Telefonica - new protocol (Sejin)

Post by **The Robman** » Mon Jun 30, 2008 1:15 pm

[Edit]The real pairs are:
1 = +310 -0
0 = +0 -310

Once you've converted the signal to binary using +310 = 1 and -310 = 0, you then use the following pattern to reduce the 4-bit binary down to 2-bit binary...

1000 = 00
0100 = 01
0010 = 10
0001 = 11

The last nibble is a checksum which is the result of adding all the other nibbles together.

You have to also skip the first lonesome +310 burst.
[/edit]

I've been sent a bunch of learns from a "Telefonica" remote. I haven't seen the protocol before. It appears to have the following 7 burst pairs:

0 = +310 -0
1 = +310 -310
2 = +310 -640
3 = +310 -970
4 = +310 -1290
5 = +310 -1620
6 = +310 -1940

It's unusual because the 1st pair has no OFF time. It's also easy to miss the last pair because it merges with the leadout time.

If anybody would like to take a crack at writing an executor for it, the following zip file includes 3 IR files which contain the learns. It also contains a spreadsheet where I've converted the learns into hex and binary.

http://www.hifi-remote.com/forums/dload ... le_id=5627

binky123 · Post by **binky123** » Mon Jul 21, 2008 9:14 am

The signals look very similar to what was seen from the NightHawk Sejin STB remote here.

Post by **The Robman** » Mon Jul 21, 2008 12:57 pm

Apart from the frequencies, the signals are identical. At a glance, the codes look the same also. Thanks for the heads-up.

binky123 · Post by **binky123** » Mon Jul 21, 2008 3:24 pm

Great. Any ideas on how the executor should work to convert the 2 data bits into the IR signals? Also, is the name mispelled in this thread(i.e. Telefonica)?

Post by **The Robman** » Tue Jul 22, 2008 12:28 pm

I have updated the zip file so that it now also includes the learns for the Sejin Nighthawk STB that uses the same protocol (with a different frequency).

http://www.hifi-remote.com/forums/dload ... le_id=5627

I have also received a doc that lists the official hex codes for the Telefonica buttons, which I have included. I have started a spreadsheet where I am attempting to convert the official hex codes into the data in the learns. I have managed to figure out the formula for 5 of the 8 nibbles, but could use some help with the remaining three nibbles.

binky123 · Post by **binky123** » Tue Jul 22, 2008 11:40 pm

It looks good. I'm not sure if you can find a formula for the missing bits because they tend to need info from the other bits. Maybe you can use a 3rd nibble(middle 4 bits)? or a rolling 4-bit window? You may have to just process every 2 bits as in the algorithm and then coalesce the values?

Post by **The Robman** » Wed Jul 23, 2008 9:41 am

binky123 wrote:You may have to just process every 2 bits as in the algorithm and then coalesce the values?

I have no idea what that sentence means!

Regardless of whether we use assembler to generate the data or just have the upgrade provide it "as is", the generation of the actual signal will be the same. My plan would be to look at the data one nibble at a time and use the value in that nibble as a counter which would control the number of occurrances of the -320 uS OFF time are sent. I don't recall which protocol it was for, but we've used this approach before.

binky123 · Post by **binky123** » Wed Jul 23, 2008 12:55 pm

I think the 1 bit should have only an ON time and the 0 bit should only have an OFF time. The executor would then convert each 2-bit data part into 4 IR signal bits and then send the signals. The protocol 0161 seems to be very similar to these signals. Protocol 0161 was also used in the URC-39860 Remote from Spain. I will try to record an 0161 device upgrade tonight.

Ok, I learned several signals from the Atlas5 1025 VCR/DVD/1962 and they looked like the Telefonica and NightHawk signals. This protocol uses 3 fixed and 1 cmd byte.

I also tried the Device Upgrade and Protocol found on the URC-39860 hidden upgrade area and it generated the same signals. This protocol uses 5 fixed and 2 cmd bytes.

THe learned data is listed at 56.338KHz.

Post by **The Robman** » Thu Jul 24, 2008 12:51 pm

You're right, it is the $0161 executor. I have received an upgrade from UEI for the Telefonica (CBL/1385) and have loaded it here:

http://www.hifi-remote.com/forums/dload ... le_id=5673

I have also attempted to document the $0161 executor here:

http://www.hifi-remote.com/forums/dload ... le_id=5674

The executor is actually too big for PB (by 1 byte) at 193 bytes, but there's a lot of code in there that's not always needed. For example, the 2nd variable byte is a control byte and it's set to zeroes for every button in the CBL/1385 code.

The last fixed bit controls whether the signal is 38kHz or 56kHz.

Post by **The Robman** » Fri Jul 25, 2008 2:27 pm

Hey Binky,
Do you know how to translate the learned signals back into the source hex code?

Rob

binky123 · Post by **binky123** » Fri Jul 25, 2008 5:54 pm

This post groups the ON/OFF signals into 4parts and converts each 4PPM symbol back to the 2-bit data. The last symbol also uses zeroes as needed from the LeadOut to make up the 4part symbol. All the 2-bit data is then formed into four bytes. I believe the 3rd byte is the OBC value. In that example, 0x59 would be the OBC.

Post by **The Robman** » Fri Jul 25, 2008 10:49 pm

I get it now. Once you've converted the signal to binary using +310 = 1 and -310 = 0, you then use the following pattern to reduce the 4-bit binary down to 2-bit binary...

1000 = 00
0100 = 01
0010 = 10
0001 = 11

The last nibble is a checksum which is the result of adding all the other nibbles together.

You have to also skip the first lonesome +310 burst.

The following spreadsheet converts all the Telefonica and Sejin learns into hex. It also calculates the checksum byte for verification purposes.

http://www.hifi-remote.com/forums/dload ... le_id=5675

csena · Post by **csena** » Tue Oct 07, 2008 1:30 pm

I'm trying to assemble a bin file for Slingbox with RemoteMaster. I have a Slingbox classic, with PL chip, and using an ADB-3800W STB, from Telefonica. I was informed that this STB uses SEJIN at 56 KHz.
The part I'm missing is the protocol ID, since I can't find SEJIN in RemoteMaster. How can I find the protocol ID?

binky123 · Post by **binky123** » Wed Dec 10, 2008 10:50 am

As a followup, this thread contains a Slingbox PL BIN file that works for the ADB3800W STB and may work for the NightHawk STB.

Post by **mathdon** » Wed Jul 29, 2009 9:31 am

I've attempted to convert this protocol into IRP notation, before attempting to add it to DecodeIR. I would be pleased if someone could confirm that I've got it (somewhere near) right:

Telefonica {56.3k,310,msb}<-1|1>(7:3,0:3,1:1,<1,-3|-1,1,-2|-2,1,-1|-3,1>(D:8,S:8,F:8,C:8))+

where the checksum C = D:4:0 + D:4:4 + S:4:0 + S:4:4 + F:4:0 + F:4:4.

Until I started trying to add protocols into DecodeIR. I hadn't appreciated just what a useful notation this is. It is so concise yet says everything one needs to know, and seems to work even for complicated situations like this one.
___________________
Graham