older lamplinc (V33) not always responding to scenes

[eddyk]

Hi,

 

I just got my V.35 switchlinc issues resolved. Smart-home replaced all 9 of them.

 

Now I still have some issues with some older lamplincs (V33). They seem to be responding reliable to "direct" ON / OFF commands given from the admin console, but they are not responding reliably to scenes. I even create a new scene with only 1 responder and no controller. When I turn the scene ON or OFF from the admin console, the lamplinc respond good about 80%-90% of the time. When it fails, the admin console reports the requested change in the device status but the status did NOT actually change.

 

I replaced some with new lamplincs (V.3b) and it corrects the problem.

 

Has anyone seen this before?

 

Is this a similar issues as with the .35 switchlincs or something else?

 

Thanks in advance.

 

Eddy K.

[oberkc]

When it fails, the admin console reports the requested change in the device status but the status did NOT actually change

I understand that the ISY assumes a status, even when a direct acknowledgement is not recieved. This sounds like the old communication problem to me.

 

I replaced some with new lamplincs (V.3b) and it corrects the problem.

I understand some devices may be better than others (manufacturing tolerances and all) but I have not noticed inherent problems with any particular version (I never had any V.35).

[Brian H]

What does a Scene Test Diagnostics show when you try it on the test scene with the LampLinc in it?

[eddyk]

What does a Scene Test Diagnostics show when you try it on the test scene with the LampLinc in it?

 

I will have to try this evening.

 

Thanks,

 

Eddy K.

[eddyk]

What does a Scene Test Diagnostics show when you try it on the test scene with the LampLinc in it?

 

Here are the scene test results for when the load responds during the test:

Tue 12/21/2010 06:38:13 PM : [GRP-RX      ] 02 61 15 13 00 06 

Tue 12/21/2010 06:38:14 PM : [CLEAN-UP-RPT] 02 58 06 

Tue 12/21/2010 06:38:14 PM : [iNST-SRX    ] 02 50 0E.A4.F9 0E.D9.09 61 13 15    LTOFFRR(15)

Tue 12/21/2010 06:38:14 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

Tue 12/21/2010 06:38:14 PM : [iNST-SRX    ] 02 50 0E.A4.F9 0E.D9.09 61 13 15    LTOFFRR(15):  Process Message: failed

Tue 12/21/2010 06:38:14 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

----- scene test Test Results -----
[succeeded] "Bird" Table Lamp (E A4 F9 1)
----- scene test Test Results -----
Tue 12/21/2010 06:38:21 PM : [iNST-ACK    ] 02 62 00.00.15 CF 13 00 06          LTOFFRR(00)

 

Here are the scene test results for when the load DOESN'T responds during the test:

 

Tue 12/21/2010 06:51:05 PM : [GRP-RX      ] 02 61 15 13 00 06 

Tue 12/21/2010 06:51:06 PM : [CLEAN-UP-RPT] 02 58 06 

Tue 12/21/2010 06:51:06 PM : [iNST-SRX    ] 02 50 0E.A4.F9 0E.D9.09 62 13 15    LTOFFRR(15)

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

Tue 12/21/2010 06:51:06 PM : [iNST-SRX    ] 02 50 0E.A4.F9 0E.D9.09 62 13 15    LTOFFRR(15):  Process Message: failed

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

----- scene test Test Results -----
[succeeded] "Bird" Table Lamp (E A4 F9 1)
----- scene test Test Results -----
Tue 12/21/2010 06:51:14 PM : [iNST-ACK    ] 02 62 00.00.15 CF 13 00 06          LTOFFRR(00)

 

See anything suspicious?

 

Thanks,

 

Eddy K.

[IndyMike]

Hello Eddy,

 

I'm nowhere near an expert at interpreting the scene test responses, but this device sure appears to be communicating that it has turned off.

 

If the device has not turned off, please try:

 

1) Plug the Lamplinc into the same circuit as the PLM and perform a device Restore. It's possible the link table for this device is not intact.

 

2) Retry your scene test with the Lamplinc on the same circuit as the PLM.

 

3) If the above works, try putting you Lamplinc back in its original location and retest.

 

If the above does not work, we may have to consider whether this device has in fact gone bad. I've never encountered a device that would communicate properly but not control the load.

 

Is this a dimmer or relay unit?

[oberkc]

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

 

Tue 12/21/2010 06:51:06 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 62 13 15 LTOFFRR(15): Process Message: failed

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

 

Tue 12/21/2010 06:51:06 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 62 13 15 LTOFFRR(15): Process Message: failed

These are the parts of the message that I find interesting. I understand the scene test is nothing more than turning a scene on or off and watching for acknowledgement. I also understand the scene test limits the number of repeats to a lower level than a normal command (it is for this, and other, reasons that a scene test may fail yet your device may still respond). I take these (no hops left, "failed") as an indication of less than robust communication amongst the ISY/PLM and whatever is device 0E.A4.F9.

 

A couple of things I have learned about scene tests. First, if you have any programs that are triggered by changes in status of devices within the scene, this can adversely affect the reliability and usefulness of a scene test. Second, I understand the scene test has been updated in the latest version of the ISY software (partially in response to some shortcomings in the scene test). I suggest updating to the latest version if you have not already done so.

 

I would take your results as an indication that you may have some problems. The typical questions to one in your situation:

 

a. is your PLM plugged into the same outlet or circuit with lots of computer stuff (UPS, surge suppressor, power supplies, etc...)?

b. what is your situation with regards to access points or dual-band devices? Do you have robust communication across the legs of your electrical system?

[eddyk]

Hello Eddy,

 

I'm nowhere near an expert at interpreting the scene test responses, but this device sure appears to be communicating that it has turned off.

 

If the device has not turned off, please try:

 

1) Plug the Lamplinc into the same circuit as the PLM and perform a device Restore. It's possible the link table for this device is not intact.

 

2) Retry your scene test with the Lamplinc on the same circuit as the PLM.

 

3) If the above works, try putting you Lamplinc back in its original location and retest.

 

If the above does not work, we may have to consider whether this device has in fact gone bad. I've never encountered a device that would communicate properly but not control the load.

 

Is this a dimmer or relay unit?

 

You may not consider yourself an expert but you sure helped me out.

 

1) Before I followed your advice, I verified that I still had the problem. The lamplinc dimmer failed to response to the scene about 20% of the time.

 

2) Then I followed your steps 1 - 3 with success.

 

3) I turned the scene ON / OFF without any failures many times.

 

I will continue to monitor this device for several days to make sure the problem is gone forever.

 

Thanks for your help!

[LeeG]

eddyk

 

In both cases, when the load turned Off and when the load did not turn Off, the LampLinc responded with an ACK indicating Scene command was received and reacted to.

 

From load turned Off trace

 

Tue 12/21/2010 06:38:14 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 61 13 15 LTOFFRR(15)

 

Tue 12/21/2010 06:38:14 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

 

From load failed to turn Off trace

 

Tue 12/21/2010 06:51:06 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 62 13 15 LTOFFRR(15)

 

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

 

However, there is a difference in the Max Hops count. Max Hops is a 1 when the load responded and Max Hops is a 2 when the load did not respond. The increase in Max Hops indicates the Controller (ISY/PLM in this case) retried the Group message which indicates a powerline issue to that device or an issue with that device specifically. Could be some noise or something attenuating the Insteon signal on that circuit or something associated with the specific load the LampLinc is controlling. Some CFLs generate noise when they are On which can interfere with the LampLinc being able to respond consistently. If not an incandescent bulb change the LampLinc load to a standard incandescent bulb and repeat the test.

 

There is another situation with the LampLinc response. The trace shows the LampLinc is sending two ACKs in both cases (when load responds and when the load does not respond). I take this to be a characteristic of the LampLinc firmware level for now and would ignore the Process Message: failed message. The ISY is simply indicating it already received the ACK from the LampLinc and is not expecting a second ACK. Since this happens in both cases I don’t think it is related to the load not responding.

 

Trace from the load responding case

 

Tue 12/21/2010 06:38:14 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=1, Hops Left=0

 

Tue 12/21/2010 06:38:14 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 61 13 15 LTOFFRR(15): Process Message: failed

 

Trace from the load not responding case

 

Tue 12/21/2010 06:51:06 PM : [standard-Cleanup Ack][0E.A4.F9-->ISY/PLM Group=0] Max Hops=2, Hops Left=0

 

Tue 12/21/2010 06:51:06 PM : [iNST-SRX ] 02 50 0E.A4.F9 0E.D9.09 62 13 15 LTOFFRR(15): Process Message: failed

 

Lee

[eddyk]

Lee,

 

Thanks for the detailed info. I also noticed the difference in hop counts, just didn't know what it meant.

 

I also think that this may boil down to firmware issues. I have 3 or 4 devices, all same firmware that have / had this problem. All other devices have been responding reliable.

 

Restoring the device made it much better but time will tell if that was a permanent fix.

 

I have other devices (newer lamplincs and switchlincs) within a few feet of the "problem child" on the same electrical circuit and they are responding fine 100% of the time. Also, the "problem child" lamplinc responds 100% to direct ON / OFF commands.

 

I don't believe that I have a lot of noise on the power line. I viewed the insteon signal with an oscilloscope. When Insteon communicates, the signal show on the scope is much stronger than any noise that might be there.

 

I'm not sure how Smart Home will respond if this is a lamplinc firmware / ISY compatibility issue. They just replaced all nine of my V.35 switchlinc dimmers which was a know problem with ISY. I have not been able to fine any similar discussions on the V.33 Lamplinc dimmers.

 

Eddy K.

[IndyMike]

Hello again Eddy,

 

I'm glad to hear that appears to have worked. Unfortunately, we may have fixed the symptom but not the underlying problem. As both Oberkc and Lee have indicated, the root cause here may be noise.

 

Let's re-cap:

1) Your scene tests indicated that the LL was responding to the scene (it acknowledged the request) but the device didn't always respond correctly. From this I'm inferring that the LL received the request, but didn't always process it properly (it performed the wrong function). I'm a little at a loss to explain how this is possible, but a "corrupted" device link table seemed a good guess.

 

2) Lee caught the fact that the "Hops" increased on one of the scene test acknowledges. As he indicated, this is symptomatic of noise on the line.

 

3) You moved the LL to the same circuit as the PLM and performed a restore. This hopefully repaired the device link table and allowed it to execute the scene properly.

 

4) You moved the LL back to the original location and it again performed well.

 

I am hoping that your problems are due to device programming problems at the "original" device location. This could be due to noise or an interaction with your (now replaced) V.35 devices. Unfortunately noise has a habit of coming and going as loads are activated/deactivate. Only time will tell whether this problem has truly been resolved.

 

One last note - please verify that you have the device programming mode set to "Automatic" under the menu item \Link Management\Advanced Options.

 

Please do report back on any further observations you may have.

[eddyk]

One last note - please verify that you have the device programming mode set to "Automatic" under the menu item \Link Management\Advanced Options.

 

Yes it is.

 

Thanks again for your support.

 

Eddy K.

[oberkc]

I am also glad you got it working. Like the others have mentioned, I have come to the belief that creating scenes in a difficult communication environment can lead to "upredictable" results. Perhaps this was a source of your problems, corrected when you restored the device while close to the PLM. I, too, share in the suspicion that you still may have something going on that may cause some communication issues down the road.

 

But, if all is working, enjoy it for now. Just watch for signs if you make changes.

[eddyk]

Hi,

 

Here are some images that show noise on my power line. It doesn't show any noticeable noise other than noise created by the dimmers themselves (triac switching)

 

Insteon traffic at zero crossing. Noise near peak due to triac switching load.

 

Insteon traffic at zero crossing.

 

No insteon traffic. Noise from triac switching. Noise is only present when load on and not at 100%.

 

I'm not an Insteon expert but it appears to me that the insteon signal strength is good compared to the noise level. Also the noise is not there during the zero crossing so it should not impact insteon communication.

 

Any thoughts?

[Brian H]

Your link didn't work but I edited the link and got to the photos.

http://www.flickr.com/photos/eddykiggen/

 

Insteon messages start 800us before Zero Crossing.

X10 starts at Zero Crossing.

 

It looks like the noise is at around AC peak if I read the photos correctly.

 

I believe the Insteon Modules also have AGC in them and the detection level maybe falsely being set incorrectly.

 

Do you know the hardware revision levels between the two different firmware LampLincs?

 

I know my older HW rev 1.3 ApplianceLincs compared to my later HW rev 4.1s are 100% different designs.

My older ones are more picky about inductive spikes from the load being switched on and off. While the 4.1s seem to be more picky to power line noise and missed commands.

[eddyk]

Your link didn't work but I edited the link and got to the photos.

http://www.flickr.com/photos/eddykiggen/

 

Insteon messages start 800us before Zero Crossing.

X10 starts at Zero Crossing.

 

It looks like the noise is at around AC peak if I read the photos correctly.

 

I believe the Insteon Modules also have AGC in them and the detection level maybe falsely being set incorrectly.

 

Do you know the hardware revision levels between the two different firmware LampLincs?

 

I know my older HW rev 1.3 ApplianceLincs compared to my later HW rev 4.1s are 100% different designs.

My older ones are more picky about inductive spikes from the load being switched on and off. While the 4.1s seem to be more picky to power line noise and missed commands.

 

Older lamplic is a 2456D3 V 4.3 0957

New lamplinc is a 2457D2 Rev 1.23 2610

[Brian H]

Oh so we are talking a later revision LampLinc and a Dual Band LampLinc.

Two completely different electrical designs. That could be the reason one works and the other doesn't. That is just a theory.

[IndyMike]

Hello Eddie,

 

I had a look at your trace over at the Flikr site - I'm assuming that you were triggering on the Insteon signal at 2V - Correct?

 

The noise that you are referring to is then the Triac impulse which occurs at roughly the peak of the 60 Hz sine - Correct?

 

That being the case, I would say that your conclusions are correct:

1) Your Insteon levels look very healthy - 5 to 6 Vp-p vs an advertised level of 3.2 Vp-p.

 

2) The impulse is well away from the zero crossing and should not affect communications.

 

Notes:

1) Your triac Impulse will move depending on the dim level. 100% on will move left toward the zero crossing. Low levels will move right toward to opposite zero crossing.

 

2) Your dimmer impulses appear to be from incandescent bulbs (rather low level). Dimmable CFL's can produce impulses of 40V an more. These can be problematic at low dim levels as the spikes approach the zero crossing.

 

3) I've never run across a good description of how Smartlabs implements it's AGC (automatic gain control). I have encountered instances where the PLM AGC appears to have "engaged" when the noise was outside the Insteon window and well outside the frequency range.

 

Beyond the above,

1) What are you using to filter out the 60Hz powerline signal (homebrew or scopetest2)?

 

2) What levels do you see around the zero crossing without Insteon communication?

 

Thanks for sharing,

IM

[eddyk]

I had a look at your trace over at the Flikr site - I'm assuming that you were triggering on the Insteon signal at 2V - Correct?

Yes

 

The noise that you are referring to is then the Triac impulse which occurs at roughly the peak of the 60 Hz sine - Correct?

Yes

 

Beyond the above,

1) What are you using to filter out the 60Hz powerline signal (homebrew or scopetest2)?

I'm using a X10 pro XPCP passive coupler. Connect L1 & N1 to the power line and L2 & N2 to the scope. the 2 pairs are electrically not connected, making it more safe. I realize that X-10 frequencies a slightly lower but it works just fine.

I also use a small 24 Volt AC transformer to safely view the power line voltage on channel 2 of the scope. I just don't like connecting power line voltage to the input of my 500 MHz memory scope.

 

Also, I tried dimming the load, moving the triac impulse closer to the zero crossing.

 

It will move (at extreme dim levels) close enough to the zero crossing where it could impact Insteon signals. Do you know if the noise signals I see are common?

 

It sure doesn't seem that I have a "noise generator" like a UPS or phone charger thats causing a problem.

 

Also, all my loads are incandescent. I have 4 Access points and 4 dual band devices.

 

Thanks,

 

Eddy K.

[IndyMike]

Eddy,

 

You have a nice (and safe) setup for monitoring the powerline signals. I don't have component values for the XPCP, but I would suspect they are similar to my ACT CP-0000 coupler. One thing to note - while these are "bandpass" couplers, signals removed from the 120Khz center frequency will be attenuated. Insteon at 130 Khz should be fine. I've found that many of my CFL's tend to resonate at around 60 KHz - the signals from these devices are attenuated by around 50% with my CP-000.

 

I would characterize the dimmer impulses you captured as rather benign. However, we can't see what is occurring around the zero crossing. Not many devices draw power in this region. That's why the communication protocols chose this time slot for communication.

 

As a reference point, I have many dimmable CFL's both inside and outside my home. These can be problematic and I have to pay attention to make sure the dim level does not get too low (the spikes begin to crowd the zero crossing)

 

The following is a trace from my basement game room - 8 - 16W CFL's at 24% level. The spike is being triggered at 40V and it's getting close to the zero crossing. Amazingly, this does not seem to impact my PLM (about 40" away). It can take down the communication to the controlling switch.

 

In this plot the yellow is the noise level at 0.02V/Div. The pink is the 60Hz crossing. The white bars are the Insteon communication window.

 

[eddyk]

Thanks,

 

I feel that I'm fine on noise levels unless someone can tell me otherwise.

 

By the way, How did you get the image in the post. The link I tried did not work.

 

Eddy K.

[IndyMike]

Not sure what to tell you on linking to your photo's. I always use the "preview" option to make sure things show up. You may need to fiddle with the link a bit.

 

You may want to try monitoring the noise levels at your PLM. It's the brain of your Insteon network. Noise at this point will have far more impact than elsewhere in the system. Beyond that, you're using incandescent lighting so short of a UPS, PC power supply (normally an absorber), or rogue cell phone charger I wouldn't think you would have issues. Unfortunately, every home has its own characteristics.