InHand Mo 62A and Mo 68A are two credit card-sized Edge AI single board computers (SBCs) powered by Texas Instruments AM62A and AM68A processors, delivering respectively 2 TOPS and 8 TOPS of AI performance.
Designed for AI Vision applications, the Mo 62A and Mo 68A SBCs support MIPI CSI camera input, Raspberry Pi HAT expansion boards, integrate 7th-generation ISP image processing capabilities, and handle H.264 / H.265 video encoding and decoding. However, they offer a different range of ports, as you’ll see in the specifications below. Target applications include industrial equipment, access control systems, robots, smart cameras, and visual inspection terminals.
InHand Mo 62A
Mo 62A specifications:
- SoC – Texas Instruments AM62A74 SoC
- Application processor – Up to four Arm Cortex- A53 cores @ 1.4GHz
- Real-time cores
- 1x Cortex-R5F @ 800MHz (MPU Channel with FFI)
- 1x Cortex-R5F @ 800MHz to support Device Management
- 3D GPU – None
- VPU – 4Kp30 H.265 and H.264 codec
- C7xV-256 + Deep Learning Accelerator up to 2 TOPS
- System Memory – 2GB, 4GB (default), or 8GB LPDDR4
- Storage – MicroSD card slot
- Video Output – Micro HDMI port
- Audio
- 3.5mm audio jack
- Digital audio via HDMI
- PCM via 40-pin GPIO header
- Camera Interface – 4-lane MIPI CSI-2 connector
- Networking
- Gigabit Ethernet RJ45 port
- WiFi 5 and Bluetooth 4.2 with on-board snap-on antenna
- USB – 4x USB 2.0 Type-A ports
- Expansion – 40-pin Raspberry Pi-compatible GPIO header with I²C, SPI, UART, PCM…; compatible with most Pi HAT expansion boards
- Debugging – TTL UART
- Security – Secure Boot, TrustZone, OP-TEE, Hardware AES-256
- Misc
- Reset button
- Power and User LEDs
- 4-pin fan connector; active fan optional
- Power Supply
- 5V/5A via USB-C port
- A 4-pin PoE header can be seen in the photos of the board, but it’s not documented.
- Dimensions – 85 x 56mm
- Weight – 47 grams
- Temperature Range – 0 to 50°C
The company provides a Debian 13.2 (Embedded Linux) image based on Linux 6.12. AI & vision capabilities are supported through TI TIDL AI Runtime with support for TFLite/ONNX, V4L2 camera framework, DRM/KMS display frameworks, and TI EdgeAI SDK. InHand also provides DeviceLive IoT device management platform for remote monitoring, edge app management, and remote maintenance of a fleet of boards. You’ll find some documentation and software development resources for the Mo 62A board on GitHub.
InHand Mo 68A
Mo 68A specifications:
- SoC – Texas Instruments AM68A SoC (TDA4VE / J721S2 family)
- Application cores – 2x Arm Cortex-A72 cores at up to 2.0 GHz with 1MB shared L2 cache with ECC
- Real-time cores – 4x Cortex-R5F real-time cores at up to 1.0 GHz
- GPU – Imagination Technologies BXS-4-64 up to 800 MHz
- VPU – 4Kp60 H.265 and H.264 codec
- AI accelerators
- Deep Learning Accelerators up to 8 TOPS
- Vision Processing Accelerators (VPAC) with Image Signal Processor (ISP)
- System Memory – 4GB or 8GB (default) LPDDR4
- Storage – MicroSD card slot
- Video Output
- MiniDP port
- Up to 2x MIPI DSI via 22-pin connectors (multiplexed with MIPI CSI)
- Audio
3.5mm audio jack- Digital audio via mini DisplayPort
- PCM and I2S via 40-pin GPIO header
- Camera Interface – Up to 2x 4-lane MIPI CSI-2 connectors (multiplexed with MIPI DSI)
- Networking
- Gigabit Ethernet RJ45 port
- WiFi 5 and Bluetooth 4.2 with on-board snap-on antenna
- USB – 4x USB 3.0 Type-A ports
- Expansion
- 40-pin Raspberry Pi-compatible GPIO header with I²C, SPI, UART, I2S, PCM…; compatible with most Pi HAT expansion boards
- 16-pin PCIe Gen 3.0 FFC connector
- Debugging – TTL UART
- Security – Secure Boot, TrustZone, OP-TEE, Hardware AES-256
- Misc
- Reset button
- Power and Status LEDs
- 4-pin fan connector; active fan recommended
- RTC with backup battery
- Power Supply
- 5V/5A via USB-C port
- A 4-pin PoE header can also be seen on the board
- Dimensions – 85 x 56mm
- Weight – 53 grams
- Temperature Range – 0 to 50°C
Software support is about the same as for the AM62A board, including a Debian 13.2 OS image based on Linux 6.12, TI TIDL AI runtime, V4L2 camera framework, DRM/KMS display frameworks, and TI EdgeAI SDK. The Debian image appears to be built with Armbian, and Python, C/C++, OpenCV, and GStreamer are also mentioned. The company has yet to share a software repository for the Mo 68A board on GitHub, but some (hardware) documentation is available there.
I’ve already highlighted the differences between the Mo 62A and Mo 68A specifications above with bold and strikethrough text, but the company also provides a more visual side-by-side comparison shown below.
InHand Networks is primarily a B2B company, although they sometimes sell samples of their products online. However, that’s not the case for the “Mo” boards, for which we don’t have any price and availability information. Additional details can be found on the company’s website and in the press release.
Thanks to TLS for the tip.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
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.
