From: wrench@engrs.unl.edu (Bradley Riensche)
Newsgroups: sci.electronics
Subject: Summary: RCA IR remote
Date: 2 Sep 1994 04:20:32 GMT

First, I'd like to thank all who replied to my request a few
weeks ago regarding RCA IR remote controls.  The following
is a summary of the results:


Mark Devries (devries@erich.triumf.ca) wrote:

    The RCA remotes used different codes depending on who made for RCA
and when it was made for them.  For example my older RCA VCR will
(the remote) will not run my newer RCA VCR.  A salesman told me this
was because one VCR was made by Hitachi and the other by panasonic.  I
(think) there may even be more than two RCA codes.  
<hr>

Dave Turner (ptsfa!dmturne@gw.PacBell.COM) wrote:

I have two RCA TV sets and each use a different scheme.

One is about 10 years old and uses a Digital Command remote control.
The channel memory is in the remote control rather than in the TV.

The other TV is three years old and uses a different scheme.

Both send a long MARK pulse followed by a long SPACE pulse.
The data is repeated twice: the second it the complement of the first.
By ANDing the two sets of data it is possible to check for transmission
errors.

The data is followed by an EOT pulse.

A data bit is a short positive pulse followed by a somewhat longer negative
pulse for a zero or an even longer negative pulse for a one bit.

The actual pulse widths depend upon the TV involved.
<hr>

I looked at my RCA remote control coding scheme using a logic analyzer
and discovered the information below. My RCA TV is about 1 year old.
From what I can gather, the coding scheme and even the IR wavelength
is highly dependent on the age of the remote.  RCA must have switched
methods _many_ times.  Here's my data:

     +------------------+         +-+ +-+   +-+         +-+         +---
     |                  |         | | | |   | |         | |         |
 ----+                  +---------+ +-+ +---+ +- . . . -+ +---------+
     |------- t1 -------|--- t2 --|T|T|T|3T |             |--- t3 --|
                                  |'0'| '1' |   

t1 = 4.2ms
t2 = 3.8ms
t3 = 7.8ms
T  = 600us

Each packet is 24 bits in length, and does not appear to exhibit the
coding scheme where the second 12 bits are the logical complement of
the first twelve. Here is an example of what I decoded by hand:

$23 $0E $CF - OFF key

It has been awhile since I did this experiment and I have misplaced 
some of my notes, so some of this is from memory.  I will post updates
if I discover any discrepencies.

B. Riensche