Re: acmi error found? - Test Equipment!?

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Terry Fritz" <twftesla-at-uswest-dot-net>

Hi All,

The error may be (probably is) the test setup!

Referring to:
http://hot-streamer-dot-com/TeslaCoils/Misc/TerryVSacmi.gif
http://hot-streamer-dot-com/TeslaCoils/Misc/TerryVSacmi2.gif

Here is the testing I ran tonight.

1.	I removed the grounding foil off the floor (sort of an accident really
:-))  This raised the primary inductance from 210.467uH to 211.688.
Probably eddy currents in the foil were having a small effect.  Note that I
am carrying digits beyond the real accuracy so percentages and other
computations are not fouled simply by too few digits.

2.	I disconnected the top meter leads with no change.
http://hot-streamer-dot-com/TeslaCoils/Temp/meter.gif

3.	I removed the top meter with no change.

4.	I disconnected the 0.1uF (I didn't have a 1.0uF handy) capacitor and 10K
resistor.  This raised the inductance reading to 221.977uH.  A shocking
4.86% Jump!  I am measuring at 34kHz so I wonder if the effect would show
up at 60Hz?  However, this does add load to the secondary that really
should be wide open theoretically.

5.	I removed the top plate and the inductance dropped slightly to
221.714uH.  This is plastic but it has a big acorn nut on it that is steel...
http://hot-streamer-dot-com/TeslaCoils/Temp/wiresf.gif

6. I removed the lead up wires from the coil terminal to the top where they
hook to the meter and the inductance dropped to 221.058.  these wires were
perpendicular and all that but they seem to be having a little effect.  So
at this point the secondary is open loaded and bare.
http://hot-streamer-dot-com/TeslaCoils/Temp/wiresc.gif

7.	I removed the secondary and the reading was still(!!) 221.058!!  Thus,
the effects of changing inductance I saw seem to be totally related to the
test equipment I had connected to the secondary coil for the coupling
measurement.  The bare unconnected secondary coil itself has no effect on
primary inductance!!  I also note that the value of 221.058 is ONLY -0.20%
lower than acmi's predicted 221.510uH!

So to make a long story short, it would appear that the error drift we see
between measurements and acmi's predictions may very well be due to the
test equipment placed into the coil's fields.  It appears the acmi may
indeed be predicting the coupling better than "I" can measure due to the
need to have test equipment "in there" while I take the real world readings. 

At this point, acmi seems to be working flawlessly and beyond the level
where I can measure the error.  I think I have an absolute error of about
0.3%, so acmi appears to be far better than is needed for any practical
coiling purpose.  I guess measuring primary to secondary coupling
coefficients with meters is an art that is no longer needed!!

I should note that my secondary has always measured 75.4mH but programs
seem to like it to be ~1% higher.  I wonder if the high loss Sonotube is
responsible.  However, this number is insensitive and does not affect the
rest of the program significantly.

I should note that I modified acmi to display many more digits of accuracy
so I can do accurate percentage differences and sensitivity analysis, but
that does not affect the numbers.  I sure would be handy if it could input
inches for the colonies!  I tried to modify mine but the program arrays the
variables or something and I could not figure out how to do it.  It could
also use a little polishing and such to make it user friendly and all, but
this program will be around coiling... forever! 

I would ask that these issues be included as a errata in the program's
documentation so 50 years from now someone will not take real measurements
again and go through all these problems all over again (like we do today :-)).

I guess it should be tested against conical primaries and solenoidal
primaries too to be sure that basically works but I don't have any
primaries like that.  Also, looking forward to hearing if the work other's
are doing confirms all this.

Great work Paul!!  This shows what a fine computer modeling program can do!
 It looks like all one needs now to measure primary to secondary coupling
is your program!

Cheers,

	Terry