Re: X10 signals can be TOO strong!

["Jeff Volp" <JeffVolp@xxxxxxx>]

"Dave Houston" <nobody@xxxxxxxxxxxx> wrote in message
news:4581ede7.129416375@xxxxxxxxxxxxxxxx
> "Jeff Volp" <JeffVolp@xxxxxxx> wrote:
>
> >I built a simulation of the AC distribution network in a typical home
using
> >inductance based on wire length.  The numbers were from the audio guys
whose
> >high power speaker cables are similar in dimensions to Romex.  I also
added
> >distributed capacitance and a few signal suckers.  It is easy to see from
> >that simulation why the signals fall off as they do.  An AC distribution
> >network is basically a very complex low-pass filter that has various
peaks
> >and nodes throughout the system.  The nodes are the "black holes" where
it
> >is virtually impossible to get a decent signal without moving the
> >transmitter.
>
> There have been numeroous white papers published by various academics and
by
> engineers at companies looking at developing powerline products that
treated
> the powerline as a communications medium. I have cited many of them many
> times in c.h.a. over the past few years.
>
> Most of them dismiss the inductance of the wire itself. They mostly felt
the
> inductance of the terminations and the capacitance of various loads were
the
> important factors.


The wire manufactures also dismiss the inductance, and at first I didn't
think it would be a factor either.  I pulled in well over 1000 feet for just
the lower level, and some of those runs have to be over 100 feet.  The specs
I found on some Monster cables indicated about .2uH per foot.  So, series
inductance in the 5 to 25 uH region is reasonable, and that is the range in
my simulation.  While distributed capacitance of the wire itself can be
neglected, devices plugged in cannot.  So I have included capacitance at
random locations ranging from .01uF to a signal sucking .1uF.  It is
interesting to note that 25uH and .1uH is resonant at 100KHz.  In my
frequency sweep test, the longer runs had peaks and nulls at various
frequencies.

> >Since Insteon is based on having multiple transmission points, signal
> >strengths throughout the network should be more constant.
>
> Agreed, but I'm sure there have been at least a few occasions in the past
30
> years with all the many millions of X-10 modules in use when an X-10
signal
> reaching a dimmer (X-10 or otherwise) was at least as high as the average
> Insteon signal. Why haven't we heard of flicker before this?

I don't expect to see flicker at normal X10 signal levels.  I just reported
this special case in case anyone else observed something similar.  The
XTB-II is normally connected at the distribution panel.  I measured about
20Vpp on those busses.  From there almost 40 circuits branch out throughout
the house, and I saw 1Vpp to 5Vpp (ESM1 max) on all circuits I tested with
the XTB at that location.  Except for the "electronics circuit" and one APC
UPS, nothing else is isolated with a filter.

I have been working on the XTB-II firmware, and have it temporarily located
on a branch circuit in the family room where the Ocelot lives.  Because of
the inductance between there and the distribution panel, the X10 signal on
that branch circuit is much higher than it would normally be, which is
inducing the flicker.  I never saw any flicker when testing the XTB at this
location, but its output is somewhat lower.

> FWIW, my ESM1 measurements around my rather large apartment do not show
the
> degree of attenuation you see. Unfortunately, I'm no longer able to do the
> amount of walking required to repeat the measurements but as best I
recall,
> using a TM751 as the source, I had at least 2-3V at all outlets after
adding
> one filter on my TV/VCR and associated gear.

This house is just under 5000 sq. ft. on two levels, and 80 ft. end to end.
The TW523 was a little marginal covering that area, so that's why I designed
the XTB.

Jeff