[Prev][Next][Index][Thread]
measuring true self capacity of a coil.
Hi Galvin, All,
Now that we are worrying about percentages in the single digits when it
comes to measuring or predicting Fo, I guess we should worry about "how" to
make the measurements.
I would guess the "ideal" would be to have the first winding of the
secondary almost directly on the ground plane with that plane being split
radially to prevent the shorted turn effect. A low impedance sine
generator would then have to energize the base of the coil and the top
would have to be terminated almost immediately at the top winding. so
pickup that would not affect the system would be needed to detect the peak
frequency.
Of course, real secondaries usually have a bare section at the top and
bottom and the top winding usually goes to a length of wire or a center
terminal. E-Tesla5 assumes a center terminal at the top of the coil.
I ran E-Tesla5 with and without the assumed top center terminal and the
frequencies were 139902.6 with and 140031.2 without the terminal. So the
error introduced by the small terminal appears to be tiny. I then raised
to bottom winding 1 inch off the plane (10.25 x 30 inch coil) and the
frequency went from 139902.6 to 140759.6 or 0.61% error.
Also one must be sure the ground plane and the terminal do not affect the
coil's inductance. In a recent actual test, my secondary went from 75.4 to
68.9mH when a large toroid was placed directly on top of it. That is a
4.61% Fo error...
Of course, you then have the "trying to tune the signal generator just
right", "far away from other objects", and "local EMI" problems too...
So measuring the self C and Fo frequencies of a coil to very high accuracy
has some subtle details that need to be considered. Although in many ways
E-Tesla5 is theoretically perfect, it does have three known sources of error.
1. It can only do about a maximum 14400 element grid in the finite element
analysis. So things get rounded to the nearest element and curved objects
are always a little "choppy".
2. The secondary voltage profile is equation matched to actual
measurements. I think the large coil I did those test on and my small coil
have about a 2%-4% difference. Perhaps the work of Robert Jones will help
this problem. The array size and programmer problems can be overcome but
getting the voltage profiles correct in all cases to within a fraction of a
percent is still a frontier.
3. Programmer eroor. :-) Although I try hard to check and test the
program, there is always the chance I messed up somewhere. I have added
diagnostics and run test grids in later versions to help weed out errors
and it has gone through a lot of iterations, but they are always there...
It is truly strange that we are now almost to the point of being able to
calculate Fo, with a number of different methods, more accurately than we
can measure it!
Cheers,
Terry
At 09:53 PM 5/30/00 +0100, you wrote:
>Hi Malcolm,
>your coil data was the only reliable measurements I could test this equation
>by. My own measurements of coils are dubious due to poor equipment and
>practical experience. As an aside, when I first got into coiling I was
>subscribing to the ITS discussion list. I needed a method to calculate the
>self capacity of a coil (this was about three years ago), you sent me a qbasic
>program that worked it out. You must have been the only one on that list who
>knew anything about TCs.
>
>Anyway, the formula looks o.k at the moment, but Bob will verify it for sure.
>
>Regards,
>
>Gavin
>
>
measuring
References: