Hi > On Mar 1, 2021, at 3:32 AM, Magnus Danielson <[email protected]> wrote: > > Hi, > > On 2021-03-01 01:23, Attila Kinali wrote: >> On Sun, 28 Feb 2021 18:35:57 -0500 >> Dan Kemppainen <[email protected]> wrote: >> >>> If any of you have any suggestions on where to find information on how >>> to get something like this to oscillate properly, guess at correct >>> parameters, or even measure any of the parameters I would really >>> appreciate it. >> The best way would be to use a network analyzer and measure >> the crystals LCR parameters. Using that you can use the >> standard harmonic oscillator literature (Parzen, Frerking,...) >> to design the circuit. >> >> I seem to have misplaced my literature on how to measure >> crystal oscillators. But if you search for "Neubig" and >> "crystal measurement", you should be able to find some of >> the nice documents that Bernd Neubig has written on the topic. > > The same Neubig made a comment that your normal network analyzer isn't > such a good tool, even a very good one. The reason being you need both a > wide and narrow sweep to make the model values accurate enough. Most > network analyzers achieve the wide sweep, few do the narrow sweep and > then having that combined to fill in the LCR parameters of a suitable > model, not so much. Things you learn by eating breakfast with him. Turns > out that my network analyzer is good for the measurement, but not for > model fitting. > > Neubig have been very much involved in standardizing measurements, and > doing those well to characterize accurate enough those high-Q resonators > have it's challenges that leads many efforts into incorrect values. > > Cheers, > Magnus >
A while back I got drafted into writing the code and validating the procedure to do this commercially. With care, it can be done and the results are quite good. As you would expect, there are various gotcha’s you run into. Along with the network analyzer (resolution and noise) , it depends a bit on the crystal you are looking at. High Q low frequency crystals are a challenge for the typical network analyzer. Something like a 150 MHz fundamental is generally pushing things like test setup (…. rigid cables are a good idea) more than the analyzer. Yes this assumes you are looking at “real” ( = expensive) analyzers rather than the low cost stuff. There is a reason people pay for those great big boxes …. :) Fun !! Bob > > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to > http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com and follow the instructions there.
