Flex Decoder data inconsistent

[js13]

I have a TPMS sensor I'm trying to decode. After running "-A" I have received the flex decoder criteria of " -X  'n=name, m=OOK_MC_ZEROBIT, s=112, l=0, r=208'." I was a bit skeptical of this decoder since I am unsure of why long has a value of 0. Regardless, I ran it, and I do receive data that looks something like this:

"00000000000000000000000096532bed832d665adc46bc"

Immediately I noticed that this data is not manchester encoded, contrary to most TPMS sensors I've seen. I have ran the flex decoder with changed values like temperature, pressure, and changing the sensor itself. Every time I run it, the data always starts with "0000000000000000000000009653," which I am assuming is a preamble/sync word. Other than than, the rest of the data seems to be completely random. Even if I run the decoder with the same temp, pressure, and sensor, the hex byte information is different, making it impossible to find patterns. 

Does anyone know what the source of inconsistent data is? Could it be the flex decoder having improper parameters? Any advice is appreciated.

[Christian Z.]

That data is already MC decoded (OOK_MC_ZEROBIT), the MC decoder can only handle s=l and ignores l. Around 100 µs is an expected length for TPMS. Get the raw data with OOK_PCM,s=100,l=100 (PCM can handle RZ, i.e. s!=l)

S.a. https://triq.org/rtl_433/PULSE_FORMATS.html

Raw data is recommended for the initial analysis to manually confirm it's really proper MC without de-sync's inserted.

[js13]

Thanks for the reply.

What do you mean raw data is recommended for the initial analysis to manually confirm it's really proper MC without de-sync's inserted? Isn't raw data necessary for documenting the data codes as per the final step of capture and analyzing a new device?

[Christian Z.]

Sorry, bad wording. I meant raw bit data, PCM data, as opposed to already decoded MC. After decoding PCM to MC the possible errors in the code are not visible anymore.

[js13]

Ok, so even after trying those new flex decoder parameters, I'm still not finding consistencies in my data. When using a TPMS, its expected that each sensor should have a unique ID, so there should be a string of bits representing the ID in the data. However, even after running the same sensor twice, I get completely different data bits. Any insight?

[Christian Z.]

Tell us about the TPMS in question. Some have encryption. An example is the Jansite TPMS and similar ones.

[js13]

The TPMS can be found at this link: https://www.ebay.com/itm/404356176555

It's unbranded, but I got it because I wanted to see if I could decode it myself. 

[Christian Z.]

That looks like the sensors we know as Jansite. Search the rtl_433 issues for that and maybe "chinese tpms" or "no name tpms" to see the decoding efforts. If you can get an alignment then the first byte should be something similar to an XOR key to the rest -- at least the entropy goes down.

[js13]

Seems like almost all other TPMS are FSK encoded, is it strange that mine is OOK? 

[Christian Z.]

It might not be OOK, you could just have tuned one of the frequencies out. Try to offset by +/- 20 k to 50 k center frequency.

[js13]

I have tried offsetting the frequency and ensured that the signal is not getting clipped... it is definitely OOK. I captures some samples with your recommended flex decoder specs " OOK_PCM,s=100,l=100." 

This was taken at 17 psi and 27 c with Right Rear sensor

This was taken at 9 psi and 27 c also with Right Rear Sensor

This was taken at 14 psi and 27c with the Front Right sensor

There's clearly a preamble/ sync word, but other than that, there is no clear sign of an ID, which I know is one of the first things to look for at a TPMS.

you mentioned that the Jansite TPMS is encrypted, is there anywhere I can find additional information on this?

[Christian Z.]

The search for all Jansite related issues should be: https://github.com/merbanan/rtl_433/issues?q=is%3Aissue+jansite

And the original one should be #1007, this: https://github.com/merbanan/rtl_433/issues/1007

But I remembered wrong, sorry, actually these are the encrypted TPMS issues:

https://github.com/merbanan/rtl_433/issues/1059
https://github.com/merbanan/rtl_433/issues/1820