On/off delay

[MarkJames]

I just installed 4 banks of LED lighting which I'm controlling through an Insteon scene. I've run into a strange issue which I've never encountered before and don't understand.

 

The lights are all connected to a 2475S2 inlinelinc. I just switched it to this from a 2476S dimmer to make sure the problem wasn't somehow ramp rate related.

 

The scene has 4 controllers - all keypadlincs - 3 are brand new kpl8's with the latest rev firmware and one is a couple of years old. The LED light switch is the only responder in the scene. When I turn the scene on from any of the controllers there's a several second lag before the switch turns on. Same thing when it turns off. It's reminiscent of how some of my x-10 used to work.

 

Any thoughts as to why there might be this lag?

 

mark

[LeeG]

MarkJames

 

With two responder devices, even of different types, having the same result I would think the problem is powerline communication related. The Controller (KPL in this case) will retry up to three times if the responder's ACK is not received. Sounds like it is taking multiple On commands for either of the responders to react. Try disconnecting the LEDs and see it the InLineLinc responds immediately. The status LED will indicate when the InLineLinc turns On. Could be load specific or something about that specific electrical circuit.

 

Lee

[MarkJames]

Thanks, Lee,

 

I'll give that a try. I thought perhaps it had something to do with the LED transformer load and maybe others had experienced a similar situation. Oddly if I turn it on and off from the ISY interface it responds quickly though repeated on/off cycles in close succession seems to 'confuse' it. That would support the idea of the ACKs not being received quickly.

 

makr

[LeeG]

The KPLs are likely on different electrical circuits than the PLM. Some folks have had success with LEDs. Others have had problems. I don't have any direct experience with LEDs but they are likely to be like CFLs. Some work fine, others destroy the ability to communicate over the powerline. If it is the transformer/LEDs disconnecting the InLineLinc load wire and watching the InLineLinc status LED should show any delay in reacting to a KPL button press.

 

There is also a difference in the Insteon commands involved between a KPL button press and a Direct command from the Admin Console. A button/paddle press starts out with a Group Broadcast which normally causes all responders to react simultaneously. That command cannot be ACKed so the Controller follows that up with a Group Cleanup Direct command to each responder. This command is ACKed. If the Group Broadcast was not received by the InLineLInc it would be the second command issued by the KPL before it was known there is a communications problem.

[MarkJames]

Well, that was interesting.

 

I disconnected the LED's and found it made no difference to the lag.

 

So I checked the leg it was connected to and noted an ioLinc plugged in to an outlet on that leg.

 

I unplugged the ioLinc and the lag went away.

 

That seems kinda counterintuitive as I would not have expected the ioLinc to cause a problem. Perhaps there's a problem with this particular ioLinc.

 

Thanks for the help, Lee. I'm not certain what caused the problem but at least it's gone away.

 

mark

[Hurting2Ride]

Do you get the lag back if you plug in the iolinc in a different location, circuit and/or phase?

[MarkJames]

Well, now I'm not entirely sure as the lag is still there, though subjectively less.

 

I think I'm gonna put the LED lights on a different circuit and see if that helps. They act very oddly on that circuit. If I cycle them on/off too quickly (and I don't mean very quickly - just on - 2 second pause - off - 2 second pause - on etc.) They fall out of sync with what I'm doing.

 

This looks very much like what you were describing, Lee. How long do Insteon switches wait for an Ack before rebroadcasting? That might shed some light on this.

 

mark

[LeeG]

MarkJames

 

I don't have a specific number the Controller waits before sending the next retry. I think it will be measured in hundreds of milliseconds and it does vary depending on the max hop count being used at the time. The larger the max hop count the slower (relative) thru put as the responder has to wait the maximum possible time it could take a message to arrive so the next available time slot to send an ACK can be calculated. This means the smaller the max hop count the sooner the Controller can reissue if the command is not ACKed. A button/paddle press command starts out with a max hop count of 1 for performance reasons. If that is not ACKed the next retry increments the max hop count, and so on.

 

It gets complex when getting into next available time slots and hop counts. There is an insteondetails.pdf document on the Smartlabs web site that goes into great detail on this subject if you want to understand the details of what is going on. Far too much information to put in a post.

 

Was not sure if you meant the delay was still there with the I/O Linc unplugged or some amount of delay came back when the I/O Linc was plugged in again.

 

Lee

[MarkJames]

Thanks, Lee,

 

I always appreciate your posts. I've read that pdf file several times - or at least tried to. It reminds me very much of Stephen Hawking's 'A Brief History of Time'. I get about halfway through thinking I've got it but then it starts to slip away.

 

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

[IndyMike]

Hi Mark,

 

I've been watching from a distance and have not had anything to add because Lee's diagnosis has (as usual) been excellent.

 

You're last post indicated that the "off command" may be working properly, while the "on command" shows the delay. This is rather odd unless the powerline is interacting with the installed devices on this circuit.

 

• [*:14x3knq3]Is this a GFCI or AFCI circuit?
  [*:14x3knq3]If turning on the LED transformer does improve communications, it could be performing the function of a line terminator. Could you try plugging in a Incandescent lamp in the IOlinc location?

 

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

[MarkJames]

Thanks, Mike,

 

I can try doing that - I can simply disconnect the LED transformers (4 of them) temporarily and hook up a trouble light. I've got a charity golf tournament to go to today and tomorrow (I love summer!) but will try and do that tomorrow eve and post back.

 

Thanks for the input,

 

mark

[ELA]

How long do Insteon switches wait for an Ack before rebroadcasting? That might shed some light on this.

mark

Hi Mark, I have been doing a lot of testing on Insteon communications recently.

From this I can tell you that I allow for 2-3.5 seconds as a worst case (that I have observed thus far) scenario between attempts to send a new message.

 

As Lee Mentioned the time can vary greatly based on the number of hops, whether standard or an extended message, and whether or not retries are at play.

 

For a worst case of an extended message, 3 hops, and 4 possible retries allow for the ~3.5 seconds.

 

It can be rather painstaking to log messages and to figure out what they are telling you. There are many here who are very good at doing this.

If you record some communications logs you could learn what to look for over time with other peoples help and use that to help gauge improvements as you move things around.

[ELA]

Slight correction I increased the numbers in my last post based on these recordings:

 

[MarkJames]

Well, my problem is now resolved.

 

First off, thanks, ELA for posting those traces. That's useful information to keep in mind while troubleshooting.

 

I disconnected the LED's and connected a standard incandescent bulb to the switch as a test. As the insteon device was a switchlinc I didn't expect this to make a difference but Insteon has taken me by surprise on more than one occasion. There was no difference in the lag with an incandescent bulb attached. I also noted that ISY was having difficulty querying this device.

 

I then disconnected the switchlinc from the circuit it was on and connected it to a different circuit. The lag immediately went away and query worked as expected.

 

So... the long and the short of it is that the circuit it was on seems to have some communications difficulty. In this case it was an easy fix as it was in my media closet not 18" from the panel so it was a simple task to switch it around. I could see this being a royal PITA were this somewhere else in the house, though. I do have other areas in the house that cause me difficulty and will keep this experience in mind as I troubleshoot.

 

It's a shame there's no easy way to connect a device somewhere in the house and diagnose Insteon connection difficulties. Between faulty insteon devices, poor communication areas, poor signal bridging, and the occasional lapse in programming/understanding it can get to be pretty time-consuming to find a problem that might not even be something that can be fixed.

 

Thanks for the help,

 

mark

[ELA]

Well, my problem is now resolved.

 

It's a shame there's no easy way to connect a device somewhere in the house and diagnose Insteon connection difficulties. Between faulty insteon devices, poor communication areas, poor signal bridging, and the occasional lapse in programming/understanding it can get to be pretty time-consuming to find a problem that might not even be something that can be fixed.

 

Thanks for the help,

 

mark

 

Glad to hear you found it Mark,

 

I have been working on a device that I think is becoming a great diagnostic tool. Don't know if I will ever be able to offer it to others but I remain hopeful.

http://forum.universal-devices.com/viewtopic.php?f=28&t=5923&start=60

 

 

This device can provide you with detailed information about how friendly any one outlet may be towards an Insteon device being located there. It identifies signal suckers immediately and quantifies them. Using this tool I am now discovering all sorts of interesting things (for an engineer geek that is

[oberkc]

So... the long and the short of it is that the circuit it was on seems to have some communications difficulty. In this case it was an easy fix as it was in my media closet not 18" from the panel so it was a simple task to switch it around. I could see this being a royal PITA were this somewhere else in the house, though.

 

Perhaps I misunderstand how you "switch it around", but this type of situation is why they make filters. Given my experience, and the perceptions from reading this forum and others, I would put a filter on your media closet devices if you have not already done so. I don't know how much power it draws, so it may take more than one, but I would do this regardless of what circuit you have put this on or how you switched it around to solve your problem. If you have not done the same thing for your computer system and peripherals, I would put a filter here, as well (just make sure your PLM is not on the filtered power).

 

I have followed this thread with interest (not posting for lack of anything meaningful to add) because I have a scene with several LED-based fixtures in my house which experiences a delay (even failure to respond, on occasion). There was no problem when these were incandescent, so I attribute this problem to the fixtures. I know that your own experiments pointed to other causes than the fixtures, but I would not completely rule this out as a possible contributor.

[MarkJames]

Well, I guess what it comes down to for me, oberkc, is a reluctance to spend any money on filters. Having had a very large X-10 installation which I've only recently completed switching over to Insteon I'm far from convinced of the efficacy of filters. Despite signal bridges, amplifiers, filters and boosters my X-10 installation never worked to my satisfaction. I have a huge collection - over 200 or so X-10 devices - all sitting in a box - all worthless - as a reminder of my days of throwing money at the problem.

 

In my particular case I was able to simply route the power source to the switchlinc from a different breaker. That resolved the problem instantly. There were numerous other devices on the first circuit that I was using but at over $30 per filter I'm not inclined to dump a few hundred dollars into this problem when an hour or so of rewiring has the desired effect.

 

Of course each person's situation is different and in my case this was the most direct approach. YMMV.

 

mark

[oberkc]

I understand your position. So long as it is working, I think I would remain unconcerned. Still, my money is on the prediction that this problem will come back for you some day in the future. My experience has been that problems don't completely go away by moving things to a different circuit. They may be marginalized, but they remain....waiting to bite. For now, though, I am guessing you will keep your eye on it.

 

I notice that you have "filters" in your list of worthless X-10 items. My suggestion would be to use one of those. Better yet, keep them in waiting. If (or when) you experience problems in the future, try one to see if it helps.

 

I am glad your system is working.

[MarkJames]

Well, you see, now there's another little 'rub' that got me a few years back.

 

I'd initially purchased quite a few boosterlincs to augment my x-10 signal. Turns out those were not insteon-friendly. Of course the fine people at Smarthome now make a new boosterlinc that IS insteon friendly. That's great but it doesn't do much for the coupla hundred dollars worth of boosterlincs I can't use. I know that being on the bleeding edge can be an expensive place to stand but one does become 'gun shy' after a few thousand dollars.

 

mark

[ELA]

I also got tired of playing with X-10 (lot of stuff in boxes) and upgraded to Insteon thinking it would have a lessor need for filtering. I have since come to the conclusion that filtering is a must (maybe not for every installation but probably for most).

 

I just dropped $90 on three more filterlincs yet to be installed, but I want them on hand as I enlarge my network into an area that did not previously have any Insteon devices. I know I have three large signal suckers in that area. An LCD TV, a Satellite receiver (& power supply) and an Air Cleaner. They are all a large distance from each other and so I cannot use one filter for all

 

I will be installing two ICon Relay switches and one Dual band Keypadlinc initially. I plan on investigating if the dual band unit will help me avoid having to install one or more filters.

 

 

I agree with oberkc that you are best to filter the offender or it will probably surface again in the future, especially if you expand your system.

[MarkJames]

I just dropped $90 on three more filterlincs yet to be installed, but I want them on hand as I enlarge my network into an area that did not previously have any Insteon devices. I know I have three large signal suckers in that area. An LCD TV, a Satellite receiver (& power supply) and an Air Cleaner. They are all a large distance from each other and so I cannot use one filter for all

 

Have you used your diagnostic tool to verify the difference between before and after installing a filterlinc? Does the filterlinc affect solely devices on the same breaker as it's installed or will it make a system-wide difference? I suppose that would depend on how many other Insteon devices are on the network repeating the messages. I guess what I'm getting at is if you had an LCD TV and Sat receiver on a home run would a filterlinc still make a difference to overall system message quality?

 

mark

[LeeG]

The FilterLinc has a filtered outlet which the appliance/device plugs into. It effectively isolates the appliance/device from the powerline, preventing noise and Insteon signal attenuation by the device(s) plugged into the FilterLinc. How much effect it has on overall powerline operation depends on how badly the now filtered device was affecting the powerline.

[MarkJames]

The FilterLinc has a filtered outlet which the appliance/device plugs into. It effectively isolates the appliance/device from the powerline, preventing noise and Insteon signal attenuation by the device(s) plugged into the FilterLinc. How much effect it has on overall powerline operation depends on how badly the now filtered device was affecting the powerline.

 

Right..

 

I'm familiar with how they're used. It's just the way I see the power flowing through the house makes me wonder how they perform.

 

I see the power running through the house as a bunch of loops which all emanate from the panel. In each loop the power flows out from the bus through a breaker, through any devices on that circuit, and back to where neutral and ground peacefully coexist at the panel. Each loop should really be its own 'entity' with the only common spot being the bus.

 

If that is the case (and I hope so as it's fundamental to my understanding of the circuits in my house) then I can see the function of a filterlinc only in the context of it preventing random device noise that looks/acts like Insteon messages from returning to the panel and distributing throughout the hosue. The notion of it preventing a device from being a 'signal sucker' only makes sense as it relates to devices downstream from the sucker on the same loop. The Insteon messages from devices on different loops (circuits) should be unaffected in strength by a device down either one.

 

So this is why I'm curious as to how a filterlinc can act to strengthen an entire network. If it's acting to prevent noise injection by some devices then OK - I can see that. But preventing a device from 'absorbing' the signal only makes sense to me within one circuit.

 

mark

[LeeG]

Powerline problems are not isolated by the circuit breaker protecting the circuit. A device that is injecting noise or absorbing the Insteon signal on the powerline can affect more than the circuit being by protected that one breaker. Depends on how far away from the breaker panel the bad device is located. How large is the noise being generated or how much Insteon signal is being absorbed. A circuit breaker has no electronics to block noise or prevent Insteon signal absorption from affecting other circuits.

 

Circuit breakers make good diagnostic tools at the gross level as they generally turn Off many appliances/devices. The circuit is actually broken, isolated from the rest of the electrical circuits in the house. Just turning Off devices is not always a good test as many devices keep their power supplies turned On so that IR devices can activate the device, clocks keep power to maintain accurate time, etc. Turning Off a UPS will not stop it from absorbing Insteon signals. Not all UPS devices do absorb Insteon signals but many do. I have my UPS and all the peripherals isolated with a single 10A FilterLinc. Often a single FilterLinc can be used to isolate all audio/video equipment which are located in a common place.

[ELA]

Hi Mark,

If you have time please read through the posts I pointed to earlier as I provided some of what you are asking there.

Yes if there are a few large sucking devices together, such as I mentioned are in my master bedroom ( especially when located at the end of the home run from the service) those effects can be measured all the way to the service.

I have measured the suck level both at the bedroom and at the service and when the problematic loads are unplugged the suck level is greatly reduced at the service as well.

 

Your home AC network becomes a much more complicated network at the 131Khz Insteon frequencies and combination's of wire inductance and signal sucker capacitance's can result in localized resonances that can have a very large effect at points a fair way away from the signal sucker itself.

The network impedance can vary greatly from one point to anther and it becomes difficult to generalize.

 

While developing my tester I ran into a very interesting situation where one particular kitchen outlet registers the largest suck level in the entire house. There is nothing plugged into that outlet and nothing in any of the immediate outlets nearby (which all register a lower suck level).

 

I believe that is due to my refrigerator causing the resonant condition I mentioned. I am yet to prove or disprove this until I am ready to move a rather large refrigerator just so that I can unplug it for further testing.

I have a pretty large 20 amp wire-in filter that I will then put in place if the refrigerator appears to require it. When talking about going one step too far in filtering I would consider this one to take the cake. I really hope I am wrong and the refrigerator is not as large of a sucker as I suspect. ( or I may find some hidden device that I have yet to realize).

 

My filterlinc etc. order did not arrive today, hopefully tomorrow. I will post my results in the other thread once I have had time to install the new devices and test the filtering that might be required.