Firefly Technology has introduced the CAM-3576 series of tiny (38 × 38 mm) SBCs based on the Rockchip RK3576 processor with a 6 TOPS NPU for AIoT, edge AI, smart vision, industrial, and automotive applications. It comes in three variants, which include the CAM-3576Q38 (commercial), the CAM-3576JQ38 (industrial), and the CAM-3576MQ38 (automotive) modules designed for smart cameras, intelligent security systems, dash cams, and private on-device AI model deployment.

The CAM-3576 series supports up to 16GB of LPDDR5 RAM, up to 256GB eMMC flash, and also includes a microSD card for expansion. Additionally, the boards feature a MIPI CSI input for up to 16MP camera sensors with HDR support, Fast Ethernet, Wi-Fi 6, USB 2.0, USB-C (device), RS-485, UART, I²C, ADC, GPIOs, audio input/output, and RTC support.

Firefly CAM-3576Q38 specifications:

• SoM – ICORE-3576Q38

• SoC – Rockchip RK3576 (Q38 – Commercial) or Rockchip RK3576J (JQ38 – Industrial) or Rockchip RK3576M (MQ38 – Automotive)

• CPU – Octa-core CPU with 4x Cortex-A72 cores at 2.2 GHz, 4x Cortex-A53 cores at 2.0 GHz (1.6GHz for Industrial and Automotive)

• GPU – Arm Mali-G52 MC3 GPU with support for OpenGL ES 1.1, 2.0, and 3.2, OpenCL 2.0, and Vulkan 1.2

• NPU – 6 TOPS (INT8) AI accelerator with support for INT4, INT8, INT16, BF16, TF32 mixed operations.

• VPU

• Video Decoder

• H.265, VP9, AV1, and AVS2 up to 8Kp30 or 4Kp120

• H.264/AVC and MJPEG up to 4Kp60

• Video Encoder – H.264, H.265, MJPEG up to 4Kp60

• System Memory – 4, 8, or 16 GB (LPDDR4/LPDDR4x/LPDDR5 options)

• Storage – Up to 256GB eMMC flash

• Host interface – 2x 100-pin high-density board-to-board connectors

• Storage – MicroSD card slot

• Display – No exposed connectors on the board (SoC supports HDMI/eDP/DP/MIPI DSI)

• Camera – 40-pin 0.5mm MIPI-CSI DPHY FPC connector (1x 4-lane or 2x 2-lane configurations)

• Audio

• 2-pin 1.25mm MIC connector

• 4-pin 1.25mm Speaker connector (supports 2x 10W @ 6Ω or 8Ω)

• Networking

• 10/100Mbps Ethernet in via 8P-1.25mm box type connector (Built-in S18610G PHY only needs an RJ45 jack with integrated magnetics)

• Dual-band Wi-Fi 6 (802.11a/b/g/n/ac/ax) module on board + antenna connector

• USB

• USB Type-C port (Device mode)

• 1x USB 2.0 interface via 24-pin expansion connector

• Serial

• 1x RS485 via 8-pin 1.25mm box-type connector

• 1x UART via 24-pin expansion connector

• Expansion – 24-pin 0.5mm FPC connector with 1x USB 2.0, 1x ADC, 1x I2C, 1x UART, 10x GPIO, and 5V power output

• Misc

• Reset and MaskRom buttons

• RTC battery connector (2-pin 1.25mm)

• Debug header (3-pin 1.25mm)

• Recovery header (2-pin 1.25mm)

• Power – 12V DC input via 8-pin 1.25mm box-type connector

• Dimensions – 38.11 x 38.09 x 11.54 mm

• Weight – ~21 grams

• Temperature Range

• Commercial – -20°C to 60°C

• Industrial / Automotive – -40°C to 85°C

While the CAM-3576Q38 is an “SBC”, it is comprised of a “SoM” on the top with the CPU, memory, eMMC storage, and PMIC, and a bottom carrier board with access to all the I/O and interfaces.

In terms of software support, the CAM-3576Q38 is compatible with RTLinux for real-time networking, hardware resource isolation, Flexbus, and DSMC, along with standard Linux and Buildroot for embedded product development. It also supports Docker container management, RKNN model import and export, and common AI frameworks such as TensorFlow, TensorFlow Lite, PyTorch, ONNX, Caffe, and Darknet, with support for custom operators. It also supports deploying Transformer-based large models and real-time AI workloads such as YOLO for on-device inference.

On the Wiki, I found that the SBC is supported by Firefly’s Linux software stack, with full SDKs, BSPs, and documentation. It supports building and flashing Linux firmware, USB and SD card upgrades, MaskRom recovery, and serial debug. The platform includes device tree (DTS) documentation, NPU and large language model support for on-device AI deployment, camera and peripheral drivers, watchdog and RTC support, and a detailed Linux user guide. More information, including source code, tools, and reference materials, is available on Firefly’s resource downloads page.

There are plenty of other SBCs, Mini PCs, and SoMs built around the Rockchip RK3576 SoC, including Radxa ROCK 4D SBC, NanoPi M5 Mini PC, and Radxa CM4 SoM, among many others. Ultra-compact SBCs include the Radxa ROCK Pi S, the Quantum Tiny, the Tiny NanoPi NEO3, and the MStar MSC313E-based BreadBee. The Firefly CAM-3576 single board computers combine the power of the Rockchip RK3576 SoC with a tiny 38x38mm form factor.

The CAM-3576Q38 RK3576 tiny AI SBCs are available on AliExpress in commercial, industrial, and automotive variants, with prices starting at US$180.47 for the 4GB/64GB commercial model, US$191.44 for the industrial, US$229.86 for the automotive, and up to US$301.20 for the 8GB/64GB industrial version. Bulk orders (10+ units) get an extra 2% discount, with taxes added at checkout. The commercial version (4GB/64GB) of the “mini single board computer” is also sold on the Firefly shop for $139, and the module itself for $129. More information is available on the product page.

Debashis Das is a technical content writer and embedded engineer with over five years of experience in the industry. With expertise in Embedded C, PCB Design, and SEO optimization, he effectively blends difficult technical topics with clear communication

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.