Getting Serious - When Comm issues strike

[ELA]

After my previous post results I began to question my test environment. Today I retested in the very same test environment using an "outdoor" applianceLinc in place of the indoor I used previously.

The results are below. Note how much better the Outdoor device behaved!!!

Now I am really question Simulcasting reliability? Would Insteon explain this as a quality control issue or is there another explanation?

Am I missing something? This concerns me greatly. Does a person need to test each new device prior to installation?

 

 

I started taking a closer look in an attempt to create a diagnostic device to assist in troubleshooting communications issues.

I seem to be getting distracted along the way.

 

I am not liking what I am finding once a person takes a closer look.

[IndyMike]

Hello ELA,

 

My apologies for the late reply. I'm currently on the road conducting development EMI testing on one of our products. It's not going particularly well...

 

Let me start by saying that you are once again ahead of me on this topic. I have not witnessed the apparent signal cancellation that you are showing, nor have I tried extended messages with multiple receivers in close proximity.

 

What I had tried previously was just the opposite. I had widely spaced I2 transceivers where I monitored the hopping/signal convolution over long distances. This appeared to work reasonably well.

 

I will agree that a zero crossing detection scheme by itself looks to be insufficient to synchronize units for "simulcasting". I had tried to do an analysis on the RF simulcasting used by Accesspoints. The numbers simply do not work at 933 Mhz. There has to be something more than simple zero cross detection going on to allow units to broadcast simultaneously on the powerline and via RF.

 

As you indicated, there is not much in the way of hardware for zero cross detection. Software (firmware) may be another matter. The Insteon unit purportedly use a Costas loop version PLL. Software is not my strong suit, but I'm thinking that this could allow SL to improve the precision of the simulcasting (and a lot of improvement would be required for simulcasted RF).

 

Could your provide the revision levels of the Applicancelincs that you tested (indoor and outdoor). The more recent units ( V5.0+) have drastically revised power supplies. It's possible they have other "enhancements" as well.

 

IM

[ELA]

Hi IM,

Thanks for your thoughts. I appreciate the input as I examine things and wonder what explains what I am seeing.

 

I have ignored the RF synchronization for now. My initial bench tests were only with power line devices.

If there is something that helps synchronize the power line signals, other than zero cross I would like to learn more.

I am not familiar with PLL types of software that would help the synchronization. I will look for info on the Costas loop version PLL.

 

I have observed this erratic signal level (from one zero cross data burst to another ) in both a small isolated network as well as on my home network.

When I trigger on individual bursts I have tried to identify 131Khz sinewave distortions (due to possible phase shifts). I have not seen this, other than at the 1/0 bit transitions.

So I am not sure this can be categorized as signal cancellations. I am just not sure what explains it? Maybe over time it will become more clear.

 

The outdoor (#2456S3E)ApplianceLinc is a Rev 2.0 and my Indoor (#2456S3) Appliancelinc is a V4.2.

 

Side note: Very interesting to hear of your EMI testing.

For several years I was involved in EMC pre-compliance testing for CE marking of equipment.

Ferrite is your Friend

[IndyMike]

Hello ELA,

 

I have also tried to view phase shifts of repeated signals. My Scope simply isn't up to the task. I can understand phase shifts over long wire lengths. Variation over short distances is a little perplexing.

 

I was hoping that you were going to tell me that your "outdoor" appliancelinc was a V5.X version device. As we've discussed elsewhere, SH has completely re-tooled the power supply in these units. I do not know what to make of your V2.0 outdoor unit. Could you provide a date code?

 

I'm beginning to wonder whether the transmitter power supply of the older units simply isn't up to driving the line at full level with extended messages. From you scope plots, the indoor V4.2 unit was already trailing off after the 5 burst standard message.

 

I have a number of the older V4.2 Appliancelincs. When (if) I get home, I try these with my newer V5.X units and see how they play.

 

On the EMI issue - I'll agree that ferrite(s) are allies (we have many). This is a composite aircraft application and the power quality/lightning levels are extreme. Lightning damage levels are at 1500V/1500A on all connector pins. We have a "mature" design that has been updated, but it is failing miserably in this new environment. This is my third pass with this unit. Based on today's results, I'll be planning another trip.

 

It's a good thing I like challenges,

 

IM

[Brian H]

IndyMike; You have ApplianceLincs that are in the revision level of 5.x?

I recently got a ApplianceLinc and it is a 4.2.

Yes the one where they changed the resume the state I was in {On or OFF] after a power loss. To default alway Off after a power loss.

[IndyMike]

Hi Brian,

 

I do not have Appliancelincs above V4.2. The V5.X units that I have are different modules (KPL, SWL, etc). Sorry if my post mislead you.

 

I have been waiting for a V5.X Appliancelinc so I can open it up and get a good look at the changes.

 

 

IndyMike; You have ApplianceLincs that are in the revision level of 5.x?

I recently got a ApplianceLinc and it is a 4.2.

Yes the one where they changed the resume the state I was in {On or OFF] after a power loss. To default alway Off after a power loss.

[Brian H]

No problem on the ApplianceLinc version. I have not seen any relay types mentioned as updated yet.

 

I have seen the reports on the latest models being much better power supply wise.

 

May have been your photo of a new SwitchLinc with the efficient power supply IC in it. Interesting to say the least.

 

With them now into the Smart Grid Arena. Guess efficient modules would be needed. Would be counterproductive if your own modules where inefficient.

[ELA]

Hello ELA,

 

 

I'm beginning to wonder whether the transmitter power supply of the older units simply isn't up to driving the line at full level with extended messages. From you scope plots, the indoor V4.2 unit was already trailing off after the 5 burst standard message.

 

IM

 

IM,

Good point on the power supply possibly trailing off. I am not anxious to open it up to test but I may at some point.

I have two new Icon Relay modules coming soon and I will test them against the appliancelinc for comparision in the same test environment.

 

Seems like once you start looking closer at this stuff that you can constantly be redirected

 

My prototype for a signal strength analyzer is working quite well thus far. I need to make a next version to clean up all the hanging wires etc. and retest. There are still lots of firmware features I still want to add.

 

I think it will be very useful at identifying signal suckers and weak signal strength areas.

 

 

Even though my home network is currently running with a high reliability ...

I knew the network still had some signal suckers and I was waiting to create a tool to better help identify/quantify them before I addressed them. Using this tool I identified four other areas that really should have filterlincs (in addition to the (3) already installed).

 

One interesting sucker I would like to mention was an Ionic Air Cleaner. I knew it was a sucker because it had been "arcing" a few months back. Not electrostatic discharge but was making power line arcing sounds. I found a bad solder joint on the "across the line" cap inside. This is a 0.68uf cap.!!! Talk about a signal sucker

 

The diagnostic tool simply quantifies it since I was already aware of it.

 

That particular sucker was not immediately on my radar because I do not have any Insteon devices in that room. A sucker that severe is still "felt" for a distance across the network though.

 

I sure hate the idea of having so many filterlincs but such is the nature of these power line communicators.

[Brian H]

Yes I have seen some across the line caps in some CLFs also.

 

On the ApplianceLinc power supplies.

There are at least two version ApplianceLincs. I know of.

The real old 1.x are completely different than the 4.x ones.

Like 120 volt AC relay in the originals. Redesigned for 24 volt DC one in the 4.x's

[ELA]

Hello ELA,

 

 

I'm beginning to wonder whether the transmitter power supply of the older units simply isn't up to driving the line at full level with extended messages. From you scope plots, the indoor V4.2 unit was already trailing off after the 5 burst standard message.

 

IM

 

IM,

Good point on the power supply possibly trailing off. I am not anxious to open it up to test but I may at some point.

I have two new Icon Relay modules coming soon and I will test them against the appliancelinc for comparision in the same test environment.

 

 

Both new Icon relays (v4.5) behaved the same as the Indoor ApplianceLinc.

 

I do not think it is the power supply "trailing off" or current limiting.

 

I did a number of tests and found that when the test network was sufficiently isolated that the appliancelincs "repeat or hop" signal strength was consistent.

On an isolated network with ,only a PLM and the appliancelinc, the appliancelinc was driving the line at >3.5Vp-p and very steady.

As soon as a second repeater is added (Switchlinc) now the hop signals became erratic.

Its as if there is some signal cancellation for some reason??

 

As part of this testing I have been able to create a very isolated network using a large choke in series and along with a filterlinc. In this test setup the appliancelinc puts out greater than 5V p-p when it is the only repeater to the PLMs test msg. So the ApplianceLincs transmitter is very strong.

 

In this "very isolated" test setup the signal levels are actually getting too strong . The signals start to get very strange (totally out of sync). I believe that at these levels the zero cross detect circuit is becoming corrupted by the large amplitude 131Khz signal forward biasing 5.1V zeners at the wrong time.

[ELA]

I have been collecting a lot of data on signal strengths on both my home network and "isolated" networks.

As in my last post I seem to finding that there is such a thing as TOO MUCH SIGNAL STRENGTH.

 

Has anyone ever heard of this possibility with Insteon devices?

 

If you look at the 2412 users guide schematic of a PLM you can see 5.1V zeners used in the zero cross circuit. It seems to me that a 131Khz signal greater than the zener voltage could adversely affect devices zero cross detection and result in erratic signal strengths.

 

It would be very cool if anyone could confirm this by there own testing or reports from Insteon??

 

I am guessing that it is not normally an issue in most installs.

I am seeing it as an issue in isolated networks.

 

Could it be an issue in networks as the number of devices grows (adding to repeat levels) or when several devices are located very close to each other in the network?

 

This symptom has distracted me from what I wanted to work on and so I want to move on. Just finding this very interesting. Any thoughts, reports, test results would be appreciated.

[Rob]

ELA and IM,

 

I enjoy reading this thread even though some of it is over my head. I think it's great there are folks like you who take the time and effort to dig into the guts of the technology.

 

So ELA, what plans do you have for your diagnostic device you've been working on? Would it be usable by someone w/o an EE degree?

 

thanks

rob

[ELA]

Hi Rob,

For the present my plan was to develope a diagnostic device that I could use myself.

One reason for undertaking the task was to learn more about Insteon communications in the process.

 

The plan is to make it such that it will not require a Oscope and that a non EE could use it.

 

When I am finished, and if it proves to be useful, I would like to make it available to others if reasonably possible.

[ELA]

When I started a project to develop an Insteon diagnostic tool I wanted to use the 2412S because I was not interested in using the RF section of these PLM devices. I also desired to use the +12V available to power my add-on PCA.

 

Most of my initial development was done using the simplehome.net # 5010K (2412S equivalent).

 

During the development I figured I had better address the fact that the 2412S might not be available someday and so I purchased a 2413U to see if I could also make my PCA work with it.

I needed to add an external power supply since +12V was not available from this unit.

 

What I did not expect to find was that the powerline transmitter section of the 2412S was very weak when compared to the 2413U.

 

I also tested my home networks 2413S for comparison.

 

Both of the 2413 units were nearly (2x) the output of the 2412. (Keep in mind we are not talking about RF here, only powerline transmissions).

 

I only have a sample size of (1) for the 2412S so I cannot say this is always true.

 

I just wanted to put it out there for anybody wanting to compare the two. If this sample test is typical then I would shy away from the 2412 in favor of the 2413 just in terms of powerline signal strength. Of course the addition of the dual band(RF) is also a plus.

 

I do wish the 2413's had a +12V output though as I hated having to add an external power supply for my add-on PCA.

 

I am just now moving from the breadboard stage to the PCA stage in order to better investigate use of the 2413 in place of the 2412.

[Brian H]

ELA; Thanks for the update.

 

Since the 2412S/U and the 2413S/U use completely different power supplies and circuitry. I can see why one may have a higher power line signal than the other.

 

Since the 2413S/U have a power line derived power supply. That could be why there is no raw +12 volts on the interface connector. No isolation from the power line.

The photos in the FCC database are actually clear enough to see the LNK354GN power supply IC.

[ELA]

Hi Brian,

I appreciate the input.

I have taken my 2413U apart and was aware the supplies were different and of the:

 

http://www.powerint.com/sites/default/files/product-docs/lnk353_354.pdf

 

used. If you look at the data sheet there is no reason that this cannot be an isolated supply. I sure hope that it is, otherwise I will have more work to do as I reference some signals on the PLM- PCA to my PCA.

 

I suspect they no longer provide external +12V so as not to have to size the larger components and the need to protect against "user shorts".

 

I will take some additional care when I get to the point of taking some measurements. I hope not to have to trace the entire power supply circuit.

It sure was nice having access to the 2412 documentation

[Brian H]

Well if you care to experiment. I believe the daughter card for the 2413S has an unoccupied Fuse position.

 

I see from the ICs data. It could be isolated.

You have a point. It may not support lots of extra power through the interface connector.

[ELA]

I did trace out the 2413U circuit and the LNK354 produces a +20V output. This output drives the transmit circuit and is also brought to pin 1 of the interface header (where the raw +12V on the 2412 is).

 

This +20V is regulated down to +5V on both the main PCB and on the USB interface daughterbrd as well.

 

So +20V is available but I am unsure yet if I will tap it for my circuit.

 

It is isolated from the line voltage.

 

I have provided for either the ability to tap a supply from the 2412/2413 main board (pin 1) or to use an external wall wart supply for my circuit. I did this because I am unsure whether or not I will include an LCD as part of the design. If I do then the wall wart may be a must (for backlighting the LCD).

[Brian H]

Thanks for the update on your findings.

[Illusion]

If it is not too difficult, I think it would be nice if the diagnostic PLM was RF capable, but with a switch to be able to disable the RF component of the device. So testing could be accomplished with and without RF. Also, there are several RF only devices.

[ELA]

Thanks for your input Illusion,

 

I agree and I have looked into that as an option. From what I understand there are no PLM commands that allow you to disable the RF section alone?

I may look into adding a switch to interrupt the line to the RF transmitter but that would be an option a ways down the line

 

I have put way too many hours into where I am currently at but pleased with the results thus far.

It is still a bit early for me to provide much more detail as I am doing some pretty extensive testing to determine its value in various troubleshooting scenarios.

 

Here is a picture of my second prototype. I just ordered another PLM to begin the third.

 

This second unit required an external power supply to power the LCD since it is based on a 2413U with no "user" power brought out like the 2412 used in my first proto.

From my investigations of the 2413U power supply it appears to be powerful enough to tap off of to feed the LCD.

The next prototype will eliminate the external supply.

However I intend to disable the LCD backlight when transmitting so it will not reduce available transmit power.

 

I am discovering all kinds of interesting things when observing my home network that each require more investigations and then tweaking of the test units firmware.

One interesting aspect is GFCI breakers sensitivity to the 131Khz signal. I have tripped one GFCI in my testing and am looking more closely at that.

My default test mode utilizes a standard message and that tripped a GFCI (when plugged directly into the GFCI outlet).

 

I suspect that a GFCI may be more likely to trip when located close the the service (this one was directly below the service). This due to the large distributed downline capacitance to ground. I am planning more testing but using extended messages as I expect the trips to be more repeatable then?

[ELA]

I have been attempting to use this device to be proactive in learning about the "communication friendliness" of areas in my home that as of yet do not contain Insteon devices. I am planning on adding more and did not want to use a trial and error method as I found that frustrating. Not saying it is not a good method, especially if you do not have another method available to you.

 

I used this device to characterize the tendency for each outlet to "suck signal" during transmit, as well as how strong the received signal was.

 

My initial readings on all bedroom outlets indicated that there were large signal suckers present in the room. I had previously identified one of them as being an AIR CLEANER. The way I knew this is that it had failed to operate normally and when I took it apart to fix it I found a 0.68uf across the line capacitor inside. It had bad solder joints which I repaired. So I knew this would be an issue in advance.

 

I am trying not to be too long winded so if I left something out I will follow up in a later post.

 

With this device I read the amount of "signal suck" during the transmit along with the signal level of the received response.

Rather than provide absolute data I will provide the "signal suck data" in terms of "1 standard Insteon load". In other words this outlet acts as if it has "this number of switchlincs" attached directly at the outlet. Since I have not yet determined a weighting scale for absolute data I feel this is the best way to communicate the data for now. The received levels are stated in a few distinct levels for now also because I have not yet established a

weighting scale.

 

Please note that I am not attempting to "detect" how many true Insteon loads are present. In this case there are none locally. I am characterizing how hard the transmitter section needs to work to produce a signal at this location. Also how much signal gets returned to this location.

 

Here is the initial data from the bedroom:

Outlet A

Acts like 4 Insteon loads, received signal strength is "very low". (very low being less than ~100mv)

 

Outlet B

Acts like 6+ Insteon loads, received signal strength is "very low". (very low being less than ~100mv)

 

Outlet C

Acts like 5+ Insteon loads, received signal strength is "very low". (very low being less than ~100mv)

 

Other outlets read similarly I will just cover these as these are where the signal suckers were located.

I then began unplugging devices in the room while monitoring the "suck level" to identify devices that caused a larger change in the suck level.

 

This lead to identification of four devices that were acting as signal suckers in the room.

1) A Satellite receiver and its dish power supply located at Outlet A

2) The Air Cleaner previous known at Outlet B

3) An LCD TV located at Outlet C.

 

I agree that these could have been identified by another method. What this does though is to gauge them as to to what level they suck signal.

 

once each offender was identified I isolated it from the home network using a Filterlinc and that allows me to measure the relative suck level of each device itself (without the combined influence of all devices and wiring in a more complicated network).

 

From this I got these results:

1) A Satellite receiver and its dish power supply = 3.5 Insteon Loads.

2) Air Cleaner = 3.7 Insteon Loads.

3) LCD TV = 3.5 Insteon Loads.

 

As a comparison a cheap Surge Suppressor plugstrip w/EMI filter(containing 0.01uf cap) "looks like" = 1/3 Insteon Load

A slightly better grade Surge Suppressor w/EMI filter (containing a 0.1uf cap) "looks like" = 1 Insteon Load.

 

Now knowing the offenders I Unplugged them all and monitored the same three outlets with these results:

Outlet A

Acts like Insteon "0.01" loads, received signal strength is "low". ( very usable but no repeaters in the immediate area)

 

Outlet B

Acts like "zero" Insteon loads, received signal strength is "low". ( very usable but no repeaters in the immediate area)

 

Outlet C

Acts like "zero" Insteon loads, received signal strength is "low". ( very usable but no repeaters in the immediate area)

 

 

So here is where I am still deciding whether or not I need to Filter all three loads or if adding devices will be reliable enough without filters on some of the loads.

 

I now used my tester to repeatedly send on/off messages every 4 seconds to a "diagnostic Switchlinc in a box with a plug" as the receiver that I can move around. The tester is located at an outlet near where my home network PLM is located (but is free of a pc connection so can also be located anywhere).

I run 100 send/receives and log the percentage that make it though.

 

Here are those results:

1) Test conditions: Standard message , RF at PLM turned off , 1 Hop

TV is plugged-in , Air Cleaner plugged-in, Receiver is plugged-in

Test Receiver at Outlet B (middle of room)

0% of the messages were received

 

2) Test conditions: Standard message , RF at PLM turned off , 1 Hop

TV is plugged-in , Air Cleaner UNplugged, Receiver is plugged-in

Test Receiver at Outlet B (middle of room)

67% of the messages were received

 

3) Test conditions: Standard message , RF at PLM turned off , 1 Hop

TV is plugged-in , Air Cleaner UNplugged, Receiver UNplugged,

Test Receiver at Outlet B (middle of room)

99% of the messages were received

 

4) 3) Test conditions: Standard message , RF at PLM turned off , 1 Hop

TV is UNplugged,, Air Cleaner UNplugged, Receiver UNplugged,

Test Receiver at Outlet B (middle of room)

100% of the messages were received

 

Just as an interesting side note I re-ran test #2 using 3 hops rather than 1 hop and the percentage was nearly equal at 73%. No big improvement from additional hops when signal suckers are the offenders (as opposed to noise).

 

I purposely used a non dual band Switchlinc in these tests. Next I would substitute a dual band lamplinc and turn the RF back on at my tester to get an idea of whether I should invest in dual band devices for this room or Filter all the offenders. At this point I lean towards filtering to get the most reliable system. In particular some of the devices I want to add are not available in dual band. The thing I like is that I can now make a more intentional design choice by running more tests to confirm which direction to go.

 

I know I left some things out but this post is getting too long. Much more testing can be done.

Lastly I want to qualify that I am just trying to share information. It will be easy to pick at this or that if you like. I am still developing capabilities and learning.

[ELA]

I wanted to follow up with another example of data I have collected.

 

After the tests above I further characterized each outlet in the house. Of particular interest were two outlets that I have directly at the service cabinet (one on each 120V leg). This electrical box originally housed my old X-10 repeater and now has two outlets I kept in its place.

 

As one would expect these both read a high xmit loading value since it "sees" all the circuits that do contain Insteon loads as well as other devices.

They read as follows:

Phase A

Acts like 6+ Insteon loads, received signal strength is "low-medium".

 

Phase B

Acts like 4.5 Insteon loads, received signal strength is "medium-high".

 

This prompted me to look closer at Phase A. By turning off breakers for loads on phase A while monitoring I was able to isolate a particular circuit that exhibited high loading. This circuit did not have any Insteon devices on it at present. It was a dedicated washer circuit and so I was surprised.

Further testing directly at that outlet I found that there was an old Belkin Surgemaster plug strip there that fed a sprinkler system transformer and a Radio shack RF remote switch module. I had forgot that these existed.

 

Now testing them each in an Isolated environment I got these results:

1) Radio Shack Remote = 1.5 Insteon Loads.

2) Belkin Plug Strip itself = 1.5 Insteon Loads.

3) transformer = minimal impact

 

Once again I know that these might have been identified by other means. I am well aware of the issue with some Surge type plugstrips w/EMI filters.

I had forgotten that this one even existed and this is an easy fix. Eliminate the plugstrip and the Radio shack device as they are not required.

 

With these devices unplugged the Phase A reading at the service was now very much inline with Phase B.

Phase A

Acts like 4+ Insteon loads, received signal strength is "medium".

 

One other note is that in the past I found fewer plugstrips that contained an internal EMI filter (capacitor). Now it seems they are much more prevalent and I always look at the specs to be sure they do not claim something like ("-43db attenuation from 150Khz- ??") That is a good indicator of an internal cap that sucks Insteon signal strength.

 

Last week I bought a surge suppressor that I was going to use in my Bedroom. The person at Radio Shack was convinced that it did not have an internal EMI filter ( I wanted MOVs for surge but no caps). The package did not state any attenuation figures.

 

When I got it home and opened the package the internal documentation did state an Attenuation figure.

I tested it on my tester and it represented at "1.2" Insteon Loads.

Of course I returned it.

 

So this is another great feature of being able to test Plug strips. I desire MOV protection in a plug strip but no EMI protection. I do not worry so much about units that only contain a 0.01uf cap but if they are 0.10 uf then I consider then to be undesirable.

[ELA]

A person in another forum felt that the LCD display was too limiting in its ability to display data.

The LCD actually came last after the serial terminal interface was done. The LCD was intended as a simple user interface. An interface that non technical people might be able to use and understand. Someone here had asked earlier if it could be a device a non engineer could use.

 

The LCD displays averaged values for XMIT and RCVD signal levels in an attempt to keep it simple.

 

When connected to a computer via the USB port I have several more technical screens that allow view what is happening at the packet level.

This first screen is just the data for each packet sent and received.

 

Shows original transmit and 2 hops using an extended msg. Note how the values vary from packet to packet. This consistent with what is seen if you were viewing the line with an oscilloscope.

 

This next one is what I would refer to as "Scope Mode" it shows the same information but graphically. In this one there are 3 hops on a standard msg.

Reads from left to right, original xmit and each hop packet. From bottom to top is amplitude. Again this is very similar to what you would see when viewed with an oscilloscope.

 

This screen was limited by the number of characters available (25 vertically). Because of this the resolution is not as good as I would like.

 

In this next screen you see the same data but rotated right to get better horizontal resolution (80) char.

Top to Bottom are the packets of the original transmit and 3 hops. Left to right is amplitude of each packet. Color change is meant to distinguish "rough levels of low, medium and high" that follow values as seen in averaging LCD screen.

 

I like this screen as it is a virtual Oscilloscope ( Although turned on its side).

 

I find it very interesting how packet levels vary from one to the next.

 

Far from perfect and a work in progress.

[ELA]

I have to post some more results. I apologize if this is pretty boring but I am excited at having this new capability to take a closer look at Insteon communications.

Especially after these tests I was able to run today.

 

As a follow up to my previous proposed, Insteon expansion into the master bedroom, I found that signal suckers in that room either had to be filtered or I had to add some RF capability to that room if I wanted to get reliable communications to any new devices.

 

I performed the same test as I did previously that had "zero" percent receipt success, when sending to a Test Switchlinc in that room, from the test PLM in another room, near where my house (ISY99)PLM is located.

Previous Test Results:

1) Test conditions: Standard message , RF at PLM turned OFF , 1 Hop

TV is plugged-in , Air Cleaner plugged-in, Receiver is plugged-in

Test (switchlinc)Receiver at Outlet B (middle of room)

0% of the messages were received

 

The change I made was to substitute a Dual Band Lamplinc in place of the Switchlinc as the receiver.

 

New Results:

99% of the messages were received

 

This showed that adding a dual band device in this room could overcome the local (signal sucker at that outlet). This was expected since the powerline communications would not be required and thus signal suckers did not matter (to that particular dual band device).

 

IN the next test I left the Lamplinc in place (as an RF repeater) and added the test Switchlinc as the receiver (at a different outlet -one with the Satellite Receiver). This to see if the single RF link in that room would drive the power line hard enough to overcome the suckers in the room.

The concept being that if one Dual band device was added to the room that possibly I could use other non dual band devices without adding filtering to the signal suckers.

 

Here are those results:

1) Test conditions: Standard message , RF at PLM turned ON , 1 Hop

TV is plugged-in , Air Cleaner plugged-in, Receiver is plugged-in

Test (switchlinc)Receiver at Outlet A (by Satellite Receiver), Lamplinc at Outlet B (as RF Linked repeater)

83% of the messages were received

 

So this showed me that I could elect a combination of one Dual band device and other additional non dual band devices but that the reliability would not be high enough at the non dual band devices.

I feel that this showed me I had better either do complete filtering and non Dual band devices, or add a dual band device and some filtering, or use all dual band devices (if available).

 

NOW FOR THE KICKER THAT WAS QUITE UNEXPECTED!

 

When re-running the first test above (to test duplicate results) I found that the reliability was going down on this test set. I came to realize that this was the result of my moving about during the test! I was walking between two rooms during the test to observe the receiver and I was interrupting the RF path!!!

The PLM transmitter was about 25ft from the test Lamplinc receiver. When I sat still out of the RF path the success rate was basically 100%.

 

When I walked in the RF path my body was blocking/reflecting the RF sufficiently for messages to fail.

 

What was most interesting is that I could watch the LED on the Lamplinc as it turned on/off with each message. When I then moved my body to a point midway between the two (BUT SLIGHTLY OUT OF THE DIRECT LINE OF SITE BETWEEN THEM) I could interrupt every message with my body! Success rate dropped to Zero!

 

This suggests something else to consider if you hope to boost reliability by adding dual band devices. Perhaps two RF paths to any receiver should be a design criteria (If the install is reliant on RF links to overcome poor powerline conditions)?