With my departure to Washington State University rapidly closing, I decided that it was about time I had a nice looking and, more importantly, a reliable demonstration coil. The first thing I had to decide was just what sort of coil I would construct. While it would be nice to have arcs a few feet long, noise seemed like it would be of significant concern. With that, it was clear that an SSTC operating on rectified mains would be the best choice.
While the output streamers would surely be visually appealing, the enclosure was also of importance. My past SSTCs, while functional, hadn't been the most attractive devices. Given my past experience with acrylic, using a clear acrylic case seemed like a Good Idea.
Lastly, a controller circuit had to be devised. I did not have good past experiences with the regular antenna feedback circuit that had become quite popular and a fixed-frequency oscillator just wasn't a good idea on such a dynamic system. Luckily, Steve Conner had recently proved that the CD4046 PLL could be used to create an excellent base-current feedback SSTC. With a few others having easily duplicated this design, it was the obvious choice. I then tacked on a staccato controller to reduce power consumption and system heating. Finally, a remote dead-man's push-button was added to enable the staccato controller (allowing for the production of streamers at the press of a button!). Since the staccato synchronized the system's enabling with zero-crossings, it would be safe to run the coil without a variac (this greatly added to the coil's portability and overall 'niceness').
Now, all that remained was to actually build the thing!
Monday evening (2004-08-16), the realization that I was leaving home in less than five days finally sunk in. At this point, it seemed that I had better get started on my SSTC if I were to have any hope of finishing it in time. Ah, procrastination... So, with that in mind, I went to work.
My first objective was to construct the new secondary coil that would be required (none of the half-dozen coils I already had were quite what I was looking for). I'm a big fan of low frequency secondaries, since they greatly reduce the stresses on all system components. Therefor, the section of 6" PVC left over from Nimiety quickly caught my attention. I had already decided that a 9" tall secondary would be reasonable for my desired streamer output. While a 6"x9" secondary does look a bit funny (some have gone so far as to call it 'cute'), the 200KHz resonant frequency was most desirable (as opposed to 350KHz for a 4"x9" secondary). After properly preparing the section of PVC, I quickly setup my usual winding rig (consisting of some wood to support the coil, and a hand-crank to spin it). Winding went fine, but drowsiness quickly set in; the rest of the winding would have to wait for the following day.
Having gone to bed at an actually reasonable hour, I woke early and went back to work. I finished winding about half of the coil, and then decided it would be good to acquire the rest of the parts I needed before it was too late in the day. I first went to TAP Plastics and purchased some additional 1/4" acrylic, a 1/32" HDPE sheet (for insulating the primary from the secondary) and some 1" acrylic square stock for the enclosure's corner supports. Next, I went to McLendon's (a local hardware store) and bought some miscellaneous parts (nylon screws, 6-32 hardware, etc.). Lastly, I went to the dreaded RadioShack and bought a low-voltage transformer and perf board for the electronics.
Once home from my excursion, I finished winding the secondary coil. I then took one of the stepper motors off of my CNC machine to spin the secondary while I varnished it (unsightly drips would not be tolerated!). Varnish was then applied. It would be 3 hours before I could give it another coat, so I went off to work on the control board.
The control electronics were the 'iffiest' part of this project. Having never personally worked with Conner's PLL scheme before (although I had used the PLL in other ways, like on Nimiety), and little time to debug, I was relying on this part of the project going off without any serious complications. Unfortunately, there was not enough time to design and mill a nice looking PCB, so I took the next best option: I would use some of the perf board I bought from RadioShack, which has the underside connected in a similar fashion to a breadboard.
Circuit construction started with the power section, and then quickly progressed to the staccato controller. Each part was checked as it was completed, to ensure functionality when the board was finished. For the most part, all of this went together with no major problems, although a few component values had to be altered for stable operation (the schematic should reflect this). The final few varnish coats were also applied to the secondary coil during this time (so that it would have plenty of time to dry over night). Laziness began to set in though, and I did not make any more progress on the circuit. On the plus side, I once again managed to get to bed early.
It was morning again, and I was back to work on the circuit. I first wired up the LM311 comparator (used for under-voltage lockout and monitoring the external push-button state) and then began on the PLL section. System noise immunity was of some concern, so I paid close attention to avoiding ground loops and installing plenty of 100nF decoupling capacitors near all of the ICs and along the power rails (these aren't all visible in the schematic though). With the PLL section half-way done, I diverted my attention to some more mechanical parts of the coil.
First, I needed to get an idea of just how big the acrylic enclosure would have to be. I placed the critical components (heatsinks, fans, control board, and low-voltage transformer) in what I thought to be a reasonable configuration. They took up approximately a 9"x9" area, with 4" of height (mainly from the fans). Since I wanted the top and bottom pieces of acrylic to overlap the sides for aesthetic reasons, they would need to be somewhat larger. Accounting for a 1/2" spacing around the edge, and thickness of the corner supports, I decided to make the top and bottom 11"x11". These pieces were carefully cut with a relatively unused 80-tooth blade on our radial arm saw to ensure high quality cuts.
Next, the corner supports had to be made. The case had been designed to have four corners with slots milled in the edges, so that the sides of the case could slide into them. Cutting the supports from the 1" square stock was not difficult. For the slots, however, I setup our router with a 3/16" bit (the thickness of the sides), and set it to cut about 1/8" into the material. There were some minor problems encountered while routing these pieces; the biggest being that the melted acrylic would tend to gunk up the bit and routed slot. As a result, several consecutive passes were required for each slot.
Finally the sides were cut. At the last minute, I decided to make them 4.5" high to allow for a little extra room beyond the fan height. I added a small amount of extra height to them as well, in order to make them slightly taller than the corners, so they would be effectively clamped between the top and bottom of the case. The case was now mostly ready for component installation, so I returned to working on the control board.
The remaining PLL components went together uneventfully, and testing commenced. First I powered the board with no feedback signal present. A clean square wave was found at the PLL's output, and adjusting the various potentiometers had the anticipated result of changing the output frequency. I then proceeded to connect a sine wave signal source to the PLL's input. After a bit of fiddling with the phase adjustment trimmer, the PLL locked onto the input signal just fine. Continuing with my sporadic work trend, I went back to the case at this point.
I first embarked on the task of cutting out the 3.5" holes required for the 92mm fans I was going to use. Not wanting to set the router up for hole cutting, I turned to a a suitably sized hole saw and the drill press. After thoroughly clamping one side down, I began the procedure. The first bit of drilling went relatively smoothly, but towards the end, the cutter began experiencing significant friction with the sides of the acrylic. It actually managed to stall the drill press' motor a couple of times. Despite this minor problem, the drilling was completed. The resulting edge was not particularly nice, so a little filing was done to clean it up some. Fortunately, it would be mostly invisible next to the fan. The second fan hole was cut with similar difficulties, but no other issues presented themselves. At this point, I was running out of energy, and decided to go to sleep.
On the final day before my departure, I woke up early and promptly went to work once again. I first began implementing the gate driver section of the control board. Having done this on many SSTCs before, this was a proverbial piece of cake. I installed some suitable driver chips, and ensured that the output waveforms were as they should be (which they were). Only one last thing remained, and that was to wire the input section for the PLL. This consisted of some diodes to clamp the PLL's input, a coupling capacitor and resistor, and a burden capacitor and resistor for the current transformer. Amazingly enough, all required components for the board fit just about perfectly in the allotted space.
More when I feel like writing it...
