On March 2010, I built a 4 inch Tesla Coil. Of course the main motivation was to see the sparks but along the way I learned how it operates and the physics of electricity. The maximum spark length I've caught on camera is about 3.5ft but the average is 2-2.5ft... still pretty good. I also attended the Western Winter Teslathon 2010 in March 2010 and was able to bring my newly finished coil (a picture of my coil operating at the teslathon). Here are pictures I took at the Teslathon. I'm looking forward to the next Teslathon hosted at the Arizona Science Center on March 5, 2011! See the link below for the video of the first successful run (thanks Ken). More videos of my coil will be available soon.
This is how the tesla coil operates: When the NST is switched on, it will begin to charge the MMC. When the voltage gets high enough, the spark gap will conduct, thus directly connecting the MMC to the primary. This will create a powerful magnetic field that will transfer energy to the secondary. The primary and secondary will resonate, transferring energy back and forth by the rising and collapsing magnetic fields and will continue to do so until the spark gap extinguishes. How many times the transferring of energy back and forth is determined by how well the spark gap is quenched. When the gap extinguishes, all of the energy is trapped in the secondary circuit and releases it into the air in the form of sparks (plasma) from the toroid.
The secondary & toroid and the primary & MMC form an LC inductor-capacitor resonant circuit. The primary and secondary circuit are tuned to the same resonant frequency so that they resonate off of each other. This creates the higher voltage that the tesla coil is famous for. The resonance is important because the tesla coil isn't very efficient compared to normal transformers. In normal transformers, at least 90% of efficiency is obtained because of the high magnetic coupling coefficient (1) but with the tesla coil, the coupling coefficient is around 0.10. This makes the tesla coil less efficient but the resonance overcomes that. You may ask, well why don't you increase the coupling for better efficiency? If the coupling is increased, then the insulation on the secondary cannot withstand the high voltage and there will be sparks between the windings... not good. I use that indicator so that I can find the best coupling. Increasing or decreasing the coupling is obtained by either raising or lowering the secondary (or the primary). For calculating inductances and resonant frequencies, I used JavaTC.
Having an RF ground is very important. You cannot use your house ground since the high RF can damage any grounded appliances. I installed an 8ft ground rod in my backyard. Everything on the tesla coil itself must be RF grounded, this includes the NST. The only thing that should be grounded to house ground is the control/switch box and the line RF filter. A good rule of thumb is to house ground anything that you might touch while the coil is running.
Tuning the tesla coil can be done in two different ways: trial & error or using a function generator & an oscilloscope. When using trial and error, just try different tap locations on the primary and see what gives the longest spark output. A function generator and a scope could also be used to find the exact resonant frequencies of the primary and secondary circuits. Later on I obtained a function generator and was able to match the primary and secondary circuits for resonance. I found it to be around 235Khz and the tap point I was using found from trial and error was spot-on. More detailed information about tuning with a function generator can be found here.
The primary coil is made of 1/4in copper tubing held in place with pvc pipes. It forms the inductor to convert the current from the MMC to an electromagnetic field. There are 10 turns that provides a maximum inductance of 34 mH. The tap is made from one half of a fuse holder. It can easily be snapped into place onto different locations on the primary to provide different inductances. A good rule of thumb is to add two or more turns than what is required.
The secondary coil is made from 4in pvc, 24 gauge enamel coated magnet wire and 10 coats of polyurethane. There is approximately 930 turns which equates to about 1100ft of wire. The calculated inductance is about 20mH. It doesn't seem like wire has much resistance in short lengths but the resistance of the secondary is about 28 ohms!
The toroid acts as a capacitor for the secondary circuit. It is made of 4 in aluminum ducting wraped into a circle and using aluminum tape to hold it together. Another one was also made from 3in ducting to add more capacitance. Both of them together provide about 15pF of capacitance. Changing the topload will affect the resonant frequency.
The static spark gap acts as the high voltage switch for the primary circuit. This design is a "Richard Quick" type that uses 8, 3 in length x1/2in dia copper tubes. Two bolts secure each pipe to the 4in pvc pipe. Two slits that are cut through 3/4 of the pipe provide a place to mount and easily adjust the gaps in between each tube. Each gap is set to around 0.03in +-0.005in using a feeler gauge. The total gap width can be adjusted by moving the wire on the right (first picture) to a different location. An overall wider gap will yeild longer sparks but the safety gap will fire more often. Once it fires too much, there is no spark output so the gap must be adjusted to where the safety only fires occasionally. Don't even think about widening the safety gap, unless you are willing to sacrifice your NST! A 120v AC fan is attached to one end that quenches the gap and also cools the pipes. The fan flushes out the ozone gas created by the spark. This is required to have some consistency in the firing voltage since the ozone gas is more conductive than normal air.
The capacitor bank is constructed with a total of 32 Cornell Dubilier .15uF 1600V DC polypropylene (942C16P15K) capacitors. There are two parallel strings with 16 capacitors in each. This gives a total capacitance of 18.75nF at 2560V DC. After the MMC was first used, the capacitance increased to 21nF according to my Fluke multimeter. Each capacitor has a 10M resistor soldered across its terminals to discharge it within a few seconds after the power is switched off. As you can see in the picture, there are two plexiglass "walls" inbetween the columns of capacitors. They were added later on because there was arcing in between the strings. It arced because I didn't have enough room on the perfboard for two parallel strings, so each string was wrapped around to the middle column. Because of this, there is a 8000+ voltage potential between the capacitor leads (the leads are 0.6in apart!). Next time I'll just get a longer perfboard to avoid this problem. By now there has been at least 45 minutes of total run time and there have been no capacitor failures.
The Terry Filter is a RC network that acts as a low-pass filter that protects the NST windings from damage from the high RF feedback from the primary. The safety gap is a three electrode gap with the middle electrode connected to RF ground. The gap is set to fire just above the operating voltage of the NST. It fires when the main spark gap does not fire and the MMC gets charged beyond the NST's voltage rating. The safety gap ultimately protects the NST's windings from internally shorting from overvoltage. The safety gap and the Terry Filter are a great combination for protecting your NST. A Terry Filter schematic can be found here (also includes the safety gap). Note: the MOVs were excluded from my filter to reduce the cost. As long as the safety gap is set correctly, the MOVs aren't necessary since they provide backup overvoltage protection. Also, everything was wired up using 10 gauge stranded copper wire with crimp on terminals. Wires that needed more insulation were sleeved with tubing.
The NST was bought from a local sign shop. This one is a Franceformer rated at 12000V AC at 60ma. I recommend that you purchase one between 9-15Kv and a 30-120ma rating. Don't buy one with a GFI protection circuit since it may trip while the coil is running.
The control box is a gutted PC power supply that houses a commercial line RF filter, 10A breaker and a switch. I can plug in an extension cord from the control box to the coil to provide a safe operating distance. The line RF filter I used was from ebay rated at 20 amps. Just wire it backwards, ie. line goes to the NST and load goes to the house 120v. This filters out any RF that makes its way back to the house 120v.
Last Updated May 2010
Contact: Tai Oliphant
Copyright © Tai Oliphant 2010