add'l schematics available
John Galvin
CM11A RS232 interface section
http://home.pacbell.net/lgalvin/CM11A_rs232.gif schematic.
Smartlinc's Powerlinc TW-523 powerline interface replacement
http://home.pacbell.net/lgalvin/pwrlinc.gif schematic and
http://home.pacbell.net/lgalvin/pwrlinc.jpg and
http://home.pacbell.net/lgalvin/pwrlinca.jpg for pictures of the circuit
board.
Elk ESM1 X-10 signal level meter http://home.pacbell.net/lgalvin/esm1.gif
schematic and http://home.pacbell.net/lgalvin/esm1.jpg and
http://home.pacbell.net/lgalvin/esm1a.jpg for pictures of the circuit board.
Herewith some comments about these designs:
The CM11A's RS232 section is a slightly interesting design, in that it
maintains full +/- voltage swings on the Txdata and Ring Indicator lines.
Referring to the schematic, whenever the Rxdata line (Tx from the host) goes
negative, it charges the 10uF capacitor connected to the anode of the 1N4148
diode and it holds the pulldown resistors R36 and R4, .6V above the level of
the Rxdata line. So, whenever the Tx and RI lines are not being driven,
they're pulled down to a level that is 1 diode drop above the voltage on the
Rx line. When Q11 or Q? are on, the lower end of the pulldown resistors is
still held at that level, so long as the Rx line remains at -12V or
whatever. If the Rx line goes high, such as when the host is sending data,
the 10uF capacitor has enough stored charge to maintain the pulldown
resistors at that negative voltage, for some period of time. All this
tricky stuff is generally unnecessary, since 99% of modern PCs do not
require negative voltage swings on RS232. I suspect that IBM is the culprit
in this case. They are/were a customer of the CM11A (HD11A) and they were
one of the last holdouts with +/- 3V thresholds on the serial port lines. I
should point out that later designed X-10 devices that were not being sold
to IBM, such as the MR26A, do not output full negative voltage swings on the
serial data line.
If the CM11A is connected directly to a microcontroller without the benefit
of RS232 line driver/receivers, the CM11A's Rx line will of course, not be
driven to a negative voltage and consequently, the CM11A will not drive the
Tx and RI lines negative when they're deasserted. This is the desired
behavior. The Tx and RI lines will be pulled down to a voltage, one diode
drop higher than whatever voltage the Rx line sits at, in the deasserted
state. That's not great threshold-wise, but probably ok. What might not be
ok, is that the Tx and RI lines when asserted, will be driven to ~5.7V,
since that is what the +5Y supply is, on the CM11A. The Ring indicator line
has a 560 ohm series resistor, so that should limit the current somewhat
into any microcontroller's input stage. The Tx line on the other hand, has
no series resistor and I'm concerned about connecting directly to an input
pin on a microcontroller. Especially in the case where the microcontroller
is powered off, yet the CM11A is still on. Quite a bit of current could
flow into the ESD protection structure for that pin on the microcontroller
and that could very well damage the part. I'm sure that the CM11A's
designer did not use a series resistor, since most host PCs would already be
well protected by an RS232 line receiver and adding the resistor would tend
to compromise speed on the Tx data line. Rxdata from a microcontroller,
that does not go negative, will generally work ok into the CM11A, since Q15
does not require a negative voltage swing, in order to turn off.
As for Smartlinc's Powerlinc TW-523 replacement, see the schematic. Just
a couple of notes here. The Powerlinc derives its zero crossing timing from
the secondary of the power transformer. This is different than what we see
from X-10 designs. It can work ok, but attention must be paid to the
specification of the transformer, else there could be phase shift that
alters the timing enough to cause trouble. X-10 sidesteps that issue on
their designs by using an opto-isolator. Apparently they feel that is
cheaper than paying for a decent power transformer. Smartlinc took the
approach of paying a little more for the power transformer. The Powerlinc
is very different in design, compared to the TW-523 that it is intended to
replace. With most of the internal circuitry isolated from the incoming AC
line, by the power transformer, there isn't any need for opto-isolators on
the interface. This also allowed Smartlinc to add an RS232 interface mode,
in addition to the TW-523 legacy mode. On power up, if the PIC senses that
pin 4 is grounded, then it assumes RS232 mode. If it doesn't, which would
be the case for the usual TW-523 application, it goes into legacy mode and
functions mostly just like a TW-523. I say mostly, since there isn't any
way that the direct connections to the PIC can perfectly emulate the
opto-isolators that a genuine TW-523 uses. Case in point: with a TW-523 you
could pull up with a resistor, the collector of the zero cross and X-10 data
out lines, to whatever voltage the opto-isolator can handle. Not so, on the
Powerlinc. Attempts to pull up to a higher voltage than +5V, will likely
result in damage to the PIC inside the Powerlinc. This is not likely to be
a problem in normal usage, but I did want to point out design differences
that could effect a small percentage of applications. What else? Big fat
diode D2 hanging across the primary of the X-10 signal trasnformer is
probably responsible for the greater loading effect of the Powerlinc on X-10
signals (perhaps the 2way Switchlincs too), compared to transmitters from
other manufacturers. Getting back to the interface side, it might have been
nice if they at least had some series resistors on the lines going to the
PIC. On the other hand, perhaps the PIC's inputs are so robust, that none
are needed for protection (yeah right). It is good design practice to use
series resistors, to limit the current that flows when the receiving device
is powered off, yet the sending device is still powered. Perhaps that
explains some of the reports of failed Powerlincs.
On to the ELK ESM1 X-10 signal level meter. See the schematic . One
interesting thing to note, is that the ESM1 receives the 120KHz X-10
signals, apparently THROUGH the wall wart power transformer. I guess I
can't argue with that, since it seems to work just fine. There is a nice
active filter that rolls off the response of the X-10 measurement circuit,
at frequencies above ~150KHz. This, in conjunction with whatever filtering
effects may be happening in the power transformer itself, seem to work well
enough at emulating the response of a real X-10 receiver module, unlike a
certain other, power line signal analyzer. After the active filter, the
X-10 signal continues on to a gain stage followed by simple half wave peak
detector with filter capacitor. From there, it's on to the PIC's A/D
convertor input. After the requisite "miracle occurs" in the firmware, the
output is displayed on the 10 segment led bargraph. The bargraph is
multiplexed in two halves, to minimize on pins needed from the PIC.
That's it for now. Happy X-10 hacking.
J.G.
Guy Lavoie
Mark Lloyd
news:9mtv4q$g21$1...@suaar1ab.prod.compuserve.com:
> The following additional schematics are now available:
>
> CM11A RS232 interface section
> http://home.pacbell.net/lgalvin/CM11A_rs232.gif schematic.
>
> Smartlinc's Powerlinc TW-523 powerline interface replacement
> http://home.pacbell.net/lgalvin/pwrlinc.gif schematic and
> http://home.pacbell.net/lgalvin/pwrlinc.jpg and
> http://home.pacbell.net/lgalvin/pwrlinca.jpg for pictures of the
> circuit board.
>
> Elk ESM1 X-10 signal level meter
> http://home.pacbell.net/lgalvin/esm1.gif schematic and
> http://home.pacbell.net/lgalvin/esm1.jpg and
> http://home.pacbell.net/lgalvin/esm1a.jpg for pictures of the circuit
> board.
>
I got the .GIF, the others are "Not Found",
--
Mark Lloyd
http://148.75.72.214
http://go.to/notstupid
Dave Houston
when I tried a direct connection.
"John Galvin" <lgalvin...@pacbell.net> wrote:
---
http://www.laser.com/dhouston/
The problem with Usenet is that so many with
nothing to say feel the necessity to say it.
John Galvin
Mark Lloyd wrote in message ...
My apologies. The Powerlinc PCB pictures should be:
http://home.pacbell.net/lgalvin/powerlinc.jpg and
http://home.pacbell.net/lgalvin/powerlinca.jpg
I just verified the esm1 pictures and they are there.
J.G.
Mark Lloyd
> http://home.pacbell.net/lgalvin/powerlinca.jpg
>
>
They're there now. Did you think they were there before?
John Galvin
Mark Lloyd wrote in message ...
>news:9murrp$t5n$2...@suaar1aa.prod.compuserve.com:
>
>> http://home.pacbell.net/lgalvin/powerlinca.jpg
>>
>>
>
>They're there now. Did you think they were there before?
>
They were there all along. I had posted the wrong filenames.
J.G.