Frist light achieved; is there an accepted guideline on properly optimizing/tuning a VTTC? / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Frist light achieved; is there an accepted guideline on properly optimizing/tuning a VTTC?

1 2 3 next

Move Thread

LAN_403

Cesiumsponge

Mon Jun 01 2009, 08:27AM

Registered Member #397 Joined: Wed Apr 19 2006, 12:56AM
Location: Western Washington
Posts: 125

I seem to be encountering some issues. Is there an accepted guideline on properly optimizing and tuning a VTTC and what general chronological order is required? From what I could dig up...there is no real definitive answer. VTTCs seem to be fairly open-ended and it seems like some of the more tenacious builders spend a large amount of time (days, weeks) to optimize these which leads me to believe hard and fast guidelines are far and few in-between.

This is with my current project. Right now I have a 700pF grid leak capacitor and an adjustable grid leak resistor which can give me 1k-9kohm overall (its around 4kohm right now), and a sliding feedback coil with 25 turns. These values are fixed for the sake of attacking one variable at a time. I don't want to change multiple variables simultaneously.

Right now I'm trying to tune my primary which is a disposable coil I tossed together (I will put in a nice looking primary when I get it tuned) consisting of 35 turns, the top 10 turns being tappable. They're whole turn taps right now; I haven't had time to strip insulation down at half turns or smaller increments. I've noticed that the better tuned it is, (30 turns seems to be the peak), the more current it draws and the longer/hotter the output gets at a given input voltage which is good, but I don't seem to get normal VTTC streamers past 100V or so. The output loses the linearity and sword-like appearance of the classical VTTC. The streamers start branching out, and the sparks are much hotter, a ghost white rather than a faint purple with a brighter core. The sound also changes but it's difficult to describe. It becomes less hissy and more bassy. This is also about the time the plate starts creeping towards glowing and I can't sustain this mode for more than 10 seconds or so without fear of risking the plate integrity. The streamer length around this level is around 16.5-17" which is about max for a single 810 tube from what I've seen so I don't know if there is much to see past 110VAC.

I have also experimented with the feedback coil location on each of the changes. I do notice that I get a 2A drop in current consumption by lowering the grid feedback coil closer to the primary, so I have done that.

Someone more familiar with VTTCs might be able to diagnose what I'm seeing based on the streamer appearance:

If I never peg the variac to run the coil at indefinite run times, then I'll never run into this problem, but it seems like a shame not being able to use the full range of the dial. I only have two MO caps used for the voltage doubler. This issue might go away if I start experimenting with the feedback coil or the grid leak resistor value, but I don't want to go about changing another variable just yet. I'd rather ask those with more experience if what I'm seeing is normal.

I've tried multiple permutations so far. That is, winding the primary clockwise and counterclockwise, then swapping the two coil connections (ie HV end goes to plate, plate end goes to HV instead). This gives me four permutations. I've also tried reversing the leads on the feedback coil which gives me two permutations. Eight in total with the primary and feedback together (I have not wound a feedback coil in reverse yet, which would give me 16 permutations)

I have a nagging feeling that if I tune the primary for longest streamer length and move onto, for example, tinkering with the feedback coil, then the tap on the primary might need to be moved to accommodate the change on the grid feedback coil. It sounds like I might need to mess with multiple variables simultaneously to find best performance. Is it possible I can't have both longest streamer formation AND sustained cool plate operation? It might be possible that I have to make a compromise somewhere. Any advice would be greatly appreciated. This thing is so close to done...I can feel the tug of freedom that'll allow me to chain myself to another project.

Steve Conner

Mon Jun 01 2009, 10:13AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Those sparks look fine to me, and I'd say 17" was a pretty decent result, it suggests >1kW peak power.

What is the resonant frequency of your setup? If it's too low, it tends to form those bushy sparks. I think I remember Steve Ward saying that you want at least 500-750kHz to get nice swords.

Yes, the variables all need optimized together. You can decouple grid current from the rest somewhat, though, by just using a grid current meter and setting up your feedback coil for the datasheet recommended value every time.

Yes, you have to compromise. The more power you want, the hotter the plate is going to get. I think the 810 is one of these tubes that can run happily with the plate cherry red, probably the greatest danger is blistering the varnish on your oak

Dr. Dark Current

Mon Jun 01 2009, 10:46AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Forgive my ignorance, but is the HV PSU halfwave or fullwave rectified? If its full wave, that could account for the branchy discharges.

Arcstarter

Mon Jun 01 2009, 03:17PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

I'd say that is a fairly decent result, man! But i know i am never satisfied either until the max length/minimum heat is achieved at full power.

But for what it is worth, you sometimes need to change a few variables at a time. If you change the primary turns or tank capacitance, that will change feedback. If you change the tank capacitor, your primary might need a new tapping point. If you change your topload, you might need to change feedback, tank capacitor, and primary tap! This is only, of course, if you want absolute optimum performance, but if you spent so much time on the thing itself, im sure awhile of tuning and experimenting would not pose much of a challenge

Also, that tank capacitor sounds a bit small. For different frequencies and inductance this does change, but maybe you should focus on that. In my single 811A lash ups that seemed to make the biggest difference right next to primary tapping. But as you said, you seem to have pretty much maxed the spark length on that tube. But im not really sure.

Good work!

Cesiumsponge

Mon Jun 01 2009, 06:18PM

Registered Member #397 Joined: Wed Apr 19 2006, 12:56AM
Location: Western Washington
Posts: 125

I am running a level shifter so its only halfwave rectified output.

It sounds like with VTTCs, there is no real trend and the ultimate answer is "your mileage may vary". It just seems hit-or-miss that one has to shift multiple variables simultaneously until one finds a happy medium. I'll have to check up on the 810 operational procedures. I was under the impression they didn't want to be run with a cherry plate so I've avoided it so far.

This is the envelope I'm getting at the 60hz feed.

This looks to be the Fres, roughly 389kHz.

Funny thing is I get the classic sword-like streamers at lower powers but once I crank it up, the sparks get angry and start loosing that linearity. Its also possible that it just doesn't look sword-like because the streamers start getting really fat and turn from a well defined purplish color to a ghost white, especially near the base. When they're faint purple, they seem to have a more refined "edge" to the spiral plasma structure that curls up the central spire. When it starts getting ghostly white, it appears to be more diffuse.

*edit

I guess I verified I'm still getting streamers with the classic sword-like structure at higher power levels but it's just diffused in appearance and they're a bit more erratic. I can see the same structure.

70VAC, 3A draw. 4.5-5" streamers.

80VAC, 4A draw. 6.5-7" streamers.

90VAC, 7A draw. 9.5-10" streamers.

100VAC, 11-11.5A draw (needle dances). 15.5-16" streamers

At 105VAC, it'll hit the 17" grounded foil sheet suspended above the breakout point and draw 13-13.5A

Here is the coil under lower power. It seems the structure is best defined at lower power levels. I'm having a hard time photographing streamers because I don't want my SLR too close or it'll start getting confused, I'll get RF burns, and my 200mm telephoto is too slow wide open, but being wide open on my aperture also kills my depth-of-field.

Proud Mary

Mon Jun 01 2009, 07:37PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Almost all the VTTC designs I have seen here and elsewhere could be described as no more than a special case of an Armstrong oscillator, and no calculation beyond those needed for this type of valve oscillator are really required.

The valve is set up to function within the design centre values defined in its data sheet (or much more if you are typical HV.orger and don't mind the life of the valve being shortened by overdriving

I'd start with a book of basic thermionic valve theory - learn how the things actually work - and then all mystery will disappear from your eyes. Without even the most basic understanding of thermionic devices, you'll have to resort to copying someone else's circuit, and blundering from one arbitrary decision about component values to another.

As with all engineering decisions, how you design something is dictated by the application. What is the circuit intended to do, and how can it best be implemented in the most efficient, reliable and cost-effective way?

Look at the data sheet of a popular VTTC power triode like 811A - why is it telling you all this stuff? What are those graphs all about?

Good Luck!

teravolt

Mon Jun 01 2009, 07:48PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

It sounds like you have the same problem as DR. Spark. With no toride Your VTTC will detune with longer sparks as shown in your second o-scope picture. If you have not read Steve Wards VTTC guidlines page already it is usefull. You probibly have frequency spliting going on so if you build the secondary on the high side and then put a small toride on the top that will help your sparks from detuning your secondary Fo so bad. I think VTTC's should be treated like DRSSTC's to some extent depending on how much cupling there is. DR. Spark's Quadzilla also has had the same problems partly with his tesla as far as I can observe. I am hoping he will chime up on the mater on Quadzilla.

Cesiumsponge

Mon Jun 01 2009, 08:09PM

Registered Member #397 Joined: Wed Apr 19 2006, 12:56AM
Location: Western Washington
Posts: 125

I have a small John Freau 6" toroid on it. I played with the idea of no toroid and putting one on. Seemed like the sparks were more linear with a toroid so I tuned for that. TC CAD predicted 411kHz without a topload and I measured 404kHz. I'll see what Dr. Spark has to say if he chimes in. I checked out the 810 datasheet and unfortunately, my plate voltage isn't even on the graph since I'm using a typical voltage doubler. I'll do some digging into tubes but a lot of terminology seems obscure or dead...things like filament quadrature excitation.

If it's detuning with longer streamers (which seems plausible), I'll try detuning it in either direction and see if it positively affects streamers in the 90-100V range. Might not do a thing but worth playing with. At this point all the hard construction work is done anyhow. It's already met my expectations, but squeaking a bit more out of it is always a nice bonus.

Herr Zapp

Mon Jun 01 2009, 08:23PM

Registered Member #480 Joined: Thu Jul 06 2006, 07:08PM
Location: North America
Posts: 644

CS -

Relative to the "diffused" appearance of your sparks at higher power levels, is there a possibility that this is really an optical artifact (related to insufficient depth of field) originating with your camera adjustments?

In the crisply-focused "coil under low power" photo (flickr title "spiral_structure"), the absolute diameter of an imaginary cylinder that would contain all the sparks is relatively small, and the camera's "depth of field" required to render all the sparks clearly focused is relatively small.

In the photo showing the "diffused" (or simply out-of-focus) sparks at higher power levels (titled "100vac, 11.5A draw, 16" streamer"), the diameter of the imaginary cylinder containing all the sparks is clearly quite a bit larger. If your camera focus and lens aperture were not carefully set to "straddle" this cylinder diameter, many of the sparks will be out of focus, creating the diffused appearance. If you look carefully at this photo, there are thin purple sparks in the center (or towards rear) of the cylinder that appear to be in much better focus than the brighter sparks.

Were these photos taken with a digital camera set on "autofocus"?

I have had similar out-of-focus results photographing Tesla coil discharges with digital cameras. Quality was improved when the camera was set to "manual" focus, and the camera's nominal focus point was carefully pre-set at the hyperfocal distance.

Another observation I have made is that conventional SLR 35mm film cameras, regardless of film speed or type, seem to be able to capture far more of the thin, filamentary sparks & streamers than digital cameras do. This may be due to some "threshold effect" inherent in CCD detectors. Also, conventional SLR film cameras generally provide more control over lens aperture and the resulting depth of field, and most quality lenses will have the depth-of-field scale directly engraved on the lens barrel.

Regards,
Herr Zapp

Dr. Drone

Mon Jun 01 2009, 09:30PM

Registered Member #290 Joined: Mon Mar 06 2006, 08:24PM
Location:
Posts: 1673

1 2 3 next

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