IOLinc Help Please

I have installed two IOLincs and both show up in the UI as 3 devices as expected but the problem I am having seems to be an extreme delay in what seems like sending the on or off command and update on status. It seems like they might be receiving the commands but a long time after they are sent. I have them on the same circuit as an EZFlora which works fine, but they don’t. Please help if you can thanks.

You probably need to define what an “extreme delay” is. Assuming you are using Altsteon, there are a few things that can build up to cause significant delays in sending commands. Internally, Altsteon queues any commands for devices because there can only be one command “in flight” at a time. If the command that Altsteon is working on isn’t ACKed within 3-5 seconds, it is sent again. This is done three times. After three times, it either discards or sends the command to the back of the queue. So, if you have a device that either doesn’t actually exist on the wire, or a wireless device that isn’t set with the proper DevCat/SubCat combination, Altsteon will get stuck on it for 9-15 seconds every time it is polled. (By default things are polled once every 10 minutes. However, there are some cases where it happens faster than that.)

The other time you will see large delays is when Altsteon first starts up. Before any “useful” commands can be sent, Altsteon needs to gather information about the available devices so that it can send the correct commands to them. IIRC, there are anywhere from 3 to 5 commands that are sent to query the device completely. If any of those commands get lost, or trashed “in flight” you will have the 3 second delay.

Insteon itself can also produce delays. The Insteon protocol attempts to retransmit a few times before failing an attempt. If a device isn’t answering, or is having problems answering, it can cause some delay. (That delay is generally absorbed by the delays in Altsteon, but there may be cases where this isn’t the case.)

Right now, the best way to troubleshoot the issue is to use the Altsteon command line to see what is happening on the wire. If you fire up the daemon, but don’t hit “Reload” in the UI, then none of the devices the Vera knows about will be pushed in to the daemon. This will allow you to manually use the CLI to add just the IOLinc. (However, if any device in the PLM’s link database sends something to the PLM, Altsteon will complain that it doesn’t recognize the device. This isn’t anything to worry about.) Once you manually add the IOLinc, wait ~20-30 seconds for the Altsteon deamon, and the Vera to finish its initial communication. Then, you can try sending commands.

You will want to watch for the status lines from the daemon that indicate that the device responded to your command. If you never see one, then there is an issue talking to your IOLinc. There may be too much noise on the line, or it may be far enough down the line that the commands can’t get to it. If there is a long delay, and then you see the response, you will want to see how many hops are left in the response. To do that, you want to look at the hex portion of the response line. The line will look something like this :

aa.bb.cc:0024,01,00 - Light brightness is 0.

This line is an event for a light switch, but the format will be the same for an IOLinc. The “aa.bb.cc” is the address that sent the message. The 0024 portion is a numeric code that identifies the event. Then, the values separated by commas are parameters. The first parameter is always the number of hops remaining. (Side note : Some commands stay local to the PLM, so they will lie and always claim 2 hops.) The hop count starts at 2 and decrements to 0 before the message is discarded. In the case of the example above, there was 1 hop remaining before it was discarded. Ideally, you want 2 hops remaining, as that would indicate that devices communicate directly with the PLM. However, in the real world you will often see 1 hop.

Here is where it starts to get weird. Depending on a bunch of different factors, you will see different hop counts for the same device. So, seeing a hop count of 0 one time may not mean your device is going to be flaky. If you run the command a bunch of times, and it is ALWAYS 0, then you run the risk of losing messages, which cause state in Altsteon to not be updated in real time. (Altsteon will pick up the state the next time is polls, assuming that communication isn’t lost too.) If you see it bouncing between 0 and 1 hops, you are probably okay. If you see it bounce between 1 and 2, you are in good shape, and if it is always 2 you are golden.

As an example, if I sit and watch the output for the devices in my house for an hour, I will see a good number of them hit a 0 hop count during that time. This will include devices that almost always show a hop count of 2.

If your hop count is consistently low, there are a few things you can do to try to help it. First, make sure you don’t have any devices plugged in to power strips. It isn’t a hard and fast rule, but power strips that have any kind of filter circuit on them will cause lots of problems with your gear. (It is possible to find some that won’t cause issues. I have several devices in my garage plugged in to a power strip. But, you will want to plug in to the bare socket first, and once everything is solid you can introduce the power strip and see what happens.) Second, if possible you can move the device to a better outlet. (Generally not possible, so I won’t dwell on this.) Third, you can install more Insteon devices which may allow the retransmissions to take fewer hops. Fourth, you can install more multiband devices. They will let commands travel over RF when possible, which may be an easier path to travel.

If none of that works, you probably need to look at environmental issues. Run around your house and unplug anything with a motor. (Fridge, washer, dryer, computer, some home theater gear, etc.) And see if it gets better. If not, and you are using multiband devices, unplug anything that might transmit in the 900Mhz frequency band. (Cordless phones, baby monitors, etc.) If that STILL doesn’t make a difference, then you are probably boned. If unplugging things does make it better, then you might be able to fix the problem with a good power strip that has a noise filter.

Oh, one other thing that just popped in to my head. Phases. If your PLM is on one phase, and everything else is on another you will see problems where sometimes commands will manage to get across, but sometimes they won’t. I hit this early on in my implementation. The best way to fix that is to make sure you have a multiband device on both phases that is close enough to one on the other phase that it can talk to it. (Generally, you can tell which phase a circuit is on by looking at the breaker box. All of the breakers on the left are in one phase, all of the breakers on the right are in the other. Assuming your panel is the same as the vast majority of the ones in the US. :slight_smile:

[quote=“fba, post:2, topic:176320”](Generally, you can tell which phase a circuit is on by looking at the breaker box. All of the breakers on the left are in one phase, all of the breakers on the right are in the other. Assuming your panel is the same as the vast majority of the ones in the US. :)[/quote]Not usually. Typically, they alternate as you go down the columns, with each row being the opposite phase of the ones above/below it, that way a double-breaker can feed something like an electric dryer or stove.

I seem to have them working for now but is it normal for the Vera UI to lag on showing the proper on or off state??

I have checked everything as FBA has suggested. I already have access points and an EZFLora to address the phase issue and all linking seems to be correct on the IOLincs and PLM with both A2 and E2 Flags showing on both devices. The IOLincs respond fine but there is a lag on the Vera UI showing proper status on the switch side. I’m not using the sensors so I can’t speak for them. What i have noticed though is that if I manually poll the iolinc the status updates right away. So could this be because the IOlinc is setup as a binary switch or a polling issue. It appears that fba has set them up as binary switches when I look at the files. Any insights?