Motion delay using integration?

[Unixprogrammer]

I edited a previous reply instead of sending a new one. The ack from the repeater is quick, but I do wonder about a delay from the repeater to the switches.

Thanks very much for your help on this! I really appreciate it!

I have tried a phantom button as I thought sending a single command would speed up the process.

If I use the lutron app on my phone, the lights seem to turn on instantly, so I was wondering if there was some other way to reduce the delay.

Regards,
Paul

[FlyingDiver]

If the app works, then it's not a repeater->switch issue, as the app is sending a command from the repeater as well.

What's the actual command string you're sending?

[Unixprogrammer]

I was previously sending:

#OUTPUT,81,1,50,00:00
#OUTPUT,82,1,50,00:00

I am now sending to phantom button that has 0 delay. I've tried both with and without the release (action id 4).
2015-10-25 09:53:23,260 Sending command '#DEVICE,1,1,3'
2015-10-25 09:53:23,261 Sending command '#DEVICE,1,1,4

[FlyingDiver]

I'm out of ideas, sorry. Maybe someone else can chime in on it.

[Unixprogrammer]

Thanks, I really appreciate all of your help.

[ultravelocity]

Which software are you using? I do the same thing with some custom software I've written and there is a small delay, but definitely not 3-5 seconds. You indicated your delay is happening after the OUTPUT command is sent? That's unusual. If you just telnet to the repeater and send the command manually, does it still take that long? The OUTPUT command should be almost instantaneous. I have less control over how long it takes to get notified of motion. If you're sure the delay is happening after the OUTPUT command is sent, I'm not sure what would cause that. If the delay might be happening before the OUTPUT command, you could check the motion sensitivity on the sensor, and make sure you're getting notified of motion as fast as possible. I believe the fastest way to get notified of motion is to look for "DEVICE" events on your connection. If you're using "GROUP" to get the status of a sensor, perhaps you (or your software) are polling the repeater occasionally, which could introduce a delay depending on how often it polls. Hope that makes sense.

[Unixprogrammer]

It's my own custom software. I am using GROUP to get the status of the motions, and it's trigger based on the controller output, not polled. I should modify it to use the DEVICE output as you suggest.

Also, I thought it was delay based on the OUTPUT command, but I put the motions into test mode and I found that it took a while for them to "see" me. Perhaps I need to adjust the motion sensors as well. I currently have them set to seated sensitivity in hopes that it would be the most sensitive. Any idea if one of the other settings provides faster triggering?

My data for determining where the latency is occurring is based on walking into the room, and watching my log. I see a fast GROUP command followed by OUTPUT, and then the delay for the lights to go on. This is definitely noticeable if I send two commands to turn on the two switches vs a single command to a virtual button.

Thanks,
Paul

[ultravelocity]

Yes I would try using DEVICE first. The DEVICE output from the controller will tell you the sensor status without having to do an additional GROUP command. Also try just sending OUTPUT without any of the other motion stuff and make sure the lights react without any delay (this should be almost instantaneous... I have built sliders that work like the iOS apps, which I believe use the same telnet protocol). Regarding your sensor, the "seated" low activity setting is the most sensitive.

[wkearney99]

It's my own custom software.

The forums here aren't configured for sending PM's. Care to discuss what you're creating? I'm likewise working on some integration scripting.

[Unixprogrammer]

Nothing too grand. I was working on time of day events for motion sensors for some latency sensitive areas (small bathroom). I may just wait now to see what kind of support the new hub has.

My current implementation works, and is based on some python code I found on github. I also implemented a version in c++, which has a very similar feature set (press different virtual buttons depending on time of day).

Regards,
Paul