Jump to content

Network Failure


scubaguyjohn

Recommended Posts

I've had a 40 device zwave network running flawlessly for a year and now many elements aren't communicating.

 

1.  Is there a best policy in how to add nodes to the network if i decide to start over (ie, closest to controller first)?

 

2.  How many 'hops" can a node use to reach the controller?  I have some devices that aren't communicating which sit very close to a repeater - I'm wondering if i have too many repeaters.

 

3.  how is a 2gig thermostat factory reset to forget the network?  I've used the reset button but it still shows "link" meaning its paired after it reboots.

 

4.  Does a network heal reroute communications based on strength of signal received by individual devices?

 

thanks for any insight!

 

John.

Link to comment
Share on other sites

Z-Wave allows at most 4 hops.

 

I don't think it's possible for a Z-Wave network to have "too many repeaters",   The controller routing table should be able to handle any size network, and at the time of the heal, chooses what it thinks is the optimal route based on what the devices report as their neighbors.   If that route is later degraded, the controller (ISY) will fall back to other less-than-optimal routes.   In all cases, the controller decides what route to use, not the devices in the path.

 

Heal polls the devices for neighbors and rebuilds the route table.  I believe signal strength is part of the calculation, but haven't dug that deep into the details of the protocol.

Link to comment
Share on other sites

1) Best practice is to get all the powered z-wave devices done first (those will function as repeaters), starting with those that are secure-capable first (and add them as secure devices), then moving on to the powered devices that can't do secure links.  This will establish your backbone, so to speak.  Make sure you run a "Heal" on your network every few devices or every device that you add.  Then move on to the battery-powered devices -- and again, heal the network each time you add a device, or move it from close to the ISY (if you had to do that in order to enroll it) to it's final location.  This will ensure that your routes are optimally determined.

 

2) I don't know the exact number allowed by the protocol -- but it's not like the Insteon "hop-based" comms.  Your devices form a mesh, and a message follows a predetermined path through that mesh each time.  If there is a failure, it falls back to another route.  But the devices don't just blindly echo whatever they hear until the hop count reaches zero.  I've never heard of a failure caused by too many repeaters.  See comments at the end.

 

3) I'll leave that for someone who has that device.  Sorry.

 

4) No, signal strength is only a part of the route determination -- the major factor determining the route selection is to minimize the number of hops while ensuring that the connections have sufficient *quality* (which is only partly determined by signal strength).  Also note that if the device in question needs a secure route (for example a barrier device for a garage door), it will determine a route that will allow it to connect securely -- meaning that it requires intermediate devices that would relay the signal to be able to repeat secure data (many don't).

 

If a network heal doesn't address your communications issues, then it's not likely that resetting everything and starting over will fix it either.  Keep in mind that more devices is not always a good thing, and the benefit of many devices stops after you reach a certain density.  Consider an analogy -- you need to relay a message across a large stadium, from on end to another.  It's difficult to shout your message from one end to the other, except when weather conditions are just right, and traffic noise is non-existent, etc.  Your signal (your shout) doesn't change, rather it's the noise that does.  So, you add a couple other people to the stadium, and position them just right, and now you can relay your shout from one end to the other regardless of wind noise and traffic outside the stadium.  Time goes on, and now a football game is in progress.  Theoretically, you should be able to get your message through easily, right?  There's way more potential relay-ers than ever before!  But what are the chances of that actually working?  The problem is that few, if any, of those other noise-makers shouting in the stadium are participating in your conversation, and therefore they won't relay -- and as such, they're just noise.  And that noise is now so loud that you can't make your signal (shout) reach 50 feet.

 

The point is that if your network quality has deteriorated over time, don't look at your z-wave devices alone.  Instead, consider what else you have in your house that's operating near that same frequency (900MHz) -- that info can be found printed on the device or on its packaging and certainly in its manuals.  Since 900MHz is an unlicensed consumer (read: Anything Goes) band, the quality of the other devices is suspect -- it could be a baby monitor, it could be some device to relay your TV remote control from one room to another, etc.)

 

Finally, do take a look at your ISY -- it's the common endpoint for all the zwave comms.  If you've (for example) "neatened" up your installation recently, and put the ISY inside your equipment cabinet along with your DVD/BlueRay/Cable/Sat TV/Audio equipment, well, it might be that it's out of the way now, but it could also be that it's now sheltered by lots of metal, wiring, and electrical noise.  Your z-wave antenna (which is inside the ISY) wants to be in the center of things, not in a box or in the basement or hiding in anything metal.

Link to comment
Share on other sites

wow, thanks guys for the detailed feedback! 

 

After two days of debugging, I was able to trace the problem to a failed switch that was playing a critical role in carrying signals to another floor.  Since there were many devices that weren't working, I looked at the network details (right click on individual nodes in ISY).n  This gave me clues as to where the failure was occurring as most nodes only had one or two "neighbors" and they all pointed back to the same switch.

 

Once i deleted the failed node and reincluded a couple of repeaters, two network heals fixed the rest of the issues.  Now when I look at the number of neighbors, its remarkable how much more robust the overall network is.

 

Again, I appreciate the insight!

 

John.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...