realized by "Lehrwerkstaette fuer Mechaniker"(LW), Basel, in a very cooperative way: Mr. Bader, chief engineer of the workshop, not only helped me to set the tolerances, but also managed to realize the gap, in time, for the "Technorama" show of Tesla coil B&W. The gap, beeing a great success in the whole, suffered a few serious beginner's mistakes: 

- The motor choosen was a very cheap 400W double-grinder motor, whose mechanical imprecision brought a hassle for reassembling, after having filed flats, in order to get it sync. It's only getting sync now, by use of a variac power supply, which, on the positive side, allows a modest setting of the phase-angle, by electrical means ( ~10 degrees).

- I got nasty carbon tracking along the surfaces of the hardpaper (~pertinax) mounting supports, 'cause of too short insulating paths.(see the pictures!). The rule of thumb "surface path = ~ 3 times the path in air" should be taken seriously! By good luck, the construction could easily be modified, by cleaning and epoxying insulating bars at the critical places of the gap. ( The HV was coming from my 16kV (RMS) -PT's, and might have been about 25..30kV peak, considering some resonant rise.)

SRSG, view from top.
Rotor d = 240mm
10mm diam. tungsten
electrodes @ d=215mm
47'709 bytes

Side view...
sorry for
pic quality :-(
34'456 bytes

Carbon track 1
30mm wide...,on top,
between HV supports
38'574 bytes

Carbon track 2:
50mm wide, hidden
between baseplates.
(look at screws..!)
44'461 bytes

Insulating bars,
hopefully preventing
new carbon tracks.
28'862 bytes

Another perspective.
23'383 bytes

SRSG, buildt into base 
of B&W; view from
"service side" of  TC.
86'479 bytes

Side view, same...
74'082 bytes
. . . .
Following are three 
snapshots of Martin
Damev's ASRSG.

47'451 bytes

39'320 bytes

fired in Technorama
with B&W-TC
17'823 bytes
. . . .
Modifying the 
armature of my 400W 
double-grinder motor
for sync operation, 
by filing two flats >>>

Motor disassembled
59'940 bytes

Armature with
25.5mm wide flat
17'635 bytes
The flats seem to
be just a little above
optimum: 24.5mm
was slightly better,
regarding torque