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

Re: Javatc inclusion of Q and Rac



Original poster: "Gerry  Reynolds" <gerryreynolds@xxxxxxxxxxxxx>

Hi Bart,

Great work and your predicted vs measured curves are very impressive. One question about spacing ratio. Is this defined as "actual_number_of_turns / maximum_number_of_turns_possible" If so, I think your spacing ratio would be = turns/(coil_length/wire_diameter).

Gerry R.


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

Hi All,

Javatc has been updated to version 10.i which now includes a secondary Q and Rac output. Also, the recommended wire size has been removed for now.

Recently, Gerry wrote a program for Q based on the Fraga equation. I've been mulling through approximately 140 Q measurements, most of which are thanks to Malcolm (Medhurst would have been very proud!). During my episode with the data, I've found that Fraga's Rac can by used with both closewound and spacewound coils. If we were to use Fraga's calculation as it is, our "real world measurements" would see very high error which is useless to coilers.

I have included a factor to Fraga's Rac to account for h/d and the spacing ratio of coils. It's very simple.
Fraga's Rac x [1+(spacing ratio/hd ratio)]. Q of course is calculated as Q = sqrt(Les/Ces)/Rac.


Where:
h/d = coil length / coil diameter
spacing ratio = turns/(wire diameter/coil length)

Here is a graph of Javatc vs. measurement of 128 loaded coils. The loaded coils include toroids, spheres, and discs.
http://www.classictesla.com/temp/Rac_Q_Chart.gif


For h/d ratios > 2, the prediction is good. Ignore any coil less than an h/d of 2. The numbers are somewhat good for loaded coils, but for bare coils, they are useless.

To put this in some perspective:
h/d ratio's of 5:1 showed that 80% of the measured coils were less than 10% error
h/d ratio's of 4:1 showed that 75% of the measured coils were less than 10% error
h/d ratio's of 3:1 showed that 50% of the measured coils were less than 10% error


Q is one of the most difficult to predict based on an initial design. Most don't even attempt it. Q is also one of the harder measurements which requires correct equipment, procedure, and experience. Since many do not have the equipment or experience, Javatc will help by providing a good ballpark figure to consider during a design process.

Many thanks to Malcolm for his excellent measurements on a "VAST" number of measurements with top of the line equipment. Measurement data like this is as good as it gets. Anyone who has ever measured Q (or even provided measurement data) can appreciate the hard work and effort Malcolm put into his measurements.

And of course, Javatc is where it always has been:

http://www.classictesla.com/java/javatc.html

Take care,
Bart