Update oregon_scientific.c #2086
Conversation
Add sanity checks
| return temp_c; | ||
| } | ||
|
|
||
| static float get_os_rain_rate(unsigned char *message) | ||
| { | ||
| float rain_rate = 0; // Nibbles 11..8 rain rate, LSD = 0.01 inches per hour | ||
| rain_rate = (((message[5] & 0x0f) * 1000) + ((message[5] >> 4) * 100) + ((message[4] & 0x0f) * 10) + ((message[4] >> 4) & 0x0f)) / 100.0F; | ||
| float rain_rate = 0.0F; // Nibbles 11..8 rain rate, LSD = 0.1 units per hour, 4321 = 123.4 units per hour |
There was a problem hiding this comment.
This is later used as "rain_rate_in_h", can you confirm that units is in/h here? Or Is a unit 0.1 in, thus the change of scale wrong here?
There was a problem hiding this comment.
Hi, good question. Not sure why I edited that section... The sensor generated rain_rate lags so much I don't use it myself (I calculate it based on when the rain_total field changes.) I used the page here: https://www.bashewa.com/wmr200-protocol.php as a starting point and compared it with the data I get from a PCR800 sensor. I believe the PCR800 reports it in units of 0.1 inches per hour.
| total_rain = (message[8] & 0x0f) * 100.0F | ||
| + ((message[8] >> 4) & 0x0f) * 10.0F + (message[7] & 0x0f) | ||
| + ((message[8] >> 4) & 0x0f) * 10.0F + (message[7] & 0x0f) |
| @@ -868,7 +922,7 @@ static char *output_fields[] = { | |||
| r_device oregon_scientific = { | |||
| .name = "Oregon Scientific Weather Sensor", | |||
| .modulation = OOK_PULSE_MANCHESTER_ZEROBIT, | |||
| .short_width = 440, // Nominal 1024Hz (488µs), but pulses are shorter than pauses | |||
| .short_width = 440, // Nominal 1024Hz (488µs), but pulses are shorter than pauses | |||
There was a problem hiding this comment.
UTF8/Latin1 conversion problems introduced an invalid code here. Best to replace µs with us
|
How did this slip through? A genuine fix for negative temps and needed sanity checks, thanks! |
zuckschwerdt
added
feedback
Welcome :) I've been running this code 24 hours a day since June and I no longer see any issues with "BCD" values outside 0 to 9 being converted to decimal and producing screwy jumps in the data. |
|
Can you fix those 3 places (revert the "units" to inches per hour, remove the trailing space, change the mangled "µ" char to "u") and just add another commit on top. Good to merge then! |
|
Merging with a force push to correct the above now. |
Hi, I've added some sanity checks to confirm BCD digits are only 0-9 and check sensor values aren't outside their maximum range. I've also corrected the flag to set the negative sign on temperature values. It was incorrectly including the bit that denotes "maximum range limit reached". I've been running the updates against my sensors for some weeks and they've stopped the screwy values I was seeing from rtl_433 every day or so.
This is my first pull request so I hope I've got the process right! :)
Thanks, Laurence