An "interesting" 3080 design decision.

[jim-easterbrook]

This came up recently on the pywws mailing list, but I thought I'd repeat it here for general interest. As I understand it, it's less of a problem for Cumulus as it doesn't rely quite so much on the station's internal logging.

The non-solar outside sensors transmit data every 48 seconds. The base station logs data every 5 minutes (if you set the logging interval to 5 minutes). As the internal and external clocks drift with respect to each other, there comes a time when a reading is due from the outside sensors at the same time as the internal unit is due to log a reading. To avoid the clash it delays this logging by an extra minute, giving a one off 6 minute logging interval. (You can find these in pywws data or EasyWeather's data file.) This is all well and good, and is a simple pragmatic solution.

Then along comes the 3080 with its separate solar sensors transmitting at 60 second intervals. This clock also drifts wrt the internal sensors. Sooner or later they coincide and the internal unit delays logging for a minute. One minute later they coincide and logging is delayed for another minute. Interesting...

[mcrossley]

So how does it get out of that situation, or doesn't it?

[jim-easterbrook]

So how does it get out of that situation, or doesn't it?

My correspondent, to whom this happened, didn't wait for the clocks to drift out of sync again.
On my advice he reset the time on the base station, shifting it by a few seconds and ending the deadlock. For the time being.