As per my usual protocol for these sorts of things, I decided to build a large spark gap Tesla coil this IAP (read: winter term) after I finally collected all of the necessary components through a couple years of hoarding and one lucky find from a dorm basement cleanout early in January. It turns out that many years ago, someone in East Campus had attempted to build a fairly large SGTC (it appears in some old i3 videos), gotten to the point where it made some amount of sparks, and then moved the pile of parts to the server room in the basement, which may or may not have contained an EE bench at the time. Fast forward to 2015 and the server room was getting cleaned out and refurbished, and there was a pile of Tesla coil parts sitting in the hallway waiting to be picked up for disposal. I took the old power supply, secondary coil, topload, and asynchronous rotary spark gap and promptly got to work after redesigning the entire system in JavaTC (see linked files at the end).
The first order of business was getting the power supply to turn on. It consisted of a 14.4kV potential transformer and a PDU (Power Distribution Unit) with a pair of variacs (one for the transformer, one for the rotary gap) and had certainly seen better days. After a day of reverse engineering the wiring and fixing broken connections inside the PDU, the electronics were restored to working order, and I could even turn the power on after the keyswitch was bypassed! (It didn’t come with a key…)
Next up was the rotary spark gap. Everything appeared to be in working order so I slowly and carefully powered it up. It promptly self-destructed, and I spent the next few days waiting for some tungsten welding rods to arrive from eBay (5/32″ x 7″ Lanthanated TIG rods) and then cutting them down on an angle grinder and rebuilding the mechanism. It turns out that the bolts holding the motor to the faceplate of the rotary gap were loose and allowed the slight, inevitable imbalance in the rotor to develop into a serious problem. Lessons learned: use lock washers, and make sure the fasteners in your high RPM mechanical assembly are tightened!
I decided to try my hand at some basic woodworking, so I routed two 32″-diameter circles of wood for the base of the coil, drilled mounting holes after mocking everything up on a computer, and then sanded, stained, and varnished the parts:
The primary supports were laser cut from 0.25″ thick acrylic sheets and the primary was wound with 0.25″ diameter flexible copper tubing ziptied to the supports.
For the primary capacitor, I used 8 parallel Ion Physics model 4006-311 low-inductance, high-voltage polypropylene film capacitors rated for 100kV at DC and several nanofarads each. The total capacitance of the bank was 54.98nF. It’s definitely overkill (and probably not quite enough capacitance), but it’s what I had on hand. I’m planning to replace it with an 0.7uF, 50kV Maxwell pulsecap from eBay, but that hasn’t arrived yet. These capacitors are gigantic, by the way:
The fully-charged bank (at 100kV) can store in excess of 250 Joules (this is not a totally inconceivable scenario if you consider resonant charging!); charged to the peak of a 14.4kV RMS sine wave, the bang energy is closer to 12J, which is reasonable for a coil of this size, but the physical size of the capacitors is mostly wasted.
The secondary coil was re-varnished, laser cut end caps were installed, and I was faced with an interesting problem. After assembling the coil in my dorm room, I realized it was far too big to fit through the door, and I didn’t have anywhere near the amount of power required to turn it on, even if I took it outside. Luckily, though the base it didn’t fit through my room’s door, it did fit through most other doors in the building and I rolled everything over to Walker Gym in building 50, which is right next to the amateur radio club and has 15A and 20A single phase and 40A three phase outlets.
Operating without a ballast (this is the maximum performance mode, as the rotary spark gap quenches very effectively), the sparks are 6 feet to a grounded target and break out in multiple directions, which suggests that the performance could be improved further with a larger topload. As I am near the maximum primary tap with my current primary capacitor, I have not yet tried this.
Setting up the coil and target:
6 foot ground strike:
If you are in the area and would like a demonstration some weekend evening, shoot me an email and perhaps something can be arranged.