Dear Attila, Thanks for the heads up.
I am currently using a HPF both in hardware (capacitive coupling into the balun driving the ADC inputs) and in software before the ZCD. This should counteract the first-order effects of this offset, although second-order effects (converter nonlinearity et al) will of course still be an issue. The plots you've quoted include (different kinds of) DC offset correction for all but the "unfiltered" data; getting an efficient DC offset correction working in real time on this 8-bit platform was indeed one of the main challenges of the software-only approach. The FPGA daughterboard is currently in production at Eurocircuits; I hope to have time to work on those the coming month. I'll also try to book some time in our climate chamber. (I've had one of our GPSDO-designs running in our general labs since before Christmas; surrounding it with bottles of water works well enough to low pass filter temperature swings, but I still see 6 degrees C swings overnight as out HVAC only runs during business hours.) To be continued, JDB. On Sun, Feb 23, 2020 at 8:11 PM Attila Kinali via time-nuts < [email protected]> wrote: > Good evening! > > I'm going through some old stuff... > > > On Wed, 27 Nov 2019 00:29:19 +0100 > Jan-Derk Bakker <[email protected]> wrote: > > > This has yielded a combined "simple" signal > > processing path of a differentiator, a double comb filter and the offset > > estimator, which is getting very close in performance to the "ideal" band > > pass filter (OADEV of 3.77e-13@tau=1s versus 3.25e-13@tau=1s for the > BPF; > > full plot: > > > http://www.lartmaker.nl/time-nuts/DMTD%20self-noise%20OADEV%20with%20PLL%20and%20various%20filters.pdf > > for this 600000-second recording: > > > http://www.lartmaker.nl/time-nuts/600ksec%20run%20with%20PLL,%2010811%20through%20splitter.png > > . OADEV past ~1000sec is severely compromised by the fact that the > > measurement setup is in my home lab which sees temperature swings of up > to > > 20 degrees C and which does get bumped from time to time. Longer runs in > a > > more controlled setting forthcoming). > > > I can offer an explanation for the large effect of the zero correction seen > here. The LTC2140 is specified to have a +/-10µV/°C drift (at 1Vpp > setting). > Converted into phase error due to zero crossing shift, this turns into > a phase shift of +/-1ps/°C @ 10MHz. Note, the shift is given as +/- and > per channel, which means, it could very well be that the channels are > not matched in their temperature characteristics and thus the total phase > shift could be +/-2ps/°C ... though total shift being closer to 0.5ps/°C is > more likely. > > Summa sumarum: DC offset correction is important if a zero crossing > detector is used. > > Attila Kinali > > -- > <JaberWorky> The bad part of Zurich is where the degenerates > throw DARK chocolate at you. > > _______________________________________________ > 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.
