Follow

@lupyuen I ran some basic tests on the BL602 and the ESP32. It's like comparing a weird WNIC to a device that was clearly designed to retrofit machinery and then upgraded.

@AmpBenzScientist Haha interesting ... I suppose we'll read more in your paper? 🙂

@lupyuen I will share the results. It was a means to verify the documentation of both modules. I put a BL602 in a closed room with small windows and concrete walls. The ESP32 was placed outside the room in an open area surrounded on three sides with concrete walls, one of which had a window. The two boards were used in BLE Mesh Node configuration (guess what I tried to test). So I used a smartphone and walked down both hallways continuously scanning for BLE devices. The BL602 has a BLE performance (on paper.)

@lupyuen *better performance on paper.

The ESP32 was only able to be detected in a very small area (<10m) while the BL602 was able to be detected up to approximately 20-30 meters in this test.

@lupyuen Not very scientific. I was trying to set up a mesh, it didn't work out. I was curious so I just carried out this crude test. The concrete walls will be a recurring theme.

@AmpBenzScientist Ah very interesting results ... Someday we should ask some students to do the complete experiment ... We can point them in the right direction

@lupyuen I used the ESP boards and thought they were okay. After the BL602 learning curve, I see two different products with two different purposes.

@lupyuen The ESP32 is the evolution of the ESP8266. The 82 was designed for serial to wifi communication. Both the 32 and 82 do this very well. The reason why it ended up in IoT was likely due to it being so cheap but reliable. I liked your take on the differences between the two.

@AmpBenzScientist I'm thinking ... Maybe RISC-V was inevitable for ESP and other mass market WiFi modules?

(1) ESP has always been hard to program because of its uncommon architecture. (LLVM and Rust took a long while to get ESP support)

(2) Licensing the IP from Arm or Tensilica was probably too expensive

(3) US-China Trade War probably pushed the China chip makers towards RISC-V.

@AmpBenzScientist What bugs me about BL602 and ESP32-C3: The Boot Code is difficult to replicate to other operating systems (other than the native FreeRTOS)

Maybe because of the EFuse Protection that protects the Flash ROM on these devices?

ESP32 is starting to get Zephyr OS support, and BL602 is getting NuttX support. Booting anything on BL602 / ESP32 bare metal is very difficult (e.g. Embedded Rust).

@lupyuen I called it lol. Ever since I got the k210 dual core 64bit RISC-V SoC in the M5 Camera, I knew that RISC-V was going to take over certain areas. I saw object detection faster than anything I have experienced before. This architecture is much like the Linux kernel of CPUs.

@AmpBenzScientist @lupyuen I was following RiscV since its inception, as part of a dream of exactly-what-we-wanted in terms of potential for freedom, modularity, but also lo-fi/frugal computing etc.

For a while i dreamt, and waited, and got convinced one future for freedom will have to be paved in RiscV chips!

I would have never expected that my first RiscV chip ever would be a... 25$ soldering iron!!! <3 amazing #pinecil

Prob a testperiment from @PINE64 towards riscV SBC (recntly anounced)

@jz @lupyuen @PINE64 I really wanted to make a half mini pcie wireless/bt card based on the BL602. It has mostly happened already.

@jz @AmpBenzScientist @PINE64 Erm GD32VF103 unfortunately has it's own headaches. And I think it's short on supply now.

Here's what I think about GD32VF103: lupyuen.github.io/articles/hey

@lupyuen With RISC-V it seems that the choice is either to pay a lot of money for open hardware or pay pocket change for a handful of odd Shenzhen Samples from good ole Guangdong, Zhongguo.

@AmpBenzScientist Another thing that bugs me about BL602 vs ESP32-C3...

(1) ESP32-C3 has a well-documented SDK. Unfortunately it seems to be force-fitted into a legacy API that's not so well designed.

(2) BL602 SDK looks more modern. But the docs are missing (and I'm filling in the gaps)

@lupyuen We just need a Japanese person to translate existing documentation. We also need a 中人吗 to actually translate. It's funny to me to interject a conversation in Kanji with Hanzi. It's like a computer glitch.

@AmpBenzScientist Actually the BL602 SDK Situation mystifies me...

(1) BL602 SDK seems to be created for AliOS. But it doesn't match the AliOS API completely. (Like the SPI functions)

(2) The SDK docs are not available in Chinese. I think most folks (including me) just copy the sample apps and tweak them.

But I'm really happy that the BL602 SDK actually works, once we take the time to figure it out.

@lupyuen I really want to help RE what remains, but I have to get results back to my professor. He wants a stable platform to build off of. I'm the hardware/firmware guy and he's the application developer.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.