Getting Serious - When Comm issues strike

[ELA]

My initial install -as an X-10 replacement went well. Then over time as I added more to the system it began to fail and become less reliable.

On my initial attempts to correct I was able to relocate a few dual band devices and add a WAP to get better reliability. With insights gained from this forum.

 

I hate the "trial and error" method of attempting to get better results.

I have the luxury of owning an Oscilloscope and have since been monitoring the signal strengths.

In many threads there have been discussions about the lack of troubleshooting tools. This has prompted me to spend way too much time thinking about diagnostic tool concepts.

 

I am envisioning a "TEST PLM".

 

I am ordering an 2412S PLM and intend to get even more serious about diagnosing comm. issues.

With this device (and a custom controller) I expect to be able to define my own commands and thus gain control over the number of HOPS allowed.

 

It is my hope to repeatedly send a "dummy extended length message" out over the network with "0" hops. I am thinking this will allow me to better view signal strengths with the oscilloscope. I further envision the ability to then plug this "Test PLM" into any outlet for diagnostic purposes.

 

Depending upon how well this first step works I may then decide to expand its capabilies further.

 

Recognizing I am still a newbie to Insteon communications ... I welcome any comments from advanced troubleshooters.

[Michel Kohanim]

Hello ELA,

 

This sounds great!

 

We would like to help with this endeavor as much as we can including a free loaner ISY. Please send an email to tech@universal-devices.com if you are interested.

 

With kind regards,

Michel

[Michel Kohanim]

Hello ELA,

 

This sounds great!

 

We would like to help with this endeavor as much as we can including a free loaner ISY. Please send an email to tech@universal-devices.com if you are interested.

 

With kind regards,

Michel

[Brian H]

The 2412S PLM has been discontinued for awhile now.

You may still find one being sold by automation vendors.

The 2413S Dual Band is the replacement.

SimpleHomeNet makes an equivalent to the 2412S.

http://www.simplehomenet.com/proddetail ... od=EZIComm

 

If you have not found it yet. The older revision of Developers Guide for the PLM is here. Along with other goodies.

http://www.madreporite.com/insteon/insteon.html

[ELA]

Thank you for the kind offer Michel,

My intent was to have a device that did not require running an ISY. I have an ISY99i for my system but I am interested in a test device I can readily move around the house.

 

Thank you Brian,

I read an old post of yours? at Smarthome yesterday and found the simplehomenet option. I ordered one from them yesterday. I have had the developers guide for a while now.

 

I was wondering if there might not be an updated version for the 2413S?

Maybe I should have spent the $199 for a developers Kit? I had considered that.

 

I had emailed Smarthome to ask if the RF portion of the 2413S could be disabled.

At the start I only want power line comm.

If I can disable the RF section via the serial interface that would be great.

 

Thank you both for your words of encouragement.

[fitzpatri8]

Are you familiar with SmartLabs' Signal Diagnostics features already built into HouseLinc 2? It sounds like it already does what you propose.

 

Using HL2 with its 2413 Dual Band PLM and a laptop computer you can ping individual, select or all your devices repeatedly with standard and extended messages (simply choose the devices you want to test with a check mark and tell HL2 how many times). HL2 will report the number of hops required to accomplish each ping and the number of signals to a particular device that are not acknowledged.

 

Since you can easily select which devices participate in a test, you can follow devices sequentially in a circuit to identify stretches of circuits where signals break down. With the extended ping test you can experiment in real-time, turning suspected devices on or off or adding a filter or unplugging a suspected noise emitter or removing a mesh repeater to measure if it improves or degrades communications.

[ELA]

thanks fitzpatri8,

I am familiar with HL2 only from posts I have read about it.

I can see that it allows more flexibility than the ISY but still not quite what I am looking for.

 

I do not want to be tied to a laptop.

 

I will initally use a terminal program(on laptop) with the 2412S and work my way towards a microcontroller embedded with the PLM for maximum portability.

[IndyMike]

Hi ELA,

 

You are definitely diving into the deep end of the Insteon pool. I also hate the trial and error methods that are required when an Insteon system is encountering problems. This is the reason why I am such a fan of passive couplers vs dualband/Accesspoints.

 

With a passive coupler installed, I know what path the signal must take and I can further test that path using standard X10 troubleshooting tools. It strikes me that an Insteon command with a hop count of zero is the functional equivalent of an X10 command:

 

1) No devices will repeat this command

2) I'm thinking that accesspoints/dual band devices will not communicate to the opposite phase (you'll need a passive coupler to transfer across phases).

 

Obviously, if you push this to a 1 hop command, repeaters and accesspoints will come into play. It will also become much more difficult to determine what you are looking at with an Oscilloscope.

 

I am honestly not trying to talk you out of anything here. Just trying to save you some time, effort and expense. I am very much looking forward to your test measurements.

 

It is my hope to repeatedly send a "dummy extended length message" out over the network with "0" hops. I am thinking this will allow me to better view signal strengths with the oscilloscope. I further envision the ability to then plug this "Test PLM" into any outlet for diagnostic purposes.

 

Depending upon how well this first step works I may then decide to expand its capabilies further.

 

Recognizing I am still a newbie to Insteon communications ... I welcome any comments from advanced troubleshooters.

[Goose66]

I have one of these:

 

http://www.smarthome.com/4814/X10-Power ... PSA/p.aspx

 

although I paid like $80 for it, not $300. Is simple to use, provides signal strenth, S/N Ratio, and shows X10 commands received and logs them.

 

Where is this device for Insteon? Could a circuit be designed easily?

[fitzpatri8]

I have one of these:

 

http://www.smarthome.com/4814/X10-Power ... PSA/p.aspx

 

although I paid like $80 for it, not $300. Is simple to use, provides signal strenth, S/N Ratio, and shows X10 commands received and logs them.

 

Where is this device for Insteon? Could a circuit be designed easily?

 

Insteon signals aren't as simple as x10, so if you want to dive into signal quality you need something more sophisticated. SmartLabs' answer to that is a laptop PC and HouseLinc 2 kit. The Signal Diagnostics features allow you to communicate with any individual or group of linked devices, run continuous ping tests for diagnosing intermittent issues, log how many hops each message requires, and test both standard and extended length messages.

[ELA]

Hi ELA,

It strikes me that an Insteon command with a hop count of zero is the functional equivalent of an X10 command:

 

1) No devices will repeat this command

2) I'm thinking that accesspoints/dual band devices will not communicate to the opposite phase (you'll need a passive coupler to transfer across phases).

 

Exactly that was the intention. I want to separate the RF from the line comm as a diagnostic tool.

I have already coded in capability to control the number of hops allowed as a next step in the troubleshooting

 

 

I am honestly not trying to talk you out of anything here. Just trying to save you some time, effort and expense.

That's good since I have already written a bit of code and my 2412S just arrived today

 

I appreciate the the input as that is why I started the thread.

 

I know the setup I am building will be helpful as long as one owns a scope. I hope to further refine it if possible to a point where is may be useful without a scope.

[Goose66]

I know the setup I am building will be helpful as long as one owns a scope

Honestly, how many people do you think that is?

[Brian H]

Most likely. Not many have a scope.

I have one and lots of spare PLMs. My guess would be even an older 417 link one would work.

[ELA]

I know the setup I am building will be helpful as long as one owns a scope

Honestly, how many people do you think that is?

 

I think maybe 4

 

As I said the ultimate goal would be to eliminate the need for a scope.

[TJF1960]

 

I think maybe 4

 

 

"4" +1

 

Keep up the good work.

 

Thanks,

Tim

[ELA]

I appreciate those who have offered encouraging words.

I am simply interested in exploring new possibilities for a diagnostic tool.

 

This from another thread that I wanted to respond to here in this thread:

 

I'm all for increasing the ease of installing and troubleshooting device control, but I still can't see how your project will be any improvement over the signal diagnostics already available in the ISY or HouseLinc 2. Of particular concern is that, if I understand the project correctly, you aren't trying to log data over an extended period of time.

 

Perhaps it would help if you stepped through how you would run a test and fix a communications problem your way as opposed to using the tools already built into HL2 or the ISY.

 

In response I will say that for those who own HouseLinc 2 _good for you. It sounds like a nice diagnostic tool and I certainly would not mind owning a copy.

 

I am not interested in competing with nor emulating HouseLinc 2.

 

I feel no need to log data over a long time. If you do and Houselinc 2 does that then that sounds great.

 

Am I wrong to assume that others are interested in addition tools that might help diagnose Insteon communications issues? That is not the impression I got when I read tons of posts here when I first joined.

 

fitzpatri8,

I am an electrical engineer. I design things. I enjoy a challenge. Insteon presents a formidable challenge when attempting to diagnose communications issues.

As an engineer with over 35 years experience I have learned to perform testing and prototyping of a proposed design before I commit to definite capabilities.

 

I will run testing to learn more about Insteon communications and then evolve what I think is a reasonable set of functions.

 

or,

 

At some point I may get too busy or loose interest.

Right now my Insteon install is performing well. I feel that is in part - by luck.

 

 

I would be happy to provide more details once I have had time to test and prove them possible (within my capabilities).

 

 

and finally ...

Whatever I do or do not come up with will not replace any other available tool, it is intended to supplement and hopefully fill in for missing capabilities in the existing tool set.

[Michel Kohanim]

Hi ELA,

 

This is quite encouraging. And if we can communicate with your tool (whenever available), we would immediatley add support for it in ISY.

 

With kind regards,

Michel

[IndyMike]

Hello all,

 

This started out to be a message to ELA, but it has expanded somewhat into a general post (lack of focus on my part).

 

I had some spare time over the weekend and decided to play with my old 2412s and the Oscilloscope. I'm not quite as adventurous as you, so rather than developing my own software interface - I cheated. I dug out an old copy of Busy_Rat's PLMTest (serial command line interface). This software is still freely available on Busy_Rat's website.

 

I focused on both Device_direct and Group_broadcast commands/responses and varied the Hop count of the commands. As usual, I learned a few things in the process (this would have been a waste had I not learned something).

 

Keep in mind that I do not use Accesspoints/Dual band devices for coupling. I have only one 4816H passive coupler at the main panel (one accesspoint of RF remote reception).

 

Device_direct commands:

 

Device direct commands are handy as they do not require you to link anything. If you have a valid device address, you can send a message.

 

I focused on an ApplianceLink at address 13 07 33. The standard message format (0X62) is shown below.

 

 

CASE 1: Direct Message with 3 hops/3 hops remaining

 

Command string: 02 62 13 07 33 0F 13 00

Address 13 07 33

Flag 0F (upper nibble 0 - direct command, lower nibble F - 3 hops remaining, 3 hop transmission.

Command1 13 (off)

Command2 00 (level?)

 

The blue trace in the below shows the 60 Hz AC powerline.

The black trace is the Insteon communication. It is constructed as follows:

Standard message : 5 packets of information (transmitted at five AC zero crossings).

Transmit length: 5 packet initial transmission + 3 hop transmissions (20 packets total separated by one zero crossing "gap").

Response length: 5 packet initial transmission + 3 hop transmissions (20 packets total separated by one zero crossing "gap").

Total transmit/response length : 40 packets transmitted over 47 AC zero crossings

 

 

CASE 2: Direct Message with 0 hops/0 hops remaining

 

Command string: 02 62 13 07 33 00 13 00

Address 13 07 33

Flag 00 (upper nibble 0 - direct command, lower nibble 0 - 0 hops remaining, 0 hop transmission.

Command1 13 (off)

Command2 00 (level?)

 

 

As before, the black trace is the Insteon communication. It is constructed as follows:

Standard message : 5 packets of information (transmitted at five AC zero crossings).

Transmit length: 5 packet initial transmission + 0 hop transmissions (5 packets total separated by one zero crossing "gap").

Response length: 5 packet initial transmission + 0 hop transmissions (5 packets total separated by one zero crossing "gap").

Total transmit/response length : 10 packets transmitted over 11 AC zero crossings

 

I found this 0 hop command to be very informative. Using this command I could interrogate every device in my home and view the response levels. All of my devices responded with zero hops. Some were very low level (distant devices on the opposite phase) but all of them responded.

 

 

CASE 3: Direct Message with 0 hops/0 hops remaining - NO RECEIVER UNINSTALLED

 

Command string: 02 62 00 00 00 00 13 00

Address 13 07 33

Flag 00 (upper nibble 0 - direct command, lower nibble 0 - 0 hops remaining, 0 hop transmission.

Command1 13 (off)

Command2 00 (level?)

 

 

This case demonstrates how the PLM behaves when a device does not respond to a direct command. As shown in the following figure, the PLM does not see a response to it's command. It then performs 3 retries of the command, bumping the hop count by 1 with each re-try.

 

 

CASE 4: Direct Message with 0 hops/0 hops remaining - distant device

 

Command string: 02 62 0a db 0b 00 13 00

Address 0A DB 0B

Flag 00 (upper nibble 0 - direct command, lower nibble 0 - 0 hops remaining, 0 hop transmission.

Command1 13 (off)

Command2 00 (level?)

 

For all intents and purposes, this is the same as Case 2 above. The difference here is that this is arguably the worst responding device in my home - my master bedroom. This isn't the most distant device from the PLM. Based on the voltage level of the response (~ 12 mv), it does appear to be on a heavily loaded circuit. I came across this entirely by accident while I was performing my 0 hop testing. I have never had a problem with devices in my master bedroom responding (don't typically use only one hop). Nonetheless, I was impressed that the PLM was able to dig the response transmission out of the noise.

 

Note that, had the device not responded, the PLM would have re-tried the communication with an additional Hop count.

 

 

Of the above, I've found the direct command with a Hop count of zero to be the most useful. I feel that this could be used within the ISY as a measurement of communication "robustness" to a single device. If the device doesn't respond on the 0 hop transmission, the PLM will bump the hop count and re-try. All of this should be visible within the ISY event monitor - it would give a user a qualitative measurement of the communication link. Since the communication is direct, programming issues should not enter into the test.

 

I have additional tests (similar) using group broadcast commands. These are a bit trickier in that the test PLM must be linked to the target device. In playing around, I found that I could link my 2412s to my house PLM as a 5 button Controlinc. This gave me 5 scenes (constructed using the ISY) that I could then exercise using the Busy_Rat tool.

 

The data, however, will have to wait. This post has gone on long enough (did I mention that I'm extremely slow?) and I have a significant other complaining about a non-functional vacuum.

 

Happy automating,

IM

[ELA]

IM,

I have done a lot of development and testing since I started this post.

I have a 2412 run by an external microcontroller. It sends adjustable hop count "broadcast" messages to an unused address. The default being zero hops.

 

As you noticed using zero hops is very informative and I feel a good way to evaluate your system exclusive of other repeats and/or hops that result in too much traffic to easily view.

 

The device I am developing seems to grow in complexity each day I have time to work on it. I am already finding it a very valuable tool for anyone with a scope. I am working to see if I cannot make it usable without a scope as well.

 

I also have spent a lot of time evaluating Insteon loads and hope to have a good way of evaluating and identifying possible signal suckers via signal strength measurements.

 

Sorry no pictures as I am too lazy to upload them right now.

[ELA]

I decided I had time to upload one scope trace below:

The Yellow Channel shows the original send and a single hop (where other devices contribute to build the signal level).

This is a extended -broadcast message to an unused address. Thus there is no acknowledgment or retries.

 

Trace #2 can be ignored for now. It is something I am working on adding.

 

I find it very interesting how the signal levels over the 11 zero cross transmits are fairly consistent from the originator. Yet in the first hop they vary greatly with many contributors.

[IndyMike]

ELA,

 

Thank you so much for the post on your progress. I was hoping to save you some time, but it looks like you've progressed far past the basics or the protocol.

 

I am (of course) extremely curious about your second trace. Looks as if you're developing a signal quality measurement (via PLL?).

 

Beyond that, I came across an "oddity" while performing group broadcasts with my 2412S. I was playing with hop counts and could not make sense out of what I was seeing on the scope.

 

I brought up the ISY interface event monitor on my other PLM and found that my 2412S was intermittently retransmitting with additional hop counts. Since there is no response to a group broadcast, this was not a re-try. Possibly a bug in the firmware of this device.

 

I'm posting because this can really mess with your head when you send a 1 hop command and receive multiple transmissions on the scope.

 

 

Event Monitor

04:13:38 PM : [standard-Group][0C.A8.B4-->Group=2] Max Hops=2, Hops Left=1 (correct transmission)

04:13:38 PM : [ C A8 B4 2] DOF 0

04:13:38 PM : [ F 82 A9 1] ST 0

04:13:38 PM : [ C A8 B4 2] ST 0

04:13:38 PM : [ 0 37 AD 1] ST 0

04:13:38 PM : [ 3 45 6B 1] ST 0

04:13:38 PM : [ 3 61 45 1] ST 0

04:13:52 PM : [iNST-SRX ] 02 50 0C.A8.B4 00.00.02 C1 13 00 LTOFFRR(00)

04:13:52 PM : [standard-Group][0C.A8.B4-->Group=2] Max Hops=1, Hops Left=0 (Correct Transmission)

04:13:52 PM : [ C A8 B4 2] DOF 0

04:13:52 PM : [iNST-SRX ] 02 50 0C.A8.B4 00.00.02 C6 13 00 LTOFFRR(00)

04:13:52 PM : [standard-Group][0C.A8.B4-->Group=2] Max Hops=2, Hops Left=1 (added transmission +1 HOP)

04:13:52 PM : [ C A8 B4 2] DOF 0

04:13:53 PM : [iNST-SRX ] 02 50 0C.A8.B4 00.00.02 CB 13 00 LTOFFRR(00)

04:13:53 PM : [standard-Group][0C.A8.B4-->Group=2] Max Hops=3, Hops Left=2 (added transmission +2 HOP)

04:13:53 PM : [ C A8 B4 2] DOF 0

[ELA]

Hello IM,

 

I believe I have also noticed the irregularity that you experienced, or something similar. I noticed it but did not attempt to understand it. It did not seem normal from anything I have read.

 

It occured when sending a Broadcast msg with 1 hop.

If there are one or more devices on-line with the PLM it works as prescribed.

 

If the PLM is the only device on the network then the PLM sends what appeared to be retries.???

At the time I was busy with other things so I elected to ignore it. It did strike me as very strange though.

 

Yes my second trace is a signal monitor of sorts. (no PLL). I am still experimenting on that. I need more time working with it to see if it will pan out.

 

Last time I worked on the device I added a scope trigger output. This output allows me to trigger anywhere in a message (original, hop1 ,hop2 etc.) This allows me to get better resolution when viewing the hops.

[IndyMike]

Hello IM,

 

I believe I have also noticed the irregularity that you experienced, or something similar. I noticed it but did not attempt to understand it. It did not seem normal from anything I have read.

 

It occured when sending a Broadcast msg with 1 hop.

If there are one or more devices on-line with the PLM it works as prescribed.

 

If the PLM is the only device on the network then the PLM sends what appeared to be retries.???

At the time I was busy with other things so I elected to ignore it. It did strike me as very strange though.

 

You've triggered some dormant memory cells. I encountered this some time ago when first experimenting the the "new" extended message format. Like you, it occurred on an isolated circuit.

 

I'm now beginning to wonder whether the PLM monitors (listens for) hops after the initial transmission. If no hop is "heard", it bumps the hop count and re-tries.

 

I've never seen anything like the above documented anywhere. Certainly nothing like this in the white paper.

[ELA]

I appreciate your level of interest in the details IM,

 

It is nice to be able to share information as I learn more.

At this point I am beginning to wonder if I would have been better off not looking so close

 

Of particular interest to me right now is simulcasting.

I posted an earlier scope trace of 1 hop and wonder how well these devices "sync up" to avoid signal cancellation.

What is the explanation for why the hops have erratic signal strengths?

 

I have looked at the zero cross detection circuit and it does not look very precise. I read what the white paper has to say about that but have yet to do the math myself.

[ELA]

In my last post I talked about how simulcasting syncronization works ?

From the little I understand about this I am thinking they must syncronize using the zero cross signal?

I have reviewed the zero cross circuit based on the 2412 schematic I have access to and this looks totally unprecise enough for that purpose.

 

I am posting a few scope traces of a standard msg with 1 hop and an extended msg with 1 hop.

Each comparision starts with just the PLM and a switchlink (to get a hop to work). I then added a Appliancelinc and then a Lamplink to "build" the signal strength via multiple devices simulcasting.

This is done a small isolated network. I would have expected more consistent results in terms of the devices adding to signal strength.

They start out strong and then what happens?

 

Why do you think the voltage levels vary so much when the additional devices are added? Are they not syncing up properly?

I am curious what others think causes the signal strength variation within from one burst to the next?

 

 

The yellow traces are the line voltage. You can ignore the blue traces once again. They are an indication of signal strength within the PLM itself and I cannot quantify this just yet. You can see that those levels directly reflect the amplitude changes seen on the line.

 

 

 

With the shorter standard msg. the signal strength does not vary so greatly but the extended msg. strengths vary a lot.

Whats up with that???

 

The Appliancelinc seemed to have the most detrimental effect. I do not have a lot of spare modules to test with in order to see if it is one particular device or a global issue?