Support of very old Oregon Scientific Sensor THGR228 possible?

[riegelbrau]

I have a very old monitoring station EMR812HG for temperature and humidity:


The station was sold together with a sensor THGR228:

(See the ancient logo of the Deutsche Bundespost, which was the former German autority for telecommunication).
Both units are working like a charm since then. Nowadays I have a Home Automation environment with several sensor integrated via RTL_433. It would be nice, if this old sensor would work for that, too. From the product numbers it seems to be made by Oregon Scientific. I found several internet sources for that assumption, where the THGR228 should be similar to the THR228N and it should use the OR protocol V2.1.

The actual release rtl_433 version 24.10-7-g54a886bf branch master at 202411201058 is running on a Raspberry Pi 4B and sends the gathered data via MQTT to openHAB. This is working good for several years. When I start rtl_433 with the 3 protocols for OregonScientific devices activated I get no resutls from my sensor. The I pulled the antenna from my rtl sdr dongle and positioned the sensor directly besides of it. Then I ran rtl_433 with this command rtl_433 -d 0 -v -R -73 -R -112 -R -165 -Y autolevel -Y magest -M noise -M level -S unknown. The following samples were written, when I saw the sensor sending data:
g003_433.92M_250k.zip
g004_433.92M_250k.zip
g007_433.92M_250k.zip

Then I ran the program with several recommended flex decoder setting, like this:
rtl_433 -d 0 -v -R 0 -X 'n=THGR228,m=OOK_PWM,s=800,l=1296,r=2724,g=0,t=0,y=128'
From the results I extracted some data received, when the sensor was sending (LED flashing) like this:

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
time      : 2024-12-23 17:34:00
model     : THGR228      count     : 2             num_rows  : 2             rows      :
len       : 106          data      : 00006c9668155109a5502642d14,
len       : 0            data      :
codes     : {106}00006c9668155109a5502642d14, {0}0
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
time      : 2024-12-23 17:34:00
model     : THGR228      count     : 2             num_rows  : 2             rows      :
len       : 106          data      : 00006c9668155109a5502642d14,
len       : 0            data      :
codes     : {106}00006c9668155109a5502642d14, {0}0
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

I found this document as of 2016 describing the protocol:
WeatherSensor-RF-Protocols-IV_2016.pdf

As the next step I tried to interpret the recorded data using that document:

{105}00006c966815540922a055a0590
{104}00006c966815540852a0a5a022
{105}00006c966815540aa9500aa1310
{105}00006c966815540922a055a0590
{108}00006c96682919a59b541694285
{103}00006c96681554084aa011088a
{106}00006c9668155109a5502642d14

Sample: {106}00006c9668155109a5502642d14
V2.1 ?
0000	--- Preamble
6		--- Sync
c		??
	Payload:	9668155109a5502642d
	9668		--- Sensor ID
	1			--- Channel
	55			--- Rolling Code
	1			--- Flags (0x4 = battery low)
	09a550264	--- Data (variable)
	2d			--- Checksum
14		--- Post-amble

I think it might be V2.1 but not for sure. I don't understand the section of the document describing the sync nibble on page 7:

2. A sync nibble (4-bits) which is “0101” in the order of transmission. With
   v2.1 sensors this actually appears as “10011001”. Since nibbles are sent
   LSB first, the preamble nibble is actually “1010” or a hexadecimal “A”.

It would be great if someone could check my findings and get an opinion if the support of THGR228 would be possible with some acceptable effort. May be I am the only one, who is intersted in that.

Regards, Christoph

[zuckschwerdt]

The signals don't look like PWM but MC -- there is data in the gaps, see https://triq.org/pdv/
The odd timings might be due to clipping, try a greater distance. You can view clipping in the pdv-successor demo at https://triq.org/spectrogram-next/ (use scroll zoom)