search

dl9rdz / rdz_ttgo_sonde

270 stars 94 forks source link

DFM-17 temp decode differences between rdz_ttgo_sonde and radiosonde_auto_rx #361

Closed markdconner closed 1 year ago

markdconner commented 1 year ago

I occasionally run both radiosonde_auto_rx and rdz_ttgo_sonde at my home near the US National Weather Service office in Valley, Nebraska to track the DFM-17 sondes launched from there. I've noticed differences in the decoded temperatures between the two software packages. An example is attached from last night's sonde (S/N 23002287, launched 11 Apr 2023 2300 UTC), which graphs the temperature decode results as archived at sondehub.org. I've also attached the radiosonde_auto_rx log file in case that's helpful - I'm not sure how to obtain the equivalent for the rdz_ttgo_sonde data. Also included is the official sounding obtained from http://weather.uwyo.edu/cgi-bin/sounding?region=naconf&TYPE=TEXT%3ALIST&YEAR=2023&MONTH=04&FROM=1200&TO=1200&STNM=72558

![2023-04-12 10_57_22-Compare Software - Grafana](https://user-images.githubusercontent.com/49458752/231514687-953bd369-3765-45aa-903b-0ece3a2ae5f OAX sounding 12 Apr 2023 00Z.txt 3.jpg) 20230411-222827_DFM-23002287_DFM17_403209_sonde.log

dl9rdz commented 1 year ago

Basically I was using exactly the same code as auto_rx for the DFM temp calculation, but it indeed there was a later change in auto_rx:

-    if (gpx->ptu_out >= 0xC) {
+    if (gpx->sensortyp0xC == 'P') {  // 0xC: "P+" DFM-09P , "T-" DFM-17TU ; 0xD: "P-" DFM-17P ?

This is likely the cause for the discrepancy...

dl9rdz commented 1 year ago

Actually, this likely is the same issue as #344. My code is assuming that 0x0C is a DFM-9P. Actually it might be either a DFM-09P or a DFM-17. I put this on my TODO list...

dl9rdz commented 1 year ago

Please check devel-20230413 if that fixes the problem.

markdconner commented 1 year ago

I have updated to this version and will have it running for the 14/00Z launch

markdconner commented 1 year ago

Temperature data matches perfectly now image

dl9rdz commented 1 year ago

Great, then I can close this issue...

darksidelemm commented 1 year ago

Will there be another master release anytime soon with this update?

There's a lot of stations running old versions of rdz_ttgo_sonde, resulting in temperature oddities...

hr-ru commented 1 year ago

Hi, yes, I guess I should just release that. I wanted to wait to get a confirmation if #344 is also fixed by that, but I guess I can just assume that it is.

darksidelemm commented 1 year ago

I think it's working. In this example we have a station running both auto_rx 1.6.1, and rdz 20230427: https://grafana.v2.sondehub.org/d/dtBeyfyVk/compare-radiosonde-callsign-telemetry?orgId=1&var-Serial=23016345&from=1687346129912&to=1687348605931

The temperature data matches!