Strange problem with Sony type signals with more than 20 bit

[jayo]

I am working on an Airsoft project where I am trying to implement the Milestag protocol which uses timings from the Sony protocol - but I have hit a small problem.

So far I have been working on 14bit "Shot" messages and things have been going fine.

But, a few days ago I started working on some of their system messages starting with their "Reload" command which is a 2400 header followed by a 3 byte command (24 bits) and I have been seeing some really strange signals very much like some of those on the AC threads with some random long bursts.

Anyway, I went back to basics and I wrote this simple sketch using IRRemote - basically it keeps sending a 24 bit code (FFFFFF) every 5 seconds or so

Code: Select all

#include <IRremote.h>

IRsend irsend;

void setup()
{}

void loop() 
{
      irsend.sendSony(0xFFFFFFFF, 24); // Sony TV power code
      delay(5000);
}

Sometimes in AnalysIR the signal received looks correct, but almost as often it will not be correct.

I have attached a screenshot of the trace below - notice the very long pulse in the middle as well as the single spike shortly after.

I exported the data which is
0000 006D 0000 0018 0060 0014 0030 0018 002F 0016 0032 0014 0030 0016 0032 0015 0030 0016 0031 0015 0030 0017 01C3 0018 002F 0017 0015 002F 0017 0030 0016 0030 0017 0030 0015 0030 0017 0031 0014 0030 0016 0030 0016 0031 0016 0031 0014 0030 0018 002F 0017 002F 0026

I have also attached a screenshot of the signal I measured on the data leg of the TSOP2238 - clear header pulse followed by 24 data pulses.

From my testing so far I am seeing this problem once the data contains more than 20 bits - 20 bits works fine and is analysed as a 20 bit Sony code, but after that it is coming up as raw with these differing signals in AnalysisIR

Any light that you could shed on this would be great

[AnalysIR]

Hi Jay

Thanks for the well presented post.

As you guessed, SONY must be 8, 12, 15 or 20 bits & is implemented as such within AnalysIR.

Have you found any corruption on the Oscilloscope with these signals? That IR receiver may not be the best? What is the modulation frequency of the source Signal - 40kHz like Sony or 38/56? If the modulation frequency doesnt match that could explain some corruption.


As we are keen to have MilesTag tested with AnalysIR, we will add in support for you & get you a preview to test with.

FIrst, I would like to get some more valid signals and some corrupted ones. You can just record them and export your Session History (Menu->File->Save Session) and just email me or upload that file here.

Once I get the signals I will send you an updated version.

PS: sorry about the delay in responding...

[AnalysIR]

Jayo

Following your post on the forum, we have added (untested) support for 24-bit SONY.

Up to 20 bits will appear as SONY & 24 bits should appear as MilesTag.


Download details have been sent to you directly via email.


Let me know if it works!

[reeta]

Hi,

Thank you for your immediate reply. I've tried it! And... IT WORKS, it works, it works!! So how can i decode correctly other signals too, e.g Turn off?? is there a pattern. thanks for you help!
Pasc

[AnalysIR]

Thank you for your immediate reply. I've tried it! And... IT WORKS, it works, it works!! So how can i decode correctly other signals too, e.g Turn off?? is there a pattern. thanks for you help!

Sorry, your question is not clear. Please provide more detail.