Hi Said, OK, I hadn't understood the full consequences of hysteresis, but yes, I've seen it. For an hour the DAC ratchets up a step every few minutes and the phase stubbornly stays put. And then, the bottom falls out and it suddenly pushes way past where you want it. Well, at least I have a better understanding of it now. I'll try to avoid any hardware changes for the next few weeks. I may even make changes that will keep the DAC stable when loading new code.
Thanks! Bob ________________________________ From: Said Jackson <saidj...@aol.com> To: Bob Stewart <b...@evoria.net>; Discussion of precise time and frequency measurement <time-nuts@febo.com> Sent: Monday, October 20, 2014 7:21 PM Subject: Re: [time-nuts] LTE-Lite module Bob, You are on the right track! Large changes in EFC can cause hysteresis, meaning you go back to an initial voltage but the crystal does not return to the exact initial frequency. It can also create dead bands in the efc vs frequency curve. Hysteresis can cause integrator wind up as the loop is chasing an ever changing OCXO.. Retrace and hysteresis are two major issues for any disciplined oscillator. Bye, Said Sent From iPhone > On Oct 20, 2014, at 17:03, Bob Stewart <b...@evoria.net> wrote: > > OK, Yahoo has done it to me again. Sent to Tom direct and not to the list. > So, repeated here: > > Hi Tom, > > One of the biggest problems I've unwittingly faces is that of retrace. I > had seen the term used several times, but hadn't looked it up until last > night. As you can imagine, with a GPSDO under development I've had to > remove power more than a few times to make hardware changes. I think > the next time power is down I'm going to try to rearrange things so that the > OCXO is permanently powered and just the board gets switched. But > then again would big jumps in the EFC cause other problems that are > almost as bad? > > > There is just so much to learn to get this going; especially without either > an engineering degree or experience in this field. Bob Camp is > definitely right that you have to put your time in - lots of it. > > Bob - AE6RV > > > ________________________________ > From: Tom Van Baak <t...@leapsecond.com> > To: Discussion of precise time and frequency measurement <time-nuts@febo.com> > Sent: Monday, October 20, 2014 6:42 PM > Subject: Re: [time-nuts] LTE-Lite module > > > To add to Bert's note... > > Realize that for a GPSDO, (linear frequency) aging-per-day is irreverent, > almost by definition. What matters is phase noise and short-term stability, > neither of which you can possibly fix with disciplining against GPS. GPS > takes care of the rest. > > Long-term stability can be critical for non-GPS applications, which is why > oscillators with daily aging rates in the -11's and -12's are so amazing. > > Consider this: if you want to run your bench with a clean 10 MHz source, > stable to 11 or 12 digits and accurate to 9 digits -- you may be much better > off with a free-running, stand-alone OCXO with an aging rate down at > 1e-11/day than using a GPSDO/TCXO. To maintain accuracy of your OCXO just > re-tune your OCXO *once a year*. Aside from ADEV plots, this is another way > to appreciate how amazing some OCXO are, any why many of us still troll eBay > for high-stability, low-noise, low-drift quartz oscillators. > > /tvb > > ----- Original Message ----- > From: "Bert Kehren via time-nuts" <time-nuts@febo.com> > To: <time-nuts@febo.com> > Sent: Monday, October 20, 2014 4:01 PM > Subject: Re: [time-nuts] LTE-Lite module > > > Allow me to clarify. > I started out with 7 MV 89 one of it a total loss. The remaining 6 after 3 > month + burn in show better than 1 E-11 aging per day, 2 closer to 5 E-12. > Only two have been tested for ADEV and are close to 1 E-12, 2X not 10 X. > Bert Kehren > > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
