Print

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

[Brian H]

132KHz. Right on top of Insteons 131.65 KHz. 

[dhouston]

132KHz. Right on top of Insteons 131.65 KHz.  

Duke uses Echelon meters but there's not a single report of smart meters causing problems on Smarthome's Insteon forum. Since Insteon also operates at ZC, it should also be affected if this is the source of the X-10 problems. Everything I've seen indicates that all of the smart meters are using Zigbee (since it's an open standard) which uses DSSS spread spectrum (centered at 915MHz but covering ALL of 902-928MHz) as well as the pulses related to usage I documented earlier. I haven't seen LonWorks mentioned in regards to the meters. Even the article referenced below seems to indicate that it's more of an adjunct for industrial/commercial applications than part of their basic meter. Also, it doesn't explain the reports related to other, non-Echelon meters.

Echelon has been doing this in Europe for years and I've seen no complaints from European X-10 users.

• http://livemetering.blogspot.com/2009/03/echelon-corp-has-released-new-version.html

Unfortunately, it appears the only way to get the documentation for Echelon's powerline protocol is to buy it after signing an NDA.

[Brian H]

Since Smartlabs is into Smart Grid Developement.
They probably have addressed any potential issues.
http://www.insteonsmartgrid.com/

[JeffVolp]

Duke uses Echelon meters but there's not a single report of smart meters causing problems on Smarthome's Insteon forum. Since Insteon also operates at ZC, it should also be affected if this is the source of the X-10 problems. Everything I've seen indicates that all of the smart meters are using Zigbee (since it's an open standard) which uses DSSS spread spectrum (centered at 915MHz but covering ALL of 902-928MHz) as well as the pulses related to usage I documented earlier. I haven't seen LonWorks mentioned in regards to the meters.users.

Here is a datasheet for one of the meters I looked at:  http://www.echelon.com/metering/datasheets/ANSI_2s.pdf

It says: “Every NES smart meter includes Echelon’s proven standards-based power line communication technology – the world’s most widely deployed signaling technology.”

Chasing that path further, the meter communicates with the NES “data concentrator” using LonWorks:

http://www.echelon.com/company/news/articles/2010/2010.04.13_SmartGridToday-2mmMeters.pdf

Then researching LonWorks tuned up the powerline transceivers:

http://www.echelon.com/support/documentation/datashts/15330.pdf

The spec sheet on that says the carrier is at 132KHz (primary) and 115KHz (secondary).

I think the problem is not the communication inside the home, which may indeed be Zigbee, it is the communication with the “data concentrator” at the utility transformer.  That is leaking over into the home wiring.

As you know X10 signals are sent as 1mS long bursts of 120KHz.  Presence of a burst is decoded as a logic “1”, and absence as a logic “0”.  In measurements here I found that as little as 30mVpp of 120KHz injected onto the powerline would prevent an X10 appliance module from decoding X10 commands.  Sensitivity decreases as the frequency deviates from 120KHz, but in other testing X10 modules could still decode commands from a Maxi Controller running at 130KHz.

If there is any significant leakage of those 115KHz or 132KHz frequencies onto the home wiring, it could certainly cause a problem for X10 communication.  Since Insteon does not need the absence of a signal to decode a “0”, it is more immune to this interference, particularly if it is at a relatively low level.

Jeff

[dhouston]

Jeff, thanks for the references. I should have remembered the Echelon chip - I looked at it (probably an earlier version) a long time ago.

I'm still skeptical. They use BPSK which Insteon also uses so it should stomp all over Insteon. However, it also has CSMA/CD which means it should sense either Insteon or X-10 and back off. Plus, it still doesn't explain the non-Echelon meters (unless they've all licensed the Echelon technology/chip or are using the same methodology).

Here's a helpful reference on Cenelec A, B & C-band...

• http://www.domologic.de/download/pdf/konnex_pl132_en.pdf

It seems Cenelec realized the error of their ways with allowing unlimited 433.92MHz RF systems.

If the pulses employed truly are 20-100 milliseconds long, I'm not sure my USB scope will see them as I'll need to connect it thru a 25VAC transformer. If they are of µS durations, they might not pass the transformer, either. And, it's not 100% clear whether these pulses are on the powerline or only on the serial or IR output.

[Backward Engineering]

Today I found another noise source in my house wiring.  The XTBM was showing a burst of noise that lasted about 1 second and occurring every 5-6 seconds.  I built a high pass coupler to use with my Fluke ScopeMeter (schematic shown below).  The noise was a burst at 71Khz and around 500mv peak-peak.  I traced the source to an ACP Back-UPS RS1500.  When I unplugged the UPS the noise went away.  However when I plugged the UPS back in to the 115 vac, the UPS began charging the battery and the noise frequency changed to 132Khz.  The XTBM then reported an Insteon signal.  I don't have Insteon equipment or this might a problem.  In this case I may add a filter inside the UPS.  The final report here is that I can find not communications coming from the new Smart Meter.  With the RV power supply out and the UPS unplugged, the line is extremely clean even with the scope at 10mv/div.  As for the RV power supply problem, it turns out that it goes into spuratic oscillations with a load of less than 1 amp.

Don

[Noam]

I've had Smarthome Filterlincs on my UPSes (well, the Belkin, APC, but NOT the one for the FiOS system) for long time. That was something I saw recommended here on the forums years ago.

[Backward Engineering]

I already have a very large inventory of ferrite cores, wire and capacitors.  I won't even need to order anything if I install the filter inside the UPS, which is where it should have been in the first place.  It doesn't take a very large inductor to filter out this high frequency component. A100uh inductor and 1uf capacitor would provide about 32db attenuation at 120Khz.  This would reduce the 500mv P-P to less than 10mv.  When ever possible, filters should be as close as possible to the switching power supply to reduce induced and radiated noise (EMI).  Those wall clocks that update automatically use VLF radio signals from the atomic clock in Bolder Colorado.   

[JeffVolp]

I just read a post on CocoonTech regarding problems with UPB switches due to the utility company communicating with its smart meters.  So the problem isn't confined to just X10.

Jeff

[Backward Engineering]

It's the "due to" part that concerns me. 
Is it a result of scientific investigation or speculation due to the timing of the problem, as it was in my case?

Don

[JeffVolp]

It sounds like a "do to" to me:

I have two clients who have a HAI system with HAI UPB switches. I'm not sure what revision switches they have either. The problem comes in with the local utility communicating with their meters over the powerline. Whenever this communication takes place, all of the HAI lights with flicker rapidly. The communication system is called Aclara.

Jeff

[dhouston]

Aclara seems to have several systems but the one for smart metering doesn't seem likely to create problems for UPB.

• http://www.aclaratech.com/AclaraRF/Brochures/STAR_ElectricMTU.pdf

The DC pulses described in earlier posts might create problems for UPB which uses high amplitude pulses in their pulse position encoding scheme.

[JeffVolp]

I agree that 450-470MHz is not a problem for any powerline control system.  However, some of these smart meters communicate with a "data concentrator" over the utility feed using signals that can interfere with automation systems.  The data concentrator then communicates with the utility network over RF.

Jeff

[Noam]

I already have a very large inventory of ferrite cores, wire and capacitors.  I won't even need to order anything if I install the filter inside the UPS, which is where it should have been in the first place.  It doesn't take a very large inductor to filter out this high frequency component. A100uh inductor and 1uf capacitor would provide about 32db attenuation at 120Khz.  This would reduce the 500mv P-P to less than 10mv.  When ever possible, filters should be as close as possible to the switching power supply to reduce induced and radiated noise (EMI).  Those wall clocks that update automatically use VLF radio signals from the atomic clock in Bolder Colorado.   

But wouldn't doing that void your warranty?

[Backward Engineering]

What is a warranty?  I buy most of my gadgets at the K6TRW ham swap meet in Redondo Beach, previously owned, no warranty.  If I can't fix it myself, I don't buy it to begin with.  That ACP UPS with a new battery cost me $40.00.  It's not defective, it's just noisy by design.  I bought 2 of them at that price and both have the same noise.  Switch mode power supplies at light load are always a problem.  My passive coupler helped but with all circuit breakers closed the X10 signal on the opposite phase is still marginal at 1.5 volts according to the XTBM.  So I will be ordering an XTB-IIR Kit to replace my failed XPCR.

Don