FSK with radiohead and Si4432

[Peter Drozda]

Hi guys,

I'm doing some experiments with intercepting the signal from my Arduino alarm project..

The alarm is using Arduino pro mini in low power mode and Si 4432 as a long range transciever..

The code is using Radiohead library with RH_RF22 driver and FSK_Rb2Fd5 mode which should be according to documentation = FSK, Whitening, Rb = 2kbs, Fd = 5kHz

Please find attached two samples in .cu8 and .sr format..

It seems that RTL_433 cannot handle that in -a nor -A mode.. it does not output any useful information..

But when I look at it in the Pulseview.. it seems that the signal makes some sense..

It is clearly FM modulation with 500us per bit which would correspond the settings of 2kbs..

It seems there is some quite long training preample and then some data follow..

Currently I'm trying to pair it with my implementation in Arduino, but maybe I will have to implement some testing mode to have less data transferred..

But the questing is... is RTL_433 able to handle such type of modulation?

And if so.. then which type of modulation would it be in terms of RLT_433?

Thanks!

[Christian Zuckschwerdt]

The FM changes look awefully smooth and are phase locked. I'd say that's GFSK and not just plain FSK. It's a very neat and bandwidth-efficient signal but the soft slopes don't work with out current level discriminator. (GFSK is on the to do list, but requires a major overhaul…)

[Peter Drozda]

 :)) LOL .. awefully smooth and very neat signal ... :) I'm getting kind of proud of the transciever I choose ;)

On Mon, Sep 24, 2018 at 11:03 AM, Christian Zuckschwerdt <chri...@zuckschwerdt.org> wrote:

The FM changes look awefully smooth and are phase locked. I'd say that's GFSK and not just plain FSK. It's a very neat and bandwidth-efficient signal but the soft slopes don't work with out current level discriminator. (GFSK is on the to do list, but requires a major overhaul…)

--
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/QcMpGdSOhIo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rtl_433+unsubscribe@googlegroups.com.
To post to this group, send email to rtl...@googlegroups.com.
Visit this group at https://groups.google.com/group/rtl_433.
To view this discussion on the web, visit https://groups.google.com/d/msgid/rtl_433/0935cfb9-e2c8-44e2-8310-d060e9da880f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Tommy Vestermark]

Hi Peter,

Looking at the signal, it should be possible to demodulate with the current code. However the detector has some basic assumptions about choosing a suitable sample frequency to get proper modulation range. To get detection, the frequency difference (Fd) should be above ~1/20 the sampling frequency (3000/65536)

When looking at the files, it looks like your signal is sampled at 250 ksps yielding a modulation ratio of only 1/50. Could you try again with a sampling frequency of 100 ksps?

Cheers,

On Monday, September 24, 2018 at 1:23:01 PM UTC+2, Peter Drozda wrote:

 :)) LOL .. awefully smooth and very neat signal ... :) I'm getting kind of proud of the transciever I choose ;)

On Mon, Sep 24, 2018 at 11:03 AM, Christian Zuckschwerdt <chri...@zuckschwerdt.org> wrote:

The FM changes look awefully smooth and are phase locked. I'd say that's GFSK and not just plain FSK. It's a very neat and bandwidth-efficient signal but the soft slopes don't work with out current level discriminator. (GFSK is on the to do list, but requires a major overhaul…)

--
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/QcMpGdSOhIo/unsubscribe.

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

To post to this group, send email to rtl...@googlegroups.com.
Visit this group at https://groups.google.com/group/rtl_433.

[Peter Drozda]

Hi Tommy,

Thanks a lot.I will try it.I captured the signal with default settings where the sampling frequency is 250kHz..

But still.. I don't understand how lowering the sampling rate can yield the better results? And what happens to other sensors when I change this setting to lower frequency?

Would it help if my transmiter uses higher bitrate?

Shouldn't it be the oposite way?

Maybe if possible, could you please give me "for dummies" explanation? :)

Thank you!

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

To post to this group, send email to rtl...@googlegroups.com.
Visit this group at https://groups.google.com/group/rtl_433.

To view this discussion on the web, visit https://groups.google.com/d/msgid/rtl_433/6801fae7-f738-47f9-964b-16d3d822fbe2%40googlegroups.com.

[Peter Drozda]

Hi Tommy,

I read your post over and I thik I understand..

Maybe I should use the following setting at my transmitter

FSK_Rb2_4Fd36 in Radiohead terms..

It would still use FSK with 2,4 kb/s but fd would be 36 khz which is above of 1/20 default sampling rate..

Currently I'm using FSK_Rb2Fd5 which has fd=5khz

Definitelly worth trying...

Thank you very much !!!

On Monday, October 8, 2018, Tommy Vestermark <tommy.ve...@gmail.com> wrote:

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

To post to this group, send email to rtl...@googlegroups.com.
Visit this group at https://groups.google.com/group/rtl_433.

To view this discussion on the web, visit https://groups.google.com/d/msgid/rtl_433/6801fae7-f738-47f9-964b-16d3d822fbe2%40googlegroups.com.

[J Dow]

Who figured out how to get rtlsdr dongles to sample at a 100 kHz rate? Last I knew 225001 was the lowest rate the dongle would accept.

{^_^}

[J Dow]

That's a rather complex subject. So far it appears nobody has put in provisions for either spread spectrum or PSK transmissions, both of which give better performance than ASK or FSK of any form. RTL_433 is a general purpose tool. So it does not do well at the edges with its brute force demodulation and decoding techniques. Fortunately at present keyfobs, wx stations, and lots of other things all seem to use simple ASK or FSK. (This no doubt pleases the FCC as they do not like difficult to decode things in the hands of GASP SHUDDER civilians. Of course, they have to hear it before they can object to it.)

{^_-}   Joanne