[vna] measurement of quartz crystal

[Tom McDermott]

Hi Koji:

The phase detector of the VNA is not capable of measuring correctly when the
input signal is attenuated a large amount (both S21 and S11 measurements).

I have also found some bugs in the software that affect crystal filters:

* Severe S21 measurement error when the time delay of the network is high
(the VNA sweeps too fast for long-delay networks).

* Error in the frequency generation routine that causes the VNA software to
have a stop_frequency error for very narrow sweeps (few kilohertz),
grid=1020, and it also forgets to set the frequency error from the detector
calibration.

I have fixed these errors, they will be incorporated in the next software
release on the TAPR website.


	-- Tom, N5EG





> -----Original Message-----
> From: vna-bounces at lists.tapr.org 
> [mailto:vna-bounces at lists.tapr.org] On Behalf Of Koji Terai
> Sent: Monday, December 05, 2005 2:07 AM
> To: TAPR/Ten-Tec VNA Users List
> Subject: Re: [vna] measurement of quartz crystal
> 
> 
> Hi Tom,
> 
> Thank you for your res and suggestions.
> 
> So far, I had no chance to use VNAs, but I read some books 
> before that 
> described importance
> of defining measurement planes and consistency of a 
> testfixture to be used. Any way I have just started to use 
> the VNA and I have to learn more about 
> measurement
> by the VNA.
> 
> The way I tried was that I measured crystals in S12 mode and 
> tried to read 
> out f(o) and f(a)
> from curves of magnitude and phase, but as for f(o), I had to expand 
> frequency pitch
> to read the frequency very precisely then the peak was 
> displayed very flat, 
> almost it turned a line
> and I could not find where the peak was. So I gave up reading 
> our f(o) from 
> a magnitude and I
> read f(o) from a curve of phase instead, but as for f(a), a 
> curve of phase 
> changed in every sweep
> so that I could not determine f(a). I think that a crystal 
> exhibits great 
> attenuation at f(a) and the VNA
> is not able to determine phase change at around f(a) due to 
> the limitation 
> of the phase detector of
> the VNA. I am not sure the way I tried was correct or not, or 
> there are some 
> better way to measure
> a crystal at this moment.
> 
> Koji/ JA3XNO
> 
> 
> ----- Original Message ----- 
> From: "Tom McDermott" <tom.mcdermott4 at verizon.net>
> To: "'TAPR/Ten-Tec VNA Users List'" <vna at lists.tapr.org>
> Sent: Sunday, December 04, 2005 8:08 AM
> Subject: RE: [vna] measurement of quartz crystal
> 
> 
> Hi Koji:
> 
> The TAPR VNA can be calibrated against an external reference 
> oscillator for good frequency accuracy. However in order to 
> determine accurately the 4 motional parameters (Ls, Rs, Cs, 
> Cp) of the crystal requires extreme precision of the fixture 
> stray reactances. I am not an expert in crystal filters, but 
> I think the 4-parameter model is used pretty extensively in 
> filter modeling. 100 Hz frequency accuracy is not a problem, 
> but stray reactances are a problem.
> 
> I think Agilent has a software package to extract the 
> motional parameters from swept S parameters. The TAPR unit 
> does not have this software, but it perhaps could be an 
> interesting exercise; certainly for those who enjoy working 
> on software.
> 
> Key to measurement is that the fixture parallel measurement 
> will be contaminated by fixture parallel capacitance. This 
> will rotate the series resonance frequency away from R + j 
> Zero X and must be compensated. The parallel anti-resonance 
> will have the same issue. S-parameters will show the crystal 
> resistance as compared to 50 ohms but the fixture will rotate 
> the phase angle. The fixture perhaps can be calibrated out of 
> the measurement if the crystal is replaced with a 50 ohm 
> resistor, but I have never tried it.
> 
> -- Tom, N5EG
> 
> 
> 
> 
> 
> 
> > -----Original Message-----
> > From: vna-bounces at lists.tapr.org 
> [mailto:vna-bounces at lists.tapr.org] 
> > On Behalf Of Koji Terai
> > Sent: Saturday, December 03, 2005 9:14 AM
> > To: vna at lists.tapr.org
> > Subject: [vna] measurement of quartz crystal
> >
> >
> > Hi, everyone! I would like to know if the VNA is able to determine 
> > resonant frq. and antiresonant frq. of a quartz crystal in 
> accuracy of 
> > 100 Hz. I plan to construct Cohn type crystal filters, and 
> I tried to 
> > measure crystals but I was not able to measure these two 
> frequencies 
> > definitely, especially ant frequency measurement was very 
> critical. I 
> > am looking forward to a good solution for this.
> >
> >  Koji
> >
> >
> >
> > _______________________________________________
> > vna mailing list
> > vna at lists.tapr.org 
> https://lists.tapr.org/cgi-bin/mailman/listinfo/vna
> >
> 
> 
> _______________________________________________
> vna mailing list
> vna at lists.tapr.org https://lists.tapr.org/cgi-bin/mailman/listinfo/vna
> 
> 
> 
> _______________________________________________
> vna mailing list
> vna at lists.tapr.org https://lists.tapr.org/cgi-bin/mailman/listinfo/vna
>