My homeassistant set up is now two days old. The core set up is
- RPi B+
- Philips Hue Hub
- Three Hue Bulbs
- some bits of discovery such as
A large house, so the connection to [bulb 3] is only reachable via a hop from the hub using either [bulb 1] or [bulb 2]. I am getting bursts of time when the lights change state. Mostly but not entirely on [bulb 3]
- 11:30 PM Hue white lamp 3 turned off
- 11:30 PM Hue white lamp 3 turned on
- 11:29 PM Hue white lamp 3 turned off
- 11:29 PM Hue white lamp 3 turned on
- 11:28 PM Hue white lamp 3 turned off
- 11:28 PM Hue white lamp 3 turned on
- 11:27 PM Hue white lamp 3 turned off
- 11:26 PM Hue white lamp 3 turned on
- 11:26 PM Hue white lamp 3 turned off
- 11:25 PM Hue white lamp 3 turned on
- 11:25 PM Hue white lamp 2 turned on <- rare bulb 2 switching
- 11:24 PM Hue white lamp 2 turned off
- 11:24 PM Hue white lamp 3 turned off
with the off/on always being in pairs. About 15 more such pairs (all [bulb 3]) until 1.30 am, then no more during the night.
I have not yet observed this directly, so not clear as to the exact off-on interval.
How have you observed it? I assume you expect it to be HA related so it does not happen when HA is not running?
The times shown are from the HA log at ip:8123/logbook (which are shown to the minute). I am presuming the HA log is telling me the truth that the lights are going off & back on again. But I have not yet personally observed the lights do this.
The times in /logbook are shown to the minute.
Did you check the HA log, any indication they get disconnected?
Am confused by your question. The entries shown are from the log. What I am not sure about is whether:
- the lights are genuinely going off and back on again
- if so, what is causing this; of
- they are staying on, but the log entries are not real.
Check the ‘home-assistant.log’ in your HA configuration folder and let me know
By the way, you can also check the log (not logbook) in the HA about menu. The part below “The following errors have been logged this session:”
Sorry, thought I had replied to this via email - but it doesn’t seem to have been posted.
the log has no entries re switching the lights. Most of the entries in the log are to do with self discovery of other kit I have in the house such as sonos and chromecast
I swapped the light bulbs round. When I did that I put [bulb 1] in the location which is not directly reachable from the hub, but the signal must hop to it via [either of] the other two light bulbs which are close enough to the hub to communicate directly. Now it is [bulb 1] that shows up as having numerous intermittent off/on transitions.
so this phenomenon appears to be associated with an indirectly connected bulb.
I’m not sure if the bulbs show ‘off’ in the logbook when they are not available, I can’t check this right now.
Because you say it’s linked to the physical location I think it’s causes by the bulb to loose connection and reconnect, probably through an other repeater. To my best knowledge there is no way to figure out how a bulb is connected to the hub. And since the hub itself has a stronger signal it might be the bulb is sometimes connected to the hub directly (or tries to connect directly) and sometimes uses an other bulb as a repeater. The official advice from Philips is to add more bulbs… To find the cause I would (temporary) move the hub closer and check if this solves the problem.
Lots of experimentation moving bulbs and hubs around. Conclusions:
whenever one of the bulbs was out of direct range of the hub, and had to hop via the other bulbs - I get these off/on entries in the log in respect of that bulb. Even if I managed to find a location for the out of range bulb that seemed to be ‘reasonably close’ to one of the bulbs in direct communication.
by rearranging where I have one of the access points for wifi in the house [the joys of an old house with stone walls] and hanging the hue hub from this access point - have now got a location from which the hue hub can directly talk to all the light bulbs.
- success - no more entries showing that the light was switching off / on
The assumption remains that this is related to loss of communication rather than the bulb actually switching off/on.
So, success from my POV - but should HA be recording these [possibly/probably] bogus state changes.
Good to hear! I just posted this in another topic: