My CM15A has been doing some real strange things.
1. If I have a light on, Like my pool table E3 and turn it on again with the PalmPad the pool table light will trun Off and then On again.
2. If I have a light on and use the blue dim buttons on the PalmPad. Normaly the light would dim a step up or down but now it runs to 100% On or 100% Off
3. For what ever reason the CM15A likes to give out an random all lights on with House code "E". Will try to change to a new house code this week.
Might I be doing somting wrong as I have not had this kind of problem before. I did send back 1 CM15A for replacment do to this problem, But it has not changed with the new one.
When I look at the Activity Monitor every thing looks and acts like it should with the buttons I push on the PalmPad remote.
When I run with out the CM15A it seems every thing works like it should, No random all lights on sent. Dimmers work etc.
Any one else seen a problem like this before? I have had this house wired as a X10 home for 6 plus years now and have not had any strange problems before adding Active Home Pro
Quote from: azbaja on November 21, 2006, 10:18:52 AM
When I run with out the CM15A it seems every thing works like it should, No random all lights on sent. Dimmers work etc.
By saying "
without the CM15A" you mean with it
unplugged, correct?
If so, do you have a
second transceiver in your set up for the
PalmPad to work?
azbaja:
When you get the E all lights on check the AM to see if there are any other signals recieved prior to this!
Then do an Advanced Search (http://www.x10community.com/forums/index.php?action=search;advanced) with PLC noise
And one with RF noise if you notice a recieve of RF before the all on!
This seems like I'm making you do a lot of work and reading ::) but the reading will help you understand what is causing this and how to fix it! ;)
I still think there is a bug in AHP when it comes to receiving RF! I have written programs that watch the CM15A interface (through AHP) and see all kinds of random garbage logged when an RF signal is received by the CM15A.
My two cents.
Bill; Could be as I have looked at the RF Receivers output and it is a constant white noise static. No squelch on the receiver.
Quote from: Puck on November 21, 2006, 02:56:30 PM
Quote from: azbaja on November 21, 2006, 10:18:52 AM
When I run with out the CM15A it seems every thing works like it should, No random all lights on sent. Dimmers work etc.
By saying "without the CM15A" you mean with it unplugged, correct?
If so, do you have a second transceiver in your set up for the PalmPad to work?
Yes a 2 RR501 that has been working fine for the last 5 or 6 years no
You could be getting multiple triggers to modules / macros in AHP.
If your remote signals are reaching the CM15A and the RR501 (that's set to the remote's house code), AHP will receive the RF first thru the CM15A, then it will receive the PLC signal (transceived thru the RR501 & picked up thru the CM15A).
When using multiple Transceivers, you have to take extra care in making sure AHP does not receive multiple triggers. One way of doing this is with flags.
Also, with regards to the other posts here, make sure noise is not your problem too. Try moving the CM15A to another electrical location (at least try to keep it off of the same one as the PC). As Tuicemen stated above, there's a lot of info on the topic of noise here.
Please post what you do and what did or didn't work for you. ;)
Also, keep an eye on this thread: ActiveHome Computer Interfaces (CM11A *AND* CM15A) Go Bananas! (http://www.x10community.com/forums/index.php?topic=10984.0)
It looks like someone else is having a simular problem.
I have done the move about 2 weeks ago, The computer was in the back room of the house, Now I have placed the computer with the CM15A in the garage.
I have alos moved the most of my house codes of from "E" and on to "F"
I hope you all understand the problem, most all the lights on house code "E" just trun on. The Activity Monitor Shows nothing being sent out from the CM15A but somthing is turning lights on and that is only with the CM15A pluged in.
I will keep you posted, Thanks
Quote from: Puck on November 22, 2006, 10:53:47 AM
You could be getting multiple triggers to modules / macros in AHP.
If your remote signals are reaching the CM15A and the RR501 (that's set to the remote's house code), AHP will receive the RF first thru the CM15A, then it will receive the PLC signal (transceived thru the RR501 & picked up thru the CM15A).
When using multiple Transceivers, you have to take extra care in making sure AHP does not receive multiple triggers. One way of doing this is with flags.
Also, with regards to the other posts here, make sure noise is not your problem too. Try moving the CM15A to another electrical location (at least try to keep it off of the same one as the PC). As Tuicemen stated above, there's a lot of info on the topic of noise here.
Please post what you do and what did or didn't work for you. ;)
Also, keep an eye on this thread: ActiveHome Computer Interfaces (CM11A *AND* CM15A) Go Bananas! (http://www.x10community.com/forums/index.php?topic=10984.0)
It looks like someone else is having a similar problem.
Having a two RR501's on the same housecode is a problem source. They keep stepping on each other. I have 1 RR501 (for the Stick-a-Switch by the front door) and another 501 in the breakfast room that also has a Stick-a-Switch. The Living Room 501 and the CM15A are in the same room and 7 feet apart. but on two different breakers. If I unplug the 501, the CM15 cannot receive the RF signals from the Stick-a-switch.
Quote from: azbaja on November 21, 2006, 10:18:52 AM
Might I be doing somting wrong as I have not had this kind of problem before. I did send back 1 CM15A for replacment do to this problem, But it has not changed with the new one.
azbaja: You had stated that everything was working
OK before, and now you are on your
2nd CM15A for this same problem. So it's unlikely the problem is the
CM15A.
Has anything new been
added / changed to your system just prior to it not working (including
AHP version)? Is the
CM15A itself the new change?
Does anything show in the
Activity Monitor when this occurs... I know you stated that is doesn't show anything being sent out, but I just want to make sure your not just referring to
nothing for house code "E", and that it's not showing anything received.
How often does this problem occur?
A
noisy RF environment could cause some strange things to happen, but always turning
house code "E" lights out??? I doubt it could be that consistant. So that is an unlikely cause.
Since this only occurs with the
CM15A plugged in, maybe try something...
Clear the Interface and
DO NOT download your Macros & Timers. Then see if the problem still occurs. If not, you may have a
corrupt ".AHX" file and have to recreate a new one from scratch. If it does, then try unplugging the
RR501's one at a time.
azbaja:
Make sure you transceivers are set up in Active Home Pro correctly!
Did this just start to happen after doing an update or adding a plug-in?
If so then it is most likely a corrupt AHX file and you'll have to rebiuld it as Puck stated!
You stated you've had a x10 home for over 6 years AHP isn't that old so when did your problem start?
Right after adding AHP? ???
If so did you manualy build the ahx file (your configuration file) or did you import your setup from active home?
If you imported the active home configuration this is the problem as errors get created in the process you may need to go through the setup and inspect each and every macro or rebuild it from scratch! ::)
Sorry for so many questions but there are so many possiblities! ;) :D
CM15a's antenna sucks. Noisy and in some cases doesn't work at all. Don't send anything from its antenna to the power line.
Noisy environments, half-dead motion sensors, etc. will consistently make the same (bogus) commands show up. All Lights On/All Off seem to be the easiest to slip by.
I use two WGL transceivers, one on the power line for remote control (for a couple of sticky switches) and one that connects to the serial port of a PC and sends the commands directly to Misterhouse (primarily for motion and brightness sensors.) That way the constant chatter from the sensors does not step on manual intervention (pressing a button on a PalmPad, camera remote, etc.) There is a very good article out there about how to pick which house codes to use. There are only six of sixteen that are comfortably spaced (has to do with the raw codes received by the transceivers and how they differentiate them.) Within these house codes, there are six units codes that are similarly spaced. So if you can get by with 36 discrete codes, you will have virtually trouble-free operation. How to get by like that? Scenes. How else? IR control (PalmPads suck!)
Oh and watch out when looking up the "two bits apart" X10 address spacing page (that should Google it) as there are two. One is correct, one is wrong. They both had the same idea, but only one guy got it right. Hopefully the other one has been taken down. Seems like A, D, F, G, H and P are the "safe" ones, but don't quote me on that. Despite using scenes and other tricks, I still can't condense my setup down to 36. Sure enough, it is the few modules that reside on the "unsafe" codes that occasionally turn on/off by themselves. The more motion sensors you have (especially the light colored models), the more "ghosts" you will have.
David Marks:
Can you identify the page with the "two bits apart" info you mention? Googling with that phrase turned up nothing related to X10 I could see. Thanks.
Charles,
A quick Google with "collisions" instead of "two bits" came up with the wrong one:
http://www.user.dccnet.com/middleton/home%20automation%20secret.htm
Doh!
The problem is the assumption he made about the codes. Look at his table with the binary representations. That is not reality.
I can't find the right one, but can recreate it:
HC UC Binary (X10 code, not literal translation!)
A 1 0110*
B 2 1110
C 3 0010
D 4 1010
E 5 0001
F 6 1001*
G 7 0101
H 8 1101
I 9 0111
J 10 1111*
K 11 0011*
L 12 1011
M 13 0000*
N 14 1000
O 15 0100
P 16 1100*
* Safe
So it looks like A, F, J, K, M, P rather than A, D, G, J, M and P. Safe units are 1, 6, 10, 11, 13, 16.
You should use these 36 codes for scenes and the individual modules within the scenes. Motion and brightness sensors can be put on any house code as you don't transceive them (or hopefully you don't!) The idea is that when you push the button to send one of these codes, chances are very slim that they will be recognized as one of the other 35. If you must transceive motion sensors (for lights that go on and off without the controller getting involved), put them on their own house code within these six (don't mix and match motions with manual and macro commands.) With the indoor models you cannot keep them segregated to the safe unit codes (except for address 10, which puts the brightness signal on 11.)
With all due respect David, I don't believe either you or that other guy you reference fully understand the X10 protocol.
He is obviously clueless when it comes to the X10 encoding of house and unit codes, which you thankfully do understand.
Now he says "... X-10 does NOT have an error correction system.", which is a true statement. However X10 does use an error detection scheme - each bit transmitted is immediately followed by its complement. If that pattern isn't detected the signal is flagged as an error and ignored. Then each signal is repeated to double the chances of the receiver getting a correct code. With this scheme the probability of noise or collided signals being received as a valid (albeit incorrect) code is very small, at least to the extent that a receiver module correctly implements the protocol. (Granted, we can't be sure of this since we don't have access to the microcontroller source code for the modules, and X-10 hasn't received many awards for bug-free programming.)
But how then to account for the instances of lights flashing on and off apparently at random?
Several reasons:
The on/off operation of these modules is nominally controlled by the microcontroller flipping on/off bits in the module's power control circuitry. But the AC power line is a pretty dirty electrical environment and it's not uncommon to have voltage spikes and surges which can induce bit-flips directly in a module's power control circuitry, bypassing the microcontroller. Evidence that the module is being controlled by other than X10 signals is that no signal is detected by a receiver like a CM11A at the time a random event occurs.
Another reason is that a flakey X10 controller is sending the wrong signals. This is easy enough to see with a CM11A by disconnecting the serial cable at the computer end and leaving it dangling from the CM11A. This happens a lot with me because I'm constantly fiddling with one or two CM11As between different PCs and serial ports. Although anything is likely to be transmitted, in my case I often see a lot of 0x00 and 0xFF being transmitted, which unfortunately are received as several of what you describe as "safe" addresses and functions - M13, M-AllOff, J10, J-StatusRequest. The TM751 Transceiver is susceptible to a mode whereby it transmits an endless stream of dim signals, thought to be due to RF feedback.
In addition to the above are the oft-heard complaints which turn out to be X10 signals sent from a neighboring house on the same power company transformer. Or the motion sensor which transmits on the programmed unit+1 on light or dark.
In short I consider that table of "safe" house and unit codes worthless. I would not use housecode A and unit 1 since that's the default for all modules and controllers and there's always the chance of a neighbor getting a new X10 toy for Christmas. I would not use housecode P, especially units 1 or 16, since they're used by X10 test equipment and by AHP at startup. In my own situation I would not use housecodes M or J for the reason stated earlier, but that wouldn't apply to most people. Aside from that I consider any particular housecode/unit as no better or no worse than any other.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM
>>With all due respect David, I don't believe either you or that other guy you reference fully understand the X10 protocol.
I think I understand it pretty well. As I said, the other guy is wrong. There is another version of the same theory out there that is accepted.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>>He is obviously clueless when it comes to the X10 encoding of house and unit codes, which you thankfully do understand.
Right.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>>Now he says "... X-10 does NOT have an error correction system.", which is a true statement. However X10 does use an error detection scheme - each bit transmitted is immediately followed by its complement. If that pattern isn't detected the signal is flagged as an error and ignored. Then each signal is repeated to double the chances of the receiver getting a correct code. With this scheme the probability of
I know about the repeated signal, but was unaware that the checksum scheme was part of the powerline transmissions. If so then in a perfect world there would be no need for a scheme like this.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>noise or collided signals being received as a valid (albeit incorrect) code is very small, at least to the extent that a receiver module correctly i>mplements the protocol. (Granted, we can't be sure of this since we don't have access to the microcontroller source code for the modules, and >X-10 hasn't received many awards for bug-free programming.)
Right. It is not a perfect world. I'm not familiar enough with the hardware/firmware end of it to comment on these articles much further. I do think that it pays to think about how you organize your units per house code, especially if you insist on transceiving motion signals.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>But how then to account for the instances of lights flashing on and off apparently at random?
I don't know about flashing on and off, but motion sensors can cause odd things to happen. Just address a unit with your remote, forget about it and walk in front of a sensor on the same house code.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PMSeveral reasons:
>The on/off operation of these modules is nominally controlled by the microcontroller flipping on/off bits in the module's power control circuitry. But the AC power line is a pretty dirty electrical environment and it's not uncommon to have voltage spikes and surges which can induce bit-flips directly in a module's power control circuitry, bypassing the microcontroller. Evidence that the module is being controlled by other than X10 signals is that no signal is detected by a receiver like a CM11A at the time a random event occurs.
Line noise. No kidding. Can definitely cause a module to come on for no good reason. Luckily I have virtually none of that here. X10 loves to blame everything on line noise when I call them though (even the macro and timer failures on the controller!)
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>Another reason is that a flakey X10 controller is sending the wrong signals. This is easy enough to see with a CM11A by disconnecting the serial cable at the computer end and leaving it dangling from the CM11A. This happens a lot with me because I'm constantly fiddling with one or two CM11As between different PCs and serial ports. Although anything is likely to be transmitted, in my case I often see a lot of 0x00 and 0xFF being transmitted, which unfortunately are received as several of what you describe as "safe" addresses and functions - M13, M-AllOff, J10, J-StatusRequest.
Well, I don't really deem them anything. I read the article years ago and it seemed there were others in the X10 community who accepted it. I had a CM11a for years but I never ran it unhooked from the PC. Are those signals really received though? Are the checksums sent as well? Regardless, leaving the serial cable hanging isn't a good way to go.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>The TM751 Transceiver is susceptible to a mode whereby it transmits an endless stream of dim signals, thought to be due to RF feedback.
That's an easy one. Plug two transceivers in on the same house code on the same phase (or anywhere if you have a coupler.) Dims go all the way to off, brights go all the way to 100% every time. Big trouble.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>In addition to the above are the oft-heard complaints which turn out to be X10 signals sent from a neighboring house on the same power company transformer. Or the motion sensor which transmits on the programmed unit+1 on light or dark.
No issues there as none of my neighbors have X10. Am quite familiar with the brightness sensors. Odd that AHP isn't.
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>In short I consider that table of "safe" house and unit codes worthless. I would not use housecode A and unit 1 since that's the default for all >modules and controllers and there's always the chance of a neighbor getting a new X10 toy for Christmas. I would not use housecode P,
If my neighbors ever get X10, I will make sure that there is no way for the signals to cross over into my house!
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>especially units 1 or 16, since they're used by X10 test equipment and by AHP at startup. In my own situation I would not use housecodes M or J
Not sure what test equipment you mean. What in God's name is AHP doing on P at startup. Never mind as I don't start it up (I shut it down.)
Quote from: Charles Sullivan on December 23, 2006, 10:23:47 PM>for the reason stated earlier, but that wouldn't apply to most people. Aside from that I consider any particular housecode/unit as no better or no worse than any other.
Could be. Misinformation is not hard to find on the Internet. Seems like some discussion in one of these forums led me to those two documents (the "right" one and the patently wrong one.)
[Edited by CS to delineate quoted from original material]
[TTA Edit: Increased SIZE of QUOTED TEXT from 8pt to 10pt so that those of us with "Advanced Eyesight" could at least read it!]
Quote from: David Mark on December 23, 2006, 11:17:00 PM
I know about the repeated signal, but was unaware that the checksum scheme was part of the powerline transmissions. If so then in a perfect world there would be no need for a scheme like this.
It's not what we normally think of as a checksum. A short 120 KHz pulse or lack thereof is transmitted at a zero crossing of the AC waveform and its complement at the next zero crossing. So 1 bit of data requires one full AC cycle.
Quote from: David Mark on December 23, 2006, 11:17:00 PM
I don't know about flashing on and off, but motion sensors can cause odd things to happen. Just address a unit with your remote, forget about it and walk in front of a sensor on the same house code.
I presume you're talking about addressing a unit with a plug-in remote like a MaxiController which can send an address without an accompanying function. X10 addressing is cumulative until the next function on the same housecode, so if you've addressed one or more units and didn't transmit a function, then the On signal from the motion detector will turn on all the addressed units. If that's not what you're talking about, then I'd venture the guess that the function part of the signal sent from the remote was corrupted or marginally too weak to be recognized as such by the module, which then remained in the cumulative address mode.
Quote from: David Mark on December 23, 2006, 11:17:00 PM
I had a CM11a for years but I never ran it unhooked from the PC. Are those signals really received though? Are the checksums sent as well?
Oh yes, the signals are received by another CM11A as well as the affected modules. I haven't put an o'scope on the power line, but assume the normal sequence of bit/~bit is there, else it the signal would be ignored by the CM11A.
Quote from: David Mark on December 23, 2006, 11:17:00 PM
Not sure what test equipment you mean. What in God's name is AHP doing on P at startup. Never mind as I don't start it up (I shut it down.)
X10 test signal generators - several I know of will repeatedly send P1 On, P1 Off. It allows you to go around the house with an X10 signal strength meter to check the level at various locations. I have no idea what X-10 had in mind for the P16 On signal transmitted by AHP at startup.
[TTA Edit: Increased font size of QUOTED TEXT to be 'readable' by our SENIOR Members. If you also agree with this increase, PLEASE post your support.]
As for the motion sensor snafus, it can be either (forgot to finish a command or the function got corrupted, leaving the unit addressed.) It isn't random, but it can seem like it to the uninitiated. I am quite familiar with X10 unit addressing (the lack of logic for it was one of my primary complaints about mh that led me to rewrite most of its X10 processing.)