Print

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

[JeffVolp]

In this era of CFLs and other electronic devices, many of you are experiencing noise problems.  The usual approach is to isolate potential noise sources with filters.  While that works well for many items, there have been cases where the noise is originating outside the home, and cannot be easily isolated with filters.  The X10 PZZ01 and equivalent Leviton 6284 are not entirely effective in blocking noise from coming in from outside the home.  Even with the 6284 we had an in-band signal coming from a neighbor’s baby monitor.

An approach I’m pursuing now is to develop a “tuned signal sucker” that will significantly attenuate out of band noise, and also provide some attenuation to in-band noise.  The reason for the in-band attenuation is to try to attenuate noise produced by things like CFLs and 120V LED bulbs that are running near the 120KHz X10 carrier frequency.

The downside is that when using this tuned signal sucker, it may be necessary to boost X10 signal levels above that produced by ordinary X10 transmitters.

The idea is to basically attenuate background noise levels, and then boost up X10 signals so they rise well above the residual noise level.  Note that if a X10 switch is controlling a strong in-band noise source, it still may need to be isolated with a filter for the X10 switch to work properly.

No active devices are needed in the tuned signal sucker, and some of you can probably build your own from scratch.  My WAG is that parts would cost about $20, including a case and PCB.  One would be plugged into a receptacle on each phase near the distribution panel.  Comments?

Jeff

[Brandt]

Since I've already have an XTB-IIR, and have used my XTBM to locate most signal suckers/noise generators inside the house and have filtered as appropriate, what would these tuned signal suckers do for me?

I have the XTB-IIR plugged into the 4-prong dryer plug in the laundry room, and 'i think' the two other 120v receptacles next to it are both on separate phases. I assume that is where these tuned signal suckers would go?

[Bill H]

I have some fluorescent fixtures/lamps that occasionally give me trouble and CFLs are in all our futures. Put me down for an order when you're ready!

[JeffVolp]

Since I've already have an XTB-IIR, and have used my XTBM to locate most signal suckers/noise generators inside the house and have filtered as appropriate, what would these tuned signal suckers do for me?

Perhaps nothing.  There are a couple of fellows who have been dealing with severe noise coming in from outside their homes.  As CFLs and 120V LEDs continue to proliferate, this will become more of a problem because everyone on the same utility transformer essentially share the same "X10 network" unless the home has a PZZ01 or Leviton 6284.  And even those do not offer complete isolation.

For noise coming in from outside the home, the best location would be near the distribution panel to try to quench the noise before it is distributed to other circuits.

This is just an idea at this point.  I have done simulation, and have a design that works well.  I'll build a prototype and see how well it does against noise that I purposely inject onto the powerline.  If there is enough interest, I'll make some PCBs.

A year ago I proposed another module to monitor when devices were on or off.  While there was interest in that device, I eventually dropped the project because it became obvious it would be too expensive for the potential market.

A tuned signal sucker just has half a dozen passive components.  The only tricky thing is that it will probably need to be tuned by selecting a capacitor to center it on 120KHz.

Jeff

[Dan Lawrence]

Interesting, Jeff, but I'm in Baltimore and have X10 since the middle 1980's.  NOBODY else on my transformer even has X10 since there are 6 other houses on it and I've NEVER had a signal problem since Day 1.  No lights going on at odd times or other problems.

[JeffVolp]

It is not interference from X10 that is the issue.  It is interference from other electrical devices producing noise on the powerline that interferes with X10 communication.  Some of the newer X10 modules incorporate AGC to deal with the issue, but there is a huge installed base of X10 devices out there that can be blocked by relatively low noise levels.

When the noise generators are in your own home, they can be isolated with filters.  There was a recent thread about one fellow who tracked down his noise problem to a CFL at his neighbor's house.  With the last incandescent bulb factory in the U.S. closing last week, CFLs and other high tech lights will continue to proliferate.  Noise problems will continue to get worse.  I'm looking into how to effectively deal with the issues.

Jeff

[JeffVolp]

For those of you who may be interested in what a tuned signal sucker could do, below is a plot from my simulation.  It sweeps a powerful 2Vpp noise signal with 10 ohms source resistance from 20KHz to 220KHz, and displays the resulting noise voltage on the line from several different attenuators.

The blue signal is the noise itself.  Green is what would happen with a typical signal sucker, such as the .1uF capacitor I found across the AC input of my computer’s power supply.  Yellow is the result of the tuned signal sucker without the added 120KHz attenuation.  Magenta adds some 120KHz attenuation.

In case you wonder why the attenuation isn’t as great above the cutoff frequency, it is because that portion of the spectrum is attenuated by the capacitor in the tuned filter.  That doesn’t have as low impedance as the inductor that covers the low end of the spectrum.  However, there is still a significant reduction in the noise voltage throughout the spectrum except near the X10 carrier frequency.

A less powerful noise source is attenuated much more than is shown by these plots.

Jeff

[pconroy]

I have no idea what you just said but that graph looks way cool!!!     

[JeffVolp]

I frequently scan other forums because people using other controllers often use X10 devices for the actual control.  I use the Ocelot myself due to its ability to process analog measurements.

While most of us know about the noise susceptibility of X10 devices, I’m running across reports of people having problems with Insteon and even UPB systems due to noise.  They may be more tolerant of noise, but they certainly are no panacea.

Prototype PCBs for the “Tuned Signal Sucker” are on the way.  I will report the preliminary test results here in the near future.

Jeff

[Brandt]

Just curious...who do you go through for PCBs?

[JeffVolp]

Just curious...who do you go through for PCBs?

For prototypes, I use Advanced Circuits Barebones service, which is under $100 for half a dozen prototype boards (no solder mask or silkscreen).

For production I had used a couple of places in China, but switched to Advanced Circuits for the XTB-II several years ago.  I may change again because the hole sizes on recent lots varied significantly across the board.  Holes made with the same drill size were too tight in one area, but were fine elsewhere on the board.  Two lots separated by about a month were similarly effected, and all the "problem" boards had tight holes at one end.  I will try them once more with the option to maintain tighter tolerances in order to avoid another setup charge elsewhere.

Jeff

[Brandt]

And what layout software do you use?

[JeffVolp]

And what layout software do you use?

CircuitMaker/TraxMaker

It's an old program, but I think the package is as good as ones costing several thousand dollars today.  Because they were strong competition, one of the "big boys" bought them out maybe 8 years ago and killed the product.  The CircuitMaker simulation tool saves a lot of work at the prototype stage.  Then after I complete a layout, TraxMaker extracts the netlist to verify it matches the schematic.  It exports the Gerber files needed by the PCB fab houses.

Jeff

[JeffVolp]

I assembled a beta version of the Tuned Signal Sucker, but have not had time to do extensive testing yet.  This is copied from a post I added to another thread that some of you may have missed:

I wanted to follow up on the noise issue.  This evening I was re-testing an early XTB-IIR on my bench that someone had returned for repair (I replaced a broken connector).  As part of my test, I send commands through the digital port to cycle a nearby appliance module on and off.  It didn't work.  Hmm….

At first I thought I sent the wrong code, but that was correct.  I had updated the firmware, so I thought maybe I made a mistake there.  Then it occurred to me that I was testing at night.  I virtually always do my testing during the day.  Since it was dark I had the set of noisemaker CFLs in that room on.  Normally I only switch them on to check the AGC response.  The noise was strong enough to totally block the appliance module from decoding even the 30Vpp signal coming from the XTB-IIR.

Checking the output of the bandpass filter in the XTB-IIR showed about 500mVpp of in-band noise.  (Those CFLs are particularly nasty because their noise is very near the 120KHz X10 carrier frequency.)  Just for the heck of it, I plugged my beta Tuned Signal Sucker into the workbench power strip.  As the simulations predicted, the in-band noise from the CFLs dropped by a factor of two.  (The TSS severely attenuates out-of-band noise, but can't do that at the X10 carrier frequency because almost nothing would be left.)

After plugging in the TSS, the appliance module worked 100% through dozens of cycles.  Unplug the TSS – no go.  Back in, it works again.  Out – no go.  I repeated this several more times.  Only on two tries of maybe a hundred did the appliance module actually switch when the TSS was unplugged.  And interestingly, in one of those cases it actually switched the WRONG WAY.

Of course, if I turn off the noisemaker CFLs, everything is fine.  I just thought I would share this with you.

Jeff

[JeffVolp]

I have a couple of beta Tuned Signal Suckers available for those trying to deal with particularly nasty noise problems.  It can be plugged into a circuit containing a noisemaker to reduce the noise getting back to the distribution panel.  Since it also loads down the X10 signal level on that circuit, it is best used in installations that contain some sort of signal booster.

Please contact me directly by a PM or email if you are interested.

Jeff