[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Mutual Inductance & K Factor
Original poster: "John H. Couture by way of Terry Fritz <twftesla-at-qwest-dot-net>" <couturejh-at-mgte-dot-com>
Bart -
Thank you for the comparison data for Kurt's coil. I am thinking of changing
the JHCTES Ver 3.3 program to bring it up to date. It obviously needs to be
changed to conform more closely with the Mandk, Acmi, and Java9.1 programs.
Someday coilers will make K Factor tests using the frequency method and with
the mutual inductance test we should have more accurate K Factor values. I
downloaded the "kurt1" data with no trouble. However, my HP 4L printer had
trouble printing out more than one or two pages.
Your comment to pay attention to OD for each turn of Kurt's coil has made me
look into the JHCTES and Java programs in more detail. To compare these two
programs with the Kurt data it is necessary to reverse engineer the data to
conform with the input requirements of the programs. This can be a problem
in finding the inputs. For example I noted your primary wire diameter plus
the wire spacing gives the center to center (c/c) distance or the edge to
edge distance = .7867. However, which Width of turns should be used, 3.1203
or 2.725 for the 3.9663 Pri turns?
c/c = .3937 + .393 = .7867 inches
Width of turns = .7867 x 3.9663 = 3.1203 ?
c/c dist. = (OD-ID)/(2 x turns)
= (25.14 - 19.69)/(2 x 3.9663) = .6870
Width of turns = .6870 x 3.9663 = 2.725 ?
-----------------------
For the JHCTES avg radius
Avg radius = (OD-ID)/4 + ID/2
= (25.14 - 19.69)/4 + (19.69)/2
= 11.21
If the Width of turns is 2.725 the program c/c = .69 inches
This works OK with the 3.97 Pri turns and the 1.97 Pri to Sec clearance of
your data.
With one Pri turn the JHCTES gives Lm = 33.98 and K = .125 which compares
favorably with the .1244 of the Mandk and Acmi programs. However, when the
Sec is raised to 3.4" above the Pri the K Factor is .097 compared to the .68
of the Mandk and Acmi. The coupling reduces very fast with the Mandk and
ACmi programs. I would be interested in seeing this verified with the
frequency test.
John Couture
----------------------------
-----Original Message-----
From: Tesla list [mailto:tesla-at-pupman-dot-com]
Sent: Wednesday, March 27, 2002 8:41 PM
To: tesla-at-pupman-dot-com
Subject: Re: Mutual Inductance & K Factor
Original poster: "Barton B. Anderson by way of Terry Fritz
<twftesla-at-qwest-dot-net>" <tesla123-at-pacbell-dot-net>
Hi John, (fixed width font needed)
Those graphs at my site were run quite a while ago. I should run the numbers
again and update because there have been many rev changes since.
For now, I pulled up Kurt's excel file and ran the numbers fresh. I've
thrown
Acmi, JavaTC, MandK, and Kurt's actual measured values in tables below.
Note, the position in the old graphs and in Kurt's file does not indicated a
measurement (inches), but rather a set of parameters for each case. Could
just
as easy be position a, b, c, etc.
I used Kurt's measurements at 0", 3.937", and -3.937". Note, for JavaTC the
~4"
heights are moving beyond it's capability, but the 0 it is of course fine.
When looking through this, keep an eye on Acmi and MandK (how close they are
to
one another, and to measurement). Categories are Lm, K, and Lp (thought it
would be good to show Lp here): Each run is by turns as measured by Kurt.
Measurement column is actual meter measurements except K of course.
Height = 0" Turns Meas JavaTC Acmi MandK
Lm(uH) 1 37.14 35.83 35.87 36.07
K 1 0.1047 0.1410 0.1244 0.1244
Lp(uH) 1 2.31 1.19 1.51 1.54
Lm(uH) 2 74.83 71.59 71.69 72.09
K 2 0.1352 0.1447 0.1375 0.1380
Lp(uH) 2 5.62 4.52 4.96 5.01
Lm(uH) 3 111.78 107.31 107.45 108.04
K 3 0.1482 0.1476 0.1441 0.1448
Lp(uH) 3 10.45 9.75 10.14 10.23
Lm(uH) 5 186.11 178.54 178.77 179.74
K 5 0.1570 0.1516 0.1512 0.1528
Lp(uH) 5 25.8 25.58 25.49 25.41
Lm(uH) 7.2 264.2 256.58 256.88 257.76
K 7.2 0.1575 0.1541 0.1545 0.1546
Lp(uH) 7.2 51.68 51.16 50.39 51.06
Lm(uH) 9 333.16 320.17 320.5 322.2
K 9 0.1619 0.1550 0.1559 0.1557
Lp(uH) 9 77.79 78.68 77.03 78.62
Now, height = 3.397" (pri above sec):
Turns Meas JavaTC Acmi MandK
Lm(uH) 1 51.86 43.44 52.29 52.43
K 1 0.1462 0.1710 0.1814 0.1808
Lp(uH) 1 2.31 1.19 1.51 1.54
Lm(uH) 2 102.71 86.82 103.11 103.44
K 2 0.1856 0.1755 0.1977 0.1980
Lp(uH) 2 5.62 4.52 4.96 5.01
Lm(uH) 3 151.09 130.12 152.66 153.19
K 3 0.2004 0.1790 0.2047 0.2053
Lp(uH) 3 10.45 9.75 10.14 10.23
Lm(uH) 5 247.22 216.5 248.55 249.55
K 5 0.2086 0.1839 0.2102 0.2121
Lp(uH) 5 25.8 25.58 25.49 25.41
Lm(uH) 7.2 344.84 311.14 350.06 350.92
K 7.2 0.2055 0.1868 0.2105 0.2104
Lp(uH) 7.2 51.68 51.16 50.39 51.06
Lm(uH) 9 431.84 388.25 430.63 432.57
K 9 0.2098 0.1880 0.2095 0.2090
Lp(uH) 9 77.79 78.68 77.03 78.62
You can see how close Acmi and MandK are to one another, how far JavaTC goes
off with this 4" height, and how they compare to Kurt's measured. This
should
give a pretty good indication how well MandK and Acmi do against real world
measurements (even a coil of this size) and why programs like JavaTC are
good
for approximation (improvement needed, but a good start).
Now, height = -3.397" (pri below sec):
Turns Meas JavaTC Acmi MandK
Lm(uH) 1 21.22 27.97 19.45 19.7
K 1 0.0598 0.1101 0.0675 0.0679
Lp(uH) 1 2.31 1.19 1.51 1.54
Lm(uH) 2 43.29 55.89 40.26 40.71
K 2 0.0782 0.1130 0.0772 0.0779
Lp(uH) 2 5.62 4.52 4.96 5.01
Lm(uH) 3 66.38 83.77 62.21 62.84
K 3 0.0880 0.1152 0.0834 0.0842
Lp(uH) 3 10.45 9.75 10.14 10.23
Lm(uH) 5 116.29 139.38 108.94 109.86
K 5 0.0981 0.1184 0.0921 0.0934
Lp(uH) 5 25.8 25.58 25.49 25.41
Lm(uH) 7.2 169.34 200.31 163.63 164.51
K 7.2 0.1009 0.1203 0.0984 0.0987
Lp(uH) 7.2 51.68 51.16 50.39 51.06
Lm(uH) 9 221.54 249.96 210.28 211.71
K 9 0.1076 0.1210 0.1023 0.1023
Lp(uH) 9 77.79 78.68 77.03 78.62
You brought up something of interest in a separate post involving frequency
affects. I can't comment on loaded coils because I haven't measured this. It
would be good to do this both at line Fr as well as high frequency near the
coils loaded Fres to see if differences pop up (including your Freq
measurement
method). Maybe someone has already made a similar check on loaded coils or
high
frequency measurements?
John, I put Kurt's coil as -kurt1- in JavaTC. You should be able to pull it
up
now (refresh the browser if it doesn't work the first time - this will show
you
my inputs - pay attention to OD for each turn of Kurt's coil - this is an
all-important parameter if you model in JHCTES).
Take care,
Bart