ESM1 X-10 Broadcast storm detector cradle

["Robert Green" <ROBERT_GREEN1963@xxxxxxxxx>]

Jeff Volp and I have been discussing, in another thread, the construction of
a docking cradle for the Elk ESM1 model X-10 power line meter that would
detect "broadcast storms" of X-10 signals.  These storms are typically
created by stuck transmitter buttons and stand-alone controllers like the
CM11A running wild.  When a broadcast storm occurs, the household lights and
X-10 devices are typically frozen in their current state and cannot be
changed remotely.  These storms, at least in my house, have a very low SAF,
especially if they occur during a critical time, like when someone's late
for work.

The ESM1 meter, as seen here:

http://www.elkproducts.com/products/esm1.htm

is designed in two pieces - a wall wart and a small, rectangular meter head.
There are two small LED's, (green for X-10 signal, yellow for power) and in
between those a ten segment bar graph.

The basic premise of the Broadcast Storm Detector Cradle (BSDC) is to
monitor the green X-10 signal LED and to sound an alarm when that LED is lit
for more than 30 seconds.  The ideal controller for such an interface seemed
to me to be the venerable 555 series IC timer package.  There are some great
sites around describing working with the various 555's.  Rob Paisley's LM555
page:

http://home.cogeco.ca/~rpaisley4/LM555.html#20

is one of them.

I was thinking of trying to tap into the ESM1's wall wart but it's probably
better to use a separate power supply that could also drive one or two loud
sirens through some fairly long cabling so I could have the alarm sound on
multiple floors.

The physical design will likely be very simple.  A small piece of plywood
large enough to fit both the meter head and the cradle's electronics.  Some
small pieces of molding will create the alignment edges. I thought a small
piece of wood exactly the size of the faceplate inset of the meter head and
covered with black felt to be mounted so it covers the faceplate when docked
would hold the phototransistor that will be "slaved" to the output of the LE
D.

When the LED lights, indicating X-10 traffic, the output of the
phototransistor starts a timing sequence.  If it ends before thirty seconds
(it doesn't need to be precise as much as it needs to be easily adjustable)
then the unit resets and nothing happens.  If the X-10 signal LED stays lit
for more than thirty seconds, a warning buzzer sounds.  The buzzer can be
regulated by a second 555 chip to time out after 5 minutes or however I
decide to handle the "storm detected" sequence.   It might be nice to run it
from a 24hr mechanical timer just to keep it from beeping late at night when
I'd rather sleep than get up and deal with locating a loco transmitter.

After some testing, it might turn out that it's more important to monitor
the signal strength bar graph because that would provide information about
noise which can indicate collisions or some sort of beat frequency
interference.  Since the display LED flickers as well, I'll need some way to
stretch the pulses coming from the phototransistor so that the continuous
flickering is considered a single, long ON signal.

Fancy features like a tiny CCD video cam in the cradle to transmit the bar
graph display on the home CCTV system should aid in troubleshooting the
source of a broadcast storm.  Adding several wallwarts throughout the house
to feed into the meter head through a rotary switch or a stepping relay
could give snapshots of different branch circuits from a single location
without having to move the meter to outlets all over the house to find the
source of the rogue signal.

This is the sort of soldering project that probably anyone (even fat
fingered me!) can do since it involves mostly discrete components with nice,
thick, easily solderable legs and not the anchovy spine-sized leads that
come off surface mount components!  (-:  It also doesn't involve any
disassembly of the ESM1.  The cradle couples to the meter head optically.

I probably should scan Google to see where the 555 hackers spend most of
their time.  and crosspost this there as well.  I'm sure this is "walk in
the park" stuff for anyone who has designed complex timing circuits with
external triggers.  I haven't wired up a 555 to anything since about 1984 so
it's bound to be more of a challenge for me than for others.

--
Bobby G.