Writing a new device driver

[Or Goshen]

Hi,

I'm trying to write a new device driver for a new device.

I know the following things about that device:

Transmission frequency: 916.3Mhz

Modulation: FSK

Codec: Manchester 1

I manage to capture signal using the command "rtl_433 -f 916.32M -S all -T 60" and analyze it using another program called URH (Universal Radio Hacker, that is how I figured it uses FSK and manchester 1).

I know exactly what a transmitted packet contains and where the fields are (Used URH).

Question is - how to I use this knowledge to write a driver ?

Right now I'm stuck at parsing a captured packet using rtl_433.

I run the following command:

"rtl_433 -A -r g001_916.3M_1000k.cu8" (same file I analysed successfully in URH), but the output doesnt make any sense.

I tried running the following command with the flex driver:

"rtl_433 -R0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=16,r=4000" -r g001_916.3M_1000k.cu8 (I guessed the s,l,r values - no idea what to assign to them)

The output doesnt look the same as it is in URH.

What am I doing wrong / what am I missing ?

Thanks

[Christian Z.]

I tried running the following command with the flex driver:

"rtl_433 -R0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=16,r=4000" -r g001_916.3M_1000k.cu8 (I guessed the s,l,r values - no idea what to assign to them)

Proper timings are the first thing to analyze. For MC and the underlying PCM (NRZ) short and long are the same value: the length of a (half-)bit.

Use a FSK_PCM decoder first and set s and l to the half-bit length. Then check if the bitstream really is valid MC and maybe switch to using MC, otherwise decode the MC later.

 

[Or Goshen]

I know for a fact that the codec used is Manchester 1.

When I'm using this tool I get correct results (as seen in attached picture).

I also attached the signal file.

The command:

rtl_433 -R0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=8,r=4000" -r g001_916.3M_1000k.cu8

.. doesn't return the same results the online tool does.

That should the "r" value be ?

[Christian Z.]

Use the minmax demod and 20 for reset. See https://github.com/merbanan/rtl_433/issues/2479 why the output has a different prefix.

rtl_433 -Y minmax -R 0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=8,r=20"

[Or Goshen]

Result of "tl_433 -Y minmax -R 0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=8,r=20" -r g001_916.3M_1000k.cu8" is:

time      : @0.045390s
model     : Test         count     : 1             num_rows  : 1             rows      :
len       : 220          data      : 7fffffffffc196f5138537b4ef079033ed786ff5fffabeed8e06870
codes     : {220}7fffffffffc196f5138537b4ef079033ed786ff5fffabeed8e06870

Result of the online tool (for same file which I attached) is:

Bits: {234} 00 00 00 00 00 00 / 06 5B D4 4E 14 DE D3 BC 1E 40 CF B5 E1 BF D7 FF EA FB B6 38 1A 1C 0

The online tool returns the correct result.

How can I make rtl_433 return same result as the online tool ?

[Christian Z.]

Except for the prefix (which is a preamble and has no data) the bits are the same, see

https://triq.net/bitbench#c=7fffffffffc196f5138537b4ef079033ed786ff5fffabeed8e06870&c=000000000000065BD44E14DED3BC1E40CFB5E1BFD7FFEAFBB6381A1C0&f=v

Also note that here actually PDV is wrong about the preamble. It is truely all 1's as in 7fffffffffc. And that is denoted by the slash ("/") in the output.

[Or Goshen]

Thanks,

I was missing the "invert".

This works:

rtl_433 -Y minmax -R 0 -X "n=Test,m=FSK_MC_ZEROBIT,s=8,l=8,r=20,invert" -r g001_916.3M_1000k.cu8