Zooz dimmer responsiveness

Preface: Wife is very sensitive to latency when operating physical dimmer/switch devices. Anything more than a couple hundred milliseconds annoys. So I'm trying to maximize responsiveness...

I recently installed a Zooz Zen72 dimmer (single-pole.) I noticed that enabling Scene Control on the device seems to have introduced a short delay on all physical button presses. Is this expected? Logically, I suspect that any scene device must wait briefly after each tap to ensure it's the final tap, and not part of a multi-tap sequence. I'm assuming that to achieve max responsiveness, I must disable scene controls :frowning:

But more concerning: I've observed that when the device has been idle for some period of time (maybe about 10 minutes), there seems to be an extra ~1/2 second delay on the first operation. It's almost as if the device is in some low-energy "sleep" mode and must reawake before processing the command. Subsequent operations are noticeably faster. That's a bummer, because that first operation response time is really crucial to the user experience -- the device is idle most of the time.

Does this sound normal? Also, as a new Hubitat user, I'm interested in learning any diagnostic troubleshooting tips (logs, ZWave analyzers, whatever) to help quantify these observations and pinpoint the root causes...

Are you noticing the delay when commanding other devices via the hub (from button presses), or is the delay for the load connected directly to the switch? I have never noticed this "sleepy" delay on any of my Zooz devices, but it might be there. After a few weeks you get used to the tiny lag time.

1 Like

It may also be related to the fade in when the switch is activated and the lowest level setting. Some dimmer allow adjustments to what is considered the lowest dim level setting. You may want to make sure that is set at the lowest value your bulbs will work with. Some will go down to 1 % and some will go off at 15% or so.

Also dont forget you can adjust how fast it will ramp up or down in the driver if it doesn't do that move fast enough for you.

Any switch set to scene mode will try to wait a moment to ensure you aren't doing more taps. In scene mode the Zooz switches can do up to a 5 tap. It will delay waiting to see if you tap more times. I have a few Zen74's which are the toggle version of the zen 72. We dont really have any delay with it. Especially from actively hitting the switch. It isn't in scene mode though.

My suggestion would be to turn off scene mode unless you really need it. You can still double tap for things like immediate 100% level and such in the regular mode.

1 Like

Lag is the reason most of my devices are Zigbee. Maybe it’s my setup, but using Zooz dimmers as button controllers has much more lag than my Zigbee button controllers do.

1 Like

Zooz has the Zen34 which is pretty reaponsive as a button device. I haven't fully tested the zooz scene control device i have but expect it to be very responsive since the buttons are single purpose

I think it is all about it having to wait to confirm if the button was pressed a number of times.

Zwave or zigbee also shouldn't matter when you talk about manual control at the switch. Just if you have other devices making changes to the switch.

Turn off scene control to keep the control local to the switch and not reliant on the hub to begin with.

I had a sofisticated setup with a zen 74 controlling a outlet and managing dim level of a wifi smart bulb in it. It worked great until i introduces the scene mode controlling myself. That created delays which we weighed and ended up keeping. But is certainly an impact, but also easily avoidable


I've reviewed some Hubitat logs from earlier in the day that seem to highlight the "sleepy" delay:

Dimmer was used moments earlier (i.e. "warmed up") and level is adjusted about ~200 ms after button event:

dev:332022-05-10 06:50:28.785 am infoTV Room Dimmer: level was set to 1% (digital) [actual: 1]
dev:332022-05-10 06:50:28.588 am infoTV Room Dimmer: button 4 pushed [down 2x]

After being idle for about 10 minutes level is adjusted about ~800 ms after button event:

dev:332022-05-10 06:59:39.257 am infoTV Room Dimmer: level was set to 100% (digital) [actual: 65]
dev:332022-05-10 06:59:38.449 am infoTV Room Dimmer: button 3 pushed [up 2x]

After being idle for another 30 minutes level is again adjusted about ~800 ms after button event:

dev:332022-05-10 07:27:05.020 am infoTV Room Dimmer: level was set to 1% (digital) [actual: 1]
dev:332022-05-10 07:27:04.239 am infoTV Room Dimmer: button 4 pushed [down 2x]

About 30 seconds later (i.e. "warmed up") level is again adjusted about ~200 ms after button event:

dev:332022-05-10 07:27:32.984 am infoTV Room Dimmer: level was set to 100% (digital) [actual: 65]
dev:332022-05-10 07:27:32.778 am infoTV Room Dimmer: button 3 pushed [up 2x]

These intervals are pretty consistent across longer log samples. "Cold" times are around 800 "warm" times around 200.

I wonder if the Hubitat is to blame here. That gap of time between "button pushed" and "level was set" could very well be hub processing time. I'm using Rule Machine to execute setLevel commands in response to button (scene) events. Maybe Hubitat is slow to process them when it has been idle for a while? I guess I need to increase logging of the rule execution to test that theory...

Edit: Hubitat may be the culprit?

Enabled logs for the Rule Machine rule and sure enough, the 800ms "sleepy" delay appears to be related to Hubitat rule execution:

dev:332022-05-10 02:23:01.384 pm infoTV Room Dimmer: switch was turned on (digital)
app:1412022-05-10 02:23:01.056 pm infoAction: Dim: TV Room Dimmer: 100 
app:1412022-05-10 02:23:00.970 pm infoTriggered
app:1412022-05-10 02:23:00.951 pm infoEvent: TV Room Dimmer pushed 3
dev:332022-05-10 02:23:00.553 pm infoTV Room Dimmer: button 3 pushed [up 2x]

Almost 400 ms elapsed between "button 3 pushed" and the "event" log entry! Another 300+ ms between the "action" log entry and the status update from the dimmer "turned on"! Once the system is warmed up, those delays are back in the 10s of milliseconds and everything is more responsive.

Is this known/expected behavior? Hubitat may exhibit processing "warm up" delays when idle for a period of time? (In my case, about 10 minutes.)

Makes sense.

We were experimenting with scene mode as a shortcut for selecting preset brightness levels. Even at the fastest configured setting, press-and-hold to change brightness level requires about 2-3 full seconds of standing there and holding the dimmer to make an adjustment from "max" to "min". My wife finds that excessive; she often reminds me that, with a dumb dimmer, she can instantly adjust brightness by adjusting a physical slider. :weary:

But the delays introduced by enabling scene mode -- both at the hardware/device level to record the extra taps, and (likely) at the hub level to execute automations -- seem to put us right back where we started, with response times of almost a full second.

I suppose the advantage of scenes that at least I can multi-tap and walk away, and trust that the lights will eventually respond, vs standing there at the dimmer with a press-and-hold for several seconds. But that requires building up some confidence in the system...

Simple answer is "Yes". When you have Central Scene enabled, the device has to be able to distinguish between one to five taps. Only a single tap should turn on the device. But when you tap the first time, the device doesn't yet know if you are getting ready to tap a second time so it has to wait a small while to see if there is another tap. If there isn't, it then knows this was a "single tap" and turns the device on (or off, as the case may be).

1 Like

Thank you for confirming. That makes perfect sense.

I think I could live with that extra short device delay if that was the only issue. When I tail the log output and physically press the device, the button event appears after a couple hundred ms. That's within the wife-acceptance threshold :slight_smile:

But tacking on the extra variable ~200-800 of (apparent) hub processing delay to execute the rule action (i.e. adjust light level) is a disappointing result -- puts the overall execution time up around 1 second and puts me into wife-rejection territory :frowning:

My hope/assumption was that, with 100% local execution, Hubitat rule processing would be significantly faster than what I'm seeing. I'm wondering if I've misconfigured something on the hub?

If you are doing basic stuff with the button presses, you might try using the built in button controller app. Would be worth a shot to see if it has a better response time, since it is more dedicated and more simple app. I use the button controller for all my button actions, have not had any issues but I have not really monitored the response time either.


If you're on a C7 hub and you haven't done the Z-wave interface firmware update yet, do that. It also reduces delays.

This is a separate step from updating the hub's firmware and has to be done from the z-wave details page.


You need to think about all the extra steps in what you are doing.

By using the switch as a sceen controller you have a several things that need to take place.

  1. The switch needs to determine how many taps you are doing which adds delays.
  2. The you have to notify the hub over zwave which adds delays. (This can be as high as half a second)
  3. Then the hub has to analyse the event and act accordingly. (This is pretty fast normally unless hub is heavily hit)
  4. Now the hub has to send a command via zwave for to the device to act on
  5. Finally the switch will process the request and change the state.

Each step adds time. Not allot but they do add time.

I know it doesn't help but saying someone can adjust a dimmer instanly means they are lying to themselves. They just dont realize it because of the difference in the engagement with the activity.

1 Like

My wife judges these smart dimmers against non-smart dimmers with physical sliders, which to be fair, are pretty instantaneous. So I've got to prove that the benefits of automation, voice control, etc. are worth sacrificing a little responsiveness. But the more sluggish it is, the harder the sell.

I'm open to suggestions here. I'm not wedded to Zooz dimmers. If there's another smart dimmer out there that supports speedier physical control for adjusting brightness, I'd love to know about it! The Lutron Aurora is a very impressive performer on our Hue bulbs; is there any load-controlling dimmer out there with similar capabilities? I.e. something that allows for speedy swipe/flick to adjust full range of brightness instead of press-and-hold for several seconds?

How often are you really adjusting the dim levels? We have a "dim" preset and full brightness and that's about all we ever use. The only time I adjusted from the paddle was to figure out where we wanted the "dim" setting to be. I have the ON to be the preferred setting (often full bright), 2x UP set to full via the firmware, and then on many devices I setup a button rule for 2X Down to be the preferred dim setting. Most of the time we are using voice control or automations anyway, I hardly ever touch a switch.

I would honestly just set some stuff up and let it settle for a few weeks. In my experience they (family) come around after getting used to it.

1 Like

Fair question! That's precisely why I've been experimenting with scene controls -- to establish a couple of preset brightness levels that can be selected quickly (vs manual up/down brightness adjust). But when the dimmer (or hub) is "cold", those scene events take about 1 second to take effect, which is why I started this topic -- to see whether that could be improved somehow.

But you've clued me in to something important here: The Zooz dimmer has a built-in double-tap up => setLevel behavior that isn't scene-based (i.e. should execute immediately). I disabled that function when I started messing with scene controls, worried that it wouldn't interact properly, but I think that a was a misunderstanding on my part.

Ideally, Zooz would also have a built-in double-tap down behavior, too, for symmetry. But alas, it does not, so accomplishing that seems to requires a scene-based solution (with the associated latency). But I think we can probably live with that single operation being a bit slow, as long as the other operations (on/off/full-bright) are fast!

The ramp up/down speed for dimming is adjustable (at least with the custom driver). I only use mine (Zen77s and Zen34s) for scene control of Hue lights and was initially very disappointed with the latency, but found some settings available in the custom driver (thanks to @jtp10181) very helpful for reducing it. It’s still not as fast as my Zigbee devices, but considering its excellent feature set and Zooz customer service, it’s a pretty good compromise (still waiting for Inovelli’s new Zigbee dimmer to be released).
Edit: I was just experimenting with this and I’m betting that disabling scene control and reducing the ramp rate will help. When the button is held, there is a small delay before the light starts changing (with scene control enabled at least). I would think this might react immediately to ramp down since there isn’t a built in double tap down without scene control.

1 Like

I also would suggest you look to make dimmer adjustments and light color based on time and other criteria.

What I mean by this is i have lights that change color temp and dimmer level based on time of day/sunset/sunrise. I also know some people use lux sensors to make this adjustment. This alows the light to be adjusted to fit the condition better.

I even wrote a rule in RM to change some bulbs based on watching a movie in netflix on roku. It would put the bulbs in movie mode when playing and as soon as I would pause or stop it the lights would return to regular brightness.

Manual control should be kept to simple tasks if possible and the let HE automate what you can. Programmatically identify with to many missed possible conditions.

I still think saying it is faster with a regular dumb switch is a little bit of a illusion. I work in it and had someone on my team years ago argue that comman line was always the fastest and best way to do things. We did a quick test just zipping a file and proved him wrong. In the end it wasn't that the other option was really much faster but allowed the person zipping the file to be more efficient in other ways. On a dimmer the only quick options are 0 and 100%. In the middle you still have to spend time dimming down or up until you find exactly what you want. That isn't instant and as the previous poster said you can adjust the ramp rate. You can tell it to go from 0% to 100% in 2 seconds. That just gives you less granular control but that is the same if you move the slider on a mechanical switch as well.


Thanks for sharing this! Your experience operating Hue lights with Zooz dimmers is very interesting to me. I haven't gotten there yet -- still in a remodel -- but we have a couple dozen Hue bulbs I intend to reintroduce, and would love to control them with Zooz dimmers so lighting controls are consistent throughout the house.

I'm curious what specific settings in the custom driver (for Zooz or for Hue?) helped you achieve better performance. I'll be going down that road pretty soon...

Also curious about your remark re Innovelli Blue. Do you expect it to out-perform Zooz for Hue Control because you'll be setting up some kind of direct association between those devices?

I was just experimenting with this and I’m betting that disabling scene control and reducing the ramp rate will help. When the button is held, there is a small delay before the light starts changing (with scene control enabled at least). I would think this might react immediately to ramp down since there isn’t a built in double tap down without scene control.

I agree that everything is a bit more responsive with scene control disabled. However, even with fastest possible ramp/dim speeds, it requires a 2-3 second press-and-hold interaction to reduce brightness from max => min. That's a pretty common operation in our household, and we're all too impatient to stand at dimmer that long :slight_smile: Hence the scene control workaround (with the understanding that it incurs a small delay on all physical operations; tradeoffs.)

Do you have any associations set up with any of them?

Great suggestion, thanks. I'm still dipping my toe into automation here. My prior experience with smart lighting has been mostly limited to voice/app control -- adjust the lights without getting up off the sofa! -- but automation is definitely the next level.

But what you're saying makes total sense: If automation can anticipate our desired light level based on time of day (or ambient light, or movie mode, etc) then it makes these manual control / responsiveness issues moot!

However, I'll probably need to proceed cautiously there -- easy to imagine these rules becoming a hindrance instead of a help. Guess it's a matter of paying close attention to what our usage habits are and then slowly build automation around it. I feel like I need some machine learning to identify my own behaviors!