Home Automation from scratch

[pknoe3lh]

Hello everyone

I'm starting a series of videos building a home automation system from scratch. I want to make it open source and maybe make an new standard for DIY home automation.

Here the introduction into the series:
https://youtu.be/hzYn_wvNji8

I want to do it as generic and open as possible.
So Feedback ist needed and welcome ;-)

Greetings
im-pro

[pknoe3lh]

The next video is online:

https://youtu.be/p9erX_hQ0Z0

[Cliff Matthews]

IMO, unless this thread is asking questions about a project or design, this thread should be moved to: https://www.eevblog.com/forum/other-blog-specific/

[pknoe3lh]

Sorry die not know that this is the wrong froum. The discription tells: "Show off your projects or other stuff you are working on. Talk about designs or ideas, ask technical questions, and share technical information. This is the big catch-all thread for anything electronic".
How do I move the thread?

[Simon]

I am moving the thread. It's not really the place to dump links to other blogs, it's more to talk about the more technical questions.

[awallin]

Is there an open-source project for a surveillance-camera like the Nest cameras?
Seems the hardware for this is sub 100 $/eur but they want 3-500 $/eur for a Nest - would be useful and interesting project IMO...

[pknoe3lh]

I don't know what the nest can do. But look at banggoods:

https://m.banggood.com/de/TTGO-T-Camera-Plus-ESP32-DOWDQ6-8MB-SPRAM-OV2640-Camera-Module-1_3-Inch-Display-With-WiFi-Bluetooth-Board-p-1426498.html?rmmds=search

It's very cheep and also a WiFi cam with face detection.

[pknoe3lh]

Now to the Network:

https://youtu.be/QIOLc_1VYxM

any ideas?

[pknoe3lh]

Now part selection:

https://youtu.be/m7WxveJLYRI

There is the List:

https://github.com/pknoe3lh/DIYsmarthome/blob/master/PCB/room/BOM.md

60€ is really nice I think :-)

[Shock]

Interested to see where this goes, a couple of design considerations if I was doing the same. Make it a true closed system, don't interface or rely on preexisting products, and if you have to then make them optional. People are going to feel differently but inviting google and the like into your home automation is a bad idea because it's exactly what they want. The other is device, network and data encryption security built into wired, wireless networking and storage.

[pknoe3lh]

Make it a true closed system, don't interface or rely on preexisting products, and if you have to then make them optional.

Yes I don't what it to rely on a random clowd service.
I want to have speak control. At first this will be done over Google or Alexa but later I think it should be possible to make this Funktion locally if you use the syntax given.

The other is device, network and data encryption security built into wired, wireless networking and storage..

The esp 32 has hardware accelerated for encryption. Do have some ideas how to implement that? How to make the key exchange? I'm not an expert in this field and if you do something wrong its wrong anyway....

[Shock]

No its been years since so I'm way out of date.

I'm not the average consumer but if this was an any off the shelf product (not just home automation) the main considerations I would have is that it can be both hardwired and wireless, is secure and can't be bypassed easily by a man in the middle type attack, give people infinite attempts to authenticate or worse, zero security. I also like devices that are largely self sustaining.

So it must not rely on any form of phoning home or requiring an app or any interface for configuration that isn't commonly available. i.e. a web interface with the exception of a secondary command line interface, it's got to be something that you don't have to dive into a crawl space to update the code or configure. You should not be locked out because you aren't running the right version of the app and your phone no longer supports it. If you have to make an app to win consumers over make it an accessory not a necessity.

Lastly and this is rare to find in a product, is a method to remove all interfaces (not just switch it off in a config) but physically any interface code from the product so once it's running as expected you have to jump through hoops to mess with it.

All this is more important if you have cameras, locks, and alarms that can be enabled and disabled.

[pknoe3lh]

The PCBs are finished! Here a timelaps video of the design process:

https://youtu.be/YQT2HXxLCkw

[pknoe3lh]

And a sort explanation of the PCB Design:

https://youtu.be/Ae0AMbXFxIA

[Shock]

Congrats on getting the first design and layout together.

[pknoe3lh]

Congrats on getting the first design and layout together.

Thanks I'm really happy that I found the time for that. It was around 40 hours. What needed the most time was the io resources problem. This was really challenging but I hop I did it right. The Esp32 iomux is really complex and limited in some ways it's really easy to overlook something ....
Root for me :-)

[Shock]

Thanks I'm really happy that I found the time for that. It was around 40 hours. What needed the most time was the io resources problem. This was really challenging but I hop I did it right. The Esp32 iomux is really complex and limited in some ways it's really easy to overlook something ....
Root for me :-)

Yeah I thought that would have taken a good chunk of time, a nice example of what is involved in getting a project of this size onto a PCB.