For my purpose a Tesla coil is for generating an aestheticaly interesting and exciting non-lethal display. I realy want to see some 10 foot sparks (Would be non-lethal?) The idea is to build a smaller, cheaper more powerfull, Tesla coil by oscillating smarter and not harder, with modern semiconductors and ferrites. Maybe I've been zapped a few times too many. I would appreciate some feedback, and hopefully can find some others who have always wanted a mega-volt multi-kilowatt Tesla coil but just can't see providing the money, space and power for one. I hope to shrink it and improve its performance by, hopefully, (now don't laugh please) 2 orders of magnitude. The info I've found and will refer to include: "A high potential Tesla coil impulse generator for lecture demonstrations and science exhibitions by Kenneth Skelden, Alastair Grant and Slan Scott, in that great little American Journal of Physics, 65 (8) August 1997, pages 744 to 754." recommended and graciously offered by Winfield Hill And many (but not all!) of the text files at ftp.funet.fi/sci/electrical/tesla/misc/ And sept 1991 RE "Solid State Tesla Coil" Duane Bylund ******** Begin by sacrificing a pole pig. Maybe the most expensive, rare, bulky, massive and nearly impossible to fabricate component. The pig (or neon, oil burner, copier transformer) used to turn 120VAC to 6 - 20 KV. The big ones are rated at 10KVA. In Bylund's RE article, he does this using a 300 watt transistor switching supply to boost 120 VAC to (is it?) 5000 volts. This is continous (CW) power at 100KHz. The secondary of this compact coil is around 10" dia tall, wrapped with #30 wire (IMHO he should have use 24 and made it 30" high, for higher Q ) It's SG3524 switcher can be tuned (YEA!) but no power control (Boo). It ONLY spits out 100KV, 7" streamers. It needs a fist-sized $40 ferrite transformer to excite the secondary, to which the ferrite transformer is directly connected. No spark gap, to waste power, or loose primary to loose more energy. Direct connection, like Tesla's "Magnifyng coil" and more familiar to my mind as a PI filter, or 1/4 wave matching transformer. Driven end is low V, hi I. Wouldn't it be nice to just switch 120 VAC directly into the secondary, and not pay for $50 bucks worth of ferrite, in ADDITION to making the secondary too? I was planning on using FETs & a 1KV microwave xfrmr. But after reading some theory, came to the conclusion that I might get a multi-kilowatt (FET's power-dependant) output, but no mega-volt >1 meter arc's, which is what I'm after. The problem here is the secondary's Q. If Q were high enough, CW power could build up high voltages. The Q is the killer here. Unless a physicaly LARGE secondary with nasty Litz and/or large gauge wires used, you loose - air core's, copper loss and small size don't all fit. What about impulses? They will give you impressive, cooler, arc's. Then you go back to needing Semiconductor un-freindly high voltage to charge a big cap to several KV. And expensive or hard to implement solid state switching. I've seen circuits to series connect, isolate BJT's & FETs. But I don't like thier complexity. A spark gap is EZ'r, but were back to another transformer power supply, and waste. And a big secondary. I propose the following design: * A simple exciter, consisting of oscillator as simple as a 555 or complex as a PIC oscillator/controller * BJT's to drive a half-bridge MOSFET's 500V 7A, which could be paralleled for upgrading *A low ESR .1uF cap to couple the bridge directly to the secondary Now the hard part... And I'll admit I'm too lazy to get out my books & programs to calculate and model this system, because I'm hoping to get a reality check from someone with experience and/or more knowledge. The Plan: A compact secondary around 4" dia by 15" long, encapsulated in two hacked-together 3-liter bottles. Maybe pressurized & dessicated, or oil filled. So now I can: Use a ferrite rod core in the secondary! This will multiply the Q by ferrites mu by increasing inductance, lowering frequency, et. Use multiple windings maybe (quit laughing, they'll be teflon/mylar insulated) Use <32 AWG copper because frequency's lower Around .5H inductance, around 200 pf capacitance, around 20KHz resonance, Q around 1000? The windings are spaced in oil. The core's insulated by wrapping with aluminized mylar sheets. Individual mylar sheets are cut to 'C' shapes to prevent magnetic coupling. Aluminization would make the dielectric self healing. Otherwise, regular mylar. Or teflon. More is at top, less at bottom to match potential gradient along coil core. (I'm guessing even "high resistance" ferrite isn't at megavoltages.) Well, even if I can shrink the secondary AND boost Q with ferrite, I could still have the CW-small, hot spark problem. Now I'm not too well magneticaly coupled to the ferrite, having an inch of teflon. So hopefully the bar core doesn't saturate. But even if I am, maybe I can use an old radar trick to turn FM CW into a high voltage impulse! In certain all-pass or delay-equalizing filters, and dispersive delay and transmission lines have the usefull property of delaying some frequency's more than others, without effecting amplitude. So imagine the driver FM chirping into such a filter. With the proper design, the chirp is compressed into an impulse, as the initial high frequencies take longer than the subsequent low frequency's to propagate through the filter/line. OK, so now I have to build not a tesla coil, but a Tesla-Chebyshev (or something) filter. Insanely difficult to make a lumped element HV filter right? I have this silly hunch, that if I make some type of "SPECIAL" core, kind of like microwave plumbing filters or wave guide, maybe a cone shaped dielectric or ferrite core or winding, maybe relying on the core saturating at higher frequencies, dissapearing like guannella transmission-line transformer cores? Or, yea, got the idea? I need a high voltage, ferrite-shrunk Tesla chirp-to-impulse matched filter. Who's done this and wrote about it at these VLF frequency's and high voltages? Then maybe I can get 6 foot sparks out of a tiny little bottle plugged (almost) right into an outlet, using a couple chips, couple FET's and ONE coil, for only a few bucks worth of surplus ferrites & MOSFETs, a PIC and an old motor or wall wart. And a few rolls of teflon tape. wouldn't you love to play with one of these? No, I don't want to wake up now :( ************************************************* Ok, so what the hell does this have to do with PIC's, the PIC'rs are wondering? Here are some requirements I began spewing out. Obviously, I need to include a kitchen sink and the DOD. But if thier is enough interest, maybe we can collaborate. Otherwise, I'll probably loose interest after my current 35W, 5"x20" spark-excited tesla kills my few remaining neurons. TESLA PIC functions- PIC has dual purpose software, 1) to generate a swept square wave for network analysis of a complex coil. Controlled by PC software, and generates smith chart.Controls 4046 VCO to generate square wave, and reads its phase detectors. Also need analog switch for VCO range. 1)Testing and Coil Evaluation: Sweep Oscillator PB1,PWM output controls 4046 VCO 1KHz to 1MHz PB2,analog switch for 4046 VCO range cap PB3, analog switch for 4046 VCO range cap Amplitude meter - diode detector, 3 port line diode/cap A/D (ya, its cheezy but I just want to see a peak Tesla Coil PIC Cheep Transformer-less cap coupled CPU power FET driver oscillator, Chirp, CW, mono-pulse, pulse width, impulse and repitition rate Power consumption (120VAC) measure E-field meter - analog switch chopping electrometer, to PIC A/D Monitor FET temp (OK, I need a muxed A/D too) Emergency shutdown solid state relay. 15 Amp :) Read safety interlock switch Read power input control - plastic knobed HOT pot Read frequency input control - plastic knob Hot pot Operation LED Standby LED Danger LED Lethal Power LED Piezo buzzer for command acknwlg & warning/fault Safe remote IR or TV/VCR format remote interface Serial IRLED or hardwire (cheap) to PC for hot op data dump **************************************** Epilogue The Burgler-Quter The 5 mega-volt pop bottle resides inside a 1meter dia plexiglass sphere 1/16" thick, filled with Krypton and neon to glow orange, and decorated like a vicious. sadistic pumkin. Two 1cm dia plastic jacketed conductors diverge from the cheaks, out 1 meter, and terminate in aluminum spheres, like catfish stingers. The sphere is suspended by a power cord on a rotating motor mount, allowing it to rotate and face objects scaned by a PIR sensor. It scans, oscillating right and left, untill an IR anomaly is detected, then homes in, powers up, zapping and emitting hilarious laughs & cackling at 120 dB. (I'll deal with the pyro-acoustic siren design another day.)