On Saturday, October 1, 2016 at 10:01:37 AM UTC-4, kobuki wrote:

> -- "... rainRate should not be in the driver at all.  the hardware reports 
> a value that is arbitrary, and the rainRate calculation is yet a different 
> arbitrary calculation ..." - well, the rf packets provide a raw value, 
> which we convert to rainfall/time period values. What do you mean by 
> arbitrary here? I agree it should be done in a single unified way, though, 
> to be able to provide comparable results across devices.
>

the algorithm implemented in meteostick.py is simply:

3600.0 / time_between_tips * rain_per_tip

that could be implemented generically in StdWXCalculate.  however, 
time_between_tips could be a problem; most hardware does not report the 
time of the last bucket tip (some sensors do not tip).

so for now leave rain_rate in meteostick.py, with the StdWXCalculate 
options of hardware/software to enable/disable it.
 

> -- We only fully process the raw radio packet data. I'm still of the 
> opinion that just because the machine readable values come handy sometimes, 
> we shouldn't keep it. It's sparsely reporting certain values and generally 
> use unknown converison algorithms (even if they are accurate and correct, 
> we have no way to tell). I'd say as part of the simplification, we should 
> drop it entirely from the driver.
>

agreed - drop it at some point.
 
 

> Not completely related, but is there an SDR driver with comparable 
> functionality? I think it wouldn't be a bad idea to use one as the radio 
> capture part and adopt the rest of the current MS driver for the decoding.
>
>
the weewx-sdr driver parses output from the rtl_433 program.  however, it 
looks like rtl_433 does not yet recognize output from davis sensors.

you might want to pick up a 20$US sdr dongle then send a pr to the rtl-433 
project with the davis decodings.

someone should make a cheap usb pressure sensor, then the pressure sensor 
plus the sdr dongle would be an inexpensive replacement/backup for just 
about any weather station console.

m 

Reply via email to