Need help with rtl_433 reading Intelbras XAS 4010 Smart sensors

[Daniel Pakuschewski]

Hello everyone, 

I just bought a SDR-RTL usb dongle and plugged it in my Home Assistant server. I thought it would be a plug-n-play solution but I have been strugling to read the data from some Intelbras XAS 4010 Smart sensors that I have around my house. 

Before going direct to Home Assistant, I have been trying to run rtl_433 direct in my shell to see if everything works. So far no luck.

First, I started to read a issue in Github from a user that have been trying to do the same:
https://github.com/merbanan/rtl_433/issues/588

Before closing the issue, it was mentioned that a -A parameter should give the right -X working spec.

I got it but Im unable to read the data coming from the device.

This is what happens when I run with the suggest -X and open/close the sensor:
https://pastebin.com/0YMDAmzV

I trying to get a readable format so I can inject it on home assistent using mqtt server (if I start the process connecting to mqtt, nothing happens when I open/close the sensor.

Any guidance that you guys can provide would be really helpful.

Thanks

[Christian Z.]

The intelbras sender should be FSK PCM at 622 µs.

In your example you are (successfully: `{38}24884e1c48`) reading OOK MC at 508 µs.

Might be some other signal or a different version of the sender.

Try to locate the signal with some SDR UI first, then grab a sample, confirm good and share for analysis.

https://triq.org/rtl_433/ANALYZE.html

[Daniel Pakuschewski]

Hey Christian, I was able to collect some data (confirmed on I/Q Spectrogram). I'm attaching the file here.

I tried to run rtl_433 -A using the file but I have no idea what to do from here (https://pastebin.com/2c4GBu7i)

Is this what you asked for?

---

Daniel Pakuschewski

--
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/BJrMiEZu1dc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rtl_433+u...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/rtl_433/cd0874d9-903c-43e8-a1aa-ca99ab950573n%40googlegroups.com.

[Karl Lohner]

I'm able to grab this data with this:

rtl_433 -r g002_433.92M_250k.cu8 -X "n=Intelbras,m=OOK_PCM,s=508,l=508,r=2032,preamble={1}8,symbol_zero={3}2,symbol_one={3}6"

You received this message because you are subscribed to the Google Groups "rtl_433" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rtl_433+u...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/rtl_433/CAEB98e8mVRkKPwPfoMr8WtnsFbyDpomBfThGg7n9wQ49vPxyNQ%40mail.gmail.com.

[Daniel Pakuschewski]

Hey Karl, thanks for it. Looks like Im able to capture it here with this config.

I just ran a test with 2 windows here and I was able to capture it. https://pastebin.com/e05t0YDb

Another dummy question not related to the analysis by itself. If I understood correctly this device will only report when the window is open, which makes sense for an alarm device. In this case it's not possible to get the windows status (open/closed) right? Am I missing something?

---

Daniel Pakuschewski

[Karl Lohner]

I'd guess it's sending an event when opened and another event when closed.  From your pastebin, it looks like one of those events is {28}da8f5ca and the other is {28}a9171ca.  You'd need to experiment to see if you get any other codes.  Bitbench is your friend if you want to try to make sense of the bits.

[Daniel Pakuschewski]

Actually it's from another sensor, I had 2 here and the ID events change based only on the device, and they always send the same data only when I open the sensor. 

I was reading the device manual and its expected, according to it, the battery level and open/close is available only on FSK modulation, luckily they have a jumper in the device to use FSK instead of OOK.

I got one device and changed the jumper and now I can get events when I open and close it. The problem is that all the progress I made was on OOK and it looks different on FSK now.

What I have done so far:

I captured a new ANALYSER using the -S unknown. (https://pastebin.com/TAdvdSsY)

The output is attached. 

Then I got the files and tried to run with -A on the file. (https://pastebin.com/rgTQGnt4) it suggested me to run using -X 'n=name,m=OOK_PWM,s=160,l=39716,r=8308,g=0,t=15822,y=0'

I tried to run (https://pastebin.com/RtSyCBa7) using this parameter and I got the same data when I close or open the sensor. The weird thing is that the data is always the same (800) for both events.

I have no clue on what I should do from here :/ 

Any thoughts?

---

Daniel Pakuschewski

[Christian Z.]

The samples are okay-ish but have to clipping (too loud), you can see this in the demo at https://triq.org/spectrogram-next/

Try to get more distance or remove the antenna, you don't want orange or red areas in the visualization.

The flex decoder look wrong. It should be something like FSK_PCM,s=600,l=600

[Daniel Pakuschewski]

Ok, so I removed my RTL-SDR dongle from my server and I connected it directly to my mac to try to get more details. I was not able to reduce the noise that you mention even removing the antenna and moving the device. I added some .cu8 samples of the latest test I ran.

I also downloaded the Universal Radio Hacker to try to understand better in a GUI scenario what's going on. Using it I was able to (I believe) get a clean signal. My recording is available here. (For some reason Google doesnt allow me to send it as attached). 

Honestly I have no idea how to configure rtl_433 to have the same sample as I got with URH.

Is this any kind of progress? Any suggestions? :)

[Christian Z.]

rtl_433 can deal with some clipping but for a good analysis we need a sample with the waveform intact (clipping rips off part of the waveform). Maybe try a fixed gain (start with -g 30) possible values should be 0 to 50.

Otherwise you might have luck with the already known FSK flex decoder.

[Karl Lohner]

FWIW, this works on one of the samples you'd provided.

rtl_433 -X "n=Intelbras_FSK,m=FSK_PCM,s=620,l=620,r=4960,preamble=aaaa2dd4" -r RTL-SDR-20250108_220216-433_883MHz-1MSps-1MHz.complex16s

On Thu, Jan 9, 2025 at 7:10 AM Christian Z. <chri...@zuckschwerdt.org> wrote:

rtl_433 can deal with some clipping but for a good analysis we need a sample with the waveform intact (clipping rips off part of the waveform). Maybe try a fixed gain (start with -g 30) possible values should be 0 to 50.

Otherwise you might have luck with the already known FSK flex decoder.

--

You received this message because you are subscribed to the Google Groups "rtl_433" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rtl_433+u...@googlegroups.com.

To view this discussion, visit https://groups.google.com/d/msgid/rtl_433/366d3895-04b5-4a79-a2c7-063c94b1e0acn%40googlegroups.com.

[Daniel Pakuschewski]

Hey guys,

rtl_433 can deal with some clipping but for a good analysis we need a sample with the waveform intact (clipping rips off part of the waveform). Maybe try a fixed gain (start with -g 30) possible values should be 0 to 50.
Otherwise you might have luck with the already known FSK flex decoder.

I tweaked my setup a bit  here and I believe I was able to get a clear signal now (attached .cu8 files). whenever I try a value on -g bigger than 0 it will not collect any signal when I open/close the device.

FWIW, this works on one of the samples you'd provided.

rtl_433 -X "n=Intelbras_FSK,m=FSK_PCM,s=620,l=620,r=4960,preamble=aaaa2dd4" -r RTL-SDR-20250108_220216-433_883MHz-1MSps-1MHz.complex16s

Indeed it works here with the sample I collected using the Universal Radio Hacker, but I can't make rtl_433 work in the same way. Any glue?

---

Daniel Pakuschewski

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/BJrMiEZu1dc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rtl_433+u...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/rtl_433/CAJ%3DPG2jmHQKRT-jvxZ5ByUgTDmyC-DWP3BpBhFaHoV61axoi%2BQ%40mail.gmail.com.

[Christian Z.]

The samples are great! Note that -g 0 is automatic gain and the default :)

The samples you got from rtl_433 do work with the flex decoder, it should also work live. Be sure to use -Y autolevel -M level -M noise

[Daniel Pakuschewski]

Hum.. I may be missing something. 

I just ran using the suggested flex decoder by pulse analysis, but it doesn't look right, looks like the data is not complete.
Here is what I tried: https://pastebin.com/r57wtiLF

If I try the flex decoder suggested by Karl, it doesn't capture anything.

rtl_433 -X "n=Intelbras_FSK,m=FSK_PCM,s=620,l=620,r=4960,preamble=aaaa2dd4" -Y autolevel -M level -M noise

---

Daniel Pakuschewski

Em dom., 12 de jan. de 2025 às 18:10, Christian Z. <chri...@zuckschwerdt.org> escreveu:

The samples are great! Note that -g 0 is automatic gain and the default :)

The samples you got from rtl_433 do work with the flex decoder, it should also work live. Be sure to use -Y autolevel -M level -M noise

--

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/BJrMiEZu1dc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rtl_433+u...@googlegroups.com.

To view this discussion, visit https://groups.google.com/d/msgid/rtl_433/60551b28-1b6b-4e4e-85e3-cc263ec19d7en%40googlegroups.com.