Multiple HE's to reduce reliability/comm issues.. thoughts?

yeah, the first production hub has zigbee, z-wave, lan cloud all automations ect, basically everything except the zigbee bulbs...
and so far the ZLL bulbs running as ZHA are playing nicely with the native ZHA bulbs, I wouldn't say with the limited number of bulbs on hub 2 that I can draw a concise conclusion that bulbs route fine for other bulbs yet, I'll report my findings as I move more bulbs over there and putz with it for a while.

1 Like

I did this over Thanksgiving. My reasoning was I replaced half of my switches with Zwave Plus between Feb (was early HE) and Nov and felt there were gremlins in my hub. I had a second dev hub already and made it primary so I could use the other hub to as reference to recreate all my automations. I followed Support/@bobbyD recommendations this go round and let the hub sit every 20 or so devices and even had no rules, apps, nothing for over 24 hours (beauty of doing it over holiday weekend). All has been rock solid since other than my Schlage locks refusing to pair. But in all fairness they were last devices in ST and that hub is now collecting dust. They just refuse to pair to HE.

1 Like

FYI, this was fixed shortly after it was reported


I didn't realize I formally reported it, but that's great news! Still totally my fault, but thanks!


I ditched mine when I move into mesh wifi.
Even my cameras are wifi enabled.

If your mesh units can do "wired backhaul" you might want to use the powerline adapters for that purpose. Just put one at the main WiFi router and the other on the mesh unit furthest away from the main.

What this does is help with the throughput between the farthest unit and the main - increasing mesh stability and speed.

Yeah just finished including all the powered z-wave switches. I tried to relocate my hub back to the most central place in my house - first floor/den closet BUT my basement YRD256 Z-Wave still refuses to pair - nothing shows up. It only works if there is almost a line of sight between it and the hub - down in the basement up in the rafters. I just purchased the Zigbee module and see if I have better luck.

The other thing since I now have a 2nd hub on the way I am probably going to put it upstairs where I have mostly Z-Wave Plus devices & a few Zigbee devices. The basement hub will have mostly Z-Wave & Zigbee devices. It worked well enough before I started messing around .

1 Like

I am really trying to get this to be as simple and reliable as possible not only for myself but also for potential home clients as well.

All these issues (most of my own making) are helping me understand the HE system and it's quirks a lot better. Again probably overkill with the 2 hubs for my house BUT the experience gained is worth it.

It can do wired backhaul. But no need I'm already getting 101% of my subscribed broadband in any spot of my house :sunglasses:

1 Like

So while I'm waiting for the 2nd HE - I'm wondering about HE placement in general.

My original device has been living in the basement but now have been experimenting with placement. In my house I started with Zigbee / Z-Wave devices before Z-Wave+ was a thing that means my downstairs is mostly plain Z-Wave. On my newly renovated 2nd floor the all the light switches are Z-Wave+ (Zooz toggle switches and dimmers).

I'm thinking that if you have to have intermediate devices then using the Z-Wave+ devices rather than the Z-Wave devices might be a better thing?

An ongoing issue I'm having is keeping my Yrd256 lock connected in the basement and the system stable. It's the reason I relocated the HE hub down there in the first place. The problem is some of the further away devices started acting a little flaky - just enough to be annoying. Seemingly random slow or no responses etc.

For fun I moved my HE to the second floor (and ran Z repair) to see if that improved anything..was actually able to pair the basement lock but it's not responsive. Seems like it has a hard time working through powered secondary devices at least the Z-wave ones.

The next thing I'm going to do is replace the Z-Wave+ module on the lock for a Zigbee one and add the 2nd HE to the second floor in a central location with the original living down in the basement.

I may be totally over engineering things but stability is my primary goal. I may have to support this for family and clients...

ZWave signal at 908mhz is going to travel through walls, floors, ceilings, etc. better than Zigbee at 2.4ghz.

The Lock needs a "beaming," always powered, device nearby, say 8-10 ft. Do you have any $1 wall sockets near? Can you replace it with a ZWave+ Outlet? You don't have to use it, just get it added to the Hub and do a repair. For you to be able to join from upstairs today, implies there is a beam'r :slight_smile: down there, but there are certainly devices that don't repeat every nuance of the packets. I'm thinking if you got a known "modern" device down there, it would be your Lock's path to the hub and improve your functionality. Remember, Joining and day-to-day have different goals for engineers. They try and protect the Join, especially on Locks. Once you have the Lock joined completely, I wouldn't experiment a lot with that portion. Move the hub, see if you can control the lock, repair if not. If you do add a modern device, see if that's controllable with the hub in it's new spot.

That's probably how I'd experiment. :smiley:

Yes that is good info to know thanks!

I do have a Zooz ZW+ outlet but it's like 15 ft away. The rest are old aeotec ZW micros that live behind the switches - some are within 2 ft of the lock itself. I could replace them I guess with the newer nanos but they have been working reliably pre-HE for several years and would love to spend the money on other fun HA stuff.

I suspect I just have to get the devices, pairing order & positioning right as well as improve my understanding of how things react with this hub. It is a lot more bare metal than ST which is awesome but perilous if like me you know just enough to be dangerous.

Also my house is as our architect described a "Frankenstein House" - a patchwork of a 120 year old base mixed in with 40's & 70's somewhat poorly constructed "enhancements". No telling what is between the floors and in the walls (some plaster lath, some drywall, beer cans, the usual stuff - mostly headache).

The Aeotec Mini Switch G2 (DSC26103-ZWUS) supports Beaming and Security. Maybe you have a different model.

There's also an issue of hops... and not the beer kind. :slight_smile: ZWave is limited to 4 hops from the Hub to a device. (Otherwise you'd be able to stretch your mesh all the way around your neighbors and have too much fun.)

Therefore, the method a specific router is selected has distance in the equation. It wants to "reach" as far as possible, so it's entirely possible the closest router is 20ft past your lock. There may be perfectly usable routers near, but doesn't get used, because that would be #5.

1 Like

So I had to go check - the one right near the lock is in fact a G2 - that leads me to think that the other one 8 ft way is also G2 since I bought them all around the same time (a PRIOR renovation).

When the hub was in my main floor closet it was also central to the house it seems unlikely that there would be any more than 3 hops but I know wireless can be kinda funky.. ZW repair wouldn't have taken care of that?

I guess that's why it's always good to start with the closest powered units and work your way out..

That's a great initial build strategy. One ZWave repair throws it out the window as they will decide on a completely different topology. During initial build out, you don't really want to wait minutes between each device add by having to not only join, but then do a repair too. Thus you 'force' the mesh by the order of deployment.

Imagine two concentric rings, a ring about 15ft from an imaginary hub and a ring 40ft away. Assuming no obstructions between the rings, the outer ring may contain all of the repeaters. Nothing in the closer ring is a repeater in this thought exercise. They all connect directly to the Hub, so really, repeaters don't get used in this example. Build a wall or some obstruction between a portion of the two rings and everything in the "shadow" of that obstruction (the Hub being the LIGHT in this metaphor) will have lost it's direct access to the hub and thus has to use a repeater to get around the obstruction.

In this example I used distance plus obstruction to cause the signal strength in the 'shadow' to be below the limit, but a ZWave repair might find that a device on the inside ring, can get through the obstruction and be the repeater to a portion of the obstruction, yet in other parts of the obstructed outer ring, the signal goes around the outside first. Two outer ring devices, 2ft apart, one goes around the outer ring, the other goes through the obstruction to a device on the inner ring.

I really, really hope that I'm helping not obscuring. The point is, the hub decides. The hub gathers the metrics from each device, of it's connectivity, and then builds a route table, with distance and 4 hops and maximum redundancy as criteria. It then sends a tiny sliver of the route table, customized, to each device. They do not have the memory to hold a substantial fraction of the table, so no attempt is made to give them one.

1 Like

Yes I think I get it (head explodes). Makes sense thinking it through thanks. Also add to the mix that various devices perform differently and have varying capabilities.

That is kinda my rationale behind adding a 2nd hub on our second floor. If I can limit the hops (:beer:) and strengthen the signal to weaker devices then it should make things more stable/faster. The hublink stuff makes it really compelling I think.

Regarding Link to Hub (and it's mate: Hub Link):

I have two hubs and my initial thinking was to have one for a Lab/experimental device. Although that remains true, it's not pure. I've spent far more time with ZWave than I have with Zigbee. Once upon a time I have 4 different Hubs here, all connected to the same ZWave network, to get the capability I wanted. Some Hubs do some work better (or at all) than others. Zigbee doesn't have that, so it was pretty easy to say to myself: No Zigbee devices.

Once I got down to one Hubitat Hub, the option of Zigbee rose again. Second Hub and 2 Zigbee motion sensors with Link to Hub got me started. Ran into Zigbee's need for repeaters and thus had to add a third Zigbee device, a peanut plug exclusively for repeating.

Fast forward to today and Link to Hub is bidirectional (both hubs are running both apps) and in one direction are 3 devices. The other has (now) 5 Zigbee motion sensors. It's a puzzle to decide, "ship zigbee to the hub with all the ZWave?" or "ship the ZWave devices that Zigbee motion will impact?"

I picked "ship Zigbee to Zwave" The constrictor on our system is the radio. ZWave and Zigbee are both slow relative to WiFi. The older Zwave communicate at 9600 baud, the best will go to 40kb or 100kb. That's still 10x slower than antique WiFi. Modern WiFi uses 'mbit' in their specs :smiley: 56mb, 300mb or more.

Two (or more) radios is good thinking. As long as you can minimize the use of Link to Hub for radio work. I cannot. I chose Zwave on one hub and Zigbee on the other. In other words status (motion) on one and activation on the other. I hope your method is superior. :slight_smile:

While true, an equilivant packet to turn on a Zigbee or Z-wave device vs Http over ethernet is at least 100x smaller...