My new Z-Net performs worse than X-10!

[tahl]

I "upgraded" to the new Z-NET Network Z-Wave Plus Controller from the old Z-troller so I could get z-wave plus. But ever since, I've been getting an extremely high levels of z-wave errors. It is now more than two a day. I have optimized my network several times and "fully optimized" the network as well. More devices have more than a dozen neighbors. The most common error is:

"Z-Wave Error; Failed sending Z-Wave command to device: ..."

This high failure rate is making the system unusable. Thoughts? Recommendations?

[sparkman]

Have you read this: https://forums.homeseer.com/forum/li...works-11-12-18? Excess z-wave traffic is a common reason. How often are you polling your devices? You don't give any info on your network so hard to provide specific advise. What version of the plugin are you running, how many and what type devices do you have, how many are battery powered versus line powered, etc.?

[AllHailJ]

I am assuming that you are on the latest z-wave plugin and there is no interference from other devices (wireless phones)??? When I get this type of behavior it usually means I have a z-wave device beginning to fail. Any other weird behavior like taking a long time to activate events? Can you hear any relays cycling at switches? Is the location of the Z-net different than the Z-Troller? This could change behavior if the conditions are just right. Are the failures at random times and on random devices? If this is not the case it may be pointing to the culprit.

[tahl]

Thanks guys for getting back.

sparkman, here's my system config:

Current Date/Time: 2/24/2020 2:00:42 PM
HomeSeer Version: HS3 Pro Edition 3.0.0.548
Operating System: Microsoft Windows 10 Pro - Work Station
System Uptime: 9 Days 2 Hours 22 Minutes 31 Seconds
IP Address: 10.0.0.6
Number of Devices: 358
Number of Events: 694
Available Threads: 1200
HSTouch Enabled: True
Event Threads: 346
Event Trigger Eval Queue: 0
Event Trigger Priority Eval Queue: 0
Device Exec Queue: 0
HSTouch Event Queue: 0
Email Send Queue: 0
Anti Virus Installed: Windows Defender Bitdefender Antivirus
In Virtual Machine: No MFG: asrock
Enabled Plug-Ins
5.0.0.58: Global Cache Pro
1.2.0.4: Zodiac Iaqualink
3.0.1.252: Z-Wave

See attachment. I have like 53 devices, a mix of z-wave and z-wave plus. I try to have as few battery-powered as possible but around 3 I think. As much as I like the modules with all the bells and whistles like power measurement, I try to constrain myself because I am concerned with the traffic as well. I have not, however, turned off polling for anything; everything is left at default. Notably, my network is not much different than when I had the Z-troller which as far as I know NEVER had a failure in many years.

Thanks for the article. I had not read that. Based on a quick read, following the instructions are fraught with risk and might be time-consuming. I'll read it more carefully.

AllHailJ, I don't have a wireless phone to interfere and the location of the Z-Net is about where the Z-troller was. There is no detectable pattern, per se, except that the HSM-200s tend to fail more often, but other than them it seems completely random. Yes, I've also occasionally experienced a longish delay (10 seconds or more) where z-wave activity should be happening (based on the events, etc.) and isn't.

[sparkman]

In my opinion the default polling is way too often and I would start by disabling all polling to see if the situation has improved. If it has, then selectively start turning on polling for the devices where you need it. Think about how often you need the info and why. Is the node number for the z-net the same as what the z-troller was? If it worked fine with the z-troller, you may be getting a lot of retries to by devices still trying to reach it through whatever routes they had to it.

PS Looks like you attached the wrong doc, so you may want to delete that one ASAP.

[tahl]

sparkman, updated document.

I can start trying to turn off polling.... Just to be clear, however.... When I think of "polling", I think of HS/Z-wave reaching out to the device to ask status. I have not turned on "polling" in that sense which is apparently the sense that the Zwave tab of the device means. I can, however, turn off "reporting". The Aerotec multi-sensor, for example, reports a fair amount of data. I can turn it off / reduce it. On other devices, like the Neo appliance module, which I love because they are so small, don't have reporting configurable.

What do you mean the "node number" for the z-net? The Z-troller was connected via serial and the z-net is an ethernet device. So pretty different. Also, the interface ID is completely different. So I'm not sure what you mean....

Thanks for your advice.

[sparkman]

sparkman, updated document.

I can start trying to turn off polling.... Just to be clear, however.... When I think of "polling", I think of HS/Z-wave reaching out to the device to ask status. I have not turned on "polling" in that sense which is apparently the sense that the Zwave tab of the device means. I can, however, turn off "reporting". The Aerotec multi-sensor, for example, reports a fair amount of data. I can turn it off / reduce it. On other devices, like the Neo appliance module, which I love because they are so small, don't have reporting configurable.

What do you mean the "node number" for the z-net? The Z-troller was connected via serial and the z-net is an ethernet device. So pretty different. Also, the interface ID is completely different. So I'm not sure what you mean....

Thanks for your advice.

Every z-wave device has a z-wave node number including the controller. In many systems, the controller is node 1. Depending on what process you used to swap the controller, the new controller could still be node 1, or could be a completely different number.

For sensors like the Aeotec, yes, reducing how much they send can have a dramatic impact on overall z-waver performance. Look at minimizing it as much as possible while still allowing you to trigger events as needed based on what they send.

[sparkman]

Looking at your PDF, it looks like your z-net is still Node 1, however, it seems to have very few direct routes.

[tahl]

it seems to have very few direct routes

Is that a good thing or bad thing?

HS Support said I need to turn off Bitdefender too although that sounds like I'll be introducing a risk....

[sparkman]

Is that a good thing or bad thing?

HS Support said I need to turn off Bitdefender too although that sounds like I'll be introducing a risk....

Bad thing. The more devices it can talk to direct, the less routing which speeds up traffic helps. Unless you are browsing sketchy sites with that PC, turning off anti-malware software should be pretty low-risk. Having said that, I can't see it causing the issues you are seeing at all.

[tahl]

Made many changes and it seems to work ... cautiously optimistic:
* Redid the routing of two nodes, using the Zwave plug-in Nodes page, as they seemed a little crazy
* Reduced module reporting (polling) when available and when it is not needed
* Removed 3 non-Plus modules; however, that still leaves 4 battery operated non-Plus and 2 mains powered non-Plus
* I've now sworn off ALL RGB z-wave bulbs since they have a very high error rate; I've tried three brands now

I still don't know why the Z-net controller is listed as 40kbps since it is Z-wave Plus. It doesn't seem to matter since many of the routings are 100k...

(PS. This whole thing is reminiscent of X-10. Started out as plug n play but eventually required a lot of "tuning" to get it to work later on....)

[racerfern]

I struggled with this when I installed a z-wave lock for one of the doors. It would regularly fail to lock/unlock automatically at a certain hour. I would occasionally get other errors but this one really annoyed me.

It was ALL solved by eliminating polling on devices. Battery operated devices are the biggest culprits because they're not going to wake up when they get polled. They're not awake so they don't know they've been polled. I have six z-wave smoke/CO2 devices and they wake up automatically so there's no need to poll them.

Get rid of polling everywhere you can! I literally went through every child device; it's amazing what you'll find.

I too came from X10. What do you mean it only works for the things on one of my two 120v legs! Just this winter I got rid of my last X10 dimmer when I ran a neutral wire to the switch. Now Qubino comes out with a no neutral needed dimmer, go figure.

[tahl]

What do you mean by "polling"? I might be misunderstanding. I think of polling as HS asking the device how it's doing (HS -> device). Under the Zwave tab of the device (root device as applicable), I do see "polling interval" and for all my devices it is set at 0 hours, 0 minutes and 0 seconds. I've never seen anything else and I've never set a polling interval.

If instead, you are referring to "reporting" (device -> HS) which is found Settings section of the Zwave tab, I have reduced as many reportings as possible but many are sensors so they have to periodically report in.

Please clarify "polling". Thanks.

[racerfern]

Here's a screenshot of a typical motion sensor. I keep it in the attic for a couple of reasons. If a squirrel or other rodent gets in there, the motion sends me an alert via an event. Also if the temp gets too high the attic fan comes on.

Normally we click on the root > z-wave tab and eliminate polling. However this device has three child devices and each of them has the ability to have the polling set on the z-wave tab. So don't just click on the root, click on every child beneath it. It's an ordeal when you have a ton of devices, but you'll see a huge difference if there's a lot of polling that you weren't aware of.

[tahl]

racerfern, thanks. I have couple of _hundred_ devices (if you include child devices) but checking a sample of sensor devices, including child devices, they _all_ have polling at 0,0,0. So I guess I'm good.