Why 2 dedicated Circuit Breakers for amplified bridge?
Bret Schuhmacher
it requires 2 dedicated 15 Amp breakers. Why? I don't have room for
this...:-( Can I connect it to 2 existing breakers that are used for
existing loads? What risks am I taking? Why does the NEC say you have to
have 2 for this application?
Why is the 6201 about $200 USD and another one (Leviton HCA02 with LEDs) is
about $90? The latter has Intellisense and the same amplification (up to
7v)... Any ideas what the differences are?
FWIW, I have a SignaLinc bridge but all of a sudden it's not working as
expected. I added 1 transceiver and my whole system is out of whack... The
shed lights (100 ft away) from my 3K ft^2 have stopped turning off in the
morning (they turn on at dusk via X10 signal just fine, though). My coffee
pot refuses to turn off (MR227A? it's a standard X10 outlet) "sort of" turns
off, then turns back on 10 minutes later...
Thanks in advance!
Bret
Pete S.
I have been in this business several years. You don't need to use dedicated
breakers. The reliability of the complete automation/lighting system may be
dependant on the repeater. If the breaker pops because of another load on
that circuit the customers whole lighting system could be disabled. This
would be disastrous if you needed a light to find the breaker panel!
The answer is you don't need dedicated breakers, but they are worth using if
you can finagle it. At least use a pair of lightly loaded breakers if you
cant dedicate two.
The 6201 does not have Intellisense. I have no idea if the CR231 will work
in the presence of noise. I have pulled out 6201's in several homes that
hadn't worked right in years and replaced them with HCA02's and have gotten
the systems working. That alone sold me, they have worked flawlessly for
me.
I never liked those bridges. They not only couple signal they couple noise.
You got some noise on each phase, all of a sudden you have twice as much
system noise when they are coupled together. The repeater isolates the
noise and only repeats the signal. And it provides the diagnostics needed.
"Bret Schuhmacher" <bsch...@starband.net> wrote in message
news:a31s1...@enews3.newsguy.com...
David M Johnson
panel. I think that is what makes the X-10 system more reliable, plus the
added "boost" it gives the signal. Of course I could be wrong. Anyone else
to comment?
Dave
"Bret Schuhmacher" <bsch...@starband.net> wrote in message
news:a31s1...@enews3.newsguy.com...
L. M. Rappaport
with the difference between the HCA02 and the 6201, but the reason
they ask for two (adjacent) breakers is so that they can couple a
signal received on one leg to the other as Dave suggests. The reason
they want you to use breakers rather than piggy-backing onto the dryer
or range circuits, for example, is because of the size of their wire
and components used. It is normally way too small, so they want you
to use a dual 15.
If you want to stay within code and don't have the room, then you
might just wire up a small sub-panel and move a few circuits over, or
better yet, have a licensed electrician do it.
Larry
--
ra...@lmr.com
"Bret Schuhmacher" <bsch...@starband.net> wrote (with possible
editing):
Robert L Bass
"slim jims" for four existing single-pole breakers. This will leave you with
two empty slots. Install a bridged, 2e-pole, 15Amp breaker there and wire your
bridge to it.
BTW, don't just hang a bridge inside the breaker panel. Instead, mount a small
box to the outside of the cabinet with a short EMT nipple. Mount the bridge in
the box and run THHN or Romex to your breakers.
Regards,
Robert L Bass
=============================>
Bass Home Electronics
The Online DIY Alarm Store
http://www.Bass-Home.com
4883 Fallcrest Circle
Sarasota, FL 34233
877-722-8900 Sales & Tech Support
941-925-9747 voice (Florida)
941-926-9857 fax
Rober...@home.com
=============================>
Now you're good for code and
L. M. Rappaport wrote:
:
: If you want to stay within code and don't have the room, then you
Brian
>BTW, don't just hang a bridge inside the breaker panel. Instead, mount a small
>box to the outside of the cabinet with a short EMT nipple. Mount the bridge in
>the box and run THHN or Romex to your breakers.
How do you propose to fit an HCA02 into a "small box"?
If you had ever seen an HCA02 in person (or actually tried to install
one) you would see that it doesn't come anywhere near properly fitting
in a "small box".
There's a difference between drawing an ideal circuit and installing
real world systems in homes.
--
HA mailing lists and tech-tips at FutureStandard.com
H...@xxxhydrologistxxx.com
message <qu4b5us9ii9ldgtqj...@4ax.com>:
An HCA02 will fit handily in a 6x6" pull-box which, as these things go, is
small in my opinion. I typically use 12x12" boxes for this sort of add-on
when space permits.
You can mount a pair of cast-aluminum receptacle boxes to the sides of the
pull box and (as best this scientoid knows) meet NEC requirements and have
both the repeater and the X-10 interfaces close to the panel which is
ideal.
Install or leave room for two separate receptacle boxes (or one
double-duplex) because a potential solution to the repeater problem is to
have a separate X-10 interface on each phase. If you derive the power for
the repeater, but not the X-10 interfaces, from the 15-amp breakers (that
is, pull 4 hot wires from the panel through the EMT connector that Robert
suggests), you can easily toggle the breakers in the panel to reset the
repeater or remove the repeater from the circuit when trouble-shooting. If
you drill a 1/4 hole though the front of pull-panel, you can see the status
lights on the HCA02.
HTH ... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
Brian
>An HCA02 will fit handily in a 6x6" pull-box which, as these things go, is
>small in my opinion. I typically use 12x12" boxes for this sort of add-on
>when space permits.
Mine is in an 8x8x8 box (all I had on hand), with the LED's remoted
and sticking and mounted to the front panel.
Personally, I don't consider a 6x6 box "small"...
>Install or leave room for two separate receptacle boxes (or one
>double-duplex) because a potential solution to the repeater problem is to
>have a separate X-10 interface on each phase.
This is another one that looks better on paper than in real life.
When a 220V appliance comes online you can get X-10 collisions and/or
double transmits if you're not careful. Especially if your controller
is *watching* for commands...
Robert L Bass
:
: How do you propose to fit an HCA02 into a "small box"?
I consider a 6" square box next to a breaker panel small, kind of like you
standing next to me at the EH Expo. :^)
: If you had ever seen an HCA02 in person (or actually
: tried to install one) you would see that it doesn't come
: anywhere near properly fitting in a "small box".
I never mentioned an HCA02. I spoke of a "bridge" and I had in mind something
like the Leviton 6299 whose dimensions are 2.693" x 4.125" x 1.387". I don't
know about you but I call that small.
Perhaps after you've been in the business a while (you *are* going to get
started soon, aren't you?) you'll become aware that there are more than one
bridge device on the market. You might even notice one day that there are lots
of auxiliary boxes bigger than a 6x6.
: There's a difference between drawing an ideal
: circuit and installing real world systems in homes.
BTW, my cousin was here this weekend. He's an EE with extensive experience in
audio as well as communications systems design. He's been working on audio
systems since the early 60's. I told him about your insistence that we use
baluns and CAT5 for short (I consider anything under 100 feet short) line level
runs. He laughed and said, "Dumb idea... probably from some kid fresh out of
school." Bingo!
Robert L Bass
:
: HA wrote:
: >
: > An HCA02 will fit handily in a 6x6" pull-box which, as
: > these things go, is small in my opinion. I typically use
: > 12x12" boxes for this sort of add-on when space
: > permits.
:
: Mine is in an 8x8x8 box (all I had on hand), with the
: LED's remoted and sticking and mounted to the front
: panel.
Your limited supplies don't make my (correct) suggestion any less viable.
: Personally, I don't consider a 6x6 box "small"...
I'll bet you say that about lots of things... :^)
: > Install or leave room for two separate receptacle
: > boxes (or one double-duplex) because a potential
: > solution to the repeater problem is to have a
: > separate X-10 interface on each phase.
:
: This is another one that looks better on paper
: than in real life...
Suggestion: Go out and get some actual experience. Then come back and we can
talk on a more equal level. Meanwhile, you can use a small box mounted to the
breaker panel in the manner I suggested if you like. You can even call small
things large for all I care.
H...@xxxhydrologistxxx.com
message <nobb5u4398n0ol5dd...@4ax.com>:
>[] [] H...@xxxhydrologistxxx.com arranged some electrons to say:
>>Install or leave room for two separate receptacle boxes (or one
>>double-duplex) because a potential solution to the repeater problem is to
^^^^^^
I prolly shoulda wrote "triple-duplex". I don't think two CM11a's will fit
side by side in a conventional double duplex.
>>have a separate X-10 interface on each phase.
>
>This is another one that looks better on paper than in real life.
>When a 220V appliance comes online you can get X-10 collisions and/or
>double transmits if you're not careful. Especially if your controller
>is *watching* for commands...
Agreed -- with respect to current devices. The X-10 interfaces and related
software would need to be aware of their lots in life for this arrangement
not to create more problems than it solved. That's why I wrote "potential
solution". As best I know, it has not (yet) been implemented.
What I'd really like is a 1-in/3-out splitter that would the route X-10
commands, address-by-address, to one of two X-10 interfaces (one for each
phase) and send ASCII commands through a third COM port to a hard-wired
controller. The splitter should at least not get confused by
inputs/reponses from the X-10 interfaces. Ideally a fourth COM port would
serve as a management interface to set up routing and permit conversations
with the X-10 traffic cop.
This splitter setup would: 1) allow for relatively painless migration from
X-10 to hardwired lighting using software that did not have an interface
specifically for hardwired lighting such as Homeseer, and 2) potentially
help the coupling/repeater problem.
I started putzing with this when the Basic-X BX-24 first came out but
serial connections beyond COM1 were problematic at the time. I decided to
wait until the Bx-xx matured more, some one else did it, or do it myself
with a DOS SBC w/ properly buffered COM1-COM4 if the need became
unavoidable. None of those things have happened.
Cheers ... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
Dave Houston
In a split phase system, using two CM11As, both interfaces see the same ZC.
Any leakage is in phase with and indistinguishable from the signal being
transmitted so there is no collision. But, there really is no chance of
double transmits. Just segregate housecodes between phases (e.g. A-H on one
and I-P on the other) and in software (or firmware) only allow each CM11A to
send to housecodes on its phase and report received commands that are for
housecodes on its phase. If there are other transmitters in the system,
there may be collisions with them but this would be no different with a
coupler/repeater.
When combined with an all housecode transceiver that also uses the CM11As to
interface to the powerline, you eliminate the need for coupler/repeaters, as
well as multiple transceivers or RF repeaters. You can even accomplish this
using the MR26A to input the RF to a PC that's running software that
controls the two CM11As.
Any interface that handles collisions in the same way that the CM11A does
would work. It is easier if the powerline interface itself can be programmed
to ignore received commands for the housecodes that are not assigned to it.
I've been waiting for Iriave Electronics to supply an SMD version of their
TR-1A and then I was going to design such an interface. But, now, I'm
leaning towards another method that would be able to accomplish the same
thing using CM11As.
I think it solves all of the commonly reported problems with both
coupler/repeaters and transceivers. There are people using the BX24-AHT with
the antenna designed for it who report RF range of 125' or more which means
there's no need for multiple transceivers or RF repeaters. Even the lower
output of the CM11A is less of a problem as each only has to reach half of
the system.
The next major phase in the BX24-AHT project will be to move the RF
reception to an FCC approved antenna/receiver unit that sends the received
codes to the main unit via RS-485. At that point, the BX24-AHT can support
two CM11As. I can do it now but with only a single software UART, I have to
poll the two CM11As and wait for them to respond. This just takes too long
to be practical as RF receptions may be missed while waiting for the CM11A
to respond to the poll.
The BX24-AHT can also communicate with other HA devices and can even handle
DIO and ADC so that everything in the HA system can pass through it. It can
also be controlled by modem or from the web with full access to everything
that it controls or communicates with. Everything passes through one central
control point.
---
http://www.laser.com/dhouston/
Brian
>Suggestion: Go out and get some actual experience. Then come back and we can
>talk on a more equal level. Meanwhile, you can use a small box mounted to the
>breaker panel in the manner I suggested if you like. You can even call small
>things large for all I care.
Robert you admitted about 2 years ago that you stopped installing
alarm equipment. You also never really did any automation
installations (as you admitted in phone conversations to me).
So what experience of yours are you referring to that is applicable to
relatively recent devices, like the HCA02? Have you ever actually
seen an HCA02 in person? Have you ever actually installed one? If
you like, you can answer in the affirmative to save face, I don't
*really* care.
Or do you think that the alarm panels you installed years ago, and
some line level cables you ran in a church several years ago will make
you an expert forever?
Again, I don't really care what your answers are, whatever you say has
no bearing on me, or anything affecting me for that matter, but I do
understand that you are still trying to make a living by drop shipping
alarms from ADI and automation supplies from Worthington and Home
Controls, so I guess it's inevitable that you try to leverage your
past experience as much as possible.
Brian
>Just segregate housecodes between phases (e.g. A-H on one
>and I-P on the other) and in software (or firmware) only allow each CM11A to
>send to housecodes on its phase and report received commands that are for
>housecodes on its phase.
This is the best approach I've seen to this tactic, however there are
very few devices that support the CM11A as a PLI, and even fewer
devices that support *2* CM11A's :)
Brian
>commands, address-by-address, to one of two X-10 interfaces (one for each
>phase) and send ASCII commands through a third COM port to a hard-wired
>controller. The splitter should at least not get confused by
>inputs/reponses from the X-10 interfaces. Ideally a fourth COM port would
>serve as a management interface to set up routing and permit conversations
>with the X-10 traffic cop.
I'm not quite sure I follow. What physical layer are you proposing
the splitter for? The powerline, the TTL out of the '523, etc?
>I started putzing with this when the Basic-X BX-24 first came out but
>serial connections beyond COM1 were problematic at the time. I decided to
>wait until the Bx-xx matured more, some one else did it, or do it myself
>with a DOS SBC w/ properly buffered COM1-COM4 if the need became
>unavoidable. None of those things have happened.
Yeah, I know what you mean. I have a handful of projects that I've
been meaning to get to.
Dave Houston
>[] [] dhou...@fuse.net (Dave Houston) arranged some electrons to say:
>>Just segregate housecodes between phases (e.g. A-H on one
>>and I-P on the other) and in software (or firmware) only allow each CM11A to
>>send to housecodes on its phase and report received commands that are for
>>housecodes on its phase.
>
>This is the best approach I've seen to this tactic, however there are
>very few devices that support the CM11A as a PLI, and even fewer
>devices that support *2* CM11A's :)
>
It only takes one. ;)
Dave Houston
>[] [] dhou...@fuse.net (Dave Houston) arranged some electrons to say:
>>Just segregate housecodes between phases (e.g. A-H on one
>>and I-P on the other) and in software (or firmware) only allow each CM11A to
>>send to housecodes on its phase and report received commands that are for
>>housecodes on its phase.
>
>This is the best approach I've seen to this tactic, however there are
>very few devices that support the CM11A as a PLI, and even fewer
>devices that support *2* CM11A's :)
>
I should have added that sending to either of two (or more) CM11As,
depending on the housecode (or any criteria), is simple even with the BX-24
software UART. It's only trying to receive from more than one that's
problematic.
Robert L Bass
:
: [] [] "Robert L Bass" <Rober...@home.com> arranged some electrons to say:
:
: >Suggestion: Go out and get some actual experience. Then come back and we
can
: >talk on a more equal level. Meanwhile, you can use a small box mounted to
the
: >breaker panel in the manner I suggested if you like. You can even call small
: >things large for all I care.
:
: Robert you admitted about 2 years ago that you stopped installing
: alarm equipment.
True, I don't like crawling around in attics any more.
: You also never really did any automation
: installations (as you admitted in phone conversations to me).
That's a flat lie, Karas. I've been doing various forms of automation all the
time. Have I installed an HCA02? Nope. I wasn't referring to an HCA02 though.
I only said a bridge and I do know precisely how to install one. You wanted
something to whine about and you picked the wrong point.
: So what experience of yours are you referring to that is applicable to
: relatively recent devices, like the HCA02? Have you ever actually
: seen an HCA02 in person?
Yes.
: Have you ever actually installed one?
I have installed signal bridges. You got hung up on the HCA02. I never
mentioned it. I said a "small box" and you pitched a fit. Sound like y'all got
a bur under your saddly, Hoss. :^)
: If you like, you can answer in the affirmative to save
: face...
No thanks. That's not my style. You're the only one here who has been caught
lying to customers, competitors, suppliers and everyone else. I prefer to tell
the truth. You and I both know that truth isn't important to you. You've
proven that to me and you've admitted it publicly in USENET.
I don't *really* care...
Oh but you most assuredly do, Karas. You've demonstrated that in the way you
clink to a hopeless position any time we disagree, despite the fact that even
other engineers have told you you're wrong.
: Or do you think that the alarm panels you installed years ago...
More specifically, I have almost 25 years of designing, installing and servicing
not just alarms, but multi-room audio and video systems, home theaters,
installing and servicing stage sound systems, etc. I've designed and installed
theatrical lighting (not just a few lamp circuits), special effects systems,
motorized drapes and both theatrical lifts to "fly" live actors, etc. I've also
been called to consult on the design of lighting systems for other venues.
Also, before I ever entered the security business, I worked on sound systems in
discos as far back as the late 60's and early 70's.
You're quite new to the field by comparison, Karas. You have a few good ideas
and you're a bright guy. But don't kid yourself that you're the resident expert
on things electrical or electronic here. There are others here with far more
experience than I have. Some are experts in areas that I've only touched on. I
read their posts to learn -- not to look for some tiny nit to jump on as you do.
I suggest you do the same.
You've done some creative things with Applied Digital products and you've got a
good understanding of data networking. Teach what you know and listen to others
in areas where you lack. You'll look like an ass a lot less often.
: and some line level cables you ran in a church several years
: ago will make you an expert forever?
Nope. That's just one small part of a carreer designing and building all kinds
of systems. Everything you do over a lifetime of work contributes to a general
knowledge base. That makes me experienced. I try to use the term expert
sparingly. There are areas where I am indeed an expert. There are lots of
others where I know what I'm doing well enough not to get into trouble.
: Again, I don't really care what your answers are...
Yeah, sure. That's why you asked, right? :)
: whatever you say has no bearing on me, or anything
: affecting me for that matter...
And yet you spend so much effort trying unsuccessfully to attack two little
words (small box) in one of my posts. It sure sounds like you care.
: ... but I do understand that you are still trying to make
: a living...
Actually, I am making a living. Do you remember the phone conversation when you
told me that my type of business was not going to make it... that Future
Standard was the business plan of the future? Still selling online, old pal?
: ... by drop shipping alarms from ADI...
Yes, ADI (not to be confused with Applied Digital) and several other
distributors plus several manufacturers. Any problem with that?
: and automation supplies from Worthington and
: Home Controls...
Wrong. I don't order from either of them. I never really did much with Home
Controls (don't recall if we ever actually placed an order with them though I
might have). I dropped Worthington after they repeatedly broke promises to me,
causing me great embarassment with a good client.
: so I guess it's inevitable that you try to leverage your
: past experience as much as possible.
I guess you could call a lifetime of work experience "leverage." Do you
begrudge me the experience I have just because you have never actually installed
any of this stuff for a living? Sorry, but I didn't know you'd be coming when I
decided to spend the past decades installing and servicing electronic systems.
--
Robert L Bass
Brian wrote:
:
: I have a handful of projects that I've
Dan Lanciani
| In a split phase system, using two CM11As, both interfaces see the same ZC.
| Any leakage is in phase with and indistinguishable from the signal being
| transmitted so there is no collision.
This would be true only if you managed to phase-lock the 120kHz carriers of
the two transmitters, possibly taking into account any phase shift induced
by the specific leakage path. Otherwise the carriers could interfere
destructively even though the high-level X10 code itself was in sync.
I was thinking about modifying a CM11a to have two sets of output drivers, one
for each leg, but still with a single 120kHz carrier source. Then I realized
that I really liked my maxi controllers to work house-wide and thus would still
need the repeater...
Dan Lanciani
ddl@danlan.*com
Brian
taunt.
Enjoy.
[] [] "Robert L Bass" <Rober...@home.com> arranged some electrons to say:
Dave Houston
>In article <3c55b6d8...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>
>| In a split phase system, using two CM11As, both interfaces see the same ZC.
>| Any leakage is in phase with and indistinguishable from the signal being
>| transmitted so there is no collision.
>
>This would be true only if you managed to phase-lock the 120kHz carriers of
>the two transmitters, possibly taking into account any phase shift induced
>by the specific leakage path. Otherwise the carriers could interfere
>destructively even though the high-level X10 code itself was in sync.
I think the probability that they're going to be exactly 180 degrees out of
phase and thus cancel at the 120KHz level is near zero. Two TM751s on
opposite phases do not cancel each other - there are problems with unit
codes 1 & 9 only (and I suspect this has something to do with the relay
which someone reported is activated before anything is sent to the
powerline).
Besides, there is no practical way that you are going to get two CM11As to
transmit at the same time, even at the high-level, whether you control them
with a PC or a microcontroller. They would have to transmit sequentially.
You might get them to transmit at exactly the same time if you used two PCs
or two microcontrollers but now we're back to the same situation as with two
TM751s but with no relay code.
One of the people who beta tested the BX24-AHT now has a second one. I'll
see if I can get him to test with one on each phase but with both using the
same antenna for RF input.
H...@xxxhydrologistxxx.com
message <hlnb5u0a3df4sb0k0...@4ax.com>:
>[] [] H...@xxxhydrologistxxx.com arranged some electrons to say:
>>What I'd really like is a 1-in/3-out splitter that would the route X-10
>>commands, address-by-address, to one of two X-10 interfaces (one for each
>>phase) and send ASCII commands through a third COM port to a hard-wired
>>controller. The splitter should at least not get confused by
>>inputs/reponses from the X-10 interfaces. Ideally a fourth COM port would
>>serve as a management interface to set up routing and permit
conversations
>>with the X-10 traffic cop.
>
>I'm not quite sure I follow. What physical layer are you proposing
>the splitter for? The powerline, the TTL out of the '523, etc?
The following is an elaboration of a suggestion I made in this newsgroup
about a year ago.
A physical "CMclops" [pronounced " cymclops ") device would be the most
straight-forward approach and the one that would be applicable to the
widest number of DIY situations. A physical CMclops would consist in a
PIC, micro, SBC, or PC with 4 to 5 RS-232 COM ports.
The input (COM1) would appear to conventional HA software (Homeseer, Hal,
CyberHouse, MisterHouse, etc) running on the main Home Automation PC
(HA-PC) to be a single conventional CM11a. COM2 and COM3 would be
connected to two actual CM11a's, one per AC phase. COM4 would send ASCII
commands, translated from the received X-10/CM11a commands, to a hard-wired
lighting controller using DIY, or published ASCII protocol for lighting
control. COM5 could be an optional management control port to which routing
commands could be sent and traffic/debugging/feedback obtained.
-HA-PC-+ +------------CMclops-------------+
| | (COM2)|--> CM11a #1 phase 0
COMm |---> |COM1 virtual CM11a input (COM3)|--> CM11a #2 phase 1
COMn |---> |COM5 mgmt port(optional) (COM4)|--> Hard-wired lighting
------ + +--------------------------------+
This would require only one or (optionally) two COM ports on the HA-PC,
would be compatible with any program and platform that supported a CM11a
(unix+MrHouse; Wintel+Homeseer etc) and would support hard-wired lighting
without modifying the HA program on the main HA-PC.
The CMclops concept could also be implemented as a program running on the
main HA-PC with some functions implemented on the HA-PC as additional
programs in lieu of physically looping out of and back into the HA-PC with
the COM ports. For example, the COMn--COM5 management control ports could
be replaced with a GUI, calls from the main HA software (Homeseer etc) to
the COMm port expecting a CM11a could be intercepted, and so on. This
CMclops implementation would be platform-specific and so less widely
applicable.
With either the PIC/SBC or HA-PC approach, the hard-wired lighting control
function (i.e., for what COM4 is included) could be partly or completely
built into a (eg) BX-24. The later and the motorized potentiometer
positioning loop is what I've begun working on now that I've finished the
wiring and most hardware for homebrew system. PIC's beckon to me.
.... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
Robert L Bass
without admitting it?
"Brian" <br...@nomail.net> wrote in message
news:9q0c5u0ut3ck7a9el...@4ax.com...
: Thanks for another evening of enjoyment. You're really quite fun to
: taunt.
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >In article <3c55b6d8...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
| >
| >| In a split phase system, using two CM11As, both interfaces see the same ZC.
| >| Any leakage is in phase with and indistinguishable from the signal being
| >| transmitted so there is no collision.
| >
| >This would be true only if you managed to phase-lock the 120kHz carriers of
| >the two transmitters, possibly taking into account any phase shift induced
| >by the specific leakage path. Otherwise the carriers could interfere
| >destructively even though the high-level X10 code itself was in sync.
|
| I think the probability that they're going to be exactly 180 degrees out of
| phase and thus cancel at the 120KHz level is near zero.
I think that the probability that they will be at or near 180 degrees out
of phase _at some point_ is quite high if they are not phase-locked. The
frequencies are bound to be a little off so it's just a question of how
fast the phase difference is changing compared to the timing of the higher-
level signal. You can imagine a situation where the null drifts through
the bits, confusing some of the clever receivers that sample at a few fixed
points per bit...
| Two TM751s on
| opposite phases do not cancel each other -
I'm sure they can, it's just that the combination of weak coupling between
phases and something like the capture effect makes it mostly a non-problem
in practice.
| Besides, there is no practical way that you are going to get two CM11As to
| transmit at the same time, even at the high-level, whether you control them
| with a PC or a microcontroller.
That may be true, but it's basically the opposite of the original argument
that they will tend to synchronize on the zero crossing. :)
| They would have to transmit sequentially.
That should work as long as there is enough leakage for them to hear each
other and do the usual collision avoidance. Or you could just never ask them
to transmit near the same time in the first place...
Dan Lanciani
ddl@danlan.*com
Brian
[] [] "Robert L Bass" <Rober...@home.com> arranged some electrons to say:
>That's your way of exiting a thread where you've been proved totally wrong
--
Robert L Bass
Brian wrote:
:
: No, you just haven't posted anything valid.
Dave Houston
>In article <3c56027f...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>| ddl@danlan.*com (Dan Lanciani) wrote:
>|
>| >In article <3c55b6d8...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>| >
>| >| In a split phase system, using two CM11As, both interfaces see the same ZC.
>| >| Any leakage is in phase with and indistinguishable from the signal being
>| >| transmitted so there is no collision.
>| >
>| >This would be true only if you managed to phase-lock the 120kHz carriers of
>| >the two transmitters, possibly taking into account any phase shift induced
>| >by the specific leakage path. Otherwise the carriers could interfere
>| >destructively even though the high-level X10 code itself was in sync.
>|
>| I think the probability that they're going to be exactly 180 degrees out of
>| phase and thus cancel at the 120KHz level is near zero.
>
>I think that the probability that they will be at or near 180 degrees out
>of phase _at some point_ is quite high if they are not phase-locked. The
>frequencies are bound to be a little off so it's just a question of how
>fast the phase difference is changing compared to the timing of the higher-
>level signal. You can imagine a situation where the null drifts through
>the bits, confusing some of the clever receivers that sample at a few fixed
>points per bit...
To cancel they have to be exactly 180 degrees out of phase and of exactly
the same amplitude. As a practical matter, it will NEVER happen.
>| Two TM751s on
>| opposite phases do not cancel each other -
>
>I'm sure they can, it's just that the combination of weak coupling between
>phases and something like the capture effect makes it mostly a non-problem
>in practice.
>
>| Besides, there is no practical way that you are going to get two CM11As to
>| transmit at the same time, even at the high-level, whether you control them
>| with a PC or a microcontroller.
>
>That may be true, but it's basically the opposite of the original argument
>that they will tend to synchronize on the zero crossing. :)
The initial argument was just to indicate to Brian that he had not given
this enough thought. It was more of a "worst case" type thing than anything
else. I have repeatedly indicated that I've planned for the BX24-AHT to do
this by segregating housecodes from the first time I mentioned it about two
years ago.
>| They would have to transmit sequentially.
>
>That should work as long as there is enough leakage for them to hear each
>other and do the usual collision avoidance. Or you could just never ask them
>to transmit near the same time in the first place...
Huh? Why is collision avoidance necessary when only one will send at any
given time? Since I plan to segregate housecodes there will never be any
need for both to send the same command. Even if you wanted both to send all
commands, one would send a command and get confirmation before the other
would send it. With macros, this would become too unwieldy and time
consuming to be practical so, for all practical purposes, it is necessary to
segregate housecodes to avoid it.
---
http://www.laser.com/dhouston/
Brian
a couple of others about 1.5 years ago.
[] [] H...@xxxhydrologistxxx.com arranged some electrons to say:
H...@xxxhydrologistxxx.com
came to naught when I abandoned it along with C, the TW523 and drivers from
Baran-Harper.
But I don't recall any discussion in this newsgroup of intercepting X-10
commands and routing a subset to a hard-wired light dimming system other
than when I brought it up before. IIRC, we discussed it a decade ago on a
pre-comp.home.automation listserve but it would seem that no actual
progress has been made.
Of course no outboard hardware would be necessary at all if the main HA
software itself split the commands. It would be a simple thing to do in
most cases, and would be a major step in making hard-wired lighting
available outside the strictures of expensive proprietary systems.
Savoy's CyberHouse partly solves this problem for my own development
purposes because it puts up with (but does not officially support) multiple
TW523 on multiple Elk MM443 modules and a single additional CM11a which I
am using as the interface to my harded-wired/DMX-512 system. Or by starting
two linked instances of the HA server (CyberHouse is designed from the
ground up to fully support multiple servers), I can split X-10 commands
between two CM11a's.
Helluvalot easier if they jist had a DMX-512 driver, but I don't know of
any affordable general purpose automation system that supports any
non-proprietary hard-wired dimming system. Premise
http://www.premisesystems.com/ may help to change that, but at ~$1000, it
isn't for everyone.
The dual problems of the lack affordable software that supports hard-wired
dimming and lack of a standardized protocol for hard-wired dimmer systems
would be addressed by adding multiple CM11a outputs to existing affordable
software and having the hard-wired dimming system itself mimic a CM11a
controller (or a superset of it). It might additionally address
never-ending X-10 phase-coupling problems. (I've already adequately
addressed that issue for my own purposes by minimizing the use of X-10,
especially devices and commands that hog bandwidth, and by moving problem
x-10 devices to the CM11a-driven phase.)
HTH ... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
On Tue, 29 Jan 2002 08:13:51 -0500, Brian <br...@nomail.net> wrote in
message <4t7d5u4vuc2tjtaea...@4ax.com>:
L. M. Rappaport
>I think that the probability that they will be at or near 180 degrees out
>of phase _at some point_ is quite high if they are not phase-locked. The
>frequencies are bound to be a little off so it's just a question of how
>fast the phase difference is changing compared to the timing of the higher-
>level signal. You can imagine a situation where the null drifts through
>the bits, confusing some of the clever receivers that sample at a few fixed
>points per bit...
Then why not simply time shift the signals to the individual phases?
This is fairly simple using zero crossing detectors. You feed a
signal to one phase, then the other, etc. for a few cycles.
Larry
--
ra...@lmr.com
H...@xxxhydrologistxxx.com
message <4t7d5u4vuc2tjtaea...@4ax.com>:
>This sounds similar to a project I was working on with Neil Cherry and
>a couple of others about 1.5 years ago.
The one CM11a/phase capability that you may be referring to might help to
drain some of the pus from the X-10 infection and put on a fresh band-aid,
but the "virtual X-10 interface" as hard-wired dimmer interface has
interested me for a decade. Lack of a standardized interface for hardwired
dimming has been an obstacle with a solution in sight for years and years.
I'd buy the first two CM11a --> DMX-512 converters (hardware or software)
sight-unseen. ABIK, there is still no link available commercially between
the de facto Power Line Control standard ( X-10) and the other two
internationally standardized light dimming systems, namely hard-wired
digital via DMX-512 (which is a variant of RS-485) and analog (0-10 VDC)
which is itself a subset of DMX-512.
Shame on us ... ;-)
... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
Dave Houston
which I first described here about two years ago and actually had running
about 1.5 years ago. ;)
For those who still think that two CM11As on opposite phases sending at the
same time (no matter how difficult it is to actually do that) will result in
cancellations, I've 'drawn' a picture. See...
http://www.laser.com/dhouston/phases.gif
DISCLAIMER REQUIRED FOR THE NITPICKERS AND HAIRSPLITTERS: I did not
literally DRAW the picture but put it together by cutting and pasting from
images captured with my scope card.
I've used square waves instead of sine waves but I think anyone should
readily see that the only condition that results in 120KHz cancellation is
when the two signals are EXACTLY 180 degrees out of phase and of EXACTLY the
same amplitude. If 180 degrees out of phase and of different amplitudes, the
resulting signal amplitude will be the difference between the two signals.
If not 180 degress out of phase, the resulting signal will have a different
duty cycle, be distorted from a sine wave, and have higher peak-to-peak
amplitude.
Brian <br...@nomail.net> wrote:
H...@xxxhydrologistxx.com
in message <3c56acc1...@nntp.fuse.net>:
>Except for the hardwired lighting, it also sounds a lot like the BX24-AHT
>which I first described here about two years ago and actually had running
>about 1.5 years ago. ;)
Yup. And as I recall, when I inquired about also grafting on the
CM11a-->RS232/hardwired interface, I was assigned the task as homework. I
don't even do all the homework I assign myself but this is getting closer
to the top of the pile every decade ;)
... Marc
Marc_F_Hult
H...@xxxhydrologistxx.com
Brian
>Except for the hardwired lighting,
No, there was a provision for a serial port (or ports) to interface to
lighting controllers and other devices with serial ports. The rough
protocol also included a means for devices to communicate with each
other using X-10 commands, but those commands wouldn't be gated to the
powerline (essentially sending a "preamble" of a specific user-defined
command would keep the following commands off the powerline, but the
device would spit them out it's other ports as if they were received
from the powerline).
>it also sounds a lot like the BX24-AHT
>which I first described here about two years ago and actually had running
>about 1.5 years ago. ;)
IIRC, this was actually before any mention of your AHT, so I may have
my dates wrong. Or, I may not :) In any case, you did release a more
real product.
Dave Houston
Well, since then I've abandoned the idea of trying to make it present a
CM11A-like image to other software as it can also report many things beyond
what any existing HA app would expect from a CM11A.
Instead, I have defined an ASCII protocol. As soon as I recover from a
health problem I'll release a firmware version that uses the ASCII protocol
for all I/O to work with the IP Commander and/or other hardware/software.
The current firmware will accept the ASCII commands for inputs but its
outputs are more verbose.
ASCII commands sent to the main port can be passed through to any of three
secondary RS-232 ports so that might be the best way to handle your
hardwired lights. There are people using a CM11A on one port, an Ocelot on
another, and the LCD+ on a third.
There are some limits (related to the single software UART) on receiving
from devices on the secondary RS-232 ports. For example, I can read the LCD+
ADC channels because they respond rapidly and I don't have to waste time in
a loop waiting for the response. I can set the relay drivers because there
is no feedback. But I cannot use the keypad because that requires listening
full time. I can poll the keypad port as digital inputs.
Someone has agreed to create an interface to HS and perhaps a Linux daemon
using the ASCII protocol.
---
http://www.laser.com/dhouston/
Dave Houston
>[] [] dhou...@fuse.net (Dave Houston) arranged some electrons to say:
>
>>Except for the hardwired lighting,
>
>No, there was a provision for a serial port (or ports) to interface to
>lighting controllers and other devices with serial ports. The rough
>protocol also included a means for devices to communicate with each
>other using X-10 commands, but those commands wouldn't be gated to the
>powerline (essentially sending a "preamble" of a specific user-defined
>command would keep the following commands off the powerline, but the
>device would spit them out it's other ports as if they were received
>from the powerline).
Which is exactly what the BX24-AHT does. Three secondary ports can interface
to any of several predefined devices or to any 9600 8N1 device. If you want
to route a ProLink command to an OmniPro on port three, you send :3 followed
by the command. There are provisions for sending binary also.
>>it also sounds a lot like the BX24-AHT
>>which I first described here about two years ago and actually had running
>>about 1.5 years ago. ;)
>
>IIRC, this was actually before any mention of your AHT, so I may have
>my dates wrong. Or, I may not :) In any case, you did release a more
>real product.
H...@xxxhydrologistxxx.com
in message <3c57d251...@nntp.fuse.net>:
>Well, since then I've abandoned the idea of trying to make it present a
>CM11A-like image to other software as it can also report many things
>beyond what any existing HA app would expect from a CM11A.
Shucks. The ability/applicability dichotomy never ends.
>Instead, I have defined an ASCII protocol.
Too bad it isn't a superset of X-10/Cm11a-speak.
Please post when able. There are quite a few protocols out there already,
but it could be very useful as is or adapted. The Centralite ASCII
protocol, for example, is simple (to a fault) and translation could be
straight-forward.
I mention Centralite yet again (and there are other excellent proprietary
systems such as TouchPlate) becasue in building my home-brew hard-wired
system, I try to keep in mind that eventually I will not support it. If I
stroke out manana, I'd prefer that those left on Earth be able to turn on
and off the lights in the house. So building things in modules that can be
replaced with commercially available products like Centralite and DMX-512
with minimal disruption is important. The solution to this problem for X-10
(implemented millions of time already) is to rip it all out. It is much
more dificult, costly and disruptive to do that with hard-wired systems.
>As soon as I recover from a health problem I'll release a firmware version
>that uses the ASCII protocol for all I/O to work with the IP Commander
>and/or other hardware/software. The current firmware will accept the ASCII
>commands for inputs but its outputs are more verbose.
>
>ASCII commands sent to the main port can be passed through to any of three
>secondary RS-232 ports so that might be the best way to handle your
>hardwired lights. There are people using a CM11A on one port, an Ocelot on
>another, and the LCD+ on a third.
>
>There are some limits (related to the single software UART) on receiving
>from devices on the secondary RS-232 ports.
Exasperating how slowly I figured that out. The lack of hardware buffers
for "COM3's" is one reason the BX-24 would not be the device of choice for
a DMX-512 translator unless DMX-512 was the primary function. Ditto for
direct phase control of triac/scr dimmers.
>For example, I can read the LCD+ ADC channels because they respond rapidly
>and I don't have to waste time in a loop waiting for the response.
>I can set the relay drivers because there is no feedback.
You already know my suggestions on how to _create_ feedback.
>But I cannot use the keypad because that requires listening full time. I
>can poll the keypad port as digital inputs.
Yup. The LCD+ turned out to be less useful for keypad input than I first
expected. I had thought that it could be an interface for the
potentiomenter-motor-Ain-dimmer and keypad for command-line control. I made
a faceplate using it for my potentiometer panel before I figgered that out.
>Someone has agreed to create an interface to HS
Great. HS now supports Napco, Slink-e, CM11a and Ocelot. It is becoming
very complete. Adding support for BX-24 implementations would make it
easier for others to roll their own. I haven't looked at Homeseer in a
while, but when Premise emerges from vapor-ware as it surely will, I may be
reassessing software again.
>and perhaps a Linux daemon
>using the ASCII protocol.
>---
>http://www.laser.com/dhouston/
Sounds wunderbar. I look forward to it all.
Take Care ...
Gotta go underground for a couple of days (literally) ... Marc
Marc_F_Hult
H...@xxxhydrologistxxx.com
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >In article <3c56027f...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
| >| ddl@danlan.*com (Dan Lanciani) wrote:
| >|
| >| >In article <3c55b6d8...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
| >| >
| >| >| In a split phase system, using two CM11As, both interfaces see the same ZC.
| >| >| Any leakage is in phase with and indistinguishable from the signal being
| >| >| transmitted so there is no collision.
| >| >
| >| >This would be true only if you managed to phase-lock the 120kHz carriers of
| >| >the two transmitters, possibly taking into account any phase shift induced
| >| >by the specific leakage path. Otherwise the carriers could interfere
| >| >destructively even though the high-level X10 code itself was in sync.
| >|
| >| I think the probability that they're going to be exactly 180 degrees out of
| >| phase and thus cancel at the 120KHz level is near zero.
| >
| >I think that the probability that they will be at or near 180 degrees out
| >of phase _at some point_ is quite high if they are not phase-locked. The
| >frequencies are bound to be a little off so it's just a question of how
| >fast the phase difference is changing compared to the timing of the higher-
| >level signal. You can imagine a situation where the null drifts through
| >the bits, confusing some of the clever receivers that sample at a few fixed
| >points per bit...
|
| To cancel they have to be exactly 180 degrees out of phase and of exactly
| the same amplitude.
This is true only if you are considering complete cancellation, but there are
infinitely many other possibilities. Consider a situation where the signal
is already near the lower limit of the receiver's sensitivity. Add another
carrier near 180 degrees out of phase and at, say, 25% of the amplitude. The
receiver will likely stop working.
| As a practical matter, it will NEVER happen.
I see no reason to believe that we would be that lucky. Consider ordinary
multi-path induced fading. Even with only one source, the phase shifts
provided by the paths themselves make it common for many radio services
and not unheard of for X10. With two unlocked oscillators you are guaranteeing
that all possible phase relationships will show up everywhere (over time).
Depending on how the beat frequency compares to the repetition rate of the
higher-level X10 signal you could see a variety of interesting effects, none
good.
| >| Besides, there is no practical way that you are going to get two CM11As to
| >| transmit at the same time, even at the high-level, whether you control them
| >| with a PC or a microcontroller.
| >
| >That may be true, but it's basically the opposite of the original argument
| >that they will tend to synchronize on the zero crossing. :)
|
| The initial argument was just to indicate to Brian that he had not given
| this enough thought. It was more of a "worst case" type thing than anything
| else.
Ok, but I still claim that the worst case is indeed a bad case. :)
| >| They would have to transmit sequentially.
| >
| >That should work as long as there is enough leakage for them to hear each
| >other and do the usual collision avoidance. Or you could just never ask them
| >to transmit near the same time in the first place...
|
| Huh? Why is collision avoidance necessary when only one will send at any
| given time?
It is necessary if you want the CM11as themselves to guarantee that the
transmissions don't overlap.
| Since I plan to segregate housecodes there will never be any
| need for both to send the same command. Even if you wanted both to send all
| commands, one would send a command and get confirmation before the other
| would send it.
As I said, ``Or you could just never ask them to transmit near the same time
in the first place...'' But all of this has little to do with the original
premise which clearly did not involve such interlocking.
Dan Lanciani
ddl@danlan.*com
Dave Houston
>As I said, ``Or you could just never ask them to transmit near the same time
>in the first place...'' But all of this has little to do with the original
>premise which clearly did not involve such interlocking.
My original premise is over two years old and has always included the
housecodes by phase concept. I have stated it many, many times.
BTW, if you take a look at...
http://www.laser.com/dhouston/phases.htm
you'll see that there's no apparent difference in the 120KHz whether two
TM751s are in phase or out of phase.
---
http://www.laser.com/dhouston/
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >As I said, ``Or you could just never ask them to transmit near the same time
| >in the first place...'' But all of this has little to do with the original
| >premise which clearly did not involve such interlocking.
|
| My original premise is over two years old and has always included the
| housecodes by phase concept. I have stated it many, many times.
I was talking about your premise that:
``In a split phase system, using two CM11As, both interfaces see the same ZC.
Any leakage is in phase with and indistinguishable from the signal being
transmitted so there is no collision.''
not about the dual-interface notion in general. Incidentally, I tried to
get Jeff what's-his-name (original Enerlogic designer) to support dual
TW523s on the 1400 back in 1993 or so. After about an hour on the phone
I gave up. :(
| BTW, if you take a look at...
|
| http://www.laser.com/dhouston/phases.htm
|
| you'll see that there's no apparent difference in the 120KHz whether two
| TM751s are in phase or out of phase.
I think you are talking about phase in the sense of which way the TM751 is
plugged in. I was talking about the relative phase of two unlocked oscillators
of near the same frequency, a constantly changing value. (Obviously reversing
a plug would instantaneously change the difference by 180 degrees, but that's
neither here nor there.) For some transceivers (notably the alarm bases that
I like) the phase in the sense you are using it does matter since they appear
to synchronize not just to the zero crossing but on the particular slope leading
thereto. Just to add further confusion. :)
Dan Lanciani
ddl@danlan.*com
Dave Houston
>In article <3c574590...@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>| ddl@danlan.*com (Dan Lanciani) wrote:
>|
>| >As I said, ``Or you could just never ask them to transmit near the same time
>| >in the first place...'' But all of this has little to do with the original
>| >premise which clearly did not involve such interlocking.
>|
>| My original premise is over two years old and has always included the
>| housecodes by phase concept. I have stated it many, many times.
>
>I was talking about your premise that:
>
>``In a split phase system, using two CM11As, both interfaces see the same ZC.
>Any leakage is in phase with and indistinguishable from the signal being
>transmitted so there is no collision.''
While choosing to ignore that, in the very same message, I said I would
segregate housecodes.
>not about the dual-interface notion in general. Incidentally, I tried to
>get Jeff what's-his-name (original Enerlogic designer) to support dual
>TW523s on the 1400 back in 1993 or so. After about an hour on the phone
>I gave up. :(
>
>| BTW, if you take a look at...
>|
>| http://www.laser.com/dhouston/phases.htm
>|
>| you'll see that there's no apparent difference in the 120KHz whether two
>| TM751s are in phase or out of phase.
>
>I think you are talking about phase in the sense of which way the TM751 is
>plugged in. I was talking about the relative phase of two unlocked oscillators
>of near the same frequency, a constantly changing value. (Obviously reversing
>a plug would instantaneously change the difference by 180 degrees, but that's
>neither here nor there.) For some transceivers (notably the alarm bases that
>I like) the phase in the sense you are using it does matter since they appear
>to synchronize not just to the zero crossing but on the particular slope leading
>thereto. Just to add further confusion. :)
No, I'm talking about the 120KHz signal which is what I captured on my scope
and which looks exactly the same whether the TM751s are on the same phase or
on opposite phases. Now, I readily admit that I did not watch for a
thousand years or so to see if the two unlocked oscillators ever slipped out
of phase.
And I'm certainly not going to use the alarm bases. Besides, how does the
way they act have anything to do with the way the CM11A acts. (Of course,
the CM11A may not act in the same way as the TM751 - I had my CM11As
otherwise occupied or I would have taken the scope shots using two of them.)
---
http://www.laser.com/dhouston/
Dave Houston
>Incidentally, I tried to
>get Jeff what's-his-name (original Enerlogic designer) to support dual
>TW523s on the 1400 back in 1993 or so. After about an hour on the phone
>I gave up. :(
Why didn't you just connect two TW523s through a splitter?
---
http://www.laser.com/dhouston/
Dan Lanciani
| >I was talking about your premise that:
| >
| >``In a split phase system, using two CM11As, both interfaces see the same ZC.
| >Any leakage is in phase with and indistinguishable from the signal being
| >transmitted so there is no collision.''
|
| While choosing to ignore that, in the very same message, I said I would
| segregate housecodes.
I ignored it because I don't see any significant issues with that approach.
But that approach entirely avoids the beating oscillators problem. The fact
that there is a way to avoid the problem and still do what you want to do
doesn't make the problem any less interesting, especially since it tends to
come up in other circumstances. I stand by my assertion that the leakage
you describe will not be indistinguishable if the two oscillators are not
locked. It may be distinguishable even if they are.
| >I think you are talking about phase in the sense of which way the TM751 is
| >plugged in. I was talking about the relative phase of two unlocked oscillators
| >of near the same frequency, a constantly changing value. (Obviously reversing
| >a plug would instantaneously change the difference by 180 degrees, but that's
| >neither here nor there.) For some transceivers (notably the alarm bases that
| >I like) the phase in the sense you are using it does matter since they appear
| >to synchronize not just to the zero crossing but on the particular slope leading
| >thereto. Just to add further confusion. :)
|
| No, I'm talking about the 120KHz signal which is what I captured on my scope
| and which looks exactly the same whether the TM751s are on the same phase or
| on opposite phases.
It still sounds like you are talking about the phase of the 60Hz power. Unless
the 120kHz oscillators are somehow synchronized to the 60Hz (perhaps in devices
where the MCU generates the carrier directly?) I don't see how this relates to
the phase of the carrier.
| Now, I readily admit that I did not watch for a
| thousand years or so to see if the two unlocked oscillators ever slipped out
| of phase.
Given the tolerances involved it seems unlikely that you would have to wait
quite that long. But waiting for them to slip out of phase implies that they
were in phase in the first place. How would that happen and how would you
confirm it?
| And I'm certainly not going to use the alarm bases. Besides, how does the
| way they act have anything to do with the way the CM11A acts.
It doesn't. I was just trying to add a little extra information. Probably
a mistake.
| (Of course,
| the CM11A may not act in the same way as the TM751 -
Right, I would think that the most interesting characteristic would be
whether the carrier oscillator is free-running or MCU-generated. We
know the CM11a is free-running I think.
| I had my CM11As
| otherwise occupied or I would have taken the scope shots using two of them.)
As you said, it could be difficult to get them to run in lockstep, but
it might be interesting...
Dan Lanciani
ddl@danlan.*com
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >Incidentally, I tried to
| >get Jeff what's-his-name (original Enerlogic designer) to support dual
| >TW523s on the 1400 back in 1993 or so. After about an hour on the phone
| >I gave up. :(
|
| Why didn't you just connect two TW523s through a splitter?
That would have gated the two unlocked oscillators at the same time resulting
in exactly the interference problem that I wanted to avoid. (I realize that
I haven't convinced you that this is a problem.) Two separately controlled
interfaces would have allowed commands to be sent alternately to the two legs
of the power line, avoiding the issue much as you plan to do with the CM11as.
Even then I considered either grafting a second output stage from one TW523
to another or borrowing the oscillator from one to drive the other. But there
were the usual isolation issues and it would have been quite a kludge, so I
dropped it. Later I considered something similar with two CM11as, but by then
reasonable repeaters were appearing...
Dan Lanciani
ddl@danlan.*com
Dave Houston
No, you haven't convinced me.
I have one TM751 plugged into a powerstrip. I have another TM751 plugged
into the same powerstrip through an extension cord that allows me to reverse
the hot and neutral to simulate in/out of phase. I have the ACT ScopeTest2
plugged into the same powerstrip. On the scope, looking at about 6 cycles of
the 120KHz carrier, I see no difference no matter which way I have the
second TM751 connected. Only the reversible TM751 is unplugged between the
tests.
I don't understand what oscillators you are talking about. I imagine the
120KHz is being generated either using PWM or merely by toggling the pin
every n clock ticks.
---
http://www.laser.com/dhouston/
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >In article <3c58e564....@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
| >| ddl@danlan.*com (Dan Lanciani) wrote:
| >|
| >| >Incidentally, I tried to
| >| >get Jeff what's-his-name (original Enerlogic designer) to support dual
| >| >TW523s on the 1400 back in 1993 or so. After about an hour on the phone
| >| >I gave up. :(
| >|
| >| Why didn't you just connect two TW523s through a splitter?
| >
| >That would have gated the two unlocked oscillators at the same time resulting
| >in exactly the interference problem that I wanted to avoid. (I realize that
| >I haven't convinced you that this is a problem.) Two separately controlled
| >interfaces would have allowed commands to be sent alternately to the two legs
| >of the power line, avoiding the issue much as you plan to do with the CM11as.
| >Even then I considered either grafting a second output stage from one TW523
| >to another or borrowing the oscillator from one to drive the other. But there
| >were the usual isolation issues and it would have been quite a kludge, so I
| >dropped it. Later I considered something similar with two CM11as, but by then
| >reasonable repeaters were appearing...
|
| No, you haven't convinced me.
Fair enough. It may take some hands-on frustration to fully appreciate
the problem. I spent a couple of years fiddling with it before I decided
that there was no practical way to work around the physics.
| I have one TM751 plugged into a powerstrip. I have another TM751 plugged
| into the same powerstrip through an extension cord that allows me to reverse
| the hot and neutral to simulate in/out of phase. I have the ACT ScopeTest2
| plugged into the same powerstrip. On the scope, looking at about 6 cycles of
| the 120KHz carrier, I see no difference no matter which way I have the
| second TM751 connected. Only the reversible TM751 is unplugged between the
| tests.
It's still not clear to me what you expect to see.
| I don't understand what oscillators you are talking about.
I'm talking about the discrete-component analog 120kHz oscillators present
in many (if not most) X10 powerline transmitters. But the argument also
applies to synthesized waveforms: the (lack of) synchronization is just
pushed off to the MPU's clock and the problem becomes harder to observe on
the bench.
| I imagine the
| 120KHz is being generated either using PWM or merely by toggling the pin
| every n clock ticks.
But it isn't in the CM11a, TW523, and RR501. Take a peek at the schematics.
They use analog oscillators. I believe the CM11a actually switches the
oscillator on and off while the other two gate an always-running oscillator.
Now that I think about it, it seems to me that the TM751 is one of the unusual
modules that *does* generate the carrier internal to the PIC. If you have
been using the TM751 exclusively for this kind of testing you might be seeing
more the exception than the rule...
Dan Lanciani
ddl@danlan.*com
Dave Houston
I don't recall seeing a full schematic for the CM11A but the schematics for
the RR501 & TM751 show an almost identical 120KHz circuit. If it were
running all the time, there would be a continuous 120KHz on the powerline.
---
http://www.laser.com/dhouston/
Dan Lanciani
| >But it isn't in the CM11a, TW523, and RR501. Take a peek at the schematics.
| >They use analog oscillators. I believe the CM11a actually switches the
| >oscillator on and off while the other two gate an always-running oscillator.
| >Now that I think about it, it seems to me that the TM751 is one of the unusual
| >modules that *does* generate the carrier internal to the PIC. If you have
| >been using the TM751 exclusively for this kind of testing you might be seeing
| >more the exception than the rule...
|
| I don't recall seeing a full schematic for the CM11A
The interesting part is here:
http://home.pacbell.net/lgalvin/CM11AX-10a.gif
| but the schematics for
| the RR501 & TM751 show an almost identical 120KHz circuit.
The RR501 schematic at:
http://www.geocities.com/ido_bartana/RR501Align/RR501Sonic.gif
appears to correspond to RR501s that I've owned. Looks to me like TR12
and L3 form a free-running oscillator which is ultimately gated to the
line by pin 17 of the PIC. On the other hand, the TM751 schematic here:
http://www.geocities.com/ido_bartana/TM751_Main_Board.gif
has a *much* simpler output circuit that to me looks like a simple
tuned coupler. I could be wrong and I have no idea if this schematic
really corresponds to a TM751 since I've avoided them as much as possible.
(As you recall, even the salvaged RF board from the one I had was pretty
useless. :()
But for an unambiguous example I still suggest you look at the TW523 (or
even PL513) schematics. There were supplied by X10 so you don't have to
worry about errors by the reverse-engineers. And we know for certain that
the input to the device at the RJ11 is a simple gate and not a 120kHz carrier.
| If it were
| running all the time, there would be a continuous 120KHz on the powerline.
Some of them run all the time, some don't. I think I got the associations
right for the versions of the devices I've seen. Maybe there are new and
different versions of some of the devices now. The only possible difference
between the ones that run all the time and the ones that are started and
stopped is that *maybe* given all the right conditions you might get a
consistent startup that synchronizes the oscillator--at least for a while--
with the MCU. As with an internally generated carrier, this will likely make
the problem harder to observe on the bench but won't solve the problem in
practice.
Note that the whole carrier interference issue is not a recent revelation.
I'm pretty sure it was well understood before I started hacking X10, and
that was ~20 years ago. I tried to work around it but I couldn't. No matter
how carefully things were arranged there was always a module somewhere that
managed to see a corrupt code maybe one try out of ten. That's way above the
error rate for otherwise-working X10 so I gave up on multiple synchronized
transmitters with unsynchronized carriers.
Dan Lanciani
ddl@danlan.*com
Dave Houston
>In article <3c58b638....@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>
>| >But it isn't in the CM11a, TW523, and RR501. Take a peek at the schematics.
>| >They use analog oscillators. I believe the CM11a actually switches the
>| >oscillator on and off while the other two gate an always-running oscillator.
>| >Now that I think about it, it seems to me that the TM751 is one of the unusual
>| >modules that *does* generate the carrier internal to the PIC. If you have
>| >been using the TM751 exclusively for this kind of testing you might be seeing
>| >more the exception than the rule...
>|
>| I don't recall seeing a full schematic for the CM11A
>
>The interesting part is here:
>
>http://home.pacbell.net/lgalvin/CM11AX-10a.gif
That, I have. John did that schematic at my request. I don't believe that is
a free running oscillator but I'll check it later today as I have a CM11A
that's still apart so it will be fairly simple to solder a couple of leads
on it so I can watch it on the scope.
>| but the schematics for
>| the RR501 & TM751 show an almost identical 120KHz circuit.
>
>The RR501 schematic at:
>
>http://www.geocities.com/ido_bartana/RR501Align/RR501Sonic.gif
>
>appears to correspond to RR501s that I've owned. Looks to me like TR12
>and L3 form a free-running oscillator which is ultimately gated to the
>line by pin 17 of the PIC. On the other hand, the TM751 schematic here:
They haven't made that one (with the RF section enclosed by a shield) in
several years. My oldest RR501 was made in 1996 but I think the new style is
older than it. The X-10 output section appears to be the same as for the
TM751. There is a schematic (by Mitch Orysh) of the later model on Ido's
site. The schematic provided by Jam a week or so ago only disagrees on a few
points.
http://free.pages.at/x10jam/x10_rr501b.htm
>
>http://www.geocities.com/ido_bartana/TM751_Main_Board.gif
>
>has a *much* simpler output circuit that to me looks like a simple
>tuned coupler. I could be wrong and I have no idea if this schematic
>really corresponds to a TM751 since I've avoided them as much as possible.
>(As you recall, even the salvaged RF board from the one I had was pretty
>useless. :()
While one I salvaged is excellent. This would seem to speak more to quality
problems that to design flaws.
>But for an unambiguous example I still suggest you look at the TW523 (or
>even PL513) schematics. There were supplied by X10 so you don't have to
>worry about errors by the reverse-engineers. And we know for certain that
>the input to the device at the RJ11 is a simple gate and not a 120kHz carrier.
The X-10 output section of the TW523 is quite similar to the TM751 and
current RR501. One of the three windings is directly connected to the
powerline. If it were a free running oscillator, it would flood the
powerline with continuous 120KHz.
>| If it were
>| running all the time, there would be a continuous 120KHz on the powerline.
>
>Some of them run all the time, some don't. I think I got the associations
>right for the versions of the devices I've seen. Maybe there are new and
>different versions of some of the devices now. The only possible difference
>between the ones that run all the time and the ones that are started and
>stopped is that *maybe* given all the right conditions you might get a
>consistent startup that synchronizes the oscillator--at least for a while--
>with the MCU. As with an internally generated carrier, this will likely make
>the problem harder to observe on the bench but won't solve the problem in
>practice.
I don't believe any current TW523s, TM751s, RR501s, or CM11As employ free
running oscillators. They are kickstarted by the MCU with the startup
sync'ed to ZC.
You've glossed over the consequences of a free running oscillator that is
always coupled to the powerline which is a powerline flooded with 120KHz,
blocking all X-10 activity.
>Note that the whole carrier interference issue is not a recent revelation.
>I'm pretty sure it was well understood before I started hacking X10, and
>that was ~20 years ago. I tried to work around it but I couldn't. No matter
>how carefully things were arranged there was always a module somewhere that
>managed to see a corrupt code maybe one try out of ten. That's way above the
>error rate for otherwise-working X10 so I gave up on multiple synchronized
>transmitters with unsynchronized carriers.
Since you're splitting hairs in debating a non-issue (I've never advocated
sending simultaneously with two CM11As.), referring to transmitters that are
~14 years older than the CM11A would seem to do nothing other than obfuscate
the issue.
And I'm not arguing that carrier interference is not a possibility with some
systems. I believe its one of the flaws in CEBus. I just don't believe that
it's an issue with current models of X-10's transmitters.
I believe someone posted that their CM11A overheated and stopped functioning
when they reversed powerline polarity so I'm not going to test two CM11As on
the same powerstrip, with one CM11A's polarity reversed, in the same way I
tested the TM751 but I will look at one on the scope. If your hypothesis is
correct, I shouldn't have to reverse the polarity of one of a pair of
CM11As. They should interfere even if on the same phase. That I can test but
I'll have to assemble another BX24-AHT in order to have a way to trigger
both at the same time.
Anyway, the CM11A detects collisions by listening to the powerline during
the time it is "sending" a 0. Unless you're going to claim that the CM11A on
the opposite phase is going to slip 1/2 of a powerline cycle, I think it's
clear that there will be no collisions detected if both send the same code
at the same time.
---
http://www.laser.com/dhouston/
Dave Houston
>In article <3c58b638....@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>
>| >But it isn't in the CM11a, TW523, and RR501. Take a peek at the schematics.
>| >They use analog oscillators. I believe the CM11a actually switches the
>| >oscillator on and off while the other two gate an always-running oscillator.
>| >Now that I think about it, it seems to me that the TM751 is one of the unusual
>| >modules that *does* generate the carrier internal to the PIC. If you have
>| >been using the TM751 exclusively for this kind of testing you might be seeing
>| >more the exception than the rule...
>|
>| I don't recall seeing a full schematic for the CM11A
>
>The interesting part is here:
>
>http://home.pacbell.net/lgalvin/CM11AX-10a.gif
I've added screenshots showing pins 7 & 10 of the CM11A while it is sending
a command to the powerline to...
http://www.laser.com/dhouston/phases.htm
Scroll down past the TM751 screenshots.
For each "1" sent to the powerline, pin 7 outputs bursts of 120KHz that last
about 1.2mS. The first of each triplet is sync'ed to ZC while the next two
are timed internally. There is no free running oscillator.
The three winding transformer appears to be a tuned coupler/filterr.
Filtering would be necessary to pass FCC limits on powerline harmonics (a
120KHz square wave would surely interfere with AM radio bands). The FCC
mandates testing "digital devices" like the CM11A by an approved lab to
"verify" it does not exceed EMI radiation limits or couple AM band noise to
the powerline.
I think the RR501, TM751, AND TW523 also synthesize the 120KHz in a similar
manner. All have nearly identical circuitry for their X-10 OUT sections and
none of those circuits look like an oscillator. I'll try to get screenshots
of each over the next couple of weeks.
---
http://www.laser.com/dhouston/
Dave Houston
http://www.laser.com/dhouston/phases.htm
with some 'scope screenshots of pin 18 of the PIC in the RR501.
Like in the CM11A, the 120KHz is generated by the PIC. Given that the TM751
and TW523 use X-10 output circuits essentially identical to the RR501, I
think they probably generate the 120KHz in a similar manner.
Neil Cherry can probably confirm or refute this for the TW523.
---
http://www.laser.com/dhouston/
Dan Lanciani
| I've added screenshots showing pins 7 & 10 of the CM11A while it is sending
| a command to the powerline to...
|
| http://www.laser.com/dhouston/phases.htm
|
| Scroll down past the TM751 screenshots.
|
| For each "1" sent to the powerline, pin 7 outputs bursts of 120KHz that last
| about 1.2mS.
You have completely misunderstood the operation of the circuit that you are
trying to analyze. Pin 7 of the PIC is used as a gate to disable the line
output driver when it is not in use. The PIC drives the pin low to disable
the driver. To enable the driver the PIC *floats* the pin. While the pin is
floating and the analog oscillator is enabled the 120kHz *from* the oscillator
naturally appears at pin 7 of the PIC. Don't believe me? There is a simple
test. Cut the trace to pin 7 of the PIC. You may be surprised to see that the
CM11a still works. It may get a little hot and the output may not be quite as
clean since in theory you can better see the oscillator's chirp, but it will
work. Want some proof that the analog oscillator is really there? Cut also the
trace to pin 10 (that's the analog oscillator enable) and tie it (not the pin,
the trace it was connected to) high. Guess what? You'll get a steady carrier
on the line. All without help from the PIC.
You say you don't want to chop up a CM11a? I'll be happy to loan you one.
It has already had those traces cut and spliced before so I don't mind. I'll
even put pigtails on them for you. Wonder why the traces are already chopped
up? It's because I went through all this back when I was thinking about
slaving one set of output drivers to another CM11a. Of course, I didn't
have a neat and almost correct (I think there is a pullup resistor missing
at the base of Q3 but don't hold me too it) schematic to start from. By the
way, the loan offer applies to anyone else who wants to play with with the
CM11a's transmitter.
| I think the RR501, TM751, AND TW523 also synthesize the 120KHz in a similar
| manner. All have nearly identical circuitry for their X-10 OUT sections
I've looked around and I can't find a TW523 schematic that differs from my
original copy. I hesitate to make definitive comments on whatever it is that
you might be looking at since you'll likely accuse me of obfuscating with
obsolete information. The original TW523 has an analog oscillator very similar
to the one in the CM11a. It has no MCU in the output path at all. Given that
the RJ11 application interface to the TW523 has not changed to require a 120kHz
input, I assume that whatever you are looking at does now have such an MCU to
"synthesize" the carrier? If it does not then I would have to believe that you
have misunderstood the circuit.
| and
| none of those circuits look like an oscillator.
What do you expect the oscillator to look like? Most of them look like the
one in the CM11a.
Dan Lanciani
ddl@danlan.*com
Dave Houston
>In article <3c5b78ce....@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
>
>| I've added screenshots showing pins 7 & 10 of the CM11A while it is sending
>| a command to the powerline to...
>|
>| http://www.laser.com/dhouston/phases.htm
>|
>| Scroll down past the TM751 screenshots.
>|
>| For each "1" sent to the powerline, pin 7 outputs bursts of 120KHz that last
>| about 1.2mS.
>
>You have completely misunderstood the operation of the circuit that you are
>trying to analyze. Pin 7 of the PIC is used as a gate to disable the line
>output driver when it is not in use. The PIC drives the pin low to disable
>the driver. To enable the driver the PIC *floats* the pin. While the pin is
>floating and the analog oscillator is enabled the 120kHz *from* the oscillator
>naturally appears at pin 7 of the PIC. Don't believe me? There is a simple
>test. Cut the trace to pin 7 of the PIC. You may be surprised to see that the
>CM11a still works. It may get a little hot and the output may not be quite as
>clean since in theory you can better see the oscillator's chirp, but it will
>work. Want some proof that the analog oscillator is really there? Cut also the
>trace to pin 10 (that's the analog oscillator enable) and tie it (not the pin,
>the trace it was connected to) high. Guess what? You'll get a steady carrier
>on the line. All without help from the PIC.
You're probably right. I've never claimed to be very good at solid state
circuit analysis. I learned electronics in pre stone age times and haven't
worked in the field in 40 odd years.
And you're right again that with pin 7 disconnected, the 120KHz still gets
to the powerline.
I'll take your word for it that supplying a continuous +5V at the point John
Galvin labeled "Osc Enable" will result in a continuous 120KHz on the
powerline.
But, I won't agree with your final statement in the above paragraph. When I
disconnect pin 10, no 120KHz makes it to the powerline. I'd say that the PIC
supplying +5V to the oscillator qualifies as "help from the PIC".
Furthermore, I'd say that pin 10 starts the oscillator. And since, for the
first burst of each triplet" pin 10 goes high at ZC, it looks like the
120KHz is synchronized to ZC. It looks to me that you've proven my point. ;)
I still don't see a free running oscillator. Perhaps your definition of a
free running oscillator is different from mine.
I've added some more screenshots to...
http://www.laser.com/dhouston/phases.htm
There's one rather interesting phenomenon that I do not know how to
interpret.
---
http://www.laser.com/dhouston/
John Galvin
I'm not sure if this is the best place for me to jump in, but here goes
anyway.
I have seen two different types of X-10 transmitters. Some like the CM11A,
TW523, early RR501s, CP290, tend to have a discrete 120KHz oscillator, while
others like the TM751, SD533, new RR501s, IR543, MC460, tend towards PIC
generated 120KHz designs. In the case of PIC generated 120KHz, it is quite
feasible to generate the 120KHz, phase locked to the 60Hz line frequency.
The TM751 definitely generates the 120KHz bursts with the PIC. If that is
phase locked to the 60Hz, then that would explain the test results that Dave
has gotten. No matter how you plug two TM751s in, you would not get
destructive coupling. Dan, we did discuss this quite some time ago. I
believe that we left it at, the difficulty of 60Hz zero crossing detection,
inducing phase shift in the PLL'd 120Khz. I'm not sure if I actually have a
working TM751 anywhere to scope, at this point. I do have a Sundowner
around here somewhere and that one has PIC generated 120KHz.
J.G.
Dave Houston wrote in message <3c5c2895...@nntp.fuse.net>...
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
|
| >In article <3c5b78ce....@nntp.fuse.net>, dhou...@fuse.net (Dave Houston) writes:
| >
| >| I've added screenshots showing pins 7 & 10 of the CM11A while it is sending
| >| a command to the powerline to...
| >|
| >| http://www.laser.com/dhouston/phases.htm
| >|
| >| Scroll down past the TM751 screenshots.
| >|
| >| For each "1" sent to the powerline, pin 7 outputs bursts of 120KHz that last
| >| about 1.2mS.
| >
| >You have completely misunderstood the operation of the circuit that you are
| >trying to analyze. Pin 7 of the PIC is used as a gate to disable the line
| >output driver when it is not in use. The PIC drives the pin low to disable
| >the driver. To enable the driver the PIC *floats* the pin. While the pin is
| >floating and the analog oscillator is enabled the 120kHz *from* the oscillator
| >naturally appears at pin 7 of the PIC. Don't believe me? There is a simple
| >test. Cut the trace to pin 7 of the PIC. You may be surprised to see that the
| >CM11a still works. It may get a little hot and the output may not be quite as
| >clean since in theory you can better see the oscillator's chirp, but it will
| >work. Want some proof that the analog oscillator is really there? Cut also the
| >trace to pin 10 (that's the analog oscillator enable) and tie it (not the pin,
| >the trace it was connected to) high. Guess what? You'll get a steady carrier
| >on the line. All without help from the PIC.
|
| You're probably right.
Probably right? What would it take to convince you that I'm actually right?
I mean, this isn't Schroedinger's cat. It's really hard to explain what's
going on if each element remains in a limbo of ambiguity.
| I've never claimed to be very good at solid state
| circuit analysis. I learned electronics in pre stone age times and haven't
| worked in the field in 40 odd years.
|
| And you're right again that with pin 7 disconnected, the 120KHz still gets
| to the powerline.
|
| I'll take your word for it that supplying a continuous +5V at the point John
| Galvin labeled "Osc Enable" will result in a continuous 120KHz on the
| powerline.
Please don't take my word for it. Try it. Otherwise in a few iterations
you may change your mind and decide that the oscillator wasn't there after
all. :( (I'm glad you noticed the helpful labels on John's schematic. They
make life much easier, no?)
| But, I won't agree with your final statement in the above paragraph. When I
| disconnect pin 10, no 120KHz makes it to the powerline.
I never said it would. I said to tie the line high. Why did you bother
disconnecting pin 10 if you weren't going to try what I suggested? Depriving
the oscillator of power tends to discourage its oscillation.
| I'd say that the PIC
| supplying +5V to the oscillator qualifies as "help from the PIC".
Come on Dave, this is beneath you. But if you want to pick that nit, the
PIC doesn't supply the +5V. The power supply does. You can totally remove
the PIC from the circuit. Desolder it from the board. The oscillator will
still oscillate without any help from the PIC. I said that if you tie the
pin high you will get a steady carrier. All without help from the PIC. I
stand by the assertion.
| Furthermore, I'd say that pin 10 starts the oscillator. And since, for the
| first burst of each triplet" pin 10 goes high at ZC, it looks like the
| 120KHz is synchronized to ZC. It looks to me that you've proven my point. ;)
I.e., even though you were completely wrong about the circuit this proves
that you were completely right in general. :) Nice try, but I don't think so.
A day or so ago you didn't even recognize that the circuit was an oscillator.
Lines above in this very message you only "probably" thought I was correct.
Now not only have you recognized the oscillator circuit but you think you
understand its chirping characteristics well enough to comment on startup
synchronization. To do that analysis on paper would be beyond my abilities.
I'd probably have to use a simulator. Even then it would be a pain to collect
all the component tolerances and temperature coefficients. Your credentials
as an expert witness on oscillator circuit characteristics are looking a tad
tarnished. Given that you won't accept the most basic assertions from me
unless I provide some sort of hands-on proof, I hope you will do me the same
courtesy and show me the computations to back up your synchronization theory.
But what the hey? Let's go ahead and assume that the analog oscillator is a
perfect design that always starts on a dime throughout the range of its
components' tolerances. It's every EE's fantasy. Let's further assume that
if you keep the carriers within 1/8 of a cycle they won't do much harm to each
other. That's a very generous assumption, but if gives a nice slop budget of
just over 1 usec, i.e., a +-0.5 usec window for each of two devices. The two
main remaining elements contributing to the slop are the tolerance of the
zero-crossing detector and the PIC instruction timing. Do the math and see
what happens.
Hint: The analysis of the tolerance in the zero crossing detector is hard.
Do the PIC analysis first to see if there is any point in continuing.
I noticed you elided the comments on the TW523. I'd still like to see
the synthesizing version that you are talking about.
Dan Lanciani
ddl@danlan.*com
Dan Lanciani
(I didn't realize that anyone else was still reading this thread.)
| I'm not sure if this is the best place for me to jump in, but here goes
| anyway.
As contributor of the nice, neat schematic you get to jump in wherever you want.
| I have seen two different types of X-10 transmitters. Some like the CM11A,
| TW523, early RR501s, CP290, tend to have a discrete 120KHz oscillator, while
| others like the TM751, SD533, new RR501s, IR543, MC460, tend towards PIC
| generated 120KHz designs. In the case of PIC generated 120KHz, it is quite
| feasible to generate the 120KHz, phase locked to the 60Hz line frequency.
It's an interesting question. (Or a non-issue depending on who you ask. :)
In effect, you are trying to phase-lock two carriers by reference to a
common and much slower intermediate frequency, i.e., the 60Hz. The problem
is that offsets that are insignificant compared to the 60Hz are very significant
to the 120kHz. In this case one microsecond plus or minus is probably the
beyond the allowable limit. Seems to me that for analysis sake we can
consider a PIC-generated carrier equivalent to Dave's idealized discrete
oscillator that starts up perfectly every time. In either case the PICs
can start the 120kHz waveforms synchronous to their instruction execution.
The three remaining asynchronous signals to consider are the 60Hz and the
clocks in the two PICs.
| The TM751 definitely generates the 120KHz bursts with the PIC.
Do you happen to know the clock speed of the PIC in the PIC version of the
TM751? My TM751 has a non-PIC part and doesn't look anything like the
available schematic. Makes me wonder just how many versions of these things
there are...
| If that is
| phase locked to the 60Hz, then that would explain the test results that Dave
| has gotten.
That was exactly my initial suggestion. However, now I'm not so sure.
| No matter how you plug two TM751s in, you would not get
| destructive coupling.
Just to revisit this polarity issue again, I think I created some confusion
a long time ago:
http://groups.google.com/groups?hl=en&selm=291%40ddlgw.UUCP
(I love having all the old news online. :)
When people first started having problems with multiple TM751s I speculated
incorrectly that the TM751 might care about the direction of the slope at the
zero crossing the way the alarm bases do. This was strictly about when the
high-level X10 code started and had nothing to do with the phase of the
carrier.
| Dan, we did discuss this quite some time ago. I
| believe that we left it at, the difficulty of 60Hz zero crossing detection,
| inducing phase shift in the PLL'd 120Khz.
Right, you need the zero-crossing detector to be accurate to microsecond levels
between the two transmitters, independent of line voltage differences. There's
no obvious reason for X10 to have designed such a good circuit, though any
two given units might just happen to have components of close enough values
that it appears so. But even if the zero-crossing detector does an adequate
job you still have a big problem with a 4MHz PIC like the one in the CM11a.
Dan Lanciani
ddl@danlan.*com
Dave Houston
No, you said it was a free running oscillator. And I said that if it were a
free running oscillator there would be 120KHz on the powerline all the time.
If you tie it to +5V it does become a free running oscillator and does, by
your own statement, put 120KHz on the powerline all the time.
If you tie your left arm to your side to immobilize it, you will have some
idea of the difficultiy I have with soldering. I'll choose what and when I
solder.
The oscillator starts when pin 10 goes high and the oscillator stops when
pin 10 goes low. Pin 10 goes high whenever there's a "1" to send. For each
"1", pin 10 goes high at ZC and at ZC + 2.778mS and at ZC + 5.556mS. Since
we're tlking about split-phase power we can ignore all but the 120KHz bursts
at ZC.
With a CM11A on either leg of the split phase power, each will see the same
ZC, pin 10 in each will go high at the same time, and the oscillator in each
will start up at the same time.
Unless one oscillator starts by swinging low while the other starts by
swinging high, the oscillators will be in phase with each other. They are
not likely to drift very far apart in the 1mS before each pin 10 goes low.
All of your assumptions about tolerances are completely irrelevant. Unless
the two carriers are 180 degrees apart and of the same amplitude, there will
not be total cancellation.
Even if the two CM11As can "hear" the other's output, there will never be a
case where a "0" gets turned into a "1" so neither will detect a collision.
An X-10 receiver located remotely from the CM11As will see a 120KHz signal.
Even if one of the 120KHz carriers gets shifted 180 degrees, the receiver
will have to be rather precisely located to see both signals at the same
amplitude. As long as the receiver sees any signal, it will be at 120KHz
although it may not have a 50% duty cycle.
My understanding is that the receiver's "window" is from about .2mS to .8mS
after ZC. If it counts 40 transistions in that window, it registers a "1".
If it counts less than 40 transitions, it registers a "0".
---
http://www.laser.com/dhouston/