I've been defeated...

[emosun]

What i love most is theres some guy in 1935 who just walked into a room , flipped on a light switch , and continued with whatever they want to do not thinking twice about the light.

so simple , so futuristic , lol

[Lurick]

2 minutes ago, emosun said:

What i love most is theres some guy in 1935 who just walked into a room , flipped on a light switch , and continued with whatever they want to do not thinking twice about the light.

so simple , so futuristic , lol

Tech on a tech forum, so future, much edge, so hip, wow.

[Sprawlie]

I've had similar frustrations getting thing integrated into my HA setup. Usually nothing to do with HA, but similar BS from multiple different vendors.

 

 

I get his anger here. I read the "I've been defeated" title and knew right away what his anger was going to be and I feel it. 

 

It's not just GE and these sorts of things. In my professional career I have constantly run into the same head banging frustration throughout the software industry as a whole. Companies are becoming far more consumer hostile towards the customers. 

 

I was defeated as well. Resulted in a mental breakdown and the inability to do my job. I don't blame Linus for his anger. If I were facing that, I'd probably have ripped every single switch by hand out myself at this point, boxed them up and dropped them on the companies front steps.

 

[WeeemRCB]

Linus called for ideas at the end of the video.

 

Just the one.

Swap the switches out for regular smart switches/dimmers (z-wave or zigbee) and use ESP32's running ESPresense in the rooms.

 

This will give you much more exact motion tracking as it works on detecting signal strength to any Bluetooth device you carry all the time. Phone, fitbit etc...

Add some simple motion detectors per room as a backup (like the Aqara FP1) for the times a guest or contactor may be moving about and you're good to go.

 

It'll be a bit more work to configure, but it'll give you a lot more "Bill Gates" level of motion tracking through your home.

 

There was a very good video on setting it up and how it works by Everything Smart Home (below)

 

 

[Keaven]

Hi Linux, Jake, and the team,

 

Personally, I think that what you are searching for is a hard thing to find.

A thing that can not only control the light/home when you ask for it but can analyze the context of your home and adapt it automatically. The struggle you had with your lighting is an excellent example.

 

For your problem, at the moment, the best solution is to have multiple device types, independent of each other (switch, dimmer, motion sensors, etc.), controlled by your Home Assistant server. So you can manually create all the rules you want based on your needs (it's long, take a UX approach, but give the best result today).

Not a lot of real device exists to do exactly what you want.

 

I'm the founder of a startup that has been working on this issue for 7 years, Evey: https://www.evey.co

Indeed, this was one of the problems we intend to solve at Evey (link), the startup I've founded. Our principle is based on context analysis to adapt the behavior of the automatic lighting control of the home. Understanding the end-user environment and habits allows the system to personalize the actions without having to program all the rules to get the desired results. 

 

Evey can accomplish that by using human presence detection and tracking the user path from room to room. It also knows when people leave the house and takes into account the time of day, such that you can save energy while you're not there, but keep some lights on when the daylight is over. Of course, the AI will take into account the time of the year and actual illuminance to better adjust light intensity. 

 

With this, we have developed a series of features currently dedicated to the elderly. Our night light tracker, a dim light that opens automatically from their bed to the bathroom, is a real success. 

 

Our human detection also allowed us to develop key safety features such as a fall detection that doesn't require the end-user to wear any kind of detector. We can also detect people's presence in their apartment in case of emergency and non-movement alerts when the AI detects that a person should have moved or is in an uncommon position. The system will soon be able to interact with voice and even get live interactions with the residents. This will open the door to communication to the outside world anywhere from your home. 

 

As stated above, we're a small startup focused only on the senior housing market for now, but we are looking for the smart home market in the future. 

Our solution is already really flexible and can control a vast product span, independent of each other (switch, dimmer, motion sensors, etc.) using WiFi, Bluetooth, Zigbee (and we can have other protocols with WiFi bridge/hub)

 

Here is a small video of what Evey can be in a smart home:

 

 

Looking forward to seeing how you'll improve your experience with your smart home. I would love to say that our product is ready for you to play with it, but we'll have to wait a bit more for that.

 

Sadly we can't all have a Jake to setup our house  
PS: Good job Jake for that dope entertainment/smart home setup! 

 

Cheers,

Keaven

[gergy008]

2 hours ago, ryankrage77 said:

Completely agree. Linus normally comes across as a nice person on camera, but this showed a whole other side of him. The condescension was completely uncalled for.

@LinusTechHonestly I have a lot of respect for you and I've been watching LTT a long, long time

 

But I agree with this chain of quotes and I recommend you make an effort to send an apology for the poor girl just doing her job. I get it, it's stupid. I agree with the entire values that this video portrays. I get that you're frustrated but it was incredibly uncomfortable watching you be so rude and borderline abusive to someone who has no control over the situation. She didn't deserve that and I hope you refrain from behaviour like this in the future.

 

Much love

[Robert_H]

@LinusTech

How about just getting the new Smart Thermostat Premium from ecobee in every room with built in radar and more, you can even add extra externa sensors.

Also please convince them to make and sell versions for the European market.

They are a perfect fit for my hydro floor heating(geothermal) that will run in my new house since the original thermostats that control the loops are also run on 24VAC. The only problem is the hz...we have 50hz in Europe, i have mailed them and asked without any progress. Maybe it will work...i'm no electrician. This guy apparently got it to work with the older version: http://toshiro.it/2016/10/13/ecobee-and-220v-european-heater/

 

/Robert

[Volcust]

Without the configuration for he can have no different thing for time. That means if one of his kids, Linus, or Yavane gets up at 3 in the morning to eat cheese from the refrigerator it will wake everyone up because it will just go at full blast.

[WeeemRCB]

4 hours ago, Caroline said:

I've reached a point where I can't even understand what the point of all this crap is, other than making people waste their money of course.

 

Something as simple as flipping a switch and having light in a room becomes a nightmare when you overcomplicate it that much.

The problem with the switches was with the manufacturer, not the end user so it looks bad in this light, as it would with any faulty product.

 

My desire is to have an automated home (not a "smart home") that is simple so it compliments your home life and be seamless in the background.

Like headlights on your car that automatically come on when it's dark enough. Such a simple thing and it quickly becomes second nature to let it do it for you.

 

Another advantage is that it saves time. A video I saw recently there was a guy saying it took him about 4 minutes at night to walk through the house switching off lights (he has kids) before he went to bed. Not a chore in itself, but over a year that works out to be 24hrs. 1 whole day of the year spent just switching off lights at night.

 

The hardest thing I found in setting up mine is re-training myself not to switch lights off when leaving a room. I set up mine about a month ago and I'm still occasionally trying to knock the light switch with my elbow when leaving the kitchen with my hands full  

 

[ToboRobot]

Having worked in call centers. I empathize with the frustration. 

The phone system UX is terrible.  The music is terrible.   The training and tools is terrible.  

I avoid calling because the experience is so bad it always leads to infuriation (especially as a power/advanced users who has to put up with tier one support).

If firmware and detailed documentation was on the website, this call could have been avoided.


I really want a full blown home automation lab of testing to move this industry along from broken early adopter to working mass market!

[dogwitch]

1 hour ago, ToboRobot said:

Having worked in call centers. I empathize with the frustration. 

The phone system UX is terrible.  The music is terrible.   The training and tools is terrible.  

I avoid calling because the experience is so bad it always leads to infuriation (especially as a power/advanced users who has to put up with tier one support).

If firmware and detailed documentation was on the website, this call could have been avoided.


I really want a full blown home automation lab of testing to move this industry along from broken early adopter to working mass market!

i mention before about my isp. the last 3 times calling them. i knew more about  issue i was having then them. like  on there end. it was dear in head lights bad.

[AnonymousGuy]

Fun fact: I got asked to do consultant work for SmartThings back when they were a startup.  My price was too high and their deadlines were unrealistic (like go to market making a hub in 6 months...nope the regulatory alone is longer than that).  I was right that it took them about a year to actually go to market and then they got bought by Samsung.  Wonder if I woulda been a millionaire from stock options...

 

Anyways I still don't believe in these zwave zigbee things.  Always felt like cludgy mid-2000s tech waiting for an Apple.  Just like how Nest destroyed every shitty smart thermostat that was before it.

[Uttamattamakin]

5 hours ago, GDRRiley said:

most people I know don't buy the hubs given the price, you can 99% of the functionality with an old PC or single board computer and a 10$ antenna
its a fairly open standard you can't claim to work with it and then go well we only work with these hubs

I said

 

5 hours ago, Uttamattamakin said:

 Since most normal sane people would just buy a hub along with some switches or sensors and live within that system. 

Surely you realize that techies like those who watch (much less work for LTT much less are Linus himself) are not sane or normal.   Most people buy this type of thing with a hub.  People who would go to github to get fimware etc are not the norm.  

Try discussing what HTML is with most people.  Even young hip people who use the net all the time do not know what HTML is. 

[tkitch]

5 hours ago, gergy008 said:

@LinusTechHonestly I have a lot of respect for you and I've been watching LTT a long, long time

 

But I agree with this chain of quotes and I recommend you make an effort to send an apology for the poor girl just doing her job. I get it, it's stupid. I agree with the entire values that this video portrays. I get that you're frustrated but it was incredibly uncomfortable watching you be so rude and borderline abusive to someone who has no control over the situation. She didn't deserve that and I hope you refrain from behaviour like this in the future.

 

Much love

I mostly disagree.

 

Was he pissed off during that call?  Yup.  He was pissed about assinine corporate policies that existed for no good reason.

 

He didn't insult her directly, or swear at her (He saved that for the company and the camera after he got off the phone) and he didn't address any of his frustrations at her, but at the company.

 

Sometimes righteous anger IS appropriate on these calls.  I've been on both sides of it over the years of helpdesk and tech support.

[coasterghost]

10 hours ago, Nuver said:

Imagine talking to a customer service agent like that. Yikes.

I'll share here what I shared on the video on Reddit.

 

They both could have handled that call better. However I took particular issues with her tone and overall company policy as well where she was just trying to give him basically the run around.

 

I have been in Linus' shoes when it comes to being run around by customer service, especially by company policy. Its like when I have to deal with spectrum. I ask for my modems power levels and am told its Proprietary information. They can't release the levels of my quality of service that I actively pay for when I was actively diagnosing network issues, because they don't want to roll a truck out when I have over thousands of data points to point otherwise. Two calls later, another tech tells me, if you want the levels, buy your own modem. Or when I request a new modem three times because the one I have is actively failing to be told on the third call that "We need to justify it to upper management." to be mailed a new one in the beginning of a pandemic (Looking at you Spectrum, April of 2020). That is overall HORRIBLE Customer service.

 

Now going back to their policy at the time of not releasing firmware updates because you don't have their particular unit. You are in for a world of hurt. All it takes is for one major CVE (Common Vulnerabilities and Exposures) disclosure and then how are you going to fix it? Are yo u going to act properly and release the fix publicly or are you not going to bother and fix it for anything who didn't buy your branded piece of equipment. When you have ANYTHING IoT, there has to be a easy way to update it, because of vulnerabilities. The Satori and Mirai malware come to mind for what happens when customers don't update and when companies won't release updates. There needs to be a common way to fix these issues, and publicly releasing these would be one for a good starter. Especially if OTA updates fail as OTAs are not bulletproof.

 

Do I agree with this actions, no. But have I been there from inept customer service and non-consumer friendly customer service, yes.

 

To add beyond my comment: If Home Automation actually wants to be a thing, there needs to be cross compatibility and not the hodgepodge of applications and hardware many of which are obviously half baked.  It's absolutely pathetic that I have to run middle man solutions because something doesn't play nicely from one ecosystem to another. A Prime example are my Philips WiZ lights. The lights are wifi connected and don't have a hub, however they are run by ESP8266s. I have to run homebridge to be graced with the ability to use them with Apple Homekit, because most rooms did not need Philips Hue so I did not pay the Hue Tax in every room. (Signify if you see this, you could have make an library or use the one that is now on Github). Secondary shout out to Samsung as well, for making the ability of having a smart stove and smart refrigerator without the ability of actually seeing your devices information such as serial number. 

[GDRRiley]

3 hours ago, Uttamattamakin said:

Surely you realize that techies like those who watch (much less work for LTT much less are Linus himself) are not sane or normal.   Most people buy this type of thing with a hub.  People who would go to github to get fimware etc are not the norm.

how many normal people do you know using Z wave or zigbee?
I don't know any.

man I'm digging into 2018 or so trying to remember what the Electrical and audio engineer I was under was talking about both standards.

[piledriven]

Thanks to LTT for exposing some of the BS these companies love to do. I'm surprised JASCO actually did something (most likely due to media pressure from LTT about this.

 

In Spoilers: Long rant and Rage at the Machine from a person working previously working commercial building automation for new/renovated buildings.

 

Spoiler

In my old job, I worked a few years as a commercial building automation programmer (some experience with HVAC and very basic lighting integration). My experience is that many of these companies are trying their best to copy Apple's latest and greatest predatory tactics of not letting the consumer delve into the "how" of controlling their products because they don't want to let you know its limitations (most often bloated, outdated/broken back-ends from lack of investment and R&D).

 

Best part of all: they don't get blamed for the failures because those companies are NOT customer-facing. Their first (and most common) response is to tell you (if you called for support) to "Go to one of our the licensed dealers you're getting support from to help you out" even if you have big bucks like Facebook, Apple, etc. (and if you did, we the dealer had to convince the fucking vendor to do modifications to actually make the products work in their modern top-of-the-line networks in the first place). You're not using the most recent software updates? Sorry, we recommend upgrading them to the newest firmware, even though they don't have the cash or hardware to support it.

 

What Linus dealt with on WAN Show regarding JASCO's customer support was the "Oh, we have new firmware, but you have to be a dealer/need to go through the dealher for you/your company to access it," which is really the same as "Our firmware is 'Proprietary Information' that we don't wish to share." Both responses are best summarized as AKA get lost and buy replacements/our better stuff. Frankly, it's abnormal to find companies that have publicly available firmware for devices than not.

 

One commonality I've noticed with both home and commercial automation is that the lack of software and firmware support (customer-side or dealer-side). Bad firmware? You're screwed, we'll come out with something soon, so roll back to an older version even though it's not secure and the security patches we use aren't released for all versions of the software. Remote updates for your devices? LOL what's that future magic you got there? A lot of the older stuff I've worked with is just not UI-friendly and needs direct PC-to-device for upgrades.

 

Example: One company's product I dealt with released final firmware for a Zigbee hub introducing NEW features (you're not supposed to release new things on final stability updates!). And was somehow less stable than the previous version. Both versions had the device flat out just randomly stop communicating with its associated thermostats and did not have an active software point that we could use to "pick up" as a display to see when it last updated, so there was no way to actually communicate properly in real-time. So you had to manually re-pair them. One-by-one. It happened A LOT.

 

This video brought back a LOT of fury and rage I forgot existed because you'd think that commercial automation would be better than home automation and then I see the video and... it isn't. Zigbee/Z-Wave/BACNET/MODBUS/KNX/CANBUS/whatever standard is barely cross-compatible with one another when there should be talks to start updating back-ends of standards so things just fucking work together without needing any research. Plug and Play is a thing for everybody else!

 

[PaulCam]

Hi.

 

When I started watching the "Linus goes smart" videos on the new house I seen the same giddy smiles about ideas for home automation, "It'll be great, we can do this and that and this and that...  yay!"

Then I seen you buying hardware based on the features it claims to have an my heart sank.

 

Fast forward to the later videos and... yep.  Confirmed.  The issue with a large portion of "off the shelf" solutions, is there are bespoke, proprietary and you also quickly learn that the cool functionality you were after is in fact hard coded and completely inflexible.  So while you can control this with that, only in the way it wants you do.  No customisation without replacing the whole system.

I came into this space a lucky way.  As a software engineer, I had been playing with automation on raspberry PIs and relays, electronics and micro-controller software.  All hobby space, play room stuff.  This was back in 2005-2010 when home auto stuff didn't exist really except in the ultra high end full install systems.  So I was already working down in the weeds with message protocols, state stores, message flows, writing all that stuff while using 100% bespoke Wifi sensors on Raspberry PIs, later ESP8266 MCs.

 

Some good things have come out of the Home Auto market expansion.  Emerging standards like MQTT and Zigbee have created some decent width and scope of inter-compatibility, at least at the transport and messaging layers.
 

The cloud is the main concern.  I have a  100% cloud ban on my home automation.  If I'm going to use a cloud, I'm going to make my own cloud in AWS or Linode etc.  So, so, so many of these "Wifi" devices...  only work when you create an account and accept the privacy policy.  They only work when you are online.  Only work via their app.  Are a MASSIVE security risk in your home.  They are a 3rd party device completely outside your control, with access to your Wifi and your Internet, sitting in your home, possibly with access to microphones and video cameras.  I mean come on, really?

 

Home Assistant.  Home assistant is great.    It provides a way for you to integrate a lot of eco-systems and use one app rather than many.  It's when you get into automation that you run into the classic problems.  You can make it simple to setup and lack functionality or any real power.  Or you make it increasingly complicated to setup and config to provide more functionality and flexibility, but in the process you start making it almost as complicated as it would be to do it in code yourself.

Today, if I want to integrate a new device or setup a new automation, I literally have access to the raw MQTT message buses (yes several), a cache of the last message on all interesting topics.... all in python code.  So I literally just start subscribing to topics, checking other topics, making decisions and sending more messages.  When you watch the MQTT topics it's like little waves flow across it as an even occurs, dozens of things respond and update sending a cascade of messages through the system in a fraction of a second while everything updates.  It's pretty cool.   

 

Heating was easy, but trust, not as easy as you think for a multi-zone, presence aware system.  Lights are more tricky.  With heating you can get it pretty much 95% automated and leave limited intervention or override controls.  Lights have too many conflicting aspirations of automation with manual control over the top.  Your examples of rolling over in bed turning the lights on, is a prime candidate.

 

In my developing system (lights is all new now)  I am decoupling the switches/buttons/sensors from the actual lights themselves.  Just like I decoupled requests for heating from actually how heating is handled.  So, keeping the rolling over in bed example; the motion sensor isn't allowed to turn the light on itself.  It is only allowed to publish an indicator of temporary presence, because as a motion sensor, that is all it knows (it could have a lux sensor too, but that, to me, is a completely separate and distinct data point).  Motions sensors don't know how to control lights and shouldn't be expected to.

It is therefore up to the automation to listen for events, check across multiple data points to determine what state change, if any, is required.  Something, some combination of time of day, pressure sensor under the mattress, NFC sensor whatever, can be used to determine if a light should be on or off, or HOW on or off, color etc.

For example, I was intending on leaving the motion sensor ON at night.  But setting the "in bed" mode to about 15% pure blue or pure red light.  The kind of light level you wouldn't notice with your eyes closed or even open, but you could see by if you got out of bed for a pee.  But you could just as easily say that "if bedTimeMode()" ignore motion sesnors.  Well it would be in the inverse really.  Every motion event triggers the automation but just doesn't turn on the light in response.  However when it recieves a "Click" event from a wall switch, it does come on.

The biggest hurdle that spans heating, lighting and most automation is state machines.  In particular the fact that 9 times out of 10 your brain only forms half of the picture of any state machine.  I'm rambling...   You have to be able to turn things OFF too.  That isn't always as apparent.  Consider your multi-light zones.  What if zone's overlap (say in hallways)?  What if zone 1 switches on it's lights from a motion sensor while zone 2 lights are already on and they share a light.  That's fine.  What happens when zone 1's motion sensor says zone 1 is empty and it switches off the lights in zone 1.... including the shared bulb from zone 2.  Hard to explain, but not only do you have to consider and deconflict competing "ON" events, but also competing "OFF" events.  As a professional software engineer that sounded like far too much real life work, so, for heating, I bypassed the whole thing and made ALL events and states expire and default to OFF, so that if the server went for a bath, the heating would still switch itself off and I didn't need to implement the "OFF" side of the state machine except for manual interventions.

 

This is exactly why, by the way, home automation software is either rubbish, very, very brittle and inflexible or hugely expensive (including paying with your data).  It's NOT easy to do.  What starts as a 10 line python script to make a light switch on from a motion sensor or 3 line HA automation.... ends up as a few dozen individual micro-service message bus handlers responding in cascades across thousands of lines of code and dozens of bus topics.

 

The better bit is, I haven't touched my heating controls in over a year, it just works.  Lights, I'm working on it.  I have full auto hall lights.  No manual intervention, fixed, coded functionality.  Bedroom, livingroom and office, have remote controlled lights, but only basic themes written so far.  I have a long way to go with lights to get where I want.

 

Not putting this here as if it's of much use, but it's mostly upto date code for my system.  It's not general use though.https://gitlab.com/paulcam/home_heating/-/tree/mqtt_conversion

 

EDIT:

Under all the fluff, this is how simple things can be.  I give you motion sensed, multi-zone, daylight sensitive light automation:
 

    def notify(self, topic):
        payload = self.cache.last_for_topic(topic)

        if topic == self.upstairs_motion_topic:
            print("Upstairs motion message, payload: {0!s}".format(payload))
            bulbs = ["upstairs_hall_bulb"]

        elif topic == self.downstairs_motion_topic:
            print("Down stairs motion message, payload: {0!s}".format(payload))
            bulbs = ["downstairs_hall_bulb"]
        elif topic == self.bathroom_motion_topic:
            print("Bathroom motion message, payload: {0!s}".format(payload))
            bulbs = [ "bathroom_bulb1", "bathroom_bulb2" ]
        else:
            return

        if "occupancy" in payload:
            occupancy = payload["occupancy"]
            if occupancy:
                panel_power = self.cache.last_for_topic(self.solar_power_topic)
                if panel_power:
                    if "value" in panel_power:
                        value = panel_power["value"]
                        print("Panel power = {0!s}".format(value))
                        if value > 0.2:
                            print("Still Daylight.")
                            return

                self.turnon(bulbs)
            else:
                self.turnoff(bulbs)

To be honest.  I do not like the if, elif, elif, else.  I would prefer to have the topic in sync by convention.  motion_something -> light_something  then use motion_* light_*   Of course this only works for tightly coupled, fully auto lights.  There is no manual intervention here.  Which is why it's so simple.

[PaulCam]

You gotta have a dashboard too!  At least one.

[PaulCam]

4 hours ago, piledriven said:

This video brought back a LOT of fury and rage I forgot existed because you'd think that commercial automation would be better than home automation and then I see the video and... it isn't. Zigbee/Z-Wave/BACNET/MODBUS/KNX/CANBUS/whatever standard is barely cross-compatible with one another when there should be talks to start updating back-ends of standards so things just fucking work together without needing any research. Plug and Play is a thing for everybody else!

You can't have your cake and eat it though.  Not all of those items your listed are even cakes, technically they are biscuits which are not supported. 

 

Which every way you slice it up, it always fall somewhere on the same chart.  A trade off in power, flexibility, complexity against ease of use, easy of setup, affordability.

 

The hardware is fine.  The layer 1 protocols are fine.  It's once you get into companies "hubs", "clouds", "apps" etc.  that it all starts to unravel.  The software and firmware side of things.  That is where the hard part lies, but also the most expensive part.  Usually the most botched part.

 

Creating a small family of devices and writing the software to be blinkered to support only a small matrix of combinations of those devices, add a few simple, rudimentary though pretty cool automations and you are done.  Blow half the software budget on a Gfx, Ux team to tart up the mobile app and you have a winner.  /sarc.

 

The first step to fixing the smart home market is separation of the gadgets and their management software.

Be they zigbee, zwave, or canbus or tasmota whatever.  "Terminate" them, not on a all-in-one hub with the automation, but just onto some middleware like MQTT or REST or RabbitMQ, whatever floats your boat.  Then enable companies to write JUST the automation software which uses this middleware.  The support lies with the automation software as to whether it supports your device.. and with your individiual device as to whether it wishes to support your middleware.

The only real back and fro on that approach is the message contents/schema.  That can be forced into the hands of the automation guys to support individual devices, but breaking the device from teh management software with middleware would make the market much more vibrant and healthy IMHO.

[RZomerman]

I understand your frustration.. which is exactly what I wanted to prevent.. 

 

.. thats why I went the Velbus way.. a hardwired CANBUS system that runs 1 (4 wire) cable around the house to all the switches..  the BUS also extends to each of the switching, dimming, input, output etc modules.. because it's hardwired and every component is updated with its own configuration (no central server required) - switching on the main power in the house ensures the domotica (read: lights, HVAC etc) in the house runs within a second.. not dependent on WIFI, Bluetooth, Zigbee, Internet or whatever.. 

 

In order to then extend the system to the outside - and create more advanced logic - you hook it up to Home Automation or OpenHab

 

for me, the core functionality of my house (switching main lights, HVAC control etc) is on the Velbus technology - accessories (bed lights, sound system etc) is on the OpenHab (Ikea / yamaha / etc bindings) - and even those I control via the Velbus input panels.. given each panel is also a thermostat - I could remove all the NEST modules too.. so it looks even better on the wall now with just 1 panel and I can measure the temperature of every room independently.. 

[PaulCam]

48 minutes ago, James Evens said:

MQTT is just a transport protocol. Think of it like the HTTP of IOT. Maybe throw XML into the mix.

Just like with the "web" there are multiple IOT protocols each of them with there own ups and down. For example there is MQTTs and COAPs. MQTTs isn't inter-compatible at all.

MQTT is not a transport protocol in the Internet model.  It's an application protocol providing middleware in the form of a message bus.  It's not an IoT protocol either.  No more than TCP is... and it IS a transport protocol.  It IS a transport protocol in the OSI layer model, but let's face it, nobody uses the OSI layer model.

 

HTTP has one advantage over MQTT, it's transactional.  Making MQTT transactional is a headache that most people just don't bother with.  Working on best effort approach is fine for 99% of things in home automation.  Although it leads to interesting bugs from time to time when messages get lost or cross on the wire resulting in unexpected state.

 

MQTT is better described as middleware.  You could use HTTP sure.  You could use JNDI, EJB, RabbitMQ, hell I used to create the data fabric/middleware from raw TCP or UDP messages all publishing to the SAME topic.  I still have 3 sensors that send data like that, so I have the hub server receiving it and publishing it on MQTT these days.

Anyway, my point was, that when you combine the bare resemblence of "standards" coming from Zigbee, a little glue code in something like zigbee2mqtt (or other gateway) and it opens things up in a way, as most automation software can support ingest of json over mqtt and publish json over mqtt, especially HA.  It thus provides a way to centrally control/monitor a much larger swath of your IoT devices from one mechanism.

Devices that aren't Zigbee, or MQTT compatible, I just write "Controllers" which handle communication with that device (or devices) and relay on MQTT anyway.  If that controller is too complex I use the one that comes in HA if it exists or fine a new device.  I have very little  proprietary management code in the house.  Shellies retain their firmware, so do zigbee, everything else is flashed and all zigbee goes through a USB dongle and zigbee2mqtt.  I don't even own a single IoT/Smart hub.  Not one.

[Required]

Why Gray Socks as an Guy? Thats so gross.... Black should be the only Colour we wear....

 

Does anyone know an European Smoke Detector why have, use a Bus System determine the exact Detector who going wild?
For example to put them inside a Network Rack, above the Fridge,... and use Multiple in Area who normally nobody get there often.

Thanks

[PaulCam]

1 hour ago, James Evens said:

Modbus over TCP might work well. Ethernet is simple/basic and error tolerant and if required you can just adapt it from TCP to RS485. Overall I like modbus but thats just my opinion formed because every program supports it and manufacturer deliver datasheet/instructions with there modbus slaves.

There are devices for this, however they are almost always sold "per device" "per manufacturer".

 

For example I monitor the solar panel over TCP.  It has an RS485 RJ45 port.  I borrow the 5V line off the RS485 to run an ESP8266 microcontroller.  Consume the RS485 registers, create (at the time) pipe delimited bulk messages containing all registers and publishing it onto the home automation data feed.

 

It's still working today.  It's been running 24/7 365 for about 4 years now.  As long as it has every barely working 2.4Ghz it will send that data every 5 seconds.

Build thread: https://www.eevblog.com/forum/projects/nodemcu-esp8266-rs485-epever-solar-monitor-diy/msg1404831/#msg1404831

 

EPEver make their own devices, but as you would expect they provide access to their UI over HTTP, rather than anything particularly useful, like publishing to MQTT or TCP.

[mattk926]

I purchased around $1400 worth of GEbyJasco switches for my house at the end of 2017. I wanted zwave(900mhz) to stay out of the 2.4 ghz wifi spectrum that zigbee used. I was looking for a combined motion and switch/dimmer in one and landed on the same 26933/26931 jasco switches for similar reasons described in the wan show. These were also the only motion built into switch for zwave at the time, and I think they still are one of the only options. I didn't want to have battery motion sensors and deal with replacing batteries multiple times per year.

 

Overall they have held up well with the few noted quirks that have been pointed out:

1. If in normal occupancy mode and you manually turn off the switch they can come back on before you have left the room etc. - This needs a firmware fix/update to be usable.

2. LED compatibility on the dimmers has been more miss than hit. I searched most local stores to find LED bulbs that wouldn't buzz/humm/flicker even though they claim to be dimmable LED. Not sure if this is using Leading edge or Trailing edge dimmer, but they need to find a better dimmer solution as most light bulbs for the past 5+ years have been LED and this shouldn't be a problem on a "smart" home device. Not sure if I was just getting some flukes, but it sounded like this was an issue some others had experience as well back in 2017 via a smartthings forum post.

3. The "switch" mode of the dimmer definitely needs some work. Should be able to disable the fadeon/fadeoff as a configuration setting even outside of switch mode. The only thing you can do currently is make the fadeon/out faster. This needs a firmware fix/update.

4. In areas like bathrooms/mudrooms/hallways they are fantastic in occupancy mode minus number 1 stated above, however in bathrooms if you are in the tub/shower and have a door/curtain closed the motion wont be detected and if you have a shorter timeout period as I believe the switches go from 5 minutes then jump to 30 minutes for a no motion timeout it can cause the lights to turn off leaving you in the dark. This could be worked around using HomeAssistant to control the motion timeout and putting the switch in manual mode, but that somewhat defeats the purpose and if HA isn't available it can leave the automated lights broken. It would be nice to have the parameter be user configurable by the minute and not just preset amounts via a zwave config push might be doable via firmware update?  I have worked around this by adding an additional multisensor for motion above the tub/shower area and using an automation in HA to keep the light on.

5. There is a light sensor the switch can use to determine if it should turn on the lights in occupancy mode, but this isn't exposed as a sensor via zwave, This would be fantastic to have this exposed as a sensor like the motion sensor that could be used for other automations in the same room. Would be nice to have this enabled via a firmware update.

6. There is an "airgap" clear plastic tab at the bottom of each switch that you can pull out which kills power to the switch. These tabs are easy to break off and from my conversation with their support they cant just send the plastic tab they need to warranty the entire unit. still functional without this, but inconvenient and shouldn't break easily. This has happened on three of the eight switches in this style, keep in mind those are the three areas where these tabs get pulled on to bypass the motion at certain times of day where the other areas we haven't use the airgap functionality as the switches have been rock solid and we aren't using the normal occupancy mode, just leveraging motion in routines/automations.

7. Doubletap functionality seems to be lacking on these specific models and would be appreciated if there is a firmware update to enable a differentiator between single/double taps so you could have the dimmers go to 100% with a double tap for example.

 

FYI their fan controller switches 55258 for ceiling fans have been rock solid. And other than the few mentioned issues with their motion dimmers/switches 

 

Hopefully you can work with jasco on getting some firmware updates on the motion activated dimmer/switches for the above mentioned issues in their current form, or work with another manufacturer to provide a better solution that incorporates a motion dimmer/light sensing, or even adding temp/humidity sensors into the motion switches as well. Would be fantastic to have a multipurpose sensor built into the switch that is powered by mains without dealing with batteries or having to run 5v usb to areas not near wall outlets for each room.