[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Tesla coil formulas



Original poster: "Barton B. Anderson" <bartb@xxxxxxxxxxxxxxxx>

Hi Terry,

Your reduction is perfect. Yea, I posted it to the list a couple of times during relative inductive discussions, but higher level programs were on the rise, so...

Take care,
Bart

Tesla list wrote:

Original poster: Terry Fritz <vardin@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Bart

EXCELLENT!!!

I never knew there was a fix to the old equation!!! It looks like it fixes the old high error problems very well!! I simplified the equation a little and I hope I did not mess it up.

The latest and now pretty finished file is at:

http://hot-streamer.com/temp/FormulasForTeslaCoils.pdf

The original base file is at:

http://hot-streamer.com/temp/FormulasForTeslaCoils.odt

Cheers,

        Terry


At 10:38 PM 12/2/2005, you wrote:

Hi Terry,

Tesla list wrote:

Original poster: Terry Fritz <vardin@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Bart,

I was going to put the cone formula in but I got worried from this post:

http://www.pupman.com/listarchives/2001/January/msg00418.html


yes, exactly! The error was unacceptable from a programming standpoint (meaning, when the error gets too high, "what's the point" in programming. This is why I started working with the equation in the first place.

It sounded like there was some question as to if the formula was valid. Has you tested it against Paul or Mark's programs?? If so, then I will put it in. If not, then maybe it needs to be checked?
I added it anyway....


I see you put in the "original" formula?. Ahh, my fault, I didn't include my factor in the puzzle. Geeze, I hate when I do that! Sorry, here it is:
My factor is 1/SQRT(sin(x)+cos(x)) and the whole beast should look like this:

L = SQRT[(L1*Sin(X))^2 + (L2*cos(X))^2] x [1/SQRT(sin(x)+cos(x))]

The difference is, well, let me show you with an 8"ID, 32"OD flat coil using .25" tubing at 24 turns.
I'll raise the outer height up in 5 degree increments maintaining 24 turns.

Goetc-L    Orig-L   Orig-err    Bart-L    Bart-err    Angle   OD
---------------------------------------------------------------------
270.6      271.7     0.40%      271.7     0.40%         0     32
269.6      274.9     1.93%      264.1     2.04%         5     31.909
267.9      280.7     4.56%      260.8     2.65%        10     31.635
262.1      285.7     8.26%      258.2     1.49%        15     31.182
255.5      288.0    11.28%      254.5     0.39%        20     30.553
247.3      286.9    13.80%      248.9     0.64%        25     29.751
237.4      282.0    15.82%      241.3     1.62%        30     28.785
226.0      273.3    17.31%      231.6     2.42%        35     27.660
213.4      261.1    18.27%      220.0     3.00%        40     26.385
200.0      245.9    18.67%      206.8     3.29%        45     24.971
185.0      228.0    18.86%      192.1     3.70%        50     23.427
169.2      208.0    18.65%      176.3     4.03%        55     21.766
153.1      186.6    17.95%      159.7     4.13%        60     20.000
136.7      164.2    16.75%      142.4     4.00%        65     18.143
120.2      141.4    14.99%      124.9     3.76%        70     16.208
104.0      118.8    12.46%      107.4     3.17%        75     14.212
88.0       96.9     9.18%       90.0     2.22%        80     12.168
72.8       76.4     4.71%       73.4     0.82%        85     10.092
58.6       59.1     0.85%       59.1     0.85%        90      8.000

This is a prime example of the error with the original equation. At only 10 or 15 degrees, the error is already up to 10% and climbs. I chose to use the angle itself for the factors reference point. It simply sharpens the inductance closer to reality and gets the error down to a usable and acceptable level. Note, this equation is fine for all 3 geometry styles.

Sorry for not including the factor when I first posted (I bet you were scratching your head on that one ;-) ). Of course, with the new programs, "out with the old and in with the new". But, for a basic differential, it's good.

I am trying to hit the basic, often used, equations that one really "needs". It won't have toroid-plat-sphere capacitance or Medhurst stuff since that really needs the computer programs now that we have good programs to figure it out right... I have not used those since the E-Tesla days...


Agreed! It's one thing to calc an objects C, but quite another to distribute it correctly.

I will put in surge impedance. "I" have never used it for much but others like it ;-))


Often it's been used in postings by various members and that was the only reason I threw it out there. So people could throw a number to it if wanted. I personally have only looked here and there to get a feel for it, nothing more.

The new document looks great! Good work.
Take care,
Bart