Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[luiz]

Hi,

It happened a couple of times recently: my CM11A stops responding and keeps sending spurious (invalid?) commands over the power line. The only way I found to reset it is to unplug and remove the batteries. I know that commands are being sent by monitoring the X10 activity leds on my phase repeater (http://www.smarthome.com/4826a.html). They just blink like crazy, not-stop, when it gets into this state. Any idea of what could be causing this and how could it be reset by software? I have my own software driver for it so any command is possible.

Regards,
Luiz

[dave w]

It may not be the CM11. What software are you using to control the CM11? Any chance you added a new macro that is looping?

[luiz]

Well, it could be that my driver put CM11A in that state, but then it was for sure acting on its own. Even when I disconnected the serial connection to the PC it continued to put garbage on the power line. I had to disconnect it from the power line and remove the batteries to reset it. I wrote myself a Java driver based on information available on the Internet (long story...)

[dave w]

Even when I disconnected the serial connection to the PC it continued to put garbage on the power line.

O-o-h that probably eliminates your software from looping because I don't think the CM11 has memory to continue looping a command when computer is removed from picture.

Well, it was a thought....

[luiz]

Thanks Dave.

I guess my two questions now are (first for curiosity, second is the critical one):

1) Is it the software (my driver) or hardware (e.g. power line fluctuations) that is causing this behaviour on CM11A?

2) I know when it's in this state as all subsequent commands will fail. Now, how can I recover from it using my driver without physically disconnecting and removing the batteries?

[Brian H]

By chance do you have any two way X10 modules like an AM14A or LM14A?
On certain addresses mine will start a Firestorm of X10 bounces with my Smarthome Repeater. The constant signals keeps any valid X10 signals from being sent as the interface is polite and waits for a clear powerline to start sending.

[luiz]

No, I don't have any two way devices. Besides, the fact that the line returns to normal (no flashing LEDs on the repeater) after CM11A is unplugged shows that the culprit is the controller itself.

I have seen reports of lock downs of this device before, which may be the same behaviour I'm experiencing, but never really found answers for my questions above.

[Charles Sullivan]

Luiz,
All it takes is a one-bit error in a macro downloaded to the CM11A memory for it to exhibit the problem you are seeing, whether or not the CM11A is connected to the PC.  Even if you haven't downloaded anything, the memory may contain pre-existing garbage which can be triggered by some X10 signal received over the power line.   Unplugging and pulling the batteries won't cure this - you'll need to make sure the memory is erased.  If your software can't do this, use ActiveHome or other software for this task.

You may also see the same problem if you disconnect the serial cable at the PC end and leave it dangling from the CM11A.  The CM11A is very sensitive to stray pickup.

[luiz]

Ohhh, I see. So maybe my driver is corrupting the macro memory or it's getting corrupted somehow by some stray signal. Then solution is than to clean up the macros? Good, I will try that and remove the connection at the PC end to try to force the problem to re-occur and test the recoverability of my driver. Thanks, Charles.

[Charles Sullivan]

Ohhh, I see. So maybe my driver is corrupting the macro memory or it's getting corrupted somehow by some stray signal. Then solution is than to clean up the macros? Good, I will try that and remove the connection at the PC end to try to force the problem to re-occur and test the recoverability of my driver. Thanks, Charles.

Luiz,
The dangling cable is a separate problem.

Does your software download timers and macros to the CM11A memory or are you just executing CM11A commands from the PC?

[Walt2]

By chance do you have any two way X10 modules like an AM14A or LM14A?
On certain addresses mine will start a Firestorm of X10 bounces with my Smarthome Repeater. The constant signals keeps any valid X10 signals from being sent as the interface is polite and waits for a clear powerline to start sending.

Yep, I just got a 4826B, and it clearly does not play well with my LM14A 2-way Lamp Modules.

No problems with any of my other modules, just the LM14A's.   

In playing around with it, the problem appears to be between the LM14A's trying to return status,
and the 4826 trying to repeat that status.   The LM14A "sees" the repeat as a collision, and subsequently
tries to return status again.  The 4826 tries to repeat it again.  Etc.  Etc.   Until you unplug either
the LM14A or the 4826. 

My suggestion, is that if you any 2-way modules, to avoid the 4826A or 4826B.

[Brian H]

AM14A two way module; also doesn't play well.

[luiz]

Interesting thing about two way modules, but none are installed in my house.

I tried to turn on debugging messages on my driver and found that about 50% of the time it hung, CM11A sent an EEPROM ADDRESS TX (0x5b) message just before the problem started. Apparently this message is sent when it's about to execute a macro, which again corroborates Charles suspicion about corrupt memory. Maybe for some reason the memory for one macro is corrupted. Now I have to figure out how to clear the EEPROM memory from my driver. Anybody out there  has a sample code for that? Is there any hardware reset?

[Charles Sullivan]

Interesting thing about two way modules, but none are installed in my house.

I tried to turn on debugging messages on my driver and found that about 50% of the time it hung, CM11A sent an EEPROM ADDRESS TX (0x5b) message just before the problem started. Apparently this message is sent when it's about to execute a macro, which again corroborates Charles suspicion about corrupt memory. Maybe for some reason the memory for one macro is corrupted. Now I have to figure out how to clear the EEPROM memory from my driver. Anybody out there  has a sample code for that? Is there any hardware reset?

Send the CM11A the following 19-byte buffer of hex bytes, read a (buffer + 1)  checksum (0xF5), then send 1 byte 0x00:
  FB 00 00 00 03 FF FF FF FF FF FF FF FF FF FF FF FF FF FF

This nulls out the EEPROM header so no timers or triggers will be recognized.

[luiz]

Thanks again, Charles. Since I don't use macros at all, the reset is part of the driver startup now. Will see how it behave during the next few days.

Assuming that the lock up problem is macro related, I'm still wondering what is triggering the macro in the first place. It happened Tuesday and Wednesday last week, at about 06:30am, normal on Thursday, happened again on Friday (about the same time), and worked as expected over the weekend. Could it be power fluctuations? My UPS didn't report any power problems during this period.