Home assistant not home assistance.
1 Like
Well we are not your search engine, but try this
@levickij Thanks mate, I’ll check it out.
@nickrout I find that comment a bit offensive mate, don’t know what’s up with that. This is not a search engine, but it is a community. I am not asking everyone to search products for me, but if they have used one such device they can come in here and share what they used with home assistance. Maybe you should know how a community site works. If you don’t have anything constructive to say, refrain from commenting. Sharing that Alibaba link with search text like you have done above is useless to me, because I have been doing that myself.
Of course there will be a smart switch, how else would Home Assistant send commands to turn the lights on or off!
Unless you are using a smart bulb of course.
Two years have passed.
How has the situation changed as regards wall-mounted POE tablets for Home Assistant?
It seems there are some tablet models from unknown (Chinese?) manufacturers on the market.
Is anyone using a POE tablet successfully?
OOps, seems it is not about a tablet.
As said earlier, aliexpress. That’s all.
Stumbled on these a while ago…
The plan is to replace all my switches with these.
(not poe though)
1 Like
Another year and not much new out there. Due to major water damage, I had to pull all drywall, flooring, cabinets, etc. The good news is that everything is exposed so I can wire anything I want. All wall switches are Decora center-off momentary switch, wth Cat5E homerunned to a central wiring panel. The switches will control latching relays (back to the 60’s:-) that switch power to lighting fixtures and switch outlets. A patch panel allows configuring 3/4/n-way light switching. This is the failsafe mode any electrician can service. Home Assistant can however, via additional HW, emulate the operation of the switches and the relays controlling power to the lighting. Lighting is all DMX controlled in this mode, allowing dimming and proper automation. An HA system failure would fail-safe to the 60’s mode. Each switch box has smurf tube home-runned to nearby the structured wiring panel. PoE will be available for the tuchpanel, whcih I DONT want to connect to with WiFi, to reduce attack surfaces.
I’m keeping my eyes open for stuff that woul dllow me to hack an HASWitchplate into a wired ethernet based gizmo using PoE. There’s not much space so I’ve been looking for components, like the power regulator, that are smaller than what HAwitchplate uses. I finalluy found a complete PoC controller/power supply that’s SMAller than the 120VAC > 5v regulator used in the HASwitchplate, Earlier searches returned no joy.
Now to the relevant part. I want touchscreens in most wall switch boxes to indicate operations selected by the switches (e.g. show a dimming slider when a long press is used to dim a light, or other useful info.) A teensy4 with ethernet would have plenty of horsepower, but its ethernet accessory dosen’t support PoE. So I’m lookin out for other MacGyvers to make this work.
Any pointers or suggestions?
Thanks. I just saw a bunch of “surplus” 48 port Cisco UPOE switches on Ebay for cheap. I’ll probably have about 90 wall switches and displays all told, so I might need a couple of these.
I was thinking of hotwiring a PoE PD controller/regulator and a different MPU, but keep the display and much of the embedded SW, just replacing the WiFi calls with ethernet. The ESP dosent have a wired ethernet option so maybe a Teensy 4.1. A single PoE controller/regulator module is available now that’s smaller than the AC->DC supply module, so that’s a step. Using PoE with Teensy remains an issue because thry use incompatible magnetics.
Power over Ethernet - PoE PoE PD Module, 6W, 3.3V Output, IEEE802.3af compliant, Thermally Protected, Smallest Outline, Surface Mount
https://www.mouser.com/ProductDetail/Silvertel/AG9903-MTB?qs=vvQtp7zwQdMW5QJXChJipQ%3D%3D
Some of us live in areas of the country where 100% of the ground beneath is solid granite. Even with a 100G fiber service, linksys webmesh routers in an open concept house… I can be 1 foot from the router and the signal drop out. It is called interference. Just because WiFi is convenient, doesn’t always make it the right application. I too am curious about finding a wall panel that is POE with android capable of voice assistant integration with my google home. They make a few, but it is a gamble because no one can tell me if it will install home assistant app, allow the microphone to be on for wake commands, and also integrate those commands from google home to home assistant.
Also to solve your coding issues. You can easily have your lights controlled by a shelly pro dimmer 2pm or relay 2pm that can allow for switch contact so that it returns to its last state by your physical switch. You still have your standard light switch but you are wiring it back to your central rack shelly device and allowing the shelly device to either be initiated by the light switch to previous state or through wifi/ethernet to be able to control on/off/brightness. Electricians love it because it has all lights on their own dedicated circuit and all low wattage.
My application that I need it for is for google voice assistant to make audible commands. Home assistant to control lights on and off and dim, also separated android POE wall panels that allow for the voice command in each room that is specific to that device so it can stream through to anthem distribution to wired speakers in that zone. That way you can use your youtube music family account to log into each device and you can have as many zones simultaneously streaming different streams as you have devices. The nice thing about touch panel POE is that it responds faster to the shelly command input and you can do it by voice or by walking to the wall to press it.
With that said, they have yc-sm55p wall panels poe with android 11. I am asking the manufacturer from China if it is capable of wake on command and installing HA app. If they say yes to both, then I will trial it because it mounts flush to the wall in it’s own box and connects and powers via POE.
Interference is when you have another signal on or off-channel that is competing with your radio receiver’s front-end. Signal attenuation is not interference. I live in New England where you don’t have to dig very far to hit solid granite, but there is no way that granite beneath your home can attenuate any RF signals above the ground.
It absolutely can. Yes, also living in New England I live on a 150ft diameter rock and cannot dig more than 6 inches in any direction for at least 1000 ft. Interference is when any angle or vector from a signal is changed by any means leading to signal degradation. If you have a device broadcasting in 360 degree array and it is bouncing off a hard surface and coming back against its’ own signal or causing other signals being broadcast to do the same and causing their vectors to change, it is absolutely interference. Attenuation would be if I was on the other side of the rock and the signal couldn’t get to it. I ruled out attenuation by explaining I can be directly in front of the broadcast and signal still drop. Keep in mind, there are thousands of signal being broadcast that are bouncing off the same surface that can cause the same interference. Without a meter I can’t tell. What I can tell is that ethernet and wired connections work flawlessly and WiFi/Wireless (3g,4g,5g, lte, etc) constantly drop out. Example: I can text message fine because the packets are sent and received whenever there is signal, I cannot make a call via WiFi or wireless because perpetual voice signal is lost. What my signal meter reads is high strength, but these packet losses are constantly causing drop outs depending on the exact timing of request and when we are talking milliseconds to make command changes, then it is bound to fail…and it fails frequently. So simply put- wired needed, wireless fails.
You may have other problems. By your postulation, WiFi would not work inside a metal building. But it does. I recently assisted with installing a POE WiFi Access Point in a steel airplane hangar. Not only was the WiFi in the hangar strong, but it was also useable outside of the hangar.
If useable outside the hangar then there is penetration through the steel building/ airplane hangar. So not all the signals are bouncing back to source.
I can stand with a satellite receiver in one spot (open sky) and have full signal and signal drops. Step 10 ft to the right. Wait several minutes and the signal comes in full and signal drops and step back to the original spot and a few minutes later the signal comes in half. And keep doing this all over the property. Functions the same as wifi and wireless. If there is a massive rock that nothing can penetrate and every single signal is refracted and reflects back up to the sky from attenuation, it can and/or must cause interference with all of the signals being broadcast to or near the surface. I understand that it isn’t supposed to but all other causes have been ruled out by myself and every carrier/telecom service that is in our area.
I am not saying this is perpetual but it does cause a cut out or “offline” response from a command a few times a day and any voice calls that are attempted to be made (even through wifi calling) are at a maximum length of 5-6 minutes before it is dropped.
Satellite? As in Starlink? A few feet will make a significant difference of signal strength. Starlink consumer is using the 37.5-42.5 GHz band and a couple of feet can indeed make the difference from no signal to full signal. I wouldn’t use a Starlink receiver as an indication of WiFi coverage.
If it is a Starlink receiver, the signal varies because you are connecting to a new satellite every few minutes. The new satellite may be at a lower azimuth at the switchover giving you a lower signal level.
Your logic is faulty in one respect. Even if you were sitting on a solid metal core that reflected 100% of any RF that reaches it, (a physical impossibility) the path attenuation would render the reflected signal so weak that the principal signal would not be significantly affected. You might get some multipath interference, but the Internet Protocol is designed to account for that. It would affect throughput but not signal level unless you are looking at a spectrum analyzer (not the fake one on your phone).
You are only comparing your own knowledge with what you have experienced. Just like you did with electrical rules in your area because they are not the same everywhere else. If every single detail you take and interpret it differently based on what you have seen then this discussion is pointless. You are comparing apples to oranges, they are fruit and that is about it.
Satellite as in radio. Which should be fluidly connection and the point was that the same location drops out and connects and it does the same for the whole property so it is never predictable where the signal is going to be strongest because it changes every few minutes.
Your example is flawed because you are talking about either radio permeable structures or now completely encased cores (like a faraday cage) and that isn’t related to either points. The example has to have open space above and impenetrable below which causes reflection and refraction of all signals back to all sources. So again, like you said initially a source of attenuation because it is creating noise, but ultimately interference because other signals are being bombarded at the source by the signals being reflected from the rock. Do you think that these wireless signals are not actually traveling through the air? They can be disrupted by similar frequencies, sound (at massive and inaudible levels), and volume. They are being shot out into the air by huge volume. If it was a faraday cage and was only closed source input then it wouldn’t be a problem because the volume of signal would be less and they would eventually be reflected back to the destination.
I am not communicating back after this because it is clearly pointless for you to even care if you are not affected by this. I for one cannot stand to have my hands full of tools entering a dark garage and a wifi motion sensor doesn’t work to turn on a light because the interference caused it to temporarily disconnect and I walk into a conveniently placed bicycle, no issue with wired ethernet devices. Hence my comment, wireless isn’t for all applications.
With some generous help from fvanroie over at the OpenHASP forums I’ve been looking at many different ways to skin the PoE powered single gang touchscreen cat, as well as how to deploy physical switches ecenomically as IoT devices. I’ve come to a few conclusions I thought I’d share for comment.
-
The touchscreen will be a Lanbon-type design that plugs into a base that will fit in a single gang NA box. I pushed on a lot of ropes and couldn’t find a way to fill the entire decorator switch opening with a touchscreen and still have it fit in a NA single gang slot. So, a custom faceplate is needed.
-
Given that, I might as well go to a 3.5" screen. That’s the biggest that will not overlap with a decorator switch in an adjacent gang.
-
There are several combinations of COTS components that can be stacked to implement PoE with a processor powerful enough to run OpenHASP. No stack yet fits in a standard wallbox approved for 120Vac because of the stack depth. A low voltage box will work however since it has no back wall. If I did a custom PCB it might be made to fit if I can replace the RJ45 connector with a push-in terminal block. That means an existing wall box has to be swapped out with a “new work” LV box. (See #3 below, but where I am, the Code would require me to remove any abandoned AC wiring. Fortunately I won’t have any.)
-
I’m doing what amounts to new construction so running smurf conduit to all wall boxes is easy (not cheap, easy). That way, it’s also easy to revert to a “conventional” 120Vac setup if I want to sell or I get hit by a bus. (All switch and lighting circuits are home-runned to a structured wiring center.)
-
The wall switches are SPDT, center-off momentary Leviton decorator switches. Normally these would be used in a commercial setting to switch 120V to the coils of remote latching relays that drive the lighting. Instead, I’ll be sensing the switch closures with a PoE MCU, which will talk to Home Assistant.
-
Only a single PoE endpoint is needed for each wall box. Unused GPIOs on the MCU for the touchscreen can connect to switches in the wallbox. If there’s no touchscreen then one MCU can service multiple switches.
-
A 3D printed case can secure both the MCU and one switch, with push-in terminals to connect it to pigtails to the additional switches.
Just to forestall going off-topic too much, the choice of wired PoE power and signaling is fixed for me. I’m even using it as much as possible for lighting. In places I’m using DMX + LED drivers. “Normal” 120V lighting will exist in only one or two locations where the local inspectors may demand it, like utility room lighting.