My Solid State Tesla Coil / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

My Solid State Tesla Coil

1 2 3 4 next

Move Thread

LAN_403

Wolfram

Mon Apr 07 2008, 10:52AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

After working on this project for a long time, I've finally decided to document it here. Since I have been working on this project over a long time, I will do the writeup in a chronological order.

This project got it's first start some time back in 2003. At that point I had already made a simple SGTC (some time in 2002, I think). I gathered many of the needed parts, and got a bit of it planned, but I was quickly sidetracked by other projects, and forgot all about it.

The real work on this project started just over a year ago. I had just returned from my first Teslathon (actually the Nottingham Gaussfest) and I was determined that I had to make a coil myself.

The secondary was wound on an 11'' piece of 3'' pipe. I wound a 10'' long winding on it with some wire from an old phone relay, estimated to be AWG30-32. The topload was a piece of flexible aluminium tubing and the primary was 6 turns of 15AWG stranded wire. The driver was my signal generator feeding a pair of UCCs, driving an IRFP450 halfbridge through a GDT. This was powered by halfwave rectified mains through a variac. Results were not very good but got a bit better when I discovered that if I tuned the oscillator to below Fres, spark loading would pull it into tune. It was still not very impressive though, partly due to the coupling being too low. Sparks were 5-6'' Video:

. I blew a lot of FETs, partly due to it being out of tune, partly because my heatsinks were too small and the thermal interface was bad (I discovered this later).

Then I changed to a feedback scheme with a single gate from a 74HC14 with an antenna on it's input driving the UCCs. This worked about as well as my previous scheme, the sparks were maybe a bit longer. Video:

. I also did some experimentation I still blew a lot of FETs. At this point, other projects took over my priorities.

In the autumn, I went to the Cambridge Teslathon. I became motivated to work on it again. I now made a new primary with much higher coupling and changed the heatsinks for some bigger ones. The primary is 5 turns of CAT5 network cable. Each pair is paralleled, and then each of these is put in series. This way it is tappable to 5, 10, 15 and 20 turns, without changing the coupling. I found out that it survived with 10 turns, but I didn't dare to go any lower. Picture:

. I also added a startup-oscillator and made it run on halfwave rectified mains again. By now it had started working quite well, sparks were around 10'' and it could run indefinitely without blowing up. I then converted the power stage to a fullbridge. Max. sparklength now was around 16''. Picture:

. From this point, I haven't blown any FETs for unknown reasons, but I have lost some during groundarcing and once when I ran the coil without the topload. I then acquired one of the beautiful spun copper toroids from Daniel Uhrenholt, and immediately the coil looked a lot better. Picture:

.

After one of the times I blew the FETs due to groundarcing, I changed the FETs for some scavenged 20N60 FETs from Infineon. Strangely, spark length was a couple of inches shorter than usual. This frustrated me a bit, so I was thinking of removing a few primary turns to get the sparklength back up. Not wanting to remake my primary, I tried tapping it to 5 turns. I didn't really want to use this few, as I was sure the FETs would blow, but I decided to try it anyways, to see how long sparks I would get before it blew. I cranked the variac up to 240V and spakrs were 20-22''. The variac amperemeter needle was pegged to the end of the scale, a bit above 8A. Mezmerized by the sparks I kept it running. After about a minute, sparklength started diminishing, and then all 4 FETs blew. I felt the heatsinks, and they were very hot. It seemed that they were the limiting factor of my system now.

The logical next step would be bigger heatsinks. I replaced the FETs (with some scavenged ST 20N60 ones) and used a seperate heatsink for each FET, no electrical insulation between FET and heatsink to minimize thermal resistance. This is where I am at now, and it is reliable. No FET deaths since I changed the heatsinks to seperate ones. Picture of the setup:

. Video:

. Frame from the YouTube video with very long and straight spark:

.

I am satisified with it's performance now. The next step is changing it to use secondary current feedback and making it into a compact, portable unit. Thread will be updated as things happen.

Anders M.

Bennem

Mon Apr 07 2008, 12:51PM

Registered Member #154 Joined: Sun Feb 12 2006, 04:28PM
Location: Westmidlands, UK
Posts: 260

I've said this before and i'll say it again,
thats one cool SSTC!!
You're pumping some 3 KW....right?

Gotta love thoughs sword-like sparks...

Mel

Wolfram

Mon Apr 07 2008, 01:02PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

I think I might be pushing around 2-3kW, but I'm not entirely sure, as my electronic power meter constantly resets and displays strange numbers when the coil is running. I should really connect up my digital scope to the line voltage and a current shunt, and set it to multiply the waveforms, and then calculate the average value. If I can find a reasonable resistor to use as a shunt I might try that later today.

Anders M.

Ken M.

Mon Apr 07 2008, 10:54PM

Registered Member #618 Joined: Sat Mar 31 2007, 04:15AM
Location: Us-Great Lakes
Posts: 628

Very Nice indeed, and the the Excalibur arc.

ShawnLG

Tue Apr 08 2008, 02:01AM

Registered Member #286 Joined: Mon Mar 06 2006, 04:52AM
Location:
Posts: 399

WOW! Those are massive size arcs you are pumping out of that small coil! Great job.

GeordieBoy

Tue Apr 08 2008, 10:18AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Nice coil Anders and good performance in the videos too. The sword-like sparks remind me of my old SSTC.

The failures with the 20N60 FETs sound like thermal-runaway to me. The TC primary acts like a constant-current load, so as MOSFETs heat up and Rds(on) increases, conduction losses increase causing further heating, etc, etc. You will see a drop in performance right before the MOSFETs blow because more and more voltage is being dropped across the switching devices instead of across the TC primary.

You certainly went down the right route with eliminating the thermal pads from beneath the MOSFETs and mounting them on seperate heatsinks. The only thing I would be tempted to do is to replace the mounting bolts with a clamp (or spring-steel clip) that provides positive pressure over where the die is located in the plastic packages. Fixing semiconductors to a heatsink using a single bolt through the hole in the package actually acts to lift the heat-producing part of the device off the heatsink. It's okay when you're only dissipating a few watts, but if you want to run the devices anywhere near their Id rating, I'd definitely mount them in a way that forces the metal backplate against the heatsink.

Switching to base-current feedback might not make much difference to performance, and could cause problems with body-diode reverse recovery if you didn't bypass these with schottky rectifier diodes?

Have you tried running the coil from full-wave rectified mains!?!? This is visually very impressive, if you are wanting things to try. The spark will be a lot more bushy and flame-like (but shorter) and it sounds more like a full throated roar than the thudding sound you get with half-wave mains bursts. Of course the MOSFETs will get a lot hotter though, so either the thermal side of things needs to be sorted first, or you need to add a few turns to the primary to reduce the load current to compensate.

Thanks for sharing the video and pics Anders,

-Richie,

Wolfram

Tue Apr 08 2008, 10:47AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Thanks for the comments everyone.

I found a set of heatsink clamps for TO247s in a power supply, and it'll be an easy task to mount them onto my heatsinks, so that'll be the next thing I'll do.

The main reason I wanted to use base current feedback was to eliminate the antenna, as I want it to be as compact as possible, and eliminate chances of sparks hitting the antenna, but if it leads to heavy forced recovery of the MOSFET body diodes I'll just stick with the antenna.

I did try fullwave rectified mains a couple of weeks ago. I didn't dare go higher than 50% on my variac though, as the sparks were pretty scary. I'll definitely try this again in the near future.

I'm planning to add a power controller to it so that I won't need the variac to vary the power. The plan is to add phase angle control. I know Richie tried powering an SSTC from a phase angle controller as a part of his SSTC experiments and the sparks ended up crooked due to the sharp rise in the RF envelope, just like with SGTCs and interrupted SSTCs. As I prefer straight sparks, I am going to cut off the end of the half cycle. This should avoid the sharp rise in the RF envelope, and hopefully not make the sparks too crooked. This could simply be implemented with a zero crossing detector and a monostable driving the enable input of my gate drivers, but I am going to implement it with a PIC12F675, which should be relatively simple. Any thoughts?

Anders M.

GeordieBoy

Tue Apr 08 2008, 12:40PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

A couple of comments...

I did try fullwave rectified mains a couple of weeks ago. I didn't dare go higher than 50% on my variac though, as the sparks were pretty scary. I'll definitely try this again in the near future.

If it survived that 50% run for any length of time then it would likely survive 100% on the variac.
It is interesting to note that after breakout the TC acts very much like a constant current sink. Therefore the current through the MOSFETs is almost constant from the point when breakout is reached, right up until full mains voltage. My point here is that the MOSFETs aren't actually stressed that much more should you have decided to increase the voltage from 50% right up to 100%. Switching losses will increase due to the increase in V. But conduction losses will only increase slightly because the actual primary current does not increase very much. (If you use a CT to view either the primary current or the secondary base current you will see what I am talking about. The current envelope is not like a sine envelope, but almost has its peaks squashed flat due to the constant current behaviour of the TC beyond breakout.)

As I prefer straight sparks, I am going to cut off the end of the half cycle. This should avoid the sharp rise in the RF envelope, and hopefully not make the sparks too crooked.

I have absolutely no idea what the sparks will look like or sound like with reverse phase-angle control!

BTW Reverse phase-angle control using IGBTs is common in high-power light dimming applications because EMI is easier to control than with conventional phase-angle control with thyristors. (You can control how quick the IGBT turns off, but you have limited control on how quickly a thyristor turns on.) So you should be able to find some design references explaining how to do it.

-Richie,

Mates

Tue Apr 08 2008, 08:39PM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Hi Anders,
I hate your coil! Once I saw your pictures in "Post your cool SSTC picture" sessions I decided redesign my own coil completely and this attempt cost me already 4 transistors and I'm still far away! You should get the bill for those poor little BUP314...

Now seriously, in my eyes you have the best SSTC I have seen.

But, I have a comment to the sparks. The length is very impressive (compared to the size of the coil) but the sword-like appearance is not as attractive as branched sparks (my personal opinion). My coil is single switch working in classE and it produces branched sparks. I took few short exposition pictures today and sending them to compare (at 1500W BTW: The easiest way of measuring the power consumption of your coil is via the speed of rotation of you home power meter. You can even buy some old one in case itâ€™s not easy to set up in your home).

The branching is a very enigmatic thing and your coil is very interesting by showing the opposite extreme...

uzzors2k

Tue Apr 08 2008, 08:50PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I see you don't need to bother with DRSSTCs to get good streamers, Anders!

Your SSTC is amazing, great job. I'm inspired to start a new coil myself. Keep up with the upgrades, it'll be interesting to see how this coil can be improved.

1 2 3 4 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob