Devolo Zwave switch activation delays

Hi all. I am looking for some advice regarding an installation using ST, Devolo switches and Fibaro relays to control lighting.

I have installed device handlers and smart apps per the following instructions:

The switches are seen by ST and I can assign buttons to activate/deactivate the relay and the lights connected to it. I can also control the lights via the relays using the IOS ST app.

The problem I’m having is responsiveness. If I use the ST IOS app to trigger the relay, the response (lights on/off) is instant. However using the Devolo switch the response can be anything from instant, to not at all, or anything in between. Occasionally I’ll push the button multiple times without response, wait 15 seconds and then have the lights go on-off-on-off…like the signals got jammed up then delivered all at once.

I assume the reason for the difference of behaviour is that the switch code is executing in the cloud, whereas the IOS app is talking directly to the ST hub so no delay?

Does anyone therefore have any suggestions for…

  1. A more reliable device handler (in case that’s the problem)
  2. A more reliable Zwave light switch (need to control 4 relays)
  3. Other suggestions to overcome the delay

Advice and thoughts appreciated! I’m in the UK if that makes any difference to product recommendations.


The Delay is caused by the following factors:

  1. The Devolo wall controllers (same is true for the Popp and variants) are woefully buggy - however this in itself is not as bad as reason 2
  2. Smartthings - the platform itself here is the problem. You press the switch, there may be a split second delay from the controller, but then it goes to your hub, then to the smartthings server where it is processed and then the correct command is then sent back to your hub, where it is then issued. If the device you are ultimately controlling is cloud controlled (e.g. LIFX) then there is even more processing where the command needs to be issued to the LIFX cloud and then finally to your lights. The speed that all of these processes take depends on your home internet connection speed, the current server load on the Smartthings server, their connection speed, and the quality of your home z-wave setup
  3. Z-wave setup. Smartthings does not tend to make a good z-wave mesh setup, you can optimise it as much as possible (lots of articles on this on the forums) but despite a lot of work on my system I often have dropped signals and lags at seemingly random times.

Unfortunately as you can see most of this is out of your control.

My advice is to try and find one of the Sunricher wall controllers (they are the same standard size as the Devolo so can be paired with nice face plates etc) - but make sure you have a variant that supports central scene commands. They seem to be less buggy, so at least the lag can’t come from there. Then spend some time optimising your z-wave network - again this can shave some time of the lag. Finally, if you still can’t stand it (I’ve managed to live with it for a while but ultimately I want to fix it for good) you’ll have to switch platforms. I ran a test with OpenHAB in my house recently where I had both the Devolo controller and LIFX lights directly connected locally to an OpenHAB hub, and there was zero lag (well from a human point of view) between pressing the button and seeing the result. Felt like actual lights! Unfortunately OpenHAB has a steep learning curve to set up fully, so I have been eyeing the Hubitat platform, which supposedly works similar to smartthings but locally.

1 Like

Hi Adam

I appreciated the fast and detailed response - thank you.

You have confirmed my suspicion that the variable delay is due to the Cloud execution. I will take a look at Zwave optimization articles, but the Hubitat does look very interesting - despite the maturity and ubiquity of Cloud, I still feel more in control when everything is on-premise and under my control (at least I know who to blame if it goes wrong!).

OpenHab I’ll take a look at as well, and see how steep the learning curve looks. I’m not trying to do anything too complicated (famous last words…) and since I already run a server at home this might be an option.

Again - thanks for the prompt and valuable answer.


There are several other systems out there that have local processing versus “cloud” processing like SmartThings. However there is a gotcha people don’t consider when thinking of the local vs cloud processing.

  1. System Resources
  2. Learning Curve in system setup

For item 1: Resources. There are generally 2 types of systems out there. “Server Based” and “Hub Based”. A Hub based system is limited and constrained to a finite set of resources (CPU, Memory, IO capability). A Server based system has no such limitation beyond your own wallet.

System resources becomes very important when thinking of “apps” and “rules/events/automations” and the numbers of devices and drivers etc. This is not often thought about with “cloud” based systems as the resources are infinite (perceiably) and people go crazy with integrations, apps, etc because it all runs in the cloud and yes there’s lag and slowness but it eventually works right? Eventually…

Now if you take all of those apps and drivers and try to shoe-horn them into a small hub based system. Well the outcome is not pretty and people then begin complaining about the “hub” not working or being unstable and all sorts of “weird” problems can start happening that are caused by those third party apps or due to a lack of resources causing unexpected system level problems.

For point 2: Learning Curve
Hub based systems are generally “easy” and “consumer friendly”. The typical Best Buy, Home Depot, Lowes shopper is the “target audience” for these systems which is why they sit next to Hue and Nest and Ring etc etc. The flip side to a Server Based system is the learning curve necessary just to get simple things working not even going into the deeper depths of capabilities that are possible. If you have the time/patience or money to learn and deploy one of these systems or pay for professional installation it’s well worth it and you will be very happy.

You’ve gotten lots of good information, but I wanted to just clear up one misconception: your phone app does not talk directly to the hub. The phone app talks to the cloud and the cloud talks to the hub. It doesn’t have to be that way, but it is the way that the smartthings platform works. If your Internet connection is not working or the SmartThings cloud is unavailable, you won’t be able to control your devices with your phone app. :disappointed_relieved:

Manual, on-demand control of a device or SmartApp through the SmartThings mobile app always requires an internet connection to the cloud and cannot be performed locally.

1 Like

That’s interesting - I hadn’t realized that. I just assumed the instant response when using phone was due to local processing. Your point means that at least some of the delay is more down to the “woefully buggy” switches (quoting Adam :slight_smile:) and might be better with other switches - I’ll give that a go too.


Good point that hubs are resource constrained. I don’t plan on scaling massively so hope that wouldn’t become a problem, but a local server would remove that potential future barrier.

I’m lucky enough to have a server on-site already and a fairly comprehensive network, so capacity is not a problem from that perspective. I also work in a technical job so am not averse to a learning curve or two!

1 Like

It is rarely planned :slight_smile:

There’s a thread on the Vera forum you may find of interest to see the “problem” first hand. Vera is a local processing Hub and when asked what should be integrated native people comes out of the wood works asking for everything you can think of…

This thread is the life and death of “Hub Based” systems that “try” to be local as they run out of resources to make everyone happy.,103048.0/topicseen.html