Terry's great theory of arcs #85692875 / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Terry's great theory of arcs #85692875

1 2 next

Move Thread

LAN_403

Terry Fritz

Sun Nov 12 2006, 12:54AM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

From an off-list thing...

Terry's great theory of arcs #85692875 of this moment. Subject to change the "next moment" :o)) It's all "pie in the sky", but it should not violate known observations other than the constants are not yet determined...

So here goes...

I think arc length is governed by "voltage/length" and "current/length" constants along with a "fixed resistance - independent of streamer length". For the sake of argument I will define them to sort of close values as:

Streamer Arc Voltage = 250kV/m
Streamer Arc Current = 2 A/m
Streamer resistance = 125kOhm

That's it!! Done!!! :o))))))

But let's look at it...

Streamer resistance is constant regardless of streamer length. From modeling and testing, going into 5 years now on many odd systems, that seems "true"....

To get twice the streamer length, you need four times the power. At 1 meter we need 250kV x 2 A = 500kW peak. At 2 meters we need 500kV at 4A = 2MW peak. "Condition satisfied"...

Suppose our coil produces only 250kV, but 50 amps and 50kW. Streamer length is still limited to "1 meter" (!) No matter how hot or how many little streamers are flying off it, we only get 1 meter... This is similar to a coil with two foot arcs running from 60 BPS to 6000 BPS with no change in streamer length... You have not got the voltage, so you only go 1 meter - no matter how many kBPS you throw at it... My DRSSTC does happily go from 60 to 6000 PBS with not much change in streamer length. The streamer sure does get "hot", but not longer... "This observation is satisfied"...

Lets go for giant voltage with no top terminal!! At the secondary voltage peak, the top current is nearing "zero"(!!).... We still have the coil's self capacitance and space capacitance of the coil's bare top to drive current. But "not much" current. As we draw power, that much smaller capacitance's voltage will drop like a rock and suddenly we don't have the needed voltage anymore... It might get off to a nice "short" start, but the streamer will starve for current and the voltage will drop out. We are saved a little as the cycle can supply "direct" current "through" the secondary's inductance, before and after the top current null... But we are starving for current and streamer length will starve too... I think this is what happens to Electrum, not enough big top load capacitance to store the current supply needed for longer streamers without loosing terminal voltage. This seems to be shown in coil's that do better with bigger top terminals that are not "too big" in that they kill top voltage. - "Seems true".

There are fun "complications"...

Streamer (leader) capacitance is dynamic and seems to be about 6.56pF/m. Leader branching gets "messy", but tends to coalesce as "one". Streamer dynamic capacitance converts current to voltage loss as the streamer grows. Not a big problem, we used to use fixed values for it (~1pF/foot). But dynamic computer simulations tell us "it's not 'fixed' dummy"... But if we guess good at the coefficient, the programs are happy and they love crunching all the numbers in 1nS increments anyway.

"Load Energy Rise Time" - A new term some may not know, but it is how fast you convert primary energy to streamer energy (officially, 10% to 90% timing of top terminal energy rise). Generally observed as "the faster the better". Less time for primary gap losses to steal energy and fewer "ring ups" that might waist power due to lower voltage and current pulses observed in very recent high speed photos. If you make a bright leader that only makes it 90% there, something has got to be "waisted"... High coupling minimizes lower power leaders, and their loss. The DRSSTC folks push coupling as high as possible for better and better streamer lengths... The SISG also like high coupling... I don't think LERT is a goal, but rather a "symptom" of other things. If you do "everything else right", LERT will be perfect.

"Ground arcs" - Used to be messy, but the programs figured that out as a massive very high speed (10nS!!!) top terminal drain to ground. It sustains until the top terminal is drained of energy and the secondary is not longer able to maintain the "comparatively feeble" secondary current to sustain the arc. When the arc dies, there may be energy left in the primary to again partially re-energize the secondary system. The ground strike models seem to predict ground stick observations perfectly!!! (Thanks to Dan M.

) The speed and power of ground arcs bring up many scary new "questions" though

Stand back "another" 10 feet...

So until theory #85692876, that is my take...

I really don't know the "constants"... Theory #85692874's constants are too buried in ambiguities to "simply" convert... But maybe with a more clearer view, that will be simpler. Supper time now here, so c'ya

)

Cheers,

Terry

ragnar

Sun Nov 12 2006, 01:13AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Theory #85692875 makes a lot of sense.. I like it. So how do I apply this to CW coils?

Tesladownunder

Sun Nov 12 2006, 01:42AM

Registered Member #10 Joined: Thu Feb 02 2006, 09:45AM
Location: Bunbury, Australia
Posts: 1424

Interesting stuff.
With ground stikes the streak camera photos (remember those!) showed that some times there was ringdown visible and sometimes not in the same run. It was generally but not always visible on the LED's though. Is this due to variable energy shots from the asynchronous gap or is something else going on.
Similarly ring up is sometimes seen well and sometimes the ground strikes just connect. Is this just the variability with a chaotic sytem?

Peter

1163295757 10 FT17879 Hvrotmirrorteslalargearrestorsmultisparkssb

Terry Fritz

Sun Nov 12 2006, 02:22AM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

So how do I apply this to CW coils?

Oh sure!! Pick on the "one" weak spot!!

The ScanTesla program is meant only for disruptive coils. But if one used a DRSSTC case with a really long dwell time, then it is CW!!! But it does not have a 60Hz drive component like most tube coils (not super hard to add though...) Calculation time would be prolonged as it got over the transient response... But not a big deal...

I have not though at all about the "theory" in the CW case though... A ripe field of study if the other theory seems to hold as valid.... Tube coils have those odd sward like streamers... Might be a "second generation" thing after the CW case seems figured out to extend things to there.

showed that some times there was ringdown visible and sometimes not in the same run.

See this odd background paper:

Once the terminal and secondary are drained of energy, the shorts goes off. But if the primary still has energy left, it can all light up again to some extent. That is all much by chance if it happens or not.

It is very interesting in that the arcs and leaders seem to be "fast" events that show up as quick arcs in the high speed photos. They seem to be "jump" events.... Not sure how all that figures in...

Cheers,

Terry

Finn Hammer

Tue Nov 28 2006, 09:37PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Terry Fritz wrote ...

So here goes...

I think arc length is governed by "voltage/length" and "current/length" constants along with a "fixed resistance - independent of streamer length". For the sake of argument I will define them to sort of close values as:

Streamer Arc Voltage = 250kV/m
Streamer Arc Current = 2 A/m
Streamer resistance = 125kOhm

That's it!! Done!!! :o))))))

Very interesting, Terry.

I am trying to model a big triggered SISG coil in Scantesla V.8.1, and would like to use this new streamer model-
But how can I access streamer arc current?. I cannot find it in the output.txt, but need it to see what I am doing wrong, because I have topload voltage enough for more than 3 meters of spark, but estimated leader length is a measly 84" with 12kW input.

My design method is this:
Chose Fres as 50kHz
Choose Vpri based on 3 MOT`s in star, with resonant charging 9kV (parallell MOT as needed)
Ipri based on a known good case with CM600`s 2500Apeak (could be higher, but playing safe for now)
This condition is met with 1ÂµF/10ÂµH primary

Chose a 10x50 secondary with a 10x40 topload
This gives a Lsec of 137mH
300BPS
input.txt

If you feel like it, it would be great if you could tweak this input for the best streamer length, and tell why, so that this could become an online tutorial in using Scantesla.
Cheers, Finn Hammer

Steve Conner

Tue Nov 28 2006, 11:42PM

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

Hi Finn

I don't know about ScanTesla, but according to my mental theory of arcs, the resonator isn't loaded heavily enough. I'd try increasing the secondary inductance quite a lot, maybe a factor of 2, and retuning the primary to match the new lower frequency. Especially if the waveforms showed several notches before the energy was used up. (I forgot if ScanTesla shows waveforms, or calculates some other number that tells how heavily the sparks load the resonator.)

Also, for a good spark length-to-power ratio you want to be running your 12kW as 100 120 joule discharges per second, not 300 40 joule ones. I learnt the hard way that the big bangs are better :(

Terry Fritz

Thu Nov 30 2006, 10:11PM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi Finn,

I am trying to model a big triggered SISG coil in Scantesla V.8.1, and would like to use this new streamer model-

ScanTesla 8.1 uses the "old" streamer model

I meant to update it, but I got way behind on things and have not yet. Sorry about that...

But how can I access streamer arc current? I cannot find it in the output.txt

Ohhhh... I guess I forgot that one

If you turn one waveforms.csv you can calculate it with a little math... But I need to fix that so you have streamer current and RMS streamer current easily. A flaw...

I don't like V8.1's streamer length at all right now, thus the newer theory, yet to be implemented.

Thanks for the input file and data. It helps me work a real world example of what you need!

Steve said:

for a good spark length-to-power ratio you want to be running your 12kW as 100 120 joule discharges per second, not 300 40 joule ones. I learnt the hard way that the big bangs are better :(

That seem to increase both voltage and power. I think there is a balance or "right proportion" there. I have no idea where that is at the moment. But if I can get tools to study it and a "theory" to go by, I can start looking

I'll see if I can get ScanTesla fixed ASAP! Hope hope...

UPDATE!!!!!!!!!!!!!!!!!!!!!!!

Hi Finn,

Update your ScanTesla 8.10 files with these new files for V 8.12:

This will allow you to see a lot more streamer current stuff!

The "streamer length model" is still "old"...

I tried to update it but it was too "wild" for the time allowed... But I saw wonderful things

Things I cannot "control" at the moment. But they seemed very "TRUE"...

You can fiddle with the factors in "control.txt" if you want to get the "right" streamer lengths("LeaderLengthVsecFactor" and "LeaderLengthPowerFactor"). That should work "good"...

I am using "streamer power" in the input.txt file for a "goal" now since it is better with a poor streamer length calculation that the program now suffers...

I also increased the range of the primary inductance since the best point is a higher inductance...

Hopefully, in a few more years

, this program will work "smooth"

It is "ultra powerful", but pretty hard to get "control off" too

But t seems to "guide" one to the "truth" about all this since few things are really "controllable"... It seems like it all "MUST" work just "ONE" way... The problem is finding that specific "one true way"...

Cheers,

Terry

Finn Hammer

Fri Dec 01 2006, 10:22PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Thanks a lot, Terry

Here is the input file of the BRISG

and here is a screenshot of it running

Notice the low Rpri, I`m sure you already spotted the 3kA Ipri

I will measure the Vpri initial tomorrow, I think it might be lower than 1000V

Another fun toy.

Cheers, Finn Hammer

HV Enthusiast

Sat Dec 02 2006, 01:22AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

A little more complicated than my theory.

Dan's great theory of arcs #664333

"They hurt like hell!"

Finn Hammer

Sat Dec 02 2006, 02:08PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Talking about "Hurt like Hell" Here is the result of a 5kVDC burn.
I now know for a fact that the arm that is attached to HVDC is denied voluntary movement. It sticks to the live circuit.
During the first frantic second I realized this. Next I felt a strong urge to remain alive, and so used the rest of my body to pull the arm away from the primary.

Lesson learned: Don`t touch an energized coil.

Cheers, Finn Hammer

1 2 next

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