[time-nuts] Strange GPS behaviour
lists at rtty.us
Fri Dec 28 22:35:24 UTC 2012
The GPS does an estimate against the local crystal frequency. It generates the PPS off of it's estimate. The less often it updates the estimate the more odd things you see as the crystal drifts.
Of course, the crystal can have trouble all it's own. If the crystal has a rapid rate of frequency change over a narrow temperature range, the GPS simply can't keep up with the crystal.
On Dec 28, 2012, at 5:27 PM, Fabio Eboli <FabioEb at quipo.it> wrote:
> Like I mentioned in a precedent message
> (answering Magnus) I'm seeing some temp
> effects on my GPS module, see this message:
> In this graph there are the FE5680 voltages
> and temperatures, and the temperature sensed
> on the PCB near the GPS:
> At time 20000s I heated the GPS receiver directly
> with hot air gun and the drift started to change
> At 25000 I heated the FE5680#2 I was using
> as reference, but no visible effects, (apart
> the slight variation in it's voltage :)
> At 30000 33000 35000 seconds I heated the
> GPS with a resistor placed near the PCB,
> this generated more gradual temperature
> variation on the GPS.
> Here can be seen the results of the heating
> on the drift, (logging GPS PPS against Rb):
> the hotair generated so much variation, that
> the script was unable to unscrable the data.
> The resistor heater generated slower temperature
> variation on the GPS, it's visible a glitch
> everytime there was a temperature variation,
> and the drift magnitude seem to follow the
> variation of the temperature in time (dT/dt).
> I will try to reduce temperature sensivity
> incrementing the thermal capacitance and
> isolating the GPS from the ambient.
> Is this normal or it's a defect ("feature") of
> my unit? I'm also curious about what internal
> structure can generate this wander in PPS.
> Like I said before it's like if the PPS pulse
> (for intervals of few 100's of nS) depends on
> something that is very temperature dependent.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts