Tuned Circuit Selectivity

On Monday, 6 May 2019 07:33:42 UTC+1, bitrex  wrote:
> On 5/6/19 12:40 AM, tabbypurr wrote:
> > On Monday, 6 May 2019 03:59:39 UTC+1, bitrex  wrote:
> >> On 5/5/19 10:49 PM, bitrex wrote:
> >>> On 5/5/19 6:13 PM, Cursitor Doom wrote:
> >>>> Gentlemen,
> >>>>
> >>>>
> >>>> I'm just trying a number of combinations of L and C to find the right
> >>>> values for resonance at around 1.35Mhz. The problem I'm having is that
> >>>> the resonance point is far from clear. It's as if the Q of the components
> >>>> is very low (even though they actually aren't). I'm trying to think of a
> >>>> way to make it more 'peaky' on the oscilloscope display to take the guess
> >>>> work out of finding that sweet spot.
> >>>> ATM the two components are in parallel, but I'm thinking maybe I'd have
> >>>> more luck if I wired them in series and increased the Zo of the signal
> >>>> generator by placing a highish value resistor into the genny's central
> >>>> output pin and feeding the tank via that.
> >>>> Would that work or has anyne got any better ideas?
> >>>>
> >>>> CD
> >>>>
> >>>>
> >>>
> >>> when you calculate the resonant Q_t of a parallel-tuned shunt LC tank
> >>> you have to include the impedance of the source as well as the load, and
> >>> the ESR of the inductor, at the resonant frequency, that ESR times Q_u^2
> >>> of the inductor, the inductor unloaded Q at the resonant frequency, all
> >>> in parallel.
> >>
> >> But with an unknown inductor how do u know precisely the inductor
> >> unloaded Q at the resonant frequency of the tank if you need to know
> >> what the inductor's unloaded Q is at that frequency to calculate
> >> precisely what the resonant frequency of the tank is? Yes it's a bit of
> >> a conundrum just do your best. It's probably about one hundred and
> >> fifty...ah....two.
> > 
> > What?? Measure f_res. Ping it & observe oscillation. Or do it actively with pfb. That approach does have nearly a century of use behind it.
> > 
> > 
> > NT
> > 
> 
> In a real circuit where the inductor has ESR pinging it will give the 
> damped resonant frequency, while driving it will give the driven 
> resonant frequency, which are different.
> 
> Usually what you're interested in is the driven resonant frequency but 
> if you want to calculate that exactly _on paper_, for a mystery 
> inductor, you need to know the unloaded Q of the inductor at the driven 
> resonant frequency, but you can't work backwards from the damped 
> resonant frequency response to get it because the damped and driven 
> resonant frequencies aren't exactly the same.
> 
> it was a bit of a joke cuz IIRC the frequency discrepancy is only 
> significant for pretty low unloaded-Q inductors like less than 10, maybe.
> 
> Anyway I'm thinking about ordering one of those HP5819As "vector 
> impedance analyzer" or whatever from the 80s. that figures all this 
> stuff out automatically. 35 years later they've come down in price a lot 
> all things come to those who wait I guess

If you want real precision, inductors aren't linear anyway, so measure rather than just calculate.


NT

On Sun, 05 May 2019 16:59:19 -0700, John Larkin wrote:

> How are you coupling the signal gen into the resonant tank? Are you
> using a 10x probe on the scope?

Yes, 10x/1x switchable. And directly coupled.

> 
> A Q of 50 should be easy at that frequency, and that would make a very
> sharp peak.

I'd have thought so, yes.

> 
> What are your L and C values?

33uH & 385pF

> 
> Here's my LC program.

I find it's easiest just to use the full features of a programmable 
scientific calculator, TBH. YMMV of course.



-- 
This message may be freely reproduced without limit or charge only via 
the Usenet protocol. Reproduction in whole or part through other 
protocols, whether for profit or not, is conditional upon a charge of 
GBP10.00 per reproduction. Publication in this manner via non-Usenet 
protocols constitutes acceptance of this condition.

On Sun, 05 May 2019 20:24:11 -0500, amdx wrote:

>   You need to lightly couple your input signal to the LC and connect the
> scope for minimum loading. I've been known to put a 1 Meg resistor
> before the scope, but often then you get 60Hz interference. Shorten up
> the leads, Ground lead too.

Yes, I think this is the area where the problem is.





-- 
This message may be freely reproduced without limit or charge only via 
the Usenet protocol. Reproduction in whole or part through other 
protocols, whether for profit or not, is conditional upon a charge of 
GBP10.00 per reproduction. Publication in this manner via non-Usenet 
protocols constitutes acceptance of this condition.

On Mon, 06 May 2019 07:28:29 +0000, Jan Panteltje wrote:

> A grid dip meter was a useful instrument long ago:
>  https://en.wikipedia.org/wiki/Grid_dip_oscillator
> Build one once.

I still have a selection of them and I built one once. NEVER had any luck 
with *any* of the damn things for some reason! I suspect they're only 
useful if you are testing really physically large combinations of C and L.



-- 
This message may be freely reproduced without limit or charge only via 
the Usenet protocol. Reproduction in whole or part through other 
protocols, whether for profit or not, is conditional upon a charge of 
GBP10.00 per reproduction. Publication in this manner via non-Usenet 
protocols constitutes acceptance of this condition.

On Sun, 05 May 2019 15:36:13 -0700, tabbypurr wrote:

> Positive feedback works wonders for Q. Just arrange it so the feeding
> back circuitry doesn't alter the tuning of your LC.

I believe there must be a simpler solution out there somewhere. Most 
probably what amdx said.



-- 
This message may be freely reproduced without limit or charge only via 
the Usenet protocol. Reproduction in whole or part through other 
protocols, whether for profit or not, is conditional upon a charge of 
GBP10.00 per reproduction. Publication in this manner via non-Usenet 
protocols constitutes acceptance of this condition.

On Mon, 6 May 2019 13:13:35 -0000 (UTC), Cursitor Doom
<curd@notformail.com> wrote:

>On Sun, 05 May 2019 16:59:19 -0700, John Larkin wrote:
>
>> How are you coupling the signal gen into the resonant tank? Are you
>> using a 10x probe on the scope?
>
>Yes, 10x/1x switchable. And directly coupled.
>
>> 
>> A Q of 50 should be easy at that frequency, and that would make a very
>> sharp peak.
>
>I'd have thought so, yes.
>
>> 
>> What are your L and C values?
>
>33uH & 385pF
>
>> 
>> Here's my LC program.
>
>I find it's easiest just to use the full features of a programmable 
>scientific calculator, TBH. YMMV of course.


Xc = Xl = 292 ohms at 1.4 MHz. If you connect a 50 ohm signal
generator across that parallel tank, Q is about 0.16. Not much of a
resonant bump. Couple gently from the generator into the tank, with a
big resistor or a tiny cap, or just proximity.


-- 

John Larkin         Highland Technology, Inc

lunatic fringe electronics 

On a sunny day (Mon, 6 May 2019 13:13:35 -0000 (UTC)) it happened Cursitor
Doom <curd@notformail.com> wrote in <qapbtv$52e$1@dont-email.me>:

>On Sun, 05 May 2019 16:59:19 -0700, John Larkin wrote:
>
>> How are you coupling the signal gen into the resonant tank? Are you
>> using a 10x probe on the scope?
>
>Yes, 10x/1x switchable. And directly coupled.
>
>> 
>> A Q of 50 should be easy at that frequency, and that would make a very
>> sharp peak.
>
>I'd have thought so, yes.
>
>> 
>> What are your L and C values?
>
>33uH & 385pF
>
>> 
>> Here's my LC program.
>
>I find it's easiest just to use the full features of a programmable 
>scientific calculator, TBH. YMMV of course.

You have internet access so:
 https://www.daycounter.com/Calculators/LC-Resonance-Calculator.phtml
 gives 1.41 MHz for your values,

The lossy part is likely the L, so measure its resistance R.
Q = w.L / R = (2 * pi * f * L) / R

3dB bandwidth B = f / Q.
If I remember my school days correctly.

Examples:
 http://physics.usask.ca/~angie/ep311/lab4.htm

On 5/6/2019 8:15 AM, Cursitor Doom wrote:
> On Sun, 05 May 2019 20:24:11 -0500, amdx wrote:
> 
>>    You need to lightly couple your input signal to the LC and connect the
>> scope for minimum loading. I've been known to put a 1 Meg resistor
>> before the scope, but often then you get 60Hz interference. Shorten up
>> the leads, Ground lead too.
> 
> Yes, I think this is the area where the problem is.
> 
> 
> 
> 
> 

  So what type of inductor? Is there a good way to couple to it?
Sometimes I just hang a clip lead from the generator near the coil.

                               Mikek

On 5/5/2019 5:13 PM, Cursitor Doom wrote:
> Gentlemen,
> 
> 
> I'm just trying a number of combinations of L and C to find the right
> values for resonance at around 1.35Mhz. The problem I'm having is that
> the resonance point is far from clear. It's as if the Q of the components
> is very low (even though they actually aren't). I'm trying to think of a
> way to make it more 'peaky' on the oscilloscope display to take the guess
> work out of finding that sweet spot.
> ATM the two components are in parallel, but I'm thinking maybe I'd have
> more luck if I wired them in series and increased the Zo of the signal
> generator by placing a highish value resistor into the genny's central
> output pin and feeding the tank via that.
> Would that work or has anyne got any better ideas?
> 
> CD

Are you wanting the loaded or unloaded Q of your network?

On a sunny day (Mon, 6 May 2019 13:18:12 -0000 (UTC)) it happened Cursitor
Doom <curd@notformail.com> wrote in <qapc6k$52e$3@dont-email.me>:

>On Mon, 06 May 2019 07:28:29 +0000, Jan Panteltje wrote:
>
>> A grid dip meter was a useful instrument long ago:
>>  https://en.wikipedia.org/wiki/Grid_dip_oscillator
>> Build one once.
>
>I still have a selection of them and I built one once. NEVER had any luck 
>with *any* of the damn things for some reason! I suspect they're only 
>useful if you are testing really physically large combinations of C and L.

Yes a bit tricky to use, oh what is large,
in the few MHz range and lower it works.

With your raspi as signal generator and a simple scope or diode voltmeter
you can find out the resonance too,
or make the LC oscillate and use a frequency counter.
Usually there is some signal, else there would not be an LC, scope it.
here a nice 25 MHz parallel LC, tunable, I like that, real silvered wire..
 http://panteltje.com/pub/25MHz_220pF_par_LC_IMG_6896.JPG

It is easy.
When scoping it, the few pF scope probe lowers frequency a bit.
Any loading increases bandwidth.
This one is just a test to filter out junk from an incoming weak signal,
before it becomes permanent on a board.