Hi all.
My house is based to a network of 12 ESP8266 sensors all powered by a 5v line and connected to a single MQTT dispatching info to HA.
I started building up my infrastrucure almost five… or six years ago.
Now I would likek to migrate step by step them to Zigbee.
Which is the best way to?
I have a DeConz key connectet ti my Pi3+ as platform.
Thankyou
s
Why would you want to do this?
1 Like
Your ESP’s do not support zigbee, so there is no other way than start everything from scratch, really.
And there are no open source zigbee stacks like esphome or arduino.
1 Like
Perfect question to ask! I did moved from a zigbee based infrastructure to a esphome wifi one and would never want to go back! More responsive and unlike to my zigbee experience it is stable and resilient!
3 Likes
Note that you can do a mesh network with wifi, too, not only with zigbee, if range is a problem…
Hardware to create the mesh is not cheap, though, while it’s built in zigbee / zwave / threads.
A downside of wifi is that there is no IoT standards, there, so every devices basically need a bespoke integration.
Not talking about esphome or devices directly supporting HA (via, e.g., MQTT), ofc. I don’t see a reason to replace a working ESP environment by zigbee, either.
True, but creating a wifi mesh is “universal” - it’s for smarthome and for wifi clients, too (phones, laptops…). But i found out that esp modules are quite sensitive and reliable - i’ve had a good and steady connection at signals as low as -90dBm.
1 Like
Just out of curiosity, why isn’t there a Zigbee stack for ESPHome or Arduino? Zigbee uses the same frequency ranges as WiFi, so presumably there’s a radio on the ESP board which can be used. Does it require totally different firmware on the WiFi chip? Or is it a licensing thing?
Not that I’d go that route anyway. I’m pretty happy using Zigbee for plug-and-play hardware, and ESPHome when I need to roll my own. But I could see a use case for building a custom Zigbee device.
2 Likes
Ok, so, all my sensors works in a wifi mesh network, I developed a simple watchdog routine in order to restart the lost ESPs, no issues…
…why move?
You’ re right.
Thankyou for clarify mi ideas.
s
1 Like
Good question, because there are ESP-boards that support Zigbee
Not all, you needs ESPs with Zigbee, and most don’t support it.
That’s definitely one of the causes:
There are also some disadvantages:
- Not free : It costs $3,500 USD at the time of writing to license the standard.
- Low data rates: It is not designed to support higher data rates.
- Star network is limited: The coordinator device supports up to 65,000 devices.
esphome has it’s own reboot system: if api is not connected or if wifi is not connected then it reboots itself after a while (time can be set, and disabled, too).
If i know correct ly then zigbee has smaller range than wifi, so a better mesh is required, and it has it’s own problems, too…
Zigbee can interfere with wifi, as i’ve read. So, for good working i think perhaps it’s not very good idea to mix it … ?
1 Like
Yes I know but the issue was in receiving and processing MQTT commands and it was not predictable when it will stop working.
So I send and MQTT message with a simple string that reset a counter on the device.
When the counter reach the set value I launch the ESP reset.
In case I want to refresh all the ESP infrastructure I cut the 5v that drives all the devices.
s
Do you have a 5v cable rail around your house, or do you power them in every room?
Just curious, i have 10 esphome devices and am thinking about power them in a centralized maner.
I am pretty sure the radio firmware on any of the following is NOT open source:
- espressif based wifi devices
- any zigbee devices
- any bluetooth devices
The ‘application’ layers above this are ‘open’ on each. Each have different levels/amounts of available ‘open’ knowledge to code at the a application level. Doing work with zigbee and bluetooth is requires more software engineering background as their lower power abilities are more interwoven with their core implementation. Even work with espressif based system becomes far more complex when you add low power/battery to the design.
I have not explored the new implementation of Zwave, so I am not sure if it’s radio firmware is open source. But there seems to be little internet based available knowledge and chips to build on.
Building and operating a DIY Zigbee device using the ptvo.info firmware will get most home automation hacker to maybe 80% of what you can do with a ESP implementation for most home automation tasks. And for a low cost you have the option for reasonable battery powered designs.
There have been ‘free’ SDK’s and examples for Zigbee stacks and chips from main line firms for some time. Texas Instruments is the only firm that I know of that went with a third party compiler that required purchase. And at a very high price, closing the door on DIY developers for their chips. Other than the fine work by ptvo to add an free ‘application’ layer on top of expensive TI based ‘application’ layer. These IDE’s, SDK’s as I have stated required a knowledge level of software engineering that most DIY folks don’t have.
Recently, several of the China based Zigbee chip manufactures have made their SDK’s and examples available. This allows full parity with WiFi Espressif DIY designs.
The new chips and SDK’s from Espressif (and I am expecting from others) to support the Zigbee and Thread radios will open up these options further.
So from my experience, I hope these points give you some information to correct the statement :
“And there are no open source zigbee stacks like esphome or arduino.”
2 Likes
From my experience, there is no ‘plug and play’ way to move between ESP, Zigbee, Bluetooth or Zwave as the based for DIY devices. You ‘application’ layer will need some rework to move among any of them.
I have found the ptvo.info based Zigbee ‘tool’ a good low cost way to build useful home automation devices based on the Texas Instruments Zigbee chips (and their Chinese clone chips). Have a look at the web site:
However, if you have something that is ‘working’, the best path is often to not mess with it. A hard thing for us DIY home automation hackers 
Good hunting!
Thank you, most interesting. I perhaps should have said that “there are no easily usable zigbee stacks like esphome or arduino”. Because there is just not the degree of DIY able software and documentation out there for zigbee.
pvto.info, while it is good as far as it goes, is NOT open source. And to leverage what (IMHO) is the great advantage of zigbee, ie very low power battery devices, pvto makes you pay for a licence that is locked to one chip. So if the magic smoke escapes, or a more powerful chip comes along, you pay again. Not entirely palatable.
2 Likes
For esp32 the last binary blob is the wifi driver which to my knowledge is mandatory if you want to get fcc approval.
Anyways the this very last blob gets already reversed engineered and it might be possible to only run open source code in the future (probably without fcc approval…)
You are using HA.
The standard IS ESPHome.
The purpose of a standard is to make things work. If ESPHome works for you, that’s all the standard you need.
2 Likes
In case you didn’t notice, a large majority of HA users are not DIYers nowadays…
If there was a de facto standard for wifi IoT, it would be Tuya, which is not to be recommended, imo.
Sorry, but lol 
A standard is just what the majority uses, and it’s definitely not necessarily the best (VHS anyone?)