[DSP-10] Winding Torroids

[Bob Larkin]

Hi Bob, Bill, Matt and others!

I think Bob and Bill's excellent discussion really covers the coil 
question.  I count turns out loud!

Yes it is not critical!  The primary defense against signals that are above 
the 24 kHz Nyquist frequency is the 19.665 MHz crystal filter.  But, there 
will be some leakage around the filter.  The digital filtering in the ADC 
will attenuate the first alias quite well. For the AD1847, this is over 80 
dB from 28.8 to 48+ kHz.  I measured one above 48 kHz and, as I remember, 
it remained 80 dB to well above 100 KHz. But there is still concern about 
the far out-of-band signals, and the simple LPF provides protection there.

The AIC23B (on the KK7P DSPx board) is not as good at alias filtering as 
the AD1847 (on the EZ-Kit), and it may benefit more from the LPF.  But, the 
xtal filter is still the primary band limiting circuit element.

And when the interfering station gets too strong and is on SSB, the 
receiver filtering can't do everything!  The transmitter  inter-modulation 
side bands will be in our pass band.  This tends to limit the amount of 
receiver filtering that is useful.

73,
Bob  W7PUA

At 06:00 PM 3/22/2004, you wrote:
>Well Bob, that "rule of thumb" is only accurate for small changes in the
>number of turns. The inductance is actually proportional to the square
>of the number of turns so if you change turns to 0.9 times the previous
>turns (-10%) the inductance actually comes out to 0.9 * 0.9 = 0.81 or
>81% which is "only" -19% change.
>
>If the coil whose "inductance is proportional to turns squared" is used
>in a simple parallel or series tuned circuit, resonance formula is one
>divided by 2*pi*square root of L*C and the "turns squared" and "square
>root of L*C" cancel each other out perfectly. A 1% turns change would
>give you a 1% frequency change! Because resonant frequency formula has
>square root of L*C on the BOTTOM of the formula there's a minus sign:
>+1% turns change gives -1% frequency change (which jibes with our gut
>feeling that "more inductance = lower frequency).
>
>It's fine to puzzle over the way the formulas interact, but don't get
>too wound up: a pi-type lowpass filter with two C's and one L is rarely
>critical. Most lowpass filters are built with 5% capacitors and they
>"just work", hi. If Bob suggested you do something special, like
>selecting capacitors to 2% accuracy or something like that, then you
>might figure the inductor is similarly critical. Absent some specific
>indication like that, just count your turns and plow onward!
>
>73 - Bill
>_______________________________________________
>DSP-10 mailing list
>DSP-10@mailman.qth.net
>http://mailman.qth.net/mailman/listinfo/dsp-10