Building My Perfect Smart Home Presence Detection Sensor!
Вставка
- Опубліковано 21 вер 2024
- A few weeks ago, I took you through a custom built room presence sensor which used mmWave to accurately detect the presence of a person in a room - even when there was no movement. The DIY sensor we built is amazing for controlling things like smart lights in rooms when you don't tend to move around much, like your living room or bedroom. So after that video, I became focussed on how to improve this sensor and what other sensors I could add in here to make MY ideal room presence sensor, everything from light levels, temperature, humidity and of course, multiple levels of occupancy and presence.
Thank you to PCBWay for sponsoring this video, check them out here: www.pcbway.com/
If you'd like to support me and future projects:
Patreon:
/ everythingsmarthome
Buy Me a Coffee:
www.buymeacoff...
Written Article:
everythingsmar...
Links:
PCB Prototype Board:
UK: amzn.to/3y2HJIy
International: amzn.to/3uaHWbk
AliExpress: bit.ly/3I8LkcI
Temperature Sensor:
UK: amzn.to/39ZD6a2
International: amzn.to/3I2Cihp
AliExpress: bit.ly/3QZ5euq
Light Sensor:
UK: amzn.to/3no3BJx
International: amzn.to/3bqNqIq
AliExpress: bit.ly/3xW7aew
ESP32:
UK: amzn.to/3NttYbp
International: amzn.to/3OxSNUP
AliExpress: bit.ly/3a4Xerp
PIR Sensor: bit.ly/3R1kCa2
mmWave Sensor:
DFRobot: bit.ly/3x3WOub
The Pi Hut: bit.ly/3NIwTgW
Mouser: bit.ly/38V35z4
Pimoroni: bit.ly/3M3hyGs
EVERYWHERE YOU CAN FIND ME:
Everything Smart Home Community Discord:
discord.everyt...
Twitter: / everysmarthome (@EverySmartHome)
Instagram: / everythingsmarthome
Facebook: / everythingsmarthomeoff...
#homeassistant #smarthome
________________________
Would love to buy a few of these without having to buy the components and assemble, but appreciate that this is such a great solution.
Well you never know...
Would be much easier to use a breadboard and jumpers than soldering. It may seem daunting but it gets pretty fun (and maybe a little addicting) after a few attempts.
oh geez me to! Id buy 4/5 right away! these things are great and i font have the time for building them.
@@EverythingSmartHome could these sensors work with hubitat too?
I would 100% buy a bunch of these!
9:00 by the way since you were speculating on this you are correct a 60ghz mmwave sensor is far more sensitive even being capable of detecting heartbeats and breathing which is why its so much more reliable than the 24ghz one when detecting sleeping humans, although it does this at a much lower range than a 24ghz module. To be fair the dfrobot sensor does specifically say in the instructions and even shows pictures to install the sensor directly above or under the bed when using it for that purpose. Seeed studio actually has a 60hz module for sale on amazon currently if you want to read about it maybe even test it. ............... seeed studio mr60fda1 60ghz mmwave sensor
Thanks!
Can I ask - because I don't see it - what the benefit is of being able to detect motion from a sleeping person?
@@JohnVanderbeck presence detection in a room after the occupant has gone motionless in bed
@@ezradmello7021 Yeah I get that but for what purpose? Just curious. Usually you use these for turning on lights but you don't want lights on when you are sleeping.
It's actually the opposite.
You'd use this as a bed occupancy sensor so you can tell when you're in bed or sleeping.
Ex: Slowly dim and turn off the lights if it's night time and presence is detected at the bed.
Great video! I loved that you built your own sensor that does everything you need. Makes me want to build one for each room in my house.
Thanks Steve, appreciate that!
This is awesome! Would be great to have some sort of documentation for this. So a buylist together with the step by step manual to put this together. Same for the STL file of printing a case. Maybe an idea when a V2 will be released? XD can't wait! Great idea!
I came back to this video for just half of this and I would be happy if that is all I found.
The parts list in the description is decent but not very detailed, once the links die it will be harder to find later.
The lack of STL files and the updated code mentioned in the video is the worst part though... I wanted to make 5 of these but make some tweaks and the lack of code really hurts. I am pretty disappointed as it seems this project is already been abandoned by Everything Smart Home.
Well, this comment didn't age well 😜 This project is very much alive as you can see :)
I will make a DIY version available at some point - truth be told this video was never intended to be a "guide" on how to do it, I more wanted to show it off - but lots have asked for one so I will chuck some stuff up at some point, I am currently very busy making this a full fledged product for people to buy, but once that's done then I will try to make a DIY guide for those who prefer that route - I don't know how long that will be though, so meantime if you have any questions at all about how to make this, Discord is the best place - there is tons of information and discussion about this sensor there and more than happy to help
@@EverythingSmartHome thanks for the great response! Awesome this project is still alive and kickin!
@@EverythingSmartHome really hope you can have a version with air quality sensor. Congrats BTW
I don't think it is so awesome - incorporating few sensors as one and using lot of mcu interfaces is overkill. Probably better idea to use radar presence detector sensors like LD2410
This is so awesome! Presence detection has always been a pain point but this seems to be the perfect solution. Cant wait for more commercially made projects to be readily available with MMW
Haha thanks, I agree!!
The fp1 is amazing. I spent months/years trying different combos of door contacts and motion detectors to detect presence in the bathroom and when in the shower, but nothing short of pressing a button to set my system into "shower mode" worked well. That all changed as soon as I received the fp1. Not only does it perfectly detect someone being in the bathroom, it detects through the shower curtain, and it even detected me in the shower when I accidentally put my clothes directly in front of the device.
I was able to remove all other devices and complex node-red flows minus one aqara motion detector to more quickly trigger the lights to turn on as it takes a few seconds longer for the fp1 to determine that there is an actual person in the room, where as a normal motion detector is nearly instant. I then only have to use the fp1's "off" mode when presence is no longer detected to turn the light off. There is no trouble in between anymore of lights going off when using the toilet or in the shower.
I would absolutely love to see a zigbee version of this! I have a bunch of ESP32-CAM cameras already, and they're a bit unreliable with regards to connectivity. The zigbee devices I have seem to be much more consistent. Also, I love the lower power draw.
If there was a zigbee version of this, I'd probably order a bunch.
You should totally design a PCB for this! I’d love to make one, but I’m not great at PCB design. I’d hate to get started just to realize I have to restart due to a bad design.
Yeah, pcbway had to be cringing at the use of a proto board in a vid they sponsored.
Great video. Do you have a diagram showing all of the connections you made from each component ?
Had this idea of combining sensor together per room since a while. The mm wave is a game changer. Thanks for the inspiration.
Thanks for watching!
This is brilliant, thank you. I know it's kinda outside the spec of this, but I'd love a battery version for the bathroom.
Air quality is something nice to add. Apart from your suggestion to add a particle sensor I would add a CO2 sensor like the SCD30. That will tell you when you have to ventilate the room.
Forgot to compliment you on this nice video…
This was essentially going to be my reply...swap out the bme280 for an scd41, and add co2 to the measurements I believe?
Other than the fact that I might have to actually get better at soldering, this video and its predecessor are right up my alley. Well, at least it's where I want to be but need to put in some effort first. Thanks so much for this. Once again Kudos to you sir!
Thanks for listening to my suggestion: This will be the most interesting upload you have done so far since this is such an important feature most smart homes lack.
Thanks, glad you liked it!
Dying for a build and programming guide on this! You're the best
Verrrry nice! Good work!
I‘d seen a link to your website with the words „written article“… but i can‘t find anything on the website…?
A wiring diagram and a .bin file for the ESP would make it very easy to make one.
I couldn’t help making a connection drawing if you want.
RF Engineer here, seems like a fun project to put into one PCB. Thanks for sharing.
Thanks, that was the whole idea! Appreciate it!
Will you write this up as a guide for others to make their own? This is just what I've been looking for!
yes please
Yes please
Yes please
Yes please
You swine!! Just after I completed the first version using the mmWave sensor, you come out with this! Seriously, a great video and another fantastic project. I will be attempting this upgrade as soon as I can.
Haha apologies! At least you have the most important part already that is gold dust 😂
@@EverythingSmartHome Just snagged another mmwave sensor. When it comes, I'll have all the parts to build this multi-sensor. All I need is the schematic and the code to make it work!!! I can't figure it out from the video alone. Any chance you might make this available soon? Many thanks.
Another awesome project Lewis, well done! 👏🏻
Apparently the IKEA Vindriktning air quality sensor is reliable and can be hacked to connect to the ESP boards.
If you don't mind the noise from the fan. Audible in a quiet room.
Thanks Rob, that's good to know! I think the added bulk would be too much but maybe there is a way!
Great appetizer! Now what would be great is to have a project page with step by step guidance and code.
I'd love to see this more compact, in a unit like the Aeotec multi-sensor. I wonder how difficult it would be to reverse engineer those boards and create a single PCB with every component neatly installed together. I also wonder if there is redundancy across the boards, as I assume they'd have circuits performing similar functions and therefore using similar components repeatedly.
Nice device. There are some sensors that I prefer to the ones you picked that might be of interest. Firstly, if you want to add PM10/PM2.5 the NOVA SDS011 is very good even compared with sensors several orders of magnitude more expensive, honestly, for £15 it's shockingly good. Second, for humidity measurement I've found the BME series sensors don't tend to give accurate humidity readings. The SI7021/HTU21D are good alternatives but lack atmospheric pressure so it's a trade off, it has a different I2C address so you could try both in the same device and see which gives the most sane readings or have both or use the BMP180 instead. For light levels the VEML6040 gives colour temperature as well as Lux which could be handy if you want your lights to match ambient conditions, unfortunately it's not built into ESPHome, but could be added as a custom sensor.
Thanks for the info, useful!
Of what benefit is a pressure sensor indoors?
Great product! I understand your reasoning behind excluding it, but I would have loved to see an air quality, or at least CO2 sensor in here as well for automating fans or air circulators.
Maybe for the v3...it'll be expensive though 😂 thanks!
Great sensor. I can see having these all over my house. Well done.
A lot of the industrial HVAC occupancy sensors use CO2. Because people breath roughly the same amount you can judge the amount of people based on the curve of CO2. It's also a fresh air thing. Cool project man. Keep up the good work
Congrats on the nice build!
A few tips:
For air quality rather than a PM sensor I use co2 sensors (the senseair s8). I use it to control my ventilation system. It also helps detect whether there are people in the house.
For the update frequency of the sensors I have the esp32 check the value every few seconds but only report this value to home Assistant if there is a change bigger than a specific value. This way I get more or less instant updates but without spamming home Assistant and my wifi network.
I've added a screen to my sensor where I can see the data of my sensors including charts. Took a while to figure out how to do it in ESPHOME but pretty cool. The screen is off by default and turns on when the PIR sensor or microwave sensors which I'm using detect me.
I also have an RGB LED that at a glance shows me the air quality (a gradient from green to red). The brightness of the LED is controlled by the light sensor.
I suspect you are not going to put a cover in front of your multisensor, but if you do: you'll need to place your BME280 outside the case, otherwise readings will be off. Same of course for light sensors etc.
What I'm struggling the most with is finding a reliable humidity/temperature sensor, but I'm pretty sure the problem is that I'm a cheapskate and am buying from AliExpress which are probably knockoffs.
Do you have an video/photo of it?
Awesome work! So when can we expect a Kickstarter project for this bad boy? 😀
Do you have your ESP firmware available to share? This is a great product, I want to test this setup and then I plan on making a nice compact custom PCB.
Cheers for the video!
Any chance Lewis? :D
Please start calling it I square c, furthermore amazing video!
This is a pretty cool all in one sensor. I've been following crlogic's work over on HA too.
I do have a few questions/challenges for you though:
Why did you spend more money on STEMA/QT compatible sensors, proceed to solder them to a proto board, and them talk about the overall footprint being too big? In an effort to appeal to a larger audience, a fair number of the Pimoroni and Adafruit sensors are much larger than some of the more basic offerings, but offer more functionality (generally have a voltage regulator and level shifter to accept both 3.3V or 5V inputs, and have STEMA/QT connectors for no-solder connection options.) For example, a generic BME280 is 11x14mm vs 19x19 for the Pimoroni, and a generic BH1750 is 18.5x14mm vs 26x18mm for the Adafruit (~50% difference!). You do have to watch out for counterfeit sensors on some of the cheaper versions though.
For a prototype board, why didn't you use headers for attaching the sensors to the board? You definitely had the room to do it for this iteration, and it would have allowed you to swap out sensors during testing and future iterations if need be.
The BME280 sensor is very precise on its own, but most of the circuit implementations suffer from heatsoak that skew the temps at least 2* higher than actual. Putting the chip in sleep mode between readings, and adjusting the heater circuity can help, but mounting it directly above an inverted ESP32 chip is not going to do anything for accuracy.
For rapid prototyping and permanent projects I got a fan of the wire wrapping and the glue gun. in comparison with a breadboard or a prototype board you can get quite quick a breakout board aggregate, works with most pin spacing and the glue gun fixes the wires on the pins and the position of the boards. Also most time no solder fumes :)
Would be great if you do a ceiling mount version of this.
Lewis amazing video! Would love a v3 and a full build guide!
Thanks Mark!
Yeah a full build would def. be appreciated. Any plans for that? :) especially the code is needed as I haven't had the chance to play around with esp home and that new Sensors so far.
I love it .. the mmwave is pure gold for presence, and getting the other sensors in there as well, really starts to make this a perfect smart home sensor solution
Thanks dude, appreciate it!
Ever considered getting a custom pcb made up so that you don’t have to stack a bunch of boards? You could even get pcway to solder it all together and make something tiny
Knocking the ball out of the park with this one! Added to my project list.
@Everything Smart Home - A lux sensor does allow you to decide whether to turn the lights on or off, as you mention. For enhancing this, consider using the lux sensor as an input to a PID regulator with whatever setpoint you'd like. This is called daylight regulation or daylight compensated lighting, and is common in commercial applications because it saves a lot of energy by only "supplementing" whatever light you need in addition to the natural light from windows and such. But it's also very comfortable since room-to-room, you'll experience less variation in brightness which makes transistions more comfortable. It also lets your lights dim slowly on as dusk falls with 100 % accuracy, rather than alternative solutions that fall short, such as sunset timers, etc.
You'd typically set a minimum brightness level, so if the PID regulator for example outputs 8 %, you might have a minimum threshold of 10 or 15 %, and thus cut off the light instead. And you'd also often need an auto/manual helper object that automatically gets changed to manual whenever physical switches or UI switches change the light level. Often, it's good to transition back to auto if no presense for X minutes (typically a much longer value though, like 60-120 minutes).
If you then want different responses from your lights in different situations (say; nightly visits to the WC), instead of implementing these as a set percentage of output, you can simply vary the setpoint of the PID regulator automatically. If you foresee the setpoints being changed while lights are on, add a numeric delay / ramp function to the setpoint input, so that it never changes faster than some rate, for example 5 lux per second.
A common problem is that you'll have less uniform light in a home and say, an office. This is "fixed" by applying a percentage factor to each light to account for the lights contributing differently. For example, a table light might contribute less than your overhead lights. So it may get a 140 % factor, which enables it to shine brighter to compensate for this difference.
Interesting read, thanks! You can create an automation that can be something like trigger when lux < x with a condition that there is someone present in the room, I think that's what he meant.
Thank you for sharing this project with us. As always, great video.
Would you mind sharing the codes with us?
Hey, do you have a written article on this project? I couldn't locate it on your website. Thank you. Your videos and tuts are much appreciated!
if you are into sensors, look at the mysensors project. it integrates well with HA. I have battery powered temp/hum/lux/pir all over my house. works great.
Yes I would like to have such sensor in my house. The only thing I'm questioning is how it would cope with pets (e.g. a cat). Is this something you would handle by software or could it be done differently? Great video, thanks!
I have ordered some of this nice thingys.. but.. temp and humidity isnt needed for alot of us.. make em cheeper plz.. nice work!
We wan't this in EU stock! such a nice idea and product
Awesome 👌
This is absolutely fantastic! I'm going to give building one a go. I was looking to do room temperature sensors, and maybe figure out room presence, but WOW this does it all, plus light level. Air quality could be useful, but like you said to add that would greatly increase the bulk of the design, at least right now.
Would you be selling the finished product?
For air quality, a CO2 sensor may be more useful than a PM sensor, and not at all bulky, but a sensor "good enough" is not the cheapest sensor.
wow this is definitely the all in one we all been looking for you are a genius. please could you do a full build and programming video on this one.
Thanks - I'll see what I can do
@@EverythingSmartHome I guess you not done with it yet you seem like you will add 10 extra features on the next one you build for another room at yours. So would be great if we can see detailed build. Thank you. Ps have not spotted in your videos what you do for living but you should be paid big big money for making these. Make your own brand 💰💰
Another great video thanks - this is one I must build
Pet friendly PIR sensor and KNX integration board.. great stuff
Very nicely done! I've the same question as some other users: could you kindly measure the actual power usage of the device? I'm not convinced the power usage is minimal, especially if you would install sensors in all rooms in the house. Things can add up pretty fast.
I'll try at some point sure!
@@EverythingSmartHome I'm also interested about energy consumption! Maybe find some ways to hibernate the sensors when possible too. Great project!
Great work, 👏 would be nice to see some content around how to setup automation between temp sensors and AC temperature control - keen to understand what would be the best approach to build such relationship considering energy efficiency for the AC
AC isn't something I have in this country unfortunately but shouldn't be too difficult...in theory!
Thanks for the informative well explained video!
I have the same experience like @mortensentim511 with the BME280 - hum% 15-25% too low and temperature about 4°C too high. Also putting apart from heat sources like the ESP and after conditioning. Maybe the special soldering process was not maintained, it was on such break-out board delivered.
Alternative #1 I tried was a SHT31. Provides much more accurate hum% and temp values.
Alternative #2 which works even better for my needs: BME680. Very accurate, plus it adds a nice extra of VOC measurement. Fully integrated in ESPHome! Recommendation - use the bme680_bsec library, not the standard one. You will enjoy the much better AQI and other calibration routines in the background and calibration status etc. Plus you get a "CO2 equivalent", which is not too far off from a real CO2 sensor, side-by-side compared. For my purpose, sitting in home office for hours and needing a kick in the pants when to open the window it is the perfect match :)
BTW, particulate matter is interesting, but it will tell you that actually opening the window will *increase* the particle concentration, and it will get better when the window is closed. This is more for a kitchen or workshop I believe.
Nice job there Lewis! Those are the things a big brand needs to get running on PoE mountable as a "wall-switch"-sensor-module. Or even ZigBee and direct wired to the mains-voltage.
I really wished you could say anything about the power-consumption of your projects, as energy is a big concern of mine when talking about those designs. Would love to hear from you!
Thanks! PoE would be cool!
I usually would talk about power consumption, but it's literally so low on these projects it's generally not a concern, we are talking a very tiny amount of power that you won't notice at all in day to day, even in the current climate!
Poe would be ideal for these, have though about something like this, ceiling mounted and fully enclosed.
Also an led chip for a nightlight. I love the very low level light a Ring Protect gives off at night when you walk under it.
Perhaps when it has wired power, a zigbee repeater(for other devicesl, if there is something suitable that could fit into it.
Enjoying all the videos related to this DIY sensor. That's a great prototype case. Are the STL available? I can see it useful to prototype many different combinations.
Super sensor! Thanks for video! I'm going to try setup: light sensor, SCD41 (CO2, temp, humidity) and mmWave Radar
FYI the temp sensor on the BME280 is not accurate at all because of self heating issues. It typically reads 2C or more over the actual air temp.
You're a fricking legend man, can't wait to try this.
Nice build ! Might also add a perfect presence detection sensor to my ToDo list. I agree that even if the price is pretty much the same, functionnalities are way better.
Thanks for the inspiration and keep up the good work !
Thanks, good luck with the To-do list 😅
Also if something goes wrong its easy to replace a part
Great video and work!! For me I would love to have an all in one solution. Your sensor paired with humidity, air quality, co2 and smoke sensor (with optional pinhole camera, where you can use the mic for Siri or Alexa). All working via a LAN cable (power and data) and all in a ceiling mounted Smoke Sensor enclosure. Of course HomeKit compatible 😀. No Switches no nothing just one sensor on the ceiling and some in wall speakers and underfloor heating
Excellent video! I’ve had bad luck with the PIRs I’ve purchased. The link in your video should get me a decent PIR. Using a HiLink HLK-LD2410C-p for microwave sensor. BME280 for temp and humidity. Then VCNL 4010 proximity, light sensor. Replacing all my esp8266 project with ESP32’s. Some with Pi Pico’s, using Micropython.
I have 2 ideas. First, have you ever heard of a volumetric sensor? it reads the volume of material in a room then triggers when that volume changes. Cars use these to set off the alarm should someone manage to get inside your vehicle without opening a door, if you've even stuck your arm through an open window of a locked car and had the alarm go off, this is the volumetric sensor doing its job.
Secondly, could you use a series of Bluetooth BLE sensors placed strategically around the house to triangulate the precise location of any phone with BLE enabled? like if you put a Bluetooth module in each corner of the house and maybe one in the centre somewhere, you could theoretically use the distance data to let home assistant know exactly where you are, it would take quite a bit of setting up with trial and error but once you've figured out the numbers, I'm sure you could have a pretty decent presence detection system.
It's an interesting idea, I think I prefer the mmWave though for movement.
And yep that's the whole idea behind Bluetooth being enabled on the device already 😅
The biggest value add for me would be using Bluetooth as a “human check” instead of a customization option. In the case of pets all of your examples would lead to a dance party in the event of cats or dogs in the house, is it possible to use Bluetooth to prevent that?
Cats are the main reason I cant do 90% of home automations :(
brilliant project. So I ordered 5 parts right away. I hope there will soon be an explanation of how to put the things into operation.
Very nice build! I ended up using the energy meter that measures solar generation as a proxy for daylight levels. That avoids the light sensor picking up the brightness of the room lighting.
Thats a good idea too!
Very cool project. I wonder if a similar project would work with battery only?
These are awesome and I would happily buy a few of these! Very Nice!
Thanks....soon!!
@@EverythingSmartHome ok understood, that's the reason why you don't react on questions like "Can we find the written article?" or "Could you please post the code?". It would be good if you openly communicate this (e.g. in your video description).
Up to now I found your videos motivating and inspiring for DIY projects. But finally it's all about making money ;).
PIR sensors should be combined with a thermal image program. A still thermal image can be triggered from a smartphone to be taken as a benchmark. Such an app must allow to set bypass at certain places of the room where there are light bulbs or the fan motor which tends to get hot in order to avoid interference with the functionality of this particular sensor. The idea here is to use as little amount of 'electric energy' as possible, thus less sensors = less idle energy consumption from such detecting systems. This is how it should work... the PIR sensor detects a movement, and switch gets turned on, it has a preset time of 4 minutes before it turns off the switch if motion has not been detected. however, on the third minute of inactivity from the PIR sensor, the camera should take a still image and notice the heat patterns and compare it with the benchmarked image which was set by the user earlier. If the set zones are hotter than the benchmarked image then have the device stay on for another 5 minutes or else turn it off. At the end of the fourth minute take a new image and again compare it with the original benchmarked image and if still heat is detected then keep the device turned on for 5 more minutes or else turn it off (keep on repeating this behavior). But whenever PIR detects a movement, it should cancel the camera sensor timer and also reset its own timer to 4 minutes (repeat this behavior). When neither of the two sensors detect any heat, then turn the desired switch to off. If one is only passing by and do not wish to turn the lights on, say because it is daytime, then a bell switch can be located nearby, when pressed it gives power to the PIR /camera sensors, the PIR/ camera sensor must turn on the appliances at startup by default (example a light bulb or whatever the user wants), the bell on the other hand must have an inching setting of 10 minutes of its own, after which it turns off the sensors. but with each detection from PIR/camera the inching must reset from the beginning. so this means if the PIR/camera did not detect movement/heat for more than 10 minutes then these sensors will shut down and only way to turn them back on again would be from the bell switch.
The only thing missing is battery power source. I would also change temperature sensor to something cheaper. This one is an overkill for controlling temperature in the room. You don’t need pressure sensor in each room ;)
Doesn't have a pressure sensor, it's temperature and humidity only ;)
Thanks! I’m looking forward to build a couple of these for my home. Do you have any idea of the power consumption of a single unit? Can it be realistically powered off replaceable or rechargeable batteries?
My god man! Clean your tip and maybe watch some soldering videos!
I mean that, but awesome job on the module. I want this for sure. Simply great.
Can you PLEASE post the code for this?
Others have asked for this as well, but I can't find it, either on your github or you main page.
I'm definitely keen to try this, but I can't get hold of the sensor. Seeed have some interesting ones that might work. Perhaps I'll do some experimenting...
Yes! It's been weeks I 'm thinking that using an ESP should be for multiple purposes. I thought of many possibilities but the idea to put all sensors around one D1 mini is excellent. Maybe I would add a microphone.
Can't do much with mics in ESPHome, yet. No i2s input support (yet). They did just add i2s output, though, so maybe not too far away?
This would make a good wall mounted sensor next to something like a light switch. Most light switches are placed in way to reach places and a great place for presence sensors. Would also be easy enough to wire.
Inclusion of lux sensor is great, i have some zwave/zigbee pir sensors with lux and this value is slow to change, so i use this value as condition in automation only for one light.
Would be great to include some battery, so it can work even when power goes out - to be able to use it as part of alarm system (alarmo) when i'm not at home.
Hi Lewis ,
that is the greatest Presence Detection Sensor ever. the ones I ordered and installed had some delays in ditecting movement and presence and most of the time did not function as i hoped for.
i wonder if I could buy from you a complete built "Smart Home Presence Detection Sensor" with the software? 🙏
I love the PIR esp8266 combo. Never thought of going w/ a higher quality sensor though. That is a huge step forward. Impressive project!!! Definitely something worth buying assembled to support/further your r&d. Well done.
I think you are close to a winner here, in terms of features. If the level of interest here in the comments is anything to go on, many people would like to buy a version of this sensor, either as a finished good or as a kit. A few thoughts on that.
First, the current appearance is not going to fly in many households. "Spouse approval factor" aside, _I_ think the current version is unattractive enough to not want to deploy it outside of a lab or test environment. So it'll need a better case , which will add to the cost.
Second, and related, with the current bill of materials (BOM), there's no way you can offer it as a finished version, unless you basically donate your time to assemble it. A few folks will pay US$75 for something like this, but probably not too many. And charging a price to make it worth your actual time and investment seems like we're in the US$100 range, as currently designed.
With those two things in mind, it seems like a v3 is inevitable. (Also, it seems like building these is fun, which similarly makes it inevitable.)
For that next version, maybe consider two things:
- What's essential, and what's nice to have? What can you take out to get the BOM down to something under, say, $US40 in parts? US
$25...?
- Similarly, what's essential if you were to offer this as a kit? Could you do "just the hard parts", and let buyers source their own components, assemble, and so on, to keep the costs (to you) down, and the price (to us) affordable?
I'm specifically thinking of another person here on UA-cam, Peter Millard, who spent a few weeks developing a design for a woodworking fixture (a DIY mount and hinge for a track saw rail). He was looking at creating it as a complete kit that you glue/screw together, but the BOM didn't pencil out. Instead of throwing in the towel, he picked out the hardest part of the design to get right, and is offering just that part as a physical product. (Plus plans, at a small cost.) It's smaller, he gets much higher yields on the CNC, and shipping costs go way down. See one of his final videos, putting them for sale, here: ua-cam.com/video/LsFN4ChQdF4/v-deo.html
So, similarly, for this sensor, there's the main board, and there's the case, both of which require either specialized tools (soldering, 3D printer), or the skill level required to get it right is high enough that it's a barrier. Most people can source the semi-generic sensor parts, and lots of us can manage with ESPHome or what have you. But I'm not buying a 3D printer any time soon, and my soldering skills make yours look terrific. ;-)
Another thought would be, could you design a custom circuit board to make assembly and soldering easier, and offer _those_ for sale?
Anyway, super interesting video series you have going here, and I'm looking forward to what comes next in this project!
Really cool! I'll probably wait to do something like this until thread is on esphome on the esp32-h2; the ZigBee functionally is pretty significant to me with an already-fun amount of devices on my wifi network.
This is really a great video on a very interesting project. It seems to me that this sensor has all the functions I need. I would like to rebuild the sensor to use it in several rooms. Unfortunately, my knowledge of electronics is not sufficient to rebuild the project with the available information. Therefore the question, will you publish a detailed plan and if so, when?
Thanks! I plan to at some point yes when I get some time, having said that...stay tuned for more!
What I missed from your built: Nicer design. Does a closed enclosure work (of course with a hole(s) for the PIR/Microwaves)?
I haven't got around to designing the faceplate for it yet. Slots are needed in the side for temperature and also light level to work
Very impressive work, your skillz are to be admired. Please enlighten us on how we can build or optionally purchase this fabulous device
Great video👍. As a tinkerer I'm would love to make a few of these rather than buy. Any chance of linking a circuit diagram in the description?
And your pilot newsletter was 👍 thanks for that! I added a few automation from it.
Just a thought the base, you could 3d print a plug mount so you can get the thing off the worktops like you can get for the Echo Dot.
Pretty cool idea and looks to be a fantastic solution, certainly will be on my upcoming build list.
🙏🏻🙌🏻
Brilliant! I really need to dig out my soldering iron, currently using Hubitat with Hue motion sensors and they do the job. But, at some point when I'm time rich, I'll be moving onto Home Assistant (assuming I can purchase the hardware, global chip shortage and all).
Thanks, look forward to seeing what you make!
Love the video. Your soldering skills need some love though...
i am inspired. Looking at this, I realise i still have a lot to learn. But I'm going to try to replicate this. i hope some more info comes out on how to put all these sensor together in esphome code.
But for now Google is my Friend ;-) and thank you for sharing this
I see it's sold out. Will you be making anymore of these?
Thank you for this very helpful video! I've not tried this setup yet but on many occasions struggle to solder properly due to the solders I have had or tried, would be helpful if you have a link to the one you use or one similar to get. Thanks
Really nice work! I’ve been looking for a good presence solution for a while now and I was thinking about building an air quality sensor. Now I want to build both in the same device… and both are far beyond my skill level, but now I’m going to try anyway.
Good luck!
This sensor is genius, thanks for posting. I guess it's your fault these millimeter wave sensors are sold out. 😁
I'd like to ask you if you could review how the bed occupancy works with the millimeter wave if it is mounted under the bed pointing to the floor. The reason I ask is because i don't want cabling on the floor, if not for other reasons it's the vacuum cleaner. Would be really nice if you could test this.
Excellent. Seems like you should start commercial production!
Great video as always, just wished it was more detailed. For Example on what Pins was connect to what on the esp-wroom-32 and the mmsensor :/
But I look forward to the writting artikel :D
Have one question:
The mmsensor needs 3,6-5v, but the esp-wroom-32 only has 3,3v output pin, so did you connect the voldtage pin on the mmsensor to something else? 🤔
Ps. I have never done this before so might have misunderatood something :p
Outstanding video, wow! So many ideas. Thanks!!
My perfect room sensor also needs sound, fire and O2/air sensors... O2 for healthy air monitoring and sound to pinpoint unusual events when room is vacant, limit music volumes or disable unused speakers... Fire/smoke sensors for the ever present battery dangers.
Might take a look at the BME680 gives you temp, pressure and air quality.
Every time I see a video about motion sensors I want to get one, then I remember I have a cat.
I'd love to get the kit but I noticed shipping adds 50% to the price! I'm in central Europe, I don't think shipping from UK/Ire should be that expensive
Otherwise I'd buy the kit in a minute, it's a perfect solution for the living room
There is nothing I would love more than to ship it for cheaper, but Brexit, War and the price of fuel make the cost more expensive unfortunately.
No problem, next time maybe 😅
@@EverythingSmartHome yeah, I'm the same guy that you responded to on facebook :D I'll keep the bookmark somewhere and keep checking
Wow! That's just great! It is really hard to imagine more functionality than you just added. Rather expensive, but every cent you spent has its reason explained. Congrats. I will definitely assemble one of this (and probably a cheaper version to compare if it fits my needs). A question here: have you thought of ways to compare human presence vs pet presence? There are some automations I would not like to have my 40 Kg golden retriever triggering...
BTW, I know bluetooth can differentiate pets, but I'm usually not carrying devices at home...
Awesome! I really like your DIY videos! Thanks!