-
Notifications
You must be signed in to change notification settings - Fork 179
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Hardware and Functional Requirements, and Integration of Matter with an Existing Project (CON-1433) #1171
Comments
I'm not Espressif employee, but I have been working with esp-matter for a long time.... It is very easy to switch CPUs with the esp-idf, you just set-target and recompile. So I would develop on an ESP32-S3 with PSRAM initially. It is a lot of hassle dealing with low RAM conditions so just avoid that initially. Then after you get your code working see if it fits on something like the ESP32-C3. All of the basic Matter devices will fit onto the ESP32-C3, so this is a function of how much other stuff you add. For example, my system implements a bridge and a lot of other stuff and it requires PSRAM. But that's just because all of the other stuff I added needs about 300KB of additional memory, if I removed the stuff I added it wouldn't need PSRAM. As for the other things, I run esp-matter in one task and I run esp-idf code in another task. As far as I know you can do anything you want in the other task as long don't do something like change the wifi mode when the other task is using it. I am also running an http server and esp-matter is basically unaware that I am doing that. You just need to implement some messaging between the tasks to coordinate. In my case esp-matter gets wifi running, then it I send a message which starts the http task. Look in the closed issues, other people have built custom pairing solutions. One way to do that is to get the device onto wifi using your non-matter solution. Then start Matter in on-network commissioning mode. On-network mode assumes the network is already running. I believe this is encoded into the QRCode so IOS will know to use on-network instead of Bluetooth. |
Hi Jon, |
At present, we are developing an ESP32S3 with LVGL screen, using 16MB of memory and 8MB of SPRAM. In terms of functionality, it includes MQTT and HTTP, BLE debugging, and OTA. The OTA I compiled is estimated to be around 6MB. If we only use HTTP functionality and the data volume is not large, the 4MB flash of ESP32S3 is sufficient. However, the premise is that after the network is distributed, BLE needs to be turned off to release about 80K of RAM |
I have a very similar system with LVGL and parallel LCD. We were running in RGB mode but that takes a lot of bus bandwidth so we are switching over to 8080 mode using the framebuffer on the LCD. Our LCD already supported both modes. To use 4MB flash with OTA your image has to be under 2MB (you have to store TWO copies of it) We are currently 2.5MB so we can't use a 4MB flash chip. We easily fit into 8MB flash and 2MB of PSRAM. We have built hardware with both the ESP32S3-PICO1-N8R2 and ESP32S3-R2 with external 8MB flash chip. Not sure yet which is cheaper in production. i was looking for better dev board and ordered one of these to try: https://www.moddable.com/moddable-six They don't have a schematic that I can find, but you have the source code for it with the LCD driver code. Systematically move everything to PSRAM and you can leave BLE on. I have 60K free internal and 1.6MB free PSRAM. That is with both wifi and BLE running. This is the largest annoyance: 40144 bytes of CHIP in internal RAM. Have not figured out how to push it into PSRAM. Another annoyance is that to use the flash driver your stack has to be in internal RAM. #1249 |
Dear Team,
We are planning to use ESP-Matter in our IoT project, and we have several questions regarding hardware and functional requirements as well as integration:
Hardware Requirements (to pass Matter certification):
Additional Functionalities:
Integration of Matter with an Existing Project:
We would greatly appreciate detailed responses to help us better plan the implementation and adaptation of our existing system to support Matter.
Kind regards,
The text was updated successfully, but these errors were encountered: