On Apr 28, 5:20*am, Bill M wrote:
K7ITM wrote:
On Apr 27, 8:02 pm, Bill M wrote:
K7ITM wrote:

To put some numbers on what Paul suggested:
If I want an unloaded coil Q of 2000 at 10MHz,
I stopped here.

I'd expect to need a....

No way you'll ever see 2000 Q at 10 Mcs. Can't happen. Some guys are
approaching 2000 at 1 MC BCB but there's lots of expensive hoops to jump
through to reach that point. *Simply cannot be had at 10 Mcs. *200-300
on a good SW coil is about all that can be achieved.
...
Interesting comment. *In the filters I build for test fixtures, I use
air-core coils that are about 1" diameter and 1" long, and they give
me Qu in excess of 300 at 10MHz. *For what I do, I don't need Qu up in
the thousands, and don't have room for really big coils, but can you
give me a reason I shouldn't expect Qu to scale linearly with size up
to the point where radiation losses become significant?

Can you tell me why I should think that the inductance calculator at
http://hamwaves.com/antennas/inductance.htmlis not giving me accurate
results when I put in, say, D=130mm, n=20, l=260mm, d=7mm, and
f=10MHz? *It agrees with other independent ways I have to estimate the
Qu of the ~1 inch coils I build, and those coils measure within
engineering tolerance of the estimates.

Self-capacitance (ultimately self-resonance) is always a contributing
factor that prevents coils from achieving their maximum theoretical Q.
Estimating coils whose Q is going to fall in the 200 range for other
reasons is relatively easy but that doesn't mean you can scale upwards
proportionately.

I build air coils in the 3 inch range along the lines of what is
pictured here, *http://www.sparkbench.com/homebrew/grebe/cr18.html

They only *measure* in the 250 range. *Silver-plated wire could
certainly improve a coil of this size but no way would you achieve
numbers like 2000.

-Bill

OK, I was curious. Was I way off-base, or is it reasonable to think
that you can get a 10MHz Qu considerably higher than 250 (and possibly
up in the stratosphere above 1000)? I'm not going to spend the time,
effort and money to build a seriously large coil as the theory
suggests I'd need, but I did wind a somewhat smaller one...

I wound 15 turns of #10 AWG (2.55mm) bare copper with about 2.25
inches ID and 3 inches long. I resonated it with 3 * 12pF C0G
capacitors; it resonates at 9.088MHz. I coupled an output to an
analyzer through 1pF tapped one turn up from the "cold" end, and
loosely coupled an input from the analyzer's source using a one-turn
loop spaced away from the coil. My back-of-the-envelope calculation
says such a coil with air insulation should have a Qu around 740 at
9MHz. In a tank circuit, the finite Q of the capacitors will lower
the tank Q below that value. What I actually measure is a 3dB
bandwidth of 15.87kHz, for a tank Q of 572. OK, so that's a bit lower
than I might have expected. BUT--this coil is wound on a length of
black ABS drain pipe, which is an absolutely terrible thing to use as
a coil form if you're trying to get the highest possible Q. (My plan
was originally to take the coil off the form after I wound it, but it
had too much of a mind of its own about what shape it was going to
assume. I've wound some smaller self-supporting coils, but this one
didn't work that way.)

I'm convinced by this little 'speriment that I could build an LC tank
resonant at 10MHz with a tank Q above 1000 with no trouble--and
probably _well_ above 1000 if I used really high Q capacitors and just
enough low-loss solid insulation to keep the coil turns properly
spaced. I don't need one at the moment, but if I ever do, I sure
won't be afraid to try it.

Cheers,
Tom