Tecknet Doorbell

[Andrew Snider]

Very new to looking at radio signals (this is my first project involving them) and am trying to detect a wireless doorbell with rtl_433 and trigger something (probably via http).  This is the product page for the specific doorbell that I am looking at: https://www.amazon.com/gp/product/B08VDCYSNQ/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&th=1 which based on the id and pictures should be this: https://fccid.io/2AK8Q-TKWD003 

When I run rtl_433 -A I'm seeing the button press be detected and I have managed to capture it to a cu8 file.  Here's the triq analyzer link: https://triq.org/pdv/#AAB1080080005C056C004401CC010C003031DC829090A0B0B0C0B0B0D0B0B0D0B2B0B0A0B0B0C0B0B0D0B0B0D0B2B0B0A0B0B0C0B0B0D0B0B0D0E755

It looks to me like the signal is repeated three times but the modulation itself is not obvious to me (again I don't have a lot of experience though).  Due to the device's relatively simple nature it would not surprise me if there is no actual data to extract?  There is a number printed on the back of the button "22 D 0230 00817" but I don't see how such a short signal could correspond to this number.  rtl_433 seems to center primarily around extracting data after detecting a modulation.  If there is no modulation to detect and no data to extract how would I go about detecting this with rtl_433? Sorry if there is a conf that I've overlooked which already shows this.

Any help/direction to look in would be much appreciated Thanks

[Christian Z.]

From doorbells we'd expect a EV1527 signal, 24 or 25 pulses, either long or short.

But the signal here looks strange. There are pulses of different lengths, which could be okay, but the very long pulse of 1388 µs (as long as the packet gap of 1388 µs) does not seem like data.

Try to grab a sample file (-S all) and share as zip. https://triq.org/rtl_433/ANALYZE.html

[Andrew Snider]

002 I had previously captured with -S all and is the analyzer link I put previously.  I tried to capture a few more examples for comparison as I've noticed the signal seems to vary slightly in its length.  006, 008, and 009 look like additional clean captures while the other files just look like noise to me. 003-006 were captured with -S all and 007-009 were captured with -S unknown as I was getting a lot of noise.  It's maybe worth noting that the button emits a similar signal on both the press and depress.  I am not sure which ones 002 or 006 were but 008 and 009 should be press/depress respectively.  

I chose this button because it seems to be piezoelectric which I thought was a cool feature.  Don't think that's important but maybe it explains a few things.  Would it help if I get another one to compare with?

[Christian Z.]

I chose this button because it seems to be piezoelectric which I thought was a cool feature.  Don't think that's important but maybe it explains a few things.

That explains the timing. Without a battery the signals are usually very fast and short. This is what we see here.

It's not one long pulse but very fast pulsing. Best to raise the sample rate (-s 1024k) and make new captures.

Also use more distance or remove the antenna. The signal is clipping (too loud).

You can see all that by using a good sample on https://triq.org/spectrogram-next/ then zoom in and play with the controls (mouse scroll).

 

 Would it help if I get another one to compare with?

You could then get a better understanding of the contained data code. But it's not required -- those buttons are not cheap.

 

[Andrew Snider]

With the antenna removed and the higher sample rate I was able to get something that looks a lot more like data.  Captures attached and should be press/depress respectively though according to analyzer they are now the same.

https://triq.org/pdv/#AAB00B040100310095060827108255+AAB02104010031009506082710818190909090909081818190908181819081818190818255+AAB02104010031009506082710818190909090909081818190908181819081818190818355

I see what you mean by clipping in the picture what tool did you use to get that? 

[Christian Z.]

The signal looks great now. The clipping was visible in the read colored lines and chopped of sines in the waveform display (below to spectrogram) on https://triq.org/spectrogram-next/

The picture is just a zoomed in and adjusted view of the spectrogram to show the pulses.

rtl_433 suggestion for a decoder is fine. You this to grab data:

rtl_433 -R 0 -X 'n=tecknet,m=OOK_PWM,s=50,l=150,r=200'

Or to match only your code:

rtl_433 -R 0 -X 'n=tecknet,m=OOK_PWM,s=50,l=150,r=200,match={23}c0e776,countonly'

If you like, write a documented decoder as seen here: https://github.com/merbanan/rtl_433/tree/master/conf

and either PR on Github or post here for inclusion in rtl_433.

[Andrew James Grant Snider]

Thanks very much for your help Christian. 

--
You received this message because you are subscribed to a topic in the Google Groups "rtl_433" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/rtl_433/Cce_l2OghZk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rtl_433+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/rtl_433/1264a6fb-046b-4aba-ba8b-6dab2f328d64n%40googlegroups.com.