On/off delay

[MarkJames]

Thanks for the responses.

 

I think you misunderstood me, Lee, regarding the breakers. I wasn't implying that a breaker would block any kind of noise. What I was trying to say was that I see the house as a series of circuits all running in parallel. When a signal is injected into the line from the PLM it will follow along the circuit it's on and return to the panel where it will be distributed through the bus to all circuits in the home. With all circuits parallel to one another I couldn't see how a device on one circuit could matter to one on another as the PLM would have injected its message into the other circuit (indeed into ALL circuits) via the bus in the same manner that it got to the circuit in question. I guess I liken it to having 20 roads leading away from a central hub. Just because there's construction on one road shouldn't block you from driving down a different one. You may not be able to get to the houses at the end of the block where the construction is but there's no reason you shouldn't be able to get to all the houses on other blocks.

 

If what ELA says is true - and there is some kind of resonant frequency or signal noise that a device generates - then all bets are off and the I can see how a device could have far reaching effects. But not so much as a signal 'sucker' but rather as a "global signal damager" if that makes sense.

 

I'll go read your other posts ELA and see what you've been up to and measuring.

 

mark

[IndyMike]

Hello Mark,

 

Let me start by saying that I like your description of the PLM communicating into the load panel. Assuming that PLM is close coupled to the panel, it will attempt to drive the loading at that point to the full 3.2 Vp-p. The load at the panel is a complex composite of all of the loads in your home reflected back to the panel. I say attempt because it's possible that the signal loading at the panel is high enough that the PLM can't drive it to full level. The signal will then fan out across the branch connections on that phase. Signal degradation on a given branch can usually be viewed on a branch by branch basis.

 

If you have a passive coupler that the panel, the PLM signal will be coupled in phase at a lower level and also fan out on that phase. If you are using a RF coupler, the PLM signal will be coupled on the next hop, but again at whatever level the RF coupler can drive into the load it "sees" (presumably 3.2Vp-p).

 

ELA is exactly correct that the signal propagation through the branch circuits can be an extremely complex "transmission" with resonances and nodes as the R-L-C characteristics interact. That said, ELA and I share similar backgrounds and tend to think in these terms (EMI/EMC background - other engineers consider us to be "geeks"). Not speaking for ELA, but I sometimes drift off into complex theories of what might happen rather than gathering all the facts and focusing on what is happening. For most installations, you can consider the powerline as a simple resistive divider circuit. The longer the line, the more series resistance. The more loads installed, the lower the resistance to return (more signal degradation).

 

This resonance condition was exactly what I was thinking of when you described the following:

 

The lag is still there regardless of the ioLinc. I'm going to move the LED's onto a different circuit and see if that matters. It's doubly confusing as I have to listen for the Switchlinc 'click' and ignore the LED lights as there's a lag between when the switch turns on, the transformer powers up, and the LED's glow. Oddly there's much less lag on an off command - almost none, really.

 

mark

 

Since your SWL responded well to an off command, I was theorizing that you had a long transmission line and the added Inductive load of the transformer(s) was changing the powerline impedance and effectively killing a resonance at that point. I had hoped that you would install a incandescent lamp at the IOLinc location (not at the switchlinc) and thereby change the R-L-C loading for both the On and Off cases. Although you mis-interpreted the intent, you may have provided enough information to disprove this theory:

 

1) Your Switchlinc is close coupled to your media center and both are close coupled to the panel (18 inches - not a long line).

2) Your PLM is close coupled to the panel.

3) Is your PLM on the same phase as the SWL, or at least bridged at the panel?

 

If your PLM is on the same phase (or coupled at the panel), your distances should be too short to set up a resonant condition (low line impedance). Adding an incandescent load would not have accomplished anything.

 

A more likely scenario would be -

1) Your media center components are significant "signal suckers". Since these are close to the panel, they could be degrading signals from distant transmitters.

2) You have some noise on this circuit. The combination of noise and "signal suck" are degrading the inbound PLM signal to the point where it is marginal. When the SWL is turned on, you are adding the inductive load of the transformer and effectively filtering some of the noise at the SWL location. This allows the device to hear to the "off" command.

 

Since your media center does appear to be a source of loading/noise (and it's close to the panel), I would agree with the others that filtering this circuit would be wise to prevent possible problems at the PLM.

 

I am also curious whether you swapped to the opposite phase when you reconnected your SWL.

 

In regard to your last comment:

If what ELA says is true - and there is some kind of resonant frequency or signal noise that a device generates - then all bets are off and the I can see how a device could have far reaching effects. But not so much as a signal 'sucker' but rather as a "global signal damager" if that makes sense.

 

What ELA stated is absolutely true in a very strict controlled environment. Resonances will happen, particularly on long lines. In the home, this is normally a "transient" condition since the impedance of our home powerlines are constantly changing (every time we switch on or plug in a device we change the impedance). While I've been able to setup and measure resonances under controlled conditions, I have never observed one in the home.

 

I have witnessed 2 instances of "global signal damagers":

1) The Boosterlinc - these devices actively interfere with insteon signals. The newer "Insteon blessed" devices are better than the old models, but still interfere with low level Insteon signals (unless there is a Rev 3.0 out there).

2) Oscillating power devices - These can be triac dimmer outputs that are "unhappy" with the load (CFL) or other power outputs that are on their last legs. While oscillating, these devices can draw significant power and radiate long distances (lower frequency than Insteon). Noise produced by these devices can activate the Automatic Gain Control (AGC) on Insteon devices and render them "deaf" to valid transmissions.

 

Not sure whether this helps or simply confuses things further...

 

IM

[ELA]

I guess I liken it to having 20 roads leading away from a central hub. Just because there's construction on one road shouldn't block you from driving down a different one. You may not be able to get to the houses at the end of the block where the construction is but there's no reason you shouldn't be able to get to all the houses on other blocks.

 

mark

 

Mark,

They often use plumbing analogies for electrical issues. In this case you might think of it as a single hose that splits into, say three other hoses at one end.

You have a certain amount of flow down the main hose that you want to maintain as even of a division of, and at a certain level of flow, in each at the end.

 

Now if one of those end hoses is leaking it has the potential to "drain" flow away from the other hoses.

 

May not be the best analogy but one that occurs to me.

[ELA]

Hello Mark,

Not speaking for ELA, but I sometimes drift off into complex theories of what might happen rather than gathering all the facts and focusing on what is happening.

 

What ELA stated is absolutely true in a very strict controlled environment. Resonances will happen, particularly on long lines. In the home, this is normally a "transient" condition since the impedance of our home powerlines are constantly changing (every time we switch on or plug in a device we change the impedance). While I've been able to setup and measure resonances under controlled conditions, I have never observed one in the home.

IM

 

Hello IndyMike,

I have also been guilty of speculating on theories and try my best not to until I have testing results to support the theory. I try to qualify statements with that in mind.

I will eventually get around to testing to prove or disprove a possible resonant condition.

 

This particular situation surprised me. It consists of an outlet in the middle of a small branch of 6 total outlets on a 20A kitchen circuit. The main load on the circuit is the refrigerator located at the first outlet off of the home run. While monitoring the "suck level" at this particular outlet it "maxes out" the measurement range. While still monitoring that outlet I have unplugged all devices on that circuit except the refrigerator. The high drain level remains.

Measuring that same circuit but at the service cabinet (looking back they other direction at the refrigerator load) measures only 1/2 of what was measured at the problem outlet.

 

Doing the calculations seemed to support a large capacitive load that the refrigerator along with the series inductance to that outlet might be just the right LRC combination to be resonant. I do qualify that I will not feel comfortable with assuming that is the case. I will remove the refrigerator one day and test further.

 

Right now I am not adding Insteon devices to that circuit and this situation does not appear to "reflect" back at the service and so it is not a large priority yet.

 

Who knows maybe I will end up finding a staple that was driven just far enough into the romex to be an issue for Insteon but not at 60Hz I really have to wait to test more.

[MarkJames]

Who knows maybe I will end up finding a staple that was driven just far enough into the romex to be an issue for Insteon but not at 60Hz I really have to wait to test more.

 

Sorry.. this post came up just as I made my last entry.

 

ELA - I have a similar situation in my home as far as an oddball run. My recent changes are the end result of a substantial renovation. We gutted our kitchen, living room, dining area etc. While the drywall was out I ran new circuits for my lighting and outlets. All are home runs back to the panel - all are 14/2 and are all grounded properly. When I went around measuring the voltages at the outlets making a map of which outlets were on which breakers I noted that one line only measured 114V while the outlet right next to it measured the usual 117. Indeed all the outlets in the room measured 117 EXCEPT for that one which measured 114. All the runs were the same length within a few feet. I could literally see the wire running all the way back to the panel with no other loads on it so I knew there was no load that was causing this.

 

I've not come up with a solution to why this is so but I speculate that the breaker on that line may not be making perfect contact. That or one of the conductors in the wire is damaged. My technical experience is in computer software and not in electrical engineering so I couldn't diagnose this if you delivered a pallet full of diagnostic equipment.

 

Anyways - I just offer that up because your story reminded me of this.

 

mark

[IndyMike]

EDIT: I've just re-read your post and am not sure whether your outlets are on the same breaker (branch) or different breakers. If the outlets are on different breakers it's also possible that you are measuring different phases. Voltage differences across the phases are not uncommon when there is a load imbalance (one phase heavily loaded, the other lightly).

 

If your comfortable with it, try measuring the voltage at the load panel for each phase. If you have no loads installed on your outlets, the voltage at the panel should be the same as the outlet voltage. If it is not, you either have a very bad connection or a device drawing significant current.

 

The resonances that ELA and I were referring to apply at the Insteon frequencies of 130 KHz. They do not apply to the 60 Hz power distribution within your home.

 

Original message below-

 

Mark,

 

I am very much hoping that you simply had a bad connection to your voltmeter when you measured the 114V. Assuming that there was no load on the outlet at the time, a 3 volt drop (117 to 114) would qualify as an extremely bad connection. Your meter should be high impedance and therefor be drawing an insignificant amount of current and no noticeable voltage drop.

 

Please inspect/tighten the connections to this outlet and retest. If the drop is occurring at the wire attachment, it will generate heat (i.e. dangerous).

 

It's possible that the receptacle "prongs" were simply oxidized and not allowing your meter to make a good connection. The acid test would be to apply a light load (60 watt incandescent) and monitor for a voltage drop. Again, a 60W load should be insignificant. If you measure a drop with the load installed, replace the outlet and re-test.

 

IM

[IndyMike]

Hello ELA,

 

It's probably very naive on my part, but I don't normally think of a refrigerator (motor load) as being a problem device for Insteon. I say naive because manufacturers are constantly adding new bells and whistles to appliances. It's very possible that the new generation refrigerators do appear capacitive as opposed to my 25 year old Kenmore. I'll find out soon enough - my better half is keen to replace the Kenmore.

 

Beyond the above, I don't doubt for an instant what you are seeing. I'm simply questioning whether a true resonant condition can exist for any significant period of time with other loads being activated. While I'm sure that it can and does happen, I've never been able the capture this condition "in the act".

 

Looking forward to your conclusion when you get to moving the refrigerator (not a task I take lightly).

 

IM

 

Hello IndyMike,

I have also been guilty of speculating on theories and try my best not to until I have testing results to support the theory. I try to qualify statements with that in mind.

I will eventually get around to testing to prove or disprove a possible resonant condition.

 

This particular situation surprised me. It consists of an outlet in the middle of a small branch of 6 total outlets on a 20A kitchen circuit. The main load on the circuit is the refrigerator located at the first outlet off of the home run. While monitoring the "suck level" at this particular outlet it "maxes out" the measurement range. While still monitoring that outlet I have unplugged all devices on that circuit except the refrigerator. The high drain level remains.

Measuring that same circuit but at the service cabinet (looking back they other direction at the refrigerator load) measures only 1/2 of what was measured at the problem outlet.

 

Doing the calculations seemed to support a large capacitive load that the refrigerator along with the series inductance to that outlet might be just the right LRC combination to be resonant. I do qualify that I will not feel comfortable with assuming that is the case. I will remove the refrigerator one day and test further.

 

Right now I am not adding Insteon devices to that circuit and this situation does not appear to "reflect" back at the service and so it is not a large priority yet.

 

Who knows maybe I will end up finding a staple that was driven just far enough into the romex to be an issue for Insteon but not at 60Hz I really have to wait to test more.

[ELA]

Thanks IndyMike,

Always a pleasure to exchange information with you.

 

I agree, and I never expected a refrigerator to be a signal sucker of much magnitude in the past (qualifying I still have not been able to measure it directly to say it is now).

I did replace a very old unit in the past year with a new one with a lot of electronics. I would not be surprised if this one had an across the line cap for filtering.

 

When I first installed the new refrig. unit I started having communications issues. At first I believed the issues to be (dual band) RF related since the new, large stainless steel unit now "blocked" a couple of RF paths. I re-arranged a couple dual band lamplincs and things got better.

 

That was about the time that I decided to "get serious about communications issues"

 

This possible resonance condition really peaked my interest. I will follow up someday in the future just because it is so interesting (even if I do not install any Insteon devices on that circuit in the future).

[MarkJames]

So what devices do you think are the most in need of filterlincs?

 

I have a few areas that I know to be weak. I've got quite a number of UPS's and computers in the house - 11 PC's and maybe 8 UPS's. I understand those to be problems for Insteon. There's a pool but I don't think there's a filterlinc style device which will handle the load of a 220v 20A pump. There's also a hot tub but again no filterlinc for 220v 15A. How about refrigerators and freezers? Plasma TV? Heat pump/AC?

 

I've tried to keep large current drain devices such as pumps and refrigerators etc. on circuits that have no Insteon devices. Indeed I have one sub panel which does most of my media stuff and heating/air conditioning that I've got no Insteon on at all. But if these devices can deteriorate communication along circuits other than their own then I'm not sure if my efforts to isolate Insteon from other devices has been of any use.

 

I'm trying not to allow this to become like the access point nightmare that plagues Insteon users just starting out. It seems that when people first get their Insteon network started they think that adding more and more access points will fix their problems when in fact that's not the case. I don't want to do the same thing with Filterlincs.

 

mark

[ELA]

Hi Mark,

I can only speak to those devices that I have measured personally. There are a lot of posts here about people getting resolution to communication issues by filtering this or that.

 

I was unhappy with the trial and error method. For that reason I developed the tester that I did. It quantifies just how much signal any one device "sucks", or how low of an impedance it appears to be at 131Khz.

I posted several results for some newly tested loads in that other post I pointed to. (sorry but I do not have a pool)

 

Having started out with X-10 I already had 3 filters (Computer, battery charger, LCD TV #1) left from those days that I did not remove when I changed to Insteon.

 

While I am happy to now be aware of just how bad each "signal sucker" is, I would prefer not to blanket filter them all, even if they are above a certain "suck" level. I simply do not want to incur the expense and the ugly wall warts if I do not have to.

 

I have just installed three new devices. I purposely picked one to be a dual band in hopes that the RF section would help me overcome one or more of the known suckers in that area.

I am in the process of doing message reliability checking to help me determine if I can avoid installing some filters.

 

Do you use Houselinc? I do not own a copy but hear it is a good test tool for message reliability testing. I have a similar feature in my personal tester. You can then add/subtract filtering and repeat reliability testing to evaluate the filtering effectiveness.

 

All said I am not aware of any major problem with installing more filters if you feel you need them. (within limits of course).

 

I do feel it is important to make the the power line carrier aspect of the dual mesh network as reliable as you can. I do not want to become too reliant on the RF section or visa versa. I like the idea that each is as strong as it can be in order to make the overall communications as reliable as possible.

 

In my current install I am hoping that between adding devices (thus strengthening the power line via hop usage) and more dual band devices to beef up the RF, that I can avoid some filters on what are known and substantial signal suckers.