That was kind of like my suggestion. You can't back out random noise. But you can measure noise.
I think not using the data when it is noisy is the extreme case of what I suggested which was to get the loop gain based on the noise. Not looking at the data is the same as setting gain to zero. I'd only set it all the way to zero if the location was off by some huge amount, like maybe 100 meters. But at 15 feet the timing error would be at worst 15 nS which is likely better than the noise on the timing signal even when the location is spot-on. In other words a nav receiver might have a 100 nS one signal error on the PPS. Speed of light is about 1nS per foot so a 10 foot position error is like an order of magnitude less than 100 nS and is "nothing". But as the position error gets "large" you might guess the timing error is also large and set the gain down lower so that the GPSDO effectively uses a longer averaging time. At some very large position error you set the gain to zero (ignore the data) and then you are effectively in "holdover mode". You'd have to experiment to see when hold over moe gives beer results then using data with some error. My guess is the error needs to be very large One other point: Can't you move the antenna above the roof? At least tie it to a plumbing vent with a zip tie or hose clamp. On Sun, Mar 23, 2014 at 11:31 PM, Hal Murray <hmur...@megapathdsl.net>wrote:. > > > If I was doing something like that, I'd skip the samples where the location > was outside a range rather than trying to correct. (even if the GPS marked > them as "good") And maybe skip several samples each side of the ones out > of > range. > > If you have a good reference, you should be able to collect some data and > see > how it looks. > > > > -- > These are my opinions. I hate spam. > > > > _______________________________________________ > 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. > -- Chris Albertson Redondo Beach, California _______________________________________________ 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.