Impedance Stuff++ - was Thinner wire and more turns

Hi All and Richie,

	First, I am sorry for suddenly being so quiet on this impedance thing.  I
got swamped at work with another "impedance matching issue" that took all
my time for the last few days.  I will explain how I find output impedance
in another post ASAP!  It is rather involved and I didn't want to just
throw it out without some thought as to making it clear and without time to
respond to questions...

	Second, we (Chip and me a bit too) have been having some problems with the
list connections the last few days.  So, if your mail got bounced or you
didn't get mail for a day (15th), it was on our side.  I don't think any
mail got lost, just delayed.

	We now have 700 members!!  NEAT!!!  So if all 700 of us send a post,
Chip's PC will send out 490700 E-mails to his ISP over the modem! :-))
Just joking Chip!!   Of course, we all need to be sure and snip old parts
of posts and all that, especially in these times of connection problems, to
keep the servers from getting needlessly bogged down with excess text to

	As usual off topic stuff, file attachments, and "goofy stuff" will be
rejected at the moderation phase (The Tesla list posts are "approved" by a
moderator before being posted to the whole group).  Since we have so many
new members, I should probably post the "fuzzy rules" again even though
everyone has been very good these days :-))...  I (Terry) also "happen" to
be the moderator for the list these days (for those that don't know).

So, on to the REAL stuff!!!

	Someone asked about what the estimated impedance of arcs is.  I use 220000
ohms real in series with 1pF per foot of capacitance.  Recently, I have
just assumed a 5 foot arc to get 5pF.  Boris mentioned 1.5 pF per foot but
that is as good as any number without much more measurement.  I wonder how
multiple streamers and such affect that load too.  Now that the impedance
issue is "hot", more detailed and careful measurements are needed.  Since
my fiber optic probes are now widely available, maybe I will not be the
"only" one doing this ;-)

More below...

At 05:28 PM 11/18/1999 +0000, you wrote:
>Hi John and all,
>For a while now I have thought that the efficiency of my 4" coil may
>have been limited slightly by the low coil diameter.  My spark lengths
>have always come in slightly under John's results for the same input
>power.  I was puzzled as to why this occured,  but always put it
>down to John's greater coiling expertise :-)
>John's most recent post about the increased inductance of using more
>than 1000 turns was the final straw,  and yesterday I decided to wind a
>coil of identical dimensions (w=4", h=22") to my original but with
>1500 turns instead of 1000.  ie. Thinner wire.
>Original coil:  4x22"  1017T  -at-  24swg (23awg 0.56mm)  Ls=16.8mH
>New coil:       4x22"  1503T  -at-  28swg (27awg 0.37mm)  Ls=43.1mH

When I got into this impedance issue, I was going to wind a bunch of coils.
 However, you and John have pretty much done that work for me ;-))  John
used less turns which should be "bad" and was.  You used more turns which
should be "good" and was.  So this impedance thing seems to be on track.

In your "old coil", the load would have had a resistance of around 264k
ohms and the output impedance was 105K ohms.  With "my" other numbers, that
would deliver 161 watts to the arc.

In the "new coil", the load would have a resistance of 293k ohms and the
output impedance would be 209k ohms.  That would deliver 225 watts.

So you get 40% more streamer power.  Realize that these numbers have a lot
of guesses in them but the "idea" should be right.

>The original coil resonated at 218kHz whereas the new one resonated at
>164kHz.  In order to get the primary to tune to the new frequency I
>had to remove one toroid and move the primary tap out considerably.
>I left the power supply,  rotary phase, tank cap, ballasting etc.  all
>the same,  to keep power the same for both tests.
>With my original coil I got mostly one streamer, and occasionally two,
>with frequent 36" strikes to a ground rod with 1020 Watts of measured
>input power (200BPS sync rotary.)
>I videod both tests for examination later in the house (in the warm !)
>When I first powered up the new coil,  the difference was immediately
>1. Firstly there were 5 or 6 simultaneous streamers from all around
>   the top toroid.

Probably due to more streamer power and a smaller toroid.

>2. I got frequent hits to ground at 36" and got a few rare hits to
>   other objects and the garage walls slightly further away.
>   Sole streamers were generally longer than one of many.

With 40% more power one may guess that you streamers would be 18% longer
but changing the toroid is affecting this also.

>3. With the new coil breakout started at around 30% on the varaic
>   compared to around 50% with the old coil. (Higher toroid voltage ?)

Probably sooner breakout at lower voltage due to the smaller toroid, but
some due to better power throughput too.

>4. The rotary gap seemed quieter and not as bright.  Usually it lights
>   up the whole garage.  Even the initial spitting sound was not as
>   sharp on the ears.

If the power input is the same, the gap is wasting less power and it should
be putting more power into the arc.  Typically, spark gaps eat about 50% of
a coil's energy.  Ideally, the gap would be silent, dark, and run stone
cold....  We ain't there yet...  Although, the CW coil people and their
class E RF drivers and zero cross switching stuff may get darn close!!
Since I am involved "professionally" with advanced class E high power RF
designs, I have a lot of proprietary knowledge of such things.
Unfortunately, that also prevents me from getting involved with that here...

>5. I noticed a rapid build up of ozone with the higher L coil also !
>   Maybe because of more simultaneous streamers,  or higher voltage ?

John mentioned that rough toroids make much more ozone.  Is your small
toroid rougher that the old larger one?  Perhaps the higher voltage on the
smaller torroid is causing more corona.  

BTW - I have heard that corona is really "HOT" and can be seen with IR
cameras.  My new Sony (TRV36) camcorder has all that night vision 0.4 LUX
stuff so I wonder if it can see such things??  It is fascinating that it
can see MUCH better in low (no) light than I can and I wonder how that
ability may be put to good use...

>6. Spark action was quite frantic.  With rapid darting and surging
>   of sparks on either side of the toroid.  With the original coil I
>   usually got one comparatively lazy evolving streamer. (This frantic,
>   chaotic spark action reminded me of my old static gap coil with a
>   very high break rate.)
>Unfortunately,  I had to remove one toroid from the coil to bring the
>higher L coil into tune,  so this is not exactly a "controlled
>experiment".  I don't know how much this affects performance. The maths
>says that increasing Ls and decreasing Cs should produce a higher peak
>terminal voltage.

Higher frequencies need lower output impedance.  As you decrease the
secondary capacitance and increase the secondary inductance the output
impedance goes up (higher voltage, lower current).  In most of our cases,
we seem to have a fairly low output impedance driving a high arc impedance.
 Perhaps this is better for multiple streamers??

>I dont know if the change in performance is due to the lower operating
>frequency, higher theoretical output voltage,  higher pri surge imp.,
>better match to the streamers, higher coupling, or maybe just due to
>having one less toroid,  but the improvement was quite dramatic.
>I plan to carry out further investigations with this new coil tomorrow.
>In particular I plan to add back the third toroid,  and increase Cp
>to regain tune. Hopefully I can get one sole streamer of great length !
>Despite adding to Cp I will keep the input power the same to do a fair
>comparison of the two coils.  (Then I will crank it up higher to take
>best advantage of the higher Cp !)

If possible also try a still smaller toroid, Just to see what it does too.

>I know little about matching the output impedance of the sec to the
>streamer load imp.,  so I cannot add much to that discussion.
>However, from an engineering viewpoint,  impedance matching comes up
>everywhere like RF amplifiers, audio amplifiers etc.  I would think it
>would be just as important for Tesla Coils in order to acheive maximum
>power transfer to the "load".

I am sure it is!  Tesla coils do not put out nice sine waves so it is far
more complex to figure out impedances, but the principles seem to still
hold just as true.  I use "average" impedance but the "real time - dynamic"
impedance may need investigation too...  thinking about that makes my brain

>Also, since small diameter coils effectively have their inductance
>crippled because of the geometry,  I see no reason not to add a few
>extra turns in order to increase inductance and ultimately let them
>fair better against larger diameter coils.

It seems to me like we need much more inductance in many cases along with
smaller toroids.  However, small toroids can reduce spark length too due to
the electrostatics involved.  So we just can't blindly tune to what the
calculations say without taking into account the voltage standoff and field
stresses.  Also, the secondary self capacitance energy may have to go
through some inductance to reach the streamer which adds another twist to
it all.  I just barely got started on trying out some new secondary models
to study that before work hit.  It is interesting that thinner and longer
secondaries my have some advantage due to their lower self capacitance...



BTW - Weather very nice in CO, haven't even got out winter coat yet...

>Comments, suggestions, flames always welcome.
>					- Richie,
>					- Freezing in a cold garage
>					  in Newcastle.