Print

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

[Noam]

I found the source of my RF problem and it had nothing to do with the Echelon meter nor the lady upstairs. It was a dead RR501 and it seemed to happen only in the evenings because that's when I usually turn the lights on in my BR.

I have an RR501 and a mini-Timer on that phase and a TM751 on the other phase. The RR501 was working a couple of months ago when I first plugged it in so I never considered it until today when I decided to get to the bottom it.

I suspect it may have been the cause of the few random ONs which were occurring a couple of weeks back but not since.

PS: Where's the sheepish grin icon?

Another one solved.
Okay, so do we still have who has problems they are attributing to their Smart Meter?

I don't see a sheepish grin icon, but the regular grin is:  .

[JeffreyB]

Oooo!  Me!  Me!  Me!!  (Has hand held up...)     

Jeff

[Backward Engineering]

I found the source of my RF problem and it had nothing to do with the Echelon meter nor the lady upstairs. It was a dead an intermittent RR501 and it seemed to happen only in the evenings because that's when I usually turn the lights on in my BR. Embarrassed[/quote

Lets give him one "ata boy" and a +1 from me. ]

[Noam]

Oooo!  Me!  Me!  Me!!  (Has hand held up...)     

Jeff

Anyone BESIDES you? (just kidding )
It is hard for me to keep track of who has which problems, and which ones were solved.
I'm sure our resident "geniuses" are still trying to help with this one.

[dhouston]

There are two in the Cincinnati area w/Echelon meters and 2-3 others where the reports are indirect (email to Jeff Volp) and with different meters (IIRC).

And I'm not dismissing any of them - it's just extremely difficult to identify anything that all have in common. So far, I cannot find anything that definitively explains anything the various smart meters all have that might affect X-10 and UPB but not Insteon.

My favorite candidates are spikes and/or brownouts (i.e. momentary voltage dips) which are a known issue for X-10. Spikes could be a problem for UPB but I've never heard of them being a problem for Insteon. Some or all of the meters output pulses meant for third-party energy monitoring devices but how these are output differs - flashing LED, serial output, ZigBee, pulsed relay... None appear to directly put pulses on the power-line but some of the relays are connected to Neutral which might cause spikes/dips when the relay coil de-energizes (inductive kickback). Since the frequency of the pulses is tied to total energy use, that could account for the apparent randomness.

And, the fact that some of the victims are long-time X-10 users might also be meaningful as some of the vulnerabilities may have been addressed in later designs (it was just sloppy design).

I believe Jeff favors a different candidate.

Some meters send data over the powerline to an aggregator at the transformer. This might be leaking into the household wiring but, again, the details/mechanism are a mystery.

[Noam]

Oh, I wasn't trying to discount anyone's claims that their SmartMeters were causing problems. However, we have seen a few reports from users of OTHER issues they found (perhaps caused by the change to the SmartMeter, perhaps a coincidence), that the user was able to track down and resolve. Perhaps the SmartMeter (or the process of shutting off the power, and installing the new meter) is causing something in the home to act crazy, and produce noise or something along those lines.
The meter itself might still have been the catalyst, but that doesn't mean that it is time to give up trying to get the system working again.

[JeffVolp]

And I'm not dismissing any of them - it's just extremely difficult to identify anything that all have in common. So far, I cannot find anything that definitively explains anything the various smart meters all have that might affect X-10 and UPB but not Insteon.

My guess at this time is communition back to the "data aggregator" using Lon Works.  While developing the XTB-ANR I measured as little as 30mVpp at 120KHz injected directly onto the powerline as a CW signal blocked a nearby X10 appliance module from decoding X10 commands.  Sensitivity degrades as the frequency deviates from the center of the passband, but there is still plenty of sensitivity at 115KHz for that to be a problem.  Like I said earlier, a Maxi Controller mistuned to 130KHz was still working fine, so X10 modules might still be blocked by a strong 132KHz signal.

I am not sure of the details of the Insteon front end, but they may include AGC to raise their detection threshold above the Lon Works signal level.  Or perhaps they have some means to discriminate between the two signals like a FM radio can lock to the stronger of two stations.

I received a message from one of the people having trouble with their smart meter: 

“Mr. ------ says Duke will probably "lower the power" or drop me off the network and read our meter the old fashioned way.  He says they are working on the problem internally.”

I communicated directly with that Duke engineer, and explained how X10 works.  So this may be the first example of a utility company actually recognizing and trying to mitigate the problem.

Jeff

[dhouston]

I communicated directly with that Duke engineer, and explained how X10 works.  So this may be the first example of a utility company actually recognizing and trying to mitigate the problem.

Where is that engineer? I met the regional engineer responsible for Northern Kentucky about 3 years ago and he was very sharp (and helpful) but that was about another issue entirely. This area was a test area for the Echelon meters and, as I've noted all along, I cannot point a finger at the meter for any X10 issues though I suspect it might be related to some non-X10 woes.

In tests a few years back I found most X10 modules (I had no X10 switches) responded to 75-150kHz with some going even higher (I was injecting a fairly strong signal.) so I do not doubt that 132kHz would be a problem. But, I wonder how it leaks into the residence. Echelon would certainly have been aware that it would create X10 problems so would, I think, have blocked it from the residence. Also, any X10 residence would likely cause fits at the aggregator with X10 signals leaking in that direction.

Paul Beam (?) who designed the ESM1 also told me he had seen responses in the 75-150kHz range. I have a pre-Elk version of the ESM1 (Brian Karas was thinking of buying the rights and asked me to look at it.) which reads 10Vpp full-scale so it may not be sensitive enough to see a very low amplitude 132kHz signal but I see nothing on it.

If it is related to 132kHz, I still do not understand why Insteon appears immune at 131.5kHz. Nor does it explain non-Echelon meters where some X10 users also appear to have problems.

[dhouston]

I am not sure of the details of the Insteon front end, but they may include AGC to raise their detection threshold above the Lon Works signal level.  Or perhaps they have some means to discriminate between the two signals like a FM radio can lock to the stronger of two stations.

Actually, Insteon has much greater sensitivity (10mV) than this, but the fact that all of their devices are repeaters does mean they tend to saturate the "network" with a strong signal (3.2 Vpp into 5 Ohms) so that could explain their apparent immunity.

But, UPB uses a totally different signalling method. They discharge a capacitor to create a ~40V pulse with the pulse location in reference to ZC determining logic 0-bit or logic 1-bit.

[dhouston]

Here is a detailed description of LonTalk 110kHz-140kHz operation.

• http://www.echelon.com/support/documentation/Manuals/transceivers/078-0175-01C.pdf

[JeffVolp]

Here is a detailed description of LonTalk 110kHz-140kHz operation.

• http://www.echelon.com/support/documentation/Manuals/transceivers/078-0175-01C.pdf

I have the LonWorks transceiver manual, which covers the same material.

My customer is in Cincinnati, so that may be indeed the same Duke facility.

Jeff

[dhouston]

My customer is in Cincinnati, so that may be indeed the same Duke facility.

But not the same regional engineer.

[dhouston]

Given that the Echelon meter probably transmits to the data aggregator only sporadically, the fact that I see nothing on my ESM1 may not be significant.

My RR5x5 PCB is about to go to the board house. One of the things it can do is output the (optically isolated) 120kHz bursts for viewing on a 'scope. I can program a trigger on another output whenever there's activity and capture it on my 'scope. It should work equally well for the 115kHz and/or 132kHz used by Echelon.

I can probably even transmit a PLC command at the same time to see what happens.

I included this feature thinking it might sometimes aid in troubleshooting - I just did not anticipate it might be useful so quickly.

Hmmm, I wonder if I can make the data aggregator run backwards.  

[dhouston]

The mystery deepens...

I replaced the intermittent RR501 with an RCA version and all was OK for a day. Now the RCA transceiver has stopped responding to RF. I unplugged it to check the outlet and my Kill-A-Watt indicates ~121VAC with no apparent intermittentency. When I plugged the transceiver back in, it started working normally again.

I tried the original RR501 on the other phase. It would not respond to RF while a TM751 in the same outlet did.

At this point I'm still not at all sure this is related in any way to the Echelon meter - I'm not at all sure what is going on.

I do not have time to troubleshoot this right now. I have 14 different prototype PCBs going to the fabricator and need to concentrate on bills-of-material to be sure I have parts on hand to assemble a few when I receive them. Even the minimal physical activity required to troubleshoot the transceiver issue aggravates my spinal cord injury and I need to avoid that right now.

[JeffVolp]

Here are a couple of updates from the fellow who reported problems after Duke Energy installed the Echelon meter:

Before the installation of the new Duke meter, we were experiencing about one X10 switching failure about every two or three weeks, considering all two dozen or so circuits under control.  After the Duke meter installation we began having so many failures every day that many of the circuits were turned off or manually operated.
 
The two active filters from Jeff at JV Digital Engineering have been in service a little more than 24 hours.  All circuits were returned to automatic operation about 18 hours ago.  We have experienced one switching failure since then when a light that should turn on did not.  It will be several days before we will know if the system is back to normal but it is definitely better.  However, looking at the signal and noise levels I am surprised things seem to be working this well.  If the JVDE filters continue to show promise we may make some changes at the distribution panel to make their installation permanent.

And another update:

Here is an update to my lengthy report four hours ago.  In the last four hours we have had two more switching failures - one that turned on when it should not have when another circuit turned on, and another one about 15 minutes later that failed to turn on.

So, it looks like the XTB-ANR is helping, but things are not perfect yet.

Jeff