...btw
Here’s the link to the Statek App Note I referenced
http://www.statek.com/pdf/tn32.pdf
From: Joe Rocci
Sent: Friday, June 12, 2015 8:26 AM
To: minima@xxxxxxxxxxxxx
Subject: [minima] Fw: 40 MHz Crystal Filter.
Steve,
My original LTSpice file for this got a bit corrupted by yours truly while
trying out some ideas months ago. The attached file (5 pole, 40 Mhz) is pretty
close and should give you a good starting point for your design simulations.
Also, if you search the forum archives, you’ll find my actual measured plots.
Here are the most important things to consider about crystal filter design:
1) The single most important parameter by a very wide margin is the precise
value of the series resonant frequency (Fs). All crystals should be within a
grouping of approx 5% of the desired filter bandwidth for best results. Other
parameters have much less effect on overall outcome.
2) Assuming you’ve grouped your crystals by Fs, if you can measure Cp and Fp,
you can calculate Cm and Lm. The Fp measurement I’m speaking of is the delta
difference between Fs and Fp, and the measurement accuracy doesn’t need to be
very good at all. Once you know Cp and the delta of Fp relative to Fs (call it
Fpd), you can calculate Cm from Cp because Cm will be the same percentage
related to 2*Cp as Fpd is as a percentage of Fs. For instance. if Fs =
40.00000 mhz and Fpd = 50 khz and Cp = 2.8pf, then Cm will be
(50000/40000000)*(2*2.8pf)=.007pf. This is all described in Statek technical
note #TN-32, page 23, Eq 19. I highly recommend reading this app note.
3) In spite of all conventional wisdom, my extensive LTSpice analyses have
convinced me that exact values of Lm, Cm, Cp and Rm aren’t critically
important. What’s important is:
a) Fs be matched as closely as possible using an oscillator and a frequency
counter. The counter doesn’t have to be accurate for this matching, it just
needs lots of resolution digits.
b) The Lm and Cm values can be estimated by knowing Cp and Fs. Knowing the
exact values of Lm and Cm isn’t terribly critical so long as their resonant
relationship precisely equals Fs. Matching for Fs ensures this.
c) Matching for Rm isn’t terribly critical, but overall higher Rm leads to
higher filter loss.
d) Matching for Cp isn’t terribly critical.
So, all this said, the W3JDR method of crystal characterization is:
- Match the crystals for Fs with a counter.
- Measure the Cp of one crystal from the matched batch.
- Measure the Rm (aka “Rs”) of one crystal from the matched batch. Methods to
do this are widely available on the net. If you can’t do it, assume 10 ohms to
get you in the ballpark.
- Calculate Cm and Lm from Cp and Fs as described above.
- Plug all your values into an LTSpice model (See mine attached)
- Adjust the termination resistances in the LTSpice model for best return-loss
- Calculate matching networks to match the termination resistances to the
circuit impedances.
Note that the QER topology is very easy to scale for more sections; just add a
crystal and a cap for each additional section desired; all the shunt caps are
equal. The termination resistance might need to be adjusted when you add
sections.
Please let me know how you make out.
Joe
W3JDR
From: Steve Dick
Sent: Thursday, June 11, 2015 7:14 PM
To: Joe Rocci
Subject: Re: [minima] Re: something fresh - a long long post
Ok – really appreciate it!!
“Digital Steve” K1RF
From: Joe Rocci
Sent: Thursday, June 11, 2015 7:00 PM
To: Steve Dick
Subject: Re: [minima] Re: something fresh - a long long post
Steve
I'm on my tablet away from home right now, but I can see in my Dropbox files
that there's a design file for a 40 mhz filter. When I get on a computer
tomorrow that runs LTSpice, I'll look at it and send the design to you if it's
still intact.
Joe
Sent from my Galaxy Tab4 10.1 4G LTE Tablet
-------- Original message --------
From: Steve Dick <sbdick@xxxxxxxxxxxxx>
Date: 06/11/2015 6:04 PM (GMT-05:00)
To: joe@xxxxxxxxxx
Subject: Re: [minima] Re: something fresh - a long long post
Hello Joe. I am interested in building a general coverage receiver with a 6 to
10 KHz bandwidth using the ECS crystals sold by digikey, part number
ECS-400-20-3X-TR, digikey part number XC1786CT-ND at 56 cents each. I purchased
a bunch of
--------------------------------------------------------------------------------
This email has been checked for viruses by Avast antivirus software.
www.avast.com
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5961 / Virus Database: 4360/9994 - Release Date: 06/11/15
--------------------------------------------------------------------------------
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5961 / Virus Database: 4360/10001 - Release Date: 06/12/15
