- View SourceWell, I am attempting to look for spikes in the data that do not fit
the average readings over the past 10 minutes.
On the windspeed, I was looking for spikes above 47 mph (which is
what have been reported on the net), and if the average windspeed
over the past 10 minutes was less than 10 mph, then I assume it was a
bad reading. On version 1.30 I will lower it to 30, perhaps it may
catch those errors below 30.
I also attempt to compare other readings over the past 10 minutes:
Current temp with 10 F degrees of average.
Humidity within 20% of average.
Pressure within 5Hg of average.
And checking on the rain fall (0 - 30.39in)
The error checking was added for all fields uploded to CWOP.
I basically am attempting to catch the gross errors that are
happening with the Lacrosse station. From what I have seen,
Heavyweather does not analyze the data it receives, it just writes it
to the file.
For instance, I found this bad reading in there:
mm = "15370.0"
inch = "605.11"
Also, I have noticed, when a bad serial data is read from the unit,
it tends to affect other variables as well. The data is probably
shifted in the serial stream of data. A simple CRC in the serial data
would have prevented this probably entirely. Real shame they chose
not to add it.
Since I do not have direct access to the serial data I can only
attempt to validate what I see from the file.
I am not currently looking at the Dewpoint but I can attempt to add
some checking there as well.
If you have some ideas let me know.
--- In firstname.lastname@example.org, "Michael"
> What bad data does it prevent from being uploaded? I am usingPoint/Rh
> WeatherView 6.0 and I get frequent errors of really low Dew
> readings. It is often uploaded to Wunderground & CWOP (upgraded WV6).
- View SourceI made a typo... I was looking for wind readings > 51 mph not 47.
Now, it compares to averages if it is > 30 mph. If the average over the
past 10 minutes is < 10 mph, then the reading is assumed to be bogus.
This should handle most situations.
If you have a better idea of how to detect and correc bad readings
(wind speed or otherwise), let me know.