Signal optimization/electric panels

[ncarney]

I don't know anything about powerline signaling and I'd appreciate some guidance on optimizing the signal strength in my house.

 

Current configuration:

> Schematic of house footprint shows floor 2 (main level). MainPanels and SubPanel 1b are on floor 1; SubPanel1a is on floor 2.

> 2 - 200 amp main panels coming off the meter.

> 1 - 90 amp SubPanel 1a feeding a detached garage with office overhead, approx 20' from the main panels.

> 1 - 90 amp SubPanel 1b located in home theater area of house, approx 60' from main panels.

> Both subpanels are branched off of MainPanel1 (more large appliances on MainPanel2).

> My ethernet patch panel is next to the main panels, but there is a switch near SubPanel1b with the A/V gear.

> ISY/PLM is currently installed next to MainPanels due to signal problems.

> 1 AccessPoint in office area; 1 in home theater area.

 

Plans:

> I am planning to pull two dedicated circuits off each leg of MainPanel1 and install new AccessPoints on each, immediately adjacent to the MainPanels.

> Once the signals improve, I would like to move the ISY/PLM back to my A/V rack.

> Once I move the ISY/PLM, it would be placed on a dedicated circuit off of SubPanel 1b, along with the AccessPoint for that area.

 

Questions:

> I assume that by installing AccessPoints on each leg of MainPanel1, the signal will be bridged at the electric meter, and there would be no reason to run circuits and AccessPoints from MainPanel2. Correct?

> Do you think four AccessPoints will be adequate: 2 at MainPanels, 1 near SubPanel1a, 1 near SubPanel1b? The house is four stories tall including the finished basement, so I probably will need more AccessPoints as I add Insteon equipment to floors 3 and 4.

> The rack for home theater and whole-house A/V gear is feed by two 20 amp dedicated circuits off of SubPanel1b. I understand it may create alot of noise and/or signal absorption. If the PLM is on a separate dedicated circuit off of the same SubPanel1b, is that sufficient isolation? Or do I need to keep the PLM near the MainPanels?

> Is any of this over-engineering?

> Is there anything else I should do or consider?

 

Thanks in advance for your help. All suggestions are appreciated.

 

[garybixler]

ncarney

 

Might I suggest using an Insteon hardwired bridge at the main panels. These provide a positive transfer from one power leg to the other without using up an additional hop count. Access Points can always be placed where needed for the wireless devices or addition power leg bridging. Also using filters on electronics like A/V equipment, TVs, PCs and UPS's will keep the Insteon single from being absorbed at their locations. Also be careful when using the Compact Florescent Lights as they tend to add noise into the electrical system. I had to remove almost all of the ones that I installed. Haven't checked to see if all brands have this problem.

 

Just though I would pass along my experiences. Maybe useful.

 

Thanks Gary

[ncarney]

Thanks for the suggestions Gary. Just this weekend I noticed that they sell the hardwired bridge (SignaLinc 2406H), which seems like it would be preferable to line/RF/RF/line jump of using AccessPoints (and cheaper). I noticed that it says that the SignaLinc does not provide signal amplification, whereas perhaps by using the AccessPoints you get amplification by virtue of the signal being re-transmitted. Any thoughts on that?

 

I do use many CFLs throughout the house, although so far I have only installed one Insteon switch which has them. I purchased a SwitchLinc Relay since they are marketed as supporting CFLs, but perhaps this marketing is misleading. I suppose the non-dimmable feature supports CFLs, but it sounds like CFLs are a bad idea from a noise perspective. What is the symptom of CFL interference? Just poor communications?

 

Thanks again.

[garybixler]

Hi Neal

 

It is true that the bridge is a passive electronic device but you would get regeneration from the other Insteon devices on the different legs.

When I had a CFL turned on that was electrically close to an Insteon device I could tell immediately that communications with that device was lacking. Usually by a slower response or missed response all together. Also I have found that the dimmable CFLs were even worse. However I do have a few CFL's that don't seem to create a problem. So all CFLs may not be created equal but, I guess I will wait for LEDs to come into their own. Although it maybe worth looking into and testing the different brands of CFLs.

I plan on getting an oscilloscope at some point and then would be able to see what exactly is on the power line.

 

Gary

[IndyMike]

Hello Gary and Neil,

 

Sorry for barging in, but you've hit on one of my favorite topics. I agree that a passive Insteon coupler at the panel can be a huge asset and I also believe it saves 1 hop in coupling the phases.

 

I use a 4616H X10 coupler at the panel for my legacy X10 components. While not optimal for Insteon, I can run the entire house (4500 sq Ft/3 levels) reliably with no Accesspoints installed. Whether you can do the same depends on the type of loads installed and how far your 1st devices are from the coupled panel.

 

Over the past couple of weeks I've been playing with I2 signal distribution. I was using 1 I2 ApplianceLinc and checking signal propagation throughout the house. I have no other I2 capable devices (Accesspoints were removed) so the I2 signal was not being repeated. Without the repeat feature of the Insteon protocol, the configuration winds up being similar to my X10 system. I found that I could reliably communicate, program, and query the I2 device anywhere in the home with only the passive coupler installed. I then pulled most of my filters (except the one on the Toshiba laptop) and still had good communication and 100% scene test response.

 

My loads include PC's (6), A/V systems (3), CFL's (25), and the normal number of wall warts for phone chargers, modems, switches, etc. I disconnected only my Hafler amplifier (black hole for PLC) and filtered only the Toshiba laptop.

 

The point of the above is that Insteon can be made reliable even in extreme cases without throwing a 6-pack of Accesspoints at the installation. I tend to treat Accesspoints as nothing more than a Rf to PLC signal bridge. If they are used to "cover up" a signal problem on a circuit, that problem is likely to come back and bite you (noisy devices/signal absorbers can move).

 

A few words on using CFL's. While standard CFL's are noise producers, they tend to induce spikes near the peak of the AC sine wave. Since Insteon and X10 communication occurs at the zero crossing it is resistant to this noise. I say resistant because I have seen documented evidence (lab environment) where multiple CFL's can combine to produce "beat" frequencies that do affect zero crossings.

 

As Gary noted, dimmable CFL's (dimmer friendly) can be worse than the standard version. When operated at max level the dimmable CFL's inject noise similar to their standard brothers. As the input to the CFL is cut (triac output from the dimmer) the spike injected by the CFL grows up to 2X. I've seen 20V P-P spiked injected by a 8W dimmable CFL. This spike did not prevent Insteon communication because it occurred at the waveform peak. It did lock up the Lamplinc dimmer that I was using to control the lamp (direct noise injection).

 

I'll try to post some plots of the CFL noise injection and Insteon phase coupling when I get a chance.

 

IM

[ncarney]

Thanks so much for the feedback IndyMike.

- I think I will get the 2406H hardwire bridge.

- Bridging at the panel seems to be the key, as others have noted. I originally bridged via RF farther away. It worked, but not 100%.

- I do need a couple of AccessPoints because I am using thermostats which are RF only. Aside from that, I agree with you that it is preferable to use them as a last resort, after debugging the root cause.

- As you alluded, they might be appropriate as an enhanced repeater, since right now there is a fair distance between the panel and my first device.

 

Follow-up:

- The PLM seems to be a "less-friendly" receiver than the switches themselves.

- I would prefer not to have the PLM located at the main panels but right now I am leaning that way based on what I have observed so far.

- If I hard bridge at the main panel, and put the PLM off the subpanel, do you think that would still maintain good quality communication? The main-to-sub distance is 60' and the PLM circuit would be about 15'. I don't know the effect of the distance and the breakers on the signal strength.

 

Sorry if I should be asking some of these questions over at SmartHome; everyone here seems really knowledgeable.

[garybixler]

Hi IndyMike

 

Thanks so much for all the valuable information. I agree 100% in creating a good solid power line infrastructure where APs are only used for wireless device access. I am not a proponent of wireless anything. In my experiences I have never seen wireless replace the solid reliability of a good old hardwired infrastructure.

 

I do have a question on the CFLs. I have two Sylvania 23w CFLs in my garage that are on the same circuit as an eight button KPL. These are not controlled by the KPL but when they are on they do impede the communications to that KPL. By doing multiple querys from the ISY I see a drop in hope counts. Normally 2 but when the CFLs are on that may go down to 1 or 0 on 20% of the messages. Are you using this brand and wattage? I haven't tried any others as yet.

 

Thanks for any insight.

 

Gary

[IndyMike]

Neil,

 

The "Optimal" location for the PLM is at the main panel. This is provides a solid neutral connection and a "point" for all signals to fan out. I can't honestly say what affect the 60' will have on the signal. It depends on distributed inductance/capacitance of the cabling and the type of loads you have on the panel. My gut tells me this should be no problem.

 

I had promised a plot of the passive coupling using I2 signals.

 

 

The signals in the above:

1) First communication - 11 packet I2 transmission from the PLM

2) 2nd communication - response from Accesspoint (opposite phase) through the signal coupler.

3) 3rd communication - PLM repeat

4) 4th communication - Accesspoint repeat.

 

Now you may have noticed that the "coupled" signal is only ~ 0.7V P-P. This was a worst case test. I have only two I2 devices (PLM and Accesspoint) coupling across the phases. The PLM is at my main panel. The Accesspoint is 2 floors up at the opposite end of the house. Also, keep in mind that my coupler is designed for X10. It's giving up some level since it's not tuned to the Insteon 130Khz.

[IndyMike]

Gary,

I've actually had good luck with the Sylvania, although I do not have any of the 23W. I do have 8-15W Sylvania in my upstairs hall controlled by SWL relay - no problems. My bathroom use 8-9watt (light bars) bulbs again controlled by a relay SWL.

 

Is it possible that you have a poor connection on the circuit feeding the CFL's? Does this happen whenever the CFL's are on, or only during warmup?

 

IM

 

Hi IndyMike

 

Thanks so much for all the valuable information. I agree 100% in creating a good solid power line infrastructure where APs are only used for wireless device access. I am not a proponent of wireless anything. In my experiences I have never seen wireless replace the solid reliability of a good old hardwired infrastructure.

 

I do have a question on the CFLs. I have two Sylvania 23w CFLs in my garage that are on the same circuit as an eight button KPL. These are not controlled by the KPL but when they are on they do impede the communications to that KPL. By doing multiple querys from the ISY I see a drop in hope counts. Normally 2 but when the CFLs are on that may go down to 1 or 0 on 20% of the messages. Are you using this brand and wattage? I haven't tried any others as yet.

 

Thanks for any insight.

 

Gary

[garybixler]

Hi IM,

 

Very impressive information. I was thinking about getting an oscilloscope for around $600.00, but now I am wondering what you may be using to see the actual power line signaling. Looks PC based?

 

My problem with the 23w CFLs occurs as soon as they are turned on until they are turned off. Strangely I also have some 13 watt and a 12/19/28 3-way that don't seem to interfere at all with communications to a 6 button KPL sharing the same circuit. Maybe I will try other wattage bulbs and see what happens.

 

Thanks

Gary

 

Update: I replaced the 2 Sylvania 23 watt CFLs with a different brand of 23 watt CFL and that seems to have taken care of the communication problem. So all CFLs are not created equal.

[ncarney]

This is great info IM.

 

I too would be interested in what you are using to capture and plot the signaling.

 

I have no background in things electrical. From what I have read, the minimum transmit on the PLM is 3.2V p-p, and the minimum receive is 10 mV. Your plot seems to show that the transmit is averaging maybe around 5.0V, and the worst case receive (across phases, two floors up, no repeaters in between) is .7V, which is still much higher than the 10 mV minimum. Do you think the 10 mV is really true, or is the real minimum reliable receive level much higher? If the 10 mV is true, it seems like you still have plenty of margin even in your worst case scenario.

 

On the CFL subject, I have 6 recessed floodlights on a SWL relay using 23w CFLs made by FEIT that I purchased at Costco. No problems that I have noticed so far, but I am still pretty new to this and trying to figure out how to monitor. With the lights on, I get a good response on ISY queries with 2 hops left.

[IndyMike]

Neil and Gary,

 

The Scope is a USB interfaced DS2200 from USB instruments. These are no longer available, but they do have a new version (DSM12) for around $200.

 

I use a ACT Scope-Test2 adapter to knock down the AC line voltage to manageable levels and separate out the Insteon/X10 transmission. These are also no longer available (maybe Ebay) but there are plans available from Uncle Phil Kingery (ACT Guru). I word of caution if you do use an adapter (or build one) - the coax shield is connected to the neutral line. This can be a problem if you have a device that is drawing current and raising the neutral level. Always treat the shield as if it was hot and use an isolated (no chassis ground) oscilloscope. I use the DS2000 with a laptop to be safe.

 

As far as the Insteon sensitivity levels are concerned - I can believe 10mv in a lab environment. In a normal home environment noise would swamp out a 10mv signal. I believe the minimum "acceptable" X10 level is considered to be 100mv.

 

IM