Re: Preventing Random X-10 Code interference...

[nobody@xxxxxxxxxxxx (Dave Houston)]

"Robert Green" <ROBERT_GREEN1963@xxxxxxxxx> wrote:

>Did you ever get the screenprint I sent?  It showed a code fragment from a
>deliberately induced set of collisions.  Two palmpads both set to B sending
>"1" on and "3" on simultaneously produced a single command for the M
>housecode.  I'd sure love to figure out why we get so many reports of
>phantom codes.

I got the screenprint but trying to interpret it gave me a headache. ;)

A minor but necessary quibble: It was two transceivers that created the
collision.

>Now that you've taught me how the signal is Manchester encoded, I am in
>agreement that it's highly unlikely that phantoms codes are generated by
>noisy appliances.  I do believe that there are powerline "events" that can
>trigger X-10 transmitters simply because because unplugging and replugging
>certain modules puts lots of BSC's on the powerline, at least according to
>the Monterey.  Something's causing people to post here, week after week,
>that they are seeing phantom activations, or more popularly, phantom
>activity log entries.

With the Monterey reporting every 1110 sequence that isn't followed by a
good code as a BSC, any burst of ~120kHz that lasts for 3 or more half
cycles will create 1110 at the first clear half cycle.

Many people jump to the conclusion that an on/off event means there are
valid codes on the powerline but this is seldom the case. I always try to
determine whether or not anything is logged and then look for a transmitter
when there are real codes. However, with the proliferation of manufacturers
who have their own feature sets, we may see more and more two-way modules
that will generate codes after some transient event.

>It's hideously hard to try and map out even a single command on the Monterey
>because of the tiny readout.

That's why I never recommend it - it just seems far too costly for what it
does.

>Well, here's to your good health and the hope that you'll eventually get
>around to it.  A marriage of the Lynx's ability to show each frame and bit,
>the Monterey's ability to read the voltage of each bit and Activehome's
>logging ability would be sweet.

My plan has been to output 3 bytes per half cycle. They would give the
amplitude (8-bit resolution) at ZC+250µS & ZC+900µS as well as the number of
rising edges between those points. All would be logged to a file with a
timestamp for the start of each sequence. Logging would stop once 3 idle
cycles were seen and resume on a new line at the next "1" seen. Software
would be able to reconstruct the sequences. Even the average frequency of
the carrier or noise could be calculated.