Rev 1.40 vs 1.42 results


Ted KX4OM
 

I think I've fixed the issue with Rev 1.42 not producing results in crystal measurement mode. I produced a new set of coefficients for the crystal measurement sweep. I believe the addition of the coefficients for xtal measurments is intended to make the sweeps more accurate over a shorter range.

I swept 5 crystals on 05/17/2020 using Rev 1.42 from the same batch I previously had swept on 05/11/2020. I created a table of average values for both sets.

Rev1.40 vs 1.42

Five crystals from same batch

1.40 tested 05/11/2020; 1.42 tested 05/17/2020

Peak frequency 8.386.xxx; Xtal Meas sweep 8,385.900 to 8,388.100 MHz

Average results:

1.40
Peak f BW RS Lm Q
8384.364 127.6 13.08 47.22 195,524

1.42
8386.369 123.9 16.8 53.73 168,696


Ted KX4OM
 

There is an error in that table. I copied the data from a test file and accidentally posted without checking it. The peak frequency of the 1.40 results is 8386.364. I also meant to add that the frequency-related data, peak frequency and bandwidth were almost identical between the 2 data sets, while the RS, Lm and Q are significantly different. Is that indicative of more loss being accounted for? Also, as to the crystal measurement coefficients, are both the sets of coefficients in parameters.txt factored into the calculations in the crystal measurement mode?

I've been going back through the messages since the release of 1.41 to try to understand what is involved with the measurements.

Ted, KX4OM


EB4APL
 

Hi Ted,

The crystal measurement coefficients were added in version 1.42 (or maybe in 1.41, I can not check it now) after realizing that the "normal" set of calibration coefficients accounted for the frequency response of the Generator - Detector combination used for sweeping filters and so, but when measuring crystals there is another thing in the play that have its own frequency response: the Crystal Measurement Fixture. This new set of parameters account for the frequency response of the Generator - Detector - Fixture combination and it is used only when measuring crystals. Before this, the attenuation of the set, that had its own value stored in the configuration file, was valid for a small frequency range, requiring to be recalibrated each time the fixture was used for a new frequency, so there was a box in the program for entering the shorted fixture dBm.

Probably the differences that you see between the measurements done with version 1.40 and 1.42 are due to the fixture attenuation was not measured at the frequency of the crystal and this affects to the calculated series resistance and Q.

73 de Ignacio EB4APL

El 19/05/2020 a las 18:21, Ted KX4OM escribió:
There is an error in that table. I copied the data from a test file and accidentally posted without checking it. The peak frequency of the 1.40 results is 8386.364. I also meant to add that the frequency-related data, peak frequency and bandwidth were almost identical between the 2 data sets, while the RS, Lm and Q are significantly different. Is that indicative of more loss being accounted for? Also, as to the crystal measurement coefficients, are both the sets of coefficients in parameters.txt factored into the calculations in the crystal measurement mode?

I've been going back through the messages since the release of 1.41 to try to understand what is involved with the measurements.

Ted, KX4OM

--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
https://www.avast.com/antivirus


Ted KX4OM
 

Hi Ignacio,

I had forgotten Nick's warning in the 1.40 version manual to sweep the shorted fixture before measuring crystals. I had -12.3 in the parameters.txt file from a couple of years ago. I measured -9.6 dBm compensated today, and I put that in the 1.40 parameters.txt file. I ran the same 5 crystals I had measured a couple of days ago through 1.40 and 1.42 back to back, a few minutes apart. The results came out very close. In fact, the 1.42 measurements came out much higher that what I had measured for my previous post. I can't explain that, but I had too many changes going on with the different versions and the parameters.txt files. I really appreciate you help with this. It looks like I won't have to sweep about 100 crystals after all :)

Ted, KX4OM