X10 Signal Path Design

• • Guidelines for successful X10 Signal Paths:**

(This is written for a North American Power Line perspective)

• • Priority 1**: Ensure an X-10 phase coupling solution between the two 117V phases of N.A. power is in place (even if it seems to work without one). One Misterhouse user who "got it" wrote" Aha! To get a path you need a signal bridge." So. Where do I get one of those? :)
  • Priority 2**: Inject and receive the signal to the PC Interface (the most important X10 receiver/transmitter) as close as possible to the breaker panel, and directly into a wall or panel socket (i.e. no powerbars or 4/6 way trailing power sockets).
  • Priority 3**: Don't connect anything at the same point as the PC interface if at all possible, do not plug into the interface even if it is equiped with a socket.
  • Priority 4:** Limit the X10 signal collisions by limiting the number of transmitters - especially wireless receivers since more than one will cause signal collisions.

if you have problems:

• • Priority 5:** Then look for signal interference - things that would create line noise in the 120 KHz range.
  • Priority 6**: Then look for signal traps (surge and line noise filters) that might be diminishing the signal.

Often the only way of doing this without spending money is to unplug electrical items one at a time see if things improve. Usually if things have been working fine and then suddenly have problems, it's either the latest piece of electrical equipment that's been brought into the house, or an older item that is failing (usually the power supply).

• • Dialog and Opinions:**

Jim writes: My opinion on the few questions posted about X-10 signal problems: Background - X10 works by injecting a nominal 120 KHz signal burst at the 3 phase zero crossing points of a single phase circuit for those 60 Hz cycles it wants to mark as a digital "1" and nothing on those that are zeros. An X10 receiver listens at its zero voltage crossing for the 120 KHz burst. My opinion - when PC's power supplies interfere with the X-10 signal the overall X-10 signal distribution is probably marginal at best anyhow - so I would insure that that is fixed first. My design principles (written from a NA 117V perspective): Attempt to create a star signal path for the X-10 signal. In particular the computer running the main control program should be near the center of the star. (This might be hard for some homes.) I locate the X10 PC interface as close as possible to the 117V breaker / distribution panel. In my case it plugs into the panel plug - located within 3 feet of the panel. I have an X10 signal phase bridge between the two 117V circuits. I DO NOT depend on the signal getting reliably coupled by the street transformer from one phase to another. It may work sometimes - but after a heavy rain there may be too much attenuation in the path to the street transformer and back. I use a passive X10 phase coupler located in a dedicated 220V box on a workbench circuit also within a few feet of the breaker panel. This creates a star configuration with the X10 computer interface at the center of the star. Everyone that I have talked to has had an X-10 phase coupling problem or very few devices that just happen to be on the same phase. If you live in NA and don't have some kind of phase coupler solution you probably have a marginal X10 signal on one phase of your AC distribution. Fix that first. (Note do not be tempted to put anything within the breaker box - I once saw an electrical fault (badly stripped cable) happen in an electrical panel after a storm induced voltage surge and I was very glad there was nothing combustible in the breaker panel because the arc just burnt out after about 3 minutes of smoke - it was not a large enough arc to trip the breaker. Since by UL code there is nothing combustible in the breaker panel there was no resulting fire). Even though it is inviting I do not plug anything into socket on the X10 computer interface. Just leave it alone to send and receiver signals without a local load on the signal. Plug your PC into another circuit - even if it is just a few feet of splitter bar line cord away. (In my case the UPS is plugged into the same socket, (not the X10 interface but the adjacent socket), and the PC is into the UPS. The UPS has line noise filtering. Now lets consider what can happen to that 120KHz signal burst marking a 60 Hz phase: it can be attenuated by some signal trap - like the inductors and capacitors in the PC power supply, or some other noise filter before it gets the chance to radiate out the star, or there is so much noise near the 120 KHz band pass filter that it can not be heard, or hear other 120 KHz signals over the noise. Now to the questions: 1) How would one determine if a PC power supply is going to interfere? 2) What does a ferrite choke do? The original post implied that there was a noise source that was interfering (rather than the signal was marginal and being trapped). So the solution was to disconnect the offending device at the zero crossing points by the diode drops. 1) PC power supplies don't seem to spec the noise they generate at their switching speeds - likely because it is not usually a big RF problem. Switching supplies operate in the 30 KHz to 300 KHz range (the lower range being more common in larger devices). I would look for a good quality supply that notes that it has line noise rejection (the filter would work both ways - keeps line noise out, and switching noise in). But there is no spec I commonly see that would help you. I would operate under the assumptions that only PC power supplies on the limit either through loading or wear out are a problem for noise generation. (Let's see how many posts that statement brings - because maybe it is a bigger problem than I think.) - So I would say most of the solid PC power supplies operating below their peak spec should not be a problem. 2) How does the RF Ferrite Choke help - the simplest way to describe this is that low frequency power flows through the conductors in the cable - but everything that is higher frequency (like the X10 signal, or noise in the same band) actually propagates through the insulator around the conductors as an electro-MAGNETIC wave. The Ferrite choke is like a magnetic conductor that shorts out the electromagnetic wave preventing it from propagating along the AC power cord.