Insteon Hub

[David]

Hi Everyone,

My Insteon PLM died on me... again!! So, I bought an Insteon Hub, which I found cheaper than the PLM and I'm hoping doesn't die as often. Of course, up until a week ago, the Hub didn't work with Pytomation. It works well now, using the local API (staying out of the "cloud"). My dopeyrealm instance, available in the contrib folder uses the Hub and if you look at the Git history, the previous version uses the PLM, allowing you to see the difference fairly easily.

There are a few gotcha's so far:

1. The Hub doesn't seem too work well with reading and writing at the same time, as a write will wipe out the reply for the previous write.
   
1. As a workaround the Hub interface only allows a write, after the previous write has received a reply.
2. Upon getting a status update from a light (someone turned a light on) Insteon (even the PLM) will only say that the light was turned on, not at which level.
   
1. Both the PLM and Hub interfaces work around this by sending requesting the true on level immediately after.
   
2. This works will with the PLM, but not with the Hub.
   
1. For some reason it only works with the Hub if I'm debugging it and set a break-point (no, having the app wait/sleep for a bit doesn't help), otherwise the Hub returns an invalid message, rather than the level of the device.
3. The PLM interface erroneously thinks a KepadLinc switch is turning on, if any device on any of the other KeypadLinc buttons is turned on; the Hub Interface doesn't exhibit this behavior.
   
1. Neither currently contains the logic to capture the extra button presses (so you could have an Insteon KeypadLinc button control a Open Zwave device, for example).
   

P.S.
I know I have been unavailable for a while, but I haven't abandoned this project. A series of events has limited the time I've been able to spend on this, but I still work on it. I almost had a replacement for the web client logic, to upgrade it to new technology and make it easier to maintain, but a computer crash wiped that work before I committed it to GitHub. I'll probably pick up the web client change again sometime next year.

[Jason Sharpee]

Howdy,

   Yeah, once upon a time same situation happened to me.   I even wrote a driver for the hub as well and found the same shortcomings frustrating.   I found a good document on replacing the plm capacitors which cause the failures and performed that surgery.  My plm was brought back from the dead and has been bullet proof for the past 5 years.

Maybe that helps.  I can try and find the source if you like

-J

...

[David]

I'm kicking myself for tossing the PLM now, but the Hub is working well enough for now. It's limitation doesn't seem to affect practical use.

I might go back to a PLM in the future and if I do I'd replace those capacitors like your saying. If you can't find the instructions, I'll dig it up.

Both of PLM interfaces needs work. At this point the Hub interface works better, without the incorrect statuses the PLM interfaces have. Of course, this is due to the rewrite. If I switch back to the PLM I'll likely update the interface I just wrote to be universal. The core commands are the same, so it's very possible.