Print

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

[Noam]

In the end, I just used a long (around 75-foot) extension cord to move the CM15A closer to the PC, but still plug it into the front of my XTB-IIR booster in the other room.

Sounds like reliable way to 'place' the cm15a near the distribution point.  I had been planning to buy a pair of usb-over-Ethernet extenders until I read your post.

I'm thinking of running about 80' or so of 14-2 from an outlet near the pc to a breaker in the breaker box to achieve what you did with the extension cord.  I will just pick up one side of the 220.
 
Do you still consider this the most reliable way to 'centralize' the cm15a when the PC is remote from that location?

Thanks,
Frank

You're likely to have fewer problems extending the CM15A's power connection than you will trying to extend the USB connection. But three are no guarantees.

[frank]

You're likely to have fewer problems extending the CM15A's power connection than you will trying to extend the USB connection. But three are no guarantees.

Thanks.  That is what I was thinking as well.  I am glad I didn't buy the extenders.

[JeffVolp]

Just remember there will be some signal loss in that AC power run to the distribution panel.

Jeff

[frank]

Thanks Jeff.  The limited advantage of a direct AC run is that it, in effect, bypasses the some 25 to 30 ea. 120 vac devices that are currently plugged into the same place, from which the cm15a is powered.  And that AC run goes around walls and such spliced at several outlets on its way to the breaker box.

I figure this will at least give the cm15a signal a fighting chance.  Not sure what else to do to improve the signal.

I don't have a way to measure the X10 signal strength.  That would be a nice tool to have.

[JeffVolp]

Out of curiosity, I just checked this myself.  I plugged an old Maxi Controller into a 50 foot length of 16 gauge extension cord stretched out straight to simulate a long wire run.  I removed my XTBR, and checked signal levels with my XTBM-Pro prototype.  The readings were 9.59Vpp at the Maxi Controller, and 2.57Vpp at the outlet adjacent to the distribution panel where the extension cord was plugged in.  So almost 3/4 of the signal level was lost in that run.

Jeff

[Brian H]

Frank, Do you presently have a phase coupler or coupler/repeater between the phases?
XTB-IIR is the gold standard one. My lowest signal is about 1.5 volt as measured by a XTBM test meter. With my XTB-IIR installed.

Have you done any X10 troubleshooting. To remove a signal sucker or noise maker.

http://jvde.us/x10_troubleshooting.htm
http://www.act-remote.com/PCC/uncle.htm
http://www.davehouston.net/

[frank]

Out of curiosity, I just checked this myself.  I plugged an old Maxi Controller into a 50 foot length of 16 gauge extension cord stretched out straight to simulate a long wire run.  I removed my XTBR, and checked signal levels with my XTBM-Pro prototype.  The readings were 9.59Vpp at the Maxi Controller, and 2.57Vpp at the outlet adjacent to the distribution panel where the extension cord was plugged in.  So almost 3/4 of the signal level was lost in that run.

Jeff

Incredible!  I never would have expected that much loss in such a short & not otherwise loaded length of wire.

[JeffVolp]

It makes sense.  The distribution panel is the common junction for all the signal loading in the home.  With the distributed inductance over the length of the extension cord, it begins to look like a low-pass filter that attenuates the 120KHz X10 signal.

This is the reason I recommend locating a repeater adjacent to the distribution panel rather than at a dryer outlet.

Jeff

[frank]

Frank, Do you presently have a phase coupler or coupler/repeater between the phases?
XTB-IIR is the gold standard one. My lowest signal is about 1.5 volt as measured by a XTBM test meter. With my XTB-IIR installed.

Have you done any X10 troubleshooting. To remove a signal sucker or noise maker.

http://jvde.us/x10_troubleshooting.htm
http://www.act-remote.com/PCC/uncle.htm
http://www.davehouston.net/

I do have a phase coupler 'available' in my breaker panel between the legs.  However, currently, I have all my x10 stuff on one leg and I have the phase coupler turned off. 

The farthest device I have from the cm15a is a lamp controller for front yard landscape lights, which was on the opposite leg.  It was plugged into the same duplex outlet as my sprinkler controller and worked fine until  placed a surge arrestor at that particular outlet for my sprinkler controller.  Doing so shunted the signal to the lamp controller so I moved the lamp controller to another outlet, which is on the cm15a side.  And then I turned off the phase coupler.

While I could place a filter inline with the surge arrestor, I'm not sure whether doing so might reduce the arrestor's effectiveness.

I have found a couple signal shunters and have placed filters inline with the AC for those.  I need to do something to increase the signal level to that far point.

I hate to hear the results of JeffVolp's recent test.  That is discouraging.

[bkenobi]

Jeff's results are similar to a set of testing I did in my garage.  I have a wired motion sensor that was added to the end of the outlet run in my garage (due to convenience).  I controlled the lights on the front of the house with the motion sensor via a power flash plugged into the same outlet as the motion sensor.  My guess is that there's a 100ft run from the distribution panel to that outlet through the walls with at least 8 outlets in series.  I did extensive testing with a XTBM and found that the signal strength drops considerably along the run (checking one outlet to the next while triggering the PF).  By the time I got to the distribution panel, I was down to a signal strength <1.0 when I started out at >7.0.  As a result, I only had moderate performance from that setup.  I found that I missed up to 15% of the signals from the motion sensor resulting in no lights when I wanted them to turn on and lights staying on all night since it missed the "end motion" signal.

The solution for me was to move the power flash to an outlet next to the distribution panel and run a dry contact wire out to the motion sensor.  Since it's only opening/closing a contact, the length of wire doesn't affect performance whereas transmitting X10 over the same length certainly was!

[frank]

Jeff's results are similar to a set of testing I did in my garage.
...
The solution for me was to move the power flash to an outlet next to the distribution panel and run a dry contact wire out to the motion sensor.  Since it's only opening/closing a contact, the length of wire doesn't affect performance whereas transmitting X10 over the same length certainly was!

Again disappointing!  While running the wires like you did solved the problem, it is unfortunate that such measures have to be taken.  Perhaps the transmit levels of the x10 devices need to be cranked up by 5v or so.

I've been looking to buy an XTBM for a year or two.  I expect I would be able to, in effect, raise signal levels through the use of one. 

[dhouston]

While I've not done the calculations myself, an EE acquaintance tells me that the inductive reactance of lamp cord and most romex at 120kHz is 20,000 times what it is at 60Hz.

It's a shame that X10 went to a USB interface for the CM15A. Long runs of a low speed serial link are not much of a problem.

This thread does make me wonder how well X10 works in Europe where Cenelec limits PLC signals to ~2.5Vpp.

[frank]

Makes me wonder whether you could pick off the 120kHz signal at the cm15a outlet point with a capacitive high pass filter, maybe even a series resonant LC circuit, then use coax to run that over to the breaker panel and reinsert it there using a similar series circuit.  At the breaker panel, you could use a dedicated breaker.... or have two of the little series circuits and feed each of two breakers to catch both legs.  (not entirely legal -- so mount a dedicated 110 (220?) vac outlet right at the breaker panel and connect the LC circuit to the AC line there?)

Don't know what the loss of 120 kHz would be in coax, but it would certainly be much less than using electrical wire.  Much cheaper too.

I suppose there could be a problem with the original cm15a signal being out of phase with the signal carried on the coax.  But it would seem the coax carried one would be quite a bit higher than the cm15a signal (capture ratio helps).  Or you might just get inside the cm15a and reroute the signal away from the AC plug to a small jack on the module and just bypass the CM15a to breaker panel AC path all-together.

You would, most likely, want to be careful not to electrocute yourself or an acquaintance.

? 

[dhouston]

Or you might just get inside the cm15a and reroute the signal away from the AC plug to a small jack on the module and just bypass the CM15a to breaker panel AC path all-together.

That might be doable - the PLC output can be separated from the PLC input rather easily just by disconnecting pins 5 & 6 of TC1 and routing the signal over the coax. You should use an isolation transformer at the insertion point. There's a copy of the CM15A schematic here.

• http://davehouston.org/CM15A.pdf

[JeffVolp]

Makes me wonder whether you could pick off the 120kHz signal at the cm15a outlet point with a capacitive high pass filter, maybe even a series resonant LC circuit, then use coax to run that over to the breaker panel and reinsert it there using a similar series circuit.  At the breaker panel, you could use a dedicated breaker.... or have two of the little series circuits and feed each of two breakers to catch both legs.  (not entirely legal -- so mount a dedicated 110 (220?) vac outlet right at the breaker panel and connect the LC circuit to the AC line there?)

That would probably work pretty well - especially with series-tuned LC networks at both ends.  You could ignore possible destructive cancellation of the X10 signal by just isolating the CM15A "network" from the powerline with the XPPF filter.

Of course, just connecting a repeater at the distribution panel would work too.

Jeff