oberkc

Good to know. Yeah, I suppose. Fortunately, I have enough automation with motion and doors and location that this is rarely the case. I am sure I would find it interesting and fun as well, but I have too many other hobbies and find myself wanting to play golf, play in the wood shop, and do some other projects around the house.

It has varied, but right now it is a small Roku TV. The echo controls it via the harmony remote control. With regards to the echo, I guess I just don't value enough voice control to put the effort into optimizing it. For the things I desire to control, I find conventional methods of control equally as convenient (if not more so).

Nicely done. I hope the logic is coming together in your mind regarding scenes, responders, controllers, etc... Thanks for the education on the echo. I actually have one, and am considering the spot when it becomes available. I remain uncertain whether I like voice commands and having to remember specific syntax, so I don't do a lot with it. Mostly, I use it to turn on my TV and one hue light.

I am mostly unfamiliar with what the echo can do with insteon. Clearly it can run a program. Can it turn on devices? Scenes? Multiple things in response to a voice command? You may have to break the single scene into three: on, single shot, and double shot. On has F as controller and G/H as responder, set to zero on level. Single shot has G as controller and F/H (100%/0)as responder. Double shot has H as controller and F/G (100%/0) as responder. Use the same response levels at the scene level, also. My temptation, after that, would be to create a couple of Alexa-specific programs. One would look something like: if nothing then turn on double-shot scene run Mazzer-Grinder-double shot (then path) The other would be similar and I leave it to you to figure out.

Accomplish this entirely with scenes. Afd.both dimmera to a scene, both as controllers. Pay attention to the responder rates are comsistent for both controllers. To activate via program, be sure to activate via the scene, rather than by one of the switches.

That looks consistent with your stated goals to me. If you are still getting results consistent with post #8, I would consider the possibility that you still have some of your existing programs are affecting button operation. Are your existing programs still active? If so, delete them. Then, create a couple of simple programs: if control button H is turned on then wait 26 seconds turn off scene else nothing If control button G is turned on then wait 17 seconds turn off scene else nothing

You must also establish correct responder levels for each of the controllers in the scene...not just the scene level. When you press buttons F, G, and H, reponders will comply with levels for each of the controller buttons which may or may not be same as scene level or as each other. With your mouse, click on button F within the scene. What are the responder levels for the other devices? Are they different than the scene levels? Make them what you want. Do the same for buttons G and H.

I think like stusviews in this case. Create a single scene with F, H, and G as controllers and appliancelinc as responder. Adjust responder levels independently for each of the three controllers. As an example, for controller F, rsponder levels G and H are set to zero. For controller G, responder F set to on, and respnder G set to zero. Do similar for controller H. Then use programs triggered by G and H to turn off scene after appropriate time. There may even be value in configuring buttons G and H as non-toggle ON.

My motion sensors are the original version (not current version). Based upon my admin panel, they have only three nodes: motion, dusk/dawn, and battery. None of the nodes show a number such as the one in your image.

Perhaps I missed someone else making this suggestion.... Remember, responder levels can be different for each controller in a scene. Be sure to ensure that the ENTRY button response level to the CL2R button is not zero (or something very low). Within the scene definition, select the CL2R button. In the main window it will show responder levels for all devices in that scene. What does the response level for the ENTRY button show?

Scene controllers act as such only when controlled locally. A switch, for example, acts as a scene controller only when you physically press the switch.

Senfing an OFF command to a scene controller wont work. You can try inserting a short wait between the turning off of each individual scene. My preference would be to create a new scene containing all the devices in the four scenes conbined. Then...turn off that combined scene.

It also seems to me that speeds can also be affected by marginal communication. In my experience, if communication is taking multiple hops before success, the slower response times are noticeable.

Hopefully we will see you around here more over the next 15 years.

It sounds as if you use a scene to control the KPL button. If you really don't want to swap out the sensor as many have suggested, you will need to remove the scene and use a program to control the KPL button.

I believe your assumption is correct. The query does not recognize the "reverse" trigger setting. Whether a bug, insteon limitation, or design choice, I don't know. Regardless, I doubt that this is going to get fixed any time soon. Normal solutions are to replace the sensor with NC type. Alternatively, one could remount the sensor such that it is in "normal" position when the door is open, but this is not ideal, in my mind.

The big point is that they are different (though an ON command may also include other parameters that would exceed the 1 byte difference). Unfortunately, in post 86, dreamerI90 claimed "everything is the same expect I have JUST this in the Body field...". Clearly everything is NOT the same.

Don't forget to check your "content-length". As I stated earlier, mine is different for ON and OFF resources. For me, everything else is NOT necessarily the same.

This is what I have in my body field, and it works: {"on": false} The only thing I see different is a space between the ":" and the "f". However, I also notice that there are also some differences in the headers. Specifically, I notice that the "content-length" is different for OFF than for ON

I use the light switch that controls the flood lights as a trigger. If turned on manually, I disable the motion timer program. If control light switch is turned on and control light switch is not turned off then nothing else nothing That program is true if I manually turned on the switch, and false if I manually turn it off. Now I can simply use the true/false status as a condition of the motion timer program. If time is from dark to light and control motion sensor is on and status first program is false then turn on light wait a while turn off light else nothing Assuming that the switch is not manually turned on, the first program will be false and the timer program will operate normally. If the switch is manually turned on, the second program will evaluate false and nothing will happen (and will interrupt the timer if ongoing...a good thing). I would likely find it necessary to add a third program to handle the program timer being interrupted during the transition to "light" if time is light then turn off light Yes, you could use variables. You could also do similar logic with program folder, or by using a program to halt and disable the motion program. Whatever makes sense in your mind. Hopefully, this will give you some ideas.

Ah. I should have been more specific. Good.

Temporarily power your PLM from a long extension cord from another circuit. Does this help? Regardless, I would filter all that computer stuff. I don't care how long it has worked before. Things get old and things stop working. Your PLM is too important to risk interference from that other stuff.

This sounds like "marginal" communications, potentially. Your PLM...is it on a circuit with lots of other computer stuff?

I see nothing wrong with the program. It should work, in my estimation, without the scene. Normally, I understand direct commands are more robust than a scene. You could certainly temporarily add some wait states between each to see if lilyoyo1's theory checks out. I suppose it might be a link record issue, but this seems unlikely since they individually work from the admin panel. There seems to be a lot of talk about PLM failures. Perhaps that is a possibility here, but I have no idea how to determine that. Otherwise, I am running out of explanations.

I have had a dimmer cycle between off and on repeatedly. Is that what you are seeing? In my case, it was a device failure. Definitely check the program logs, however. Maybe Gary Funk is correct and you have a program that has found itself in some sort of loop.