Partial CM11A schematic available
John Galvin
schematic. I undertook this reverse engineering task in an effort to try to
understand why X-10 signals transmitted by the CM11A are significantly
weaker than from other transmitters. In the process, I discovered a few
interesting things.
I used 4 different methods to measure the X-10 output of the CM11A.
Monterey PLSA, ELK ESM1, Leviton passive coupler with oscilloscope and
straight across the line. All 4 methods yielded different results. The
Monterey read the peak to peak voltage across the line, as a sort of average
burst amplitude. It was also extremely sensitive to out of band, non X-10
signals. Many times this resulted in erroneous readings compared to direct
'scope measurement across the line. The ESM1 was an early model that had
been calibrated to read half the peak to peak voltage. It is my
understanding that this was changed in later production. Accounting for
that difference, the ESM1 tended to read momentary peaks and report that as
the amplitude. This was especially apparent with the tapering amplitude of
the CM11A's X-10 bursts (more on that later). The Leviton passive coupler,
which I understand is something like what is used in the ACT ScopeTest2,
yielded signifcantly higher amplitude measurements than what was actually on
the line. I can only attribute that to some amplification of the signal due
to resonance of the coupler. I do not know if the ACT ScopeTest2 also
exhibits this behavior. Using an oscilloscope straight across the line, is
an indisputable method of measuring amplitude, so that is what I settled on
using.
A standard CM11A outputs X-10 bursts of approximately 5 volts peak to peak
amplitude. This is about half what is measured for other transmitters, such
as an SD533 (Sundowner). In addition to being half the amplitude, if there
is significant loading, the burst amplitude will fade over the duration of
the burst. Depending on how heavy the load, that is how many X-10 signal
sucking devices you have on the line, the amount of tapering of the burst
amplitude varies. With the CM11A completely isolated from any X-10 signal
load, there was no tapering of the bursts at all. With a more typical load,
the bursts taper as much as 40-50% in amplitude. This tapering can be
explained by referring to the partial schematic mentioned previously.
The CM11A's X-10 transmitter section consists of a 120KHz oscillator circuit
coupled to a power driver that is connected to an output transformer.
Interestingly enough, there is also a second isolation transformer after the
output transformer. I did verify that minimal amplitude is lost in that
isolation transformer. The 120KHz oscillator is comprised of NPN transistor
Q6 and transformer X3, shown on the left side of the schematic. The CM11A's
PIC controller can, turn this oscillator on or off, by controlling the state
of pin 10. When that is low, there can be no base current in Q6, so no
oscillation. When the CM11A's firmware decides that it is time to output a
burst, it brings pin 10 high, supplying base drive to Q6 and the oscillator
starts. The collector of Q6 is connected through R16 39K resistor to the
base of Q4 PNP transistor, to provide some current gain. Q4's collector is
in turn connected to the base of NPN transistor Q3 oeprating in emitter
follower mode. The output NPN transistor Q7 is driven from the emitter of
Q3 and drives the primary of the output transistor with a current roughly
proportional to the voltage from Q3 emitter. The secondary of the output
transformer is connected to the isolation transformer and that is then
coupled to the actual AC line through a .22uF capacitor. So, what is
causing the burst amplitude to fade? The answer lies in the circuitry
surrounding Q4. The emitter of Q4 is connected to R12 100k resistor and
then to the +36V supply. There is a .22uF "bypass" capacitor from the
emitter of Q4 to ground, also. When the 120KHz oscillator is off, there is
minimal emitter current in Q4. There is a path from the base of Q4 through
R16 and back up into the +5Y supply. Due to that, the base of Q4 sits
pretty much at the same voltage as the +5Y supply. Q4's emitter will be .6V
higher, even though R12 is connected to +36V. When the oscillator is turned
on, Q4 draws enough current to significantly deplete the voltage stored on
C9 bypass capacitor. The emitter of Q4 drops in voltage, as much as 1.5V
over the duration of the burst. This severely compromises the base drive to
Q3 and that results in a tapering burst amplitude. Replacing C9 .22uF
bypass capacitor with 2.2uF, resulted in negligible droop on Q4 emitter and
subsequent elimination of the tapering burst amplitude. I also tried
replacing R12 with a 22k resistor and that too, eliminated the burst
tapering. I don't have a strong preference for either fix. They were both
very effective. After making this change, the Monterey PLSA showed an
increased amplitude, but the ESM1 did not. I attribute that to the ESM1's
peak detector charging up to 2.5V on the first part of the burst, before the
signal had faded on the unmodified CM11A. The amplitude of the first part
of the burst was not effected by the modification, only the tapering was
eliminated. So, the ESM1 registered no change. This isn't actually a
problem normally, since every other transmitter I've tested exhibits uniform
burst amplitude over the duration. The Monterey appeared to measure the
average amplitude of the bursts and so it registered an increase. There
were many instances when the Monterey gave much higher readings than what
was really there, due I'm guessing, to it's sensitivity to noise outside the
normal X-10 frequency range. It was also quite sensitive to distortion on
the X-10 signal, giving erroneous readings in the presence of significant
distortion. The ESM1 was much less sensitive to that effect.
I investigated a bit more, to see if there was any way to boost the overall
amplitude of the CM11A. I tried many different things, too numerous to
mention here. None of them was successful in increasing the basic amplitude
of the CM11A's X-10 output bursts. I suspect that the CM11A's output
transformer's turns ratio, is what causes the output to be half the
amplitude of other transmitters.
J.G.
Dave Houston
Great job!
"John Galvin" <lgalvin...@pacbell.net> wrote:
---
http://www.laser.com/dhouston/
The problem with Usenet is that so many with
nothing to say feel the necessity to say it.
John LeMay
> See http://home.pacbell.net/lgalvin/CM11AX-10a.gif for a partial CM11A
> schematic. I undertook this reverse engineering task in an effort to
Nice work John! This is the kind of work we need in the lower end home
automation field. Let us know how rewinding that transformer works out!
<g>
--
John LeMay Jr.
Senior Enterprise Consultant
NJMC, LLC.
Sent with PAN 0.10.0.90cvs running on Mandrake Linux 8.0f1