tinyVision.ai FAQ — MIPI to USB, FPGA Boards, and Engineering Services

MIPI to USB

What is a MIPI to USB converter?

A MIPI to USB converter is a hardware module that bridges MIPI CSI-2 camera output to a USB host as a UVC (USB Video Class) device. The host computer sees the camera as a standard webcam with no custom driver required. tinyVision.ai's tinyCLUNX33 SoM is a production-ready 25.4 × 25.4mm MIPI-to-USB module based on the Lattice CrossLinkU-NX33 FPGA, supporting up to 7 concurrent MIPI camera streams at 3.5 Gbps application-layer throughput.

How do I connect a MIPI camera to a USB host?

To connect a MIPI camera to a USB host you need a MIPI-to-USB bridge module. The tinyCLUNX33 Developer Kit is the fastest evaluation path: connect a 2-lane MIPI camera via an adapter board, plug into USB-C, and the device enumerates as a standard UVC webcam immediately. For production, the tinyCLUNX33 SoM embeds into your own carrier board design, handling MIPI receive through to USB enumeration and UVC streaming.

What cameras are compatible with the tinyCLUNX33?

The tinyCLUNX33 supports all MIPI CSI-2 cameras. Out-of-the-box driver support is available for the Sony IMX219 (Raspberry Pi Camera v2) and other popular sensors. tinyVision.ai can develop custom camera drivers for other MIPI sensors as part of their engineering services.

What is the maximum bandwidth of the tinyCLUNX33?

The tinyCLUNX33 delivers 3.5 Gbps at the application layer over USB3. It supports Full HD at 30 FPS as a UVC video stream and has been validated with dual HD camera streams at 4K/30fps total throughput.

What is the difference between the tinyCLUNX33 Connectivity and Compute variants?

Both variants use the same 25.4 × 25.4mm form factor and firmware stack. The Connectivity variant ($100) has 14 differential pairs and no onboard SSRAM — best for designs requiring high IO count, multiple simultaneous MIPI cameras, sensor aggregation, or display driving. The Compute variant ($110) has 6 differential pairs plus 32MB onboard SSRAM at 0.5 Gbps — best for image processing requiring large frame buffers, ISP pipelines, on-board inference, or edge AI pre-processing.

What is the development journey from prototype to production?

Stage 1 — Prototype: Use the tinyCLUNX33 Developer Kit with a camera adapter board and MIPI camera. Validates feasibility in one day. Stage 2 — Development: Use the Developer Kit as your firmware and software development platform. Write camera drivers, build Zephyr applications, and choose your SoM variant for carrier board design. Stage 3 — Production: Embed the tinyCLUNX33 SoM into your custom carrier board. A software IP license is required for production deployment. See the full tinyCLUNX33 Product Guide.

FPGA Development Boards

What FPGA development boards does tinyVision.ai make?

tinyVision.ai makes two open-source FPGA development boards: the pico2-ice ($49.99), which combines a Raspberry Pi RP2350B microcontroller with a Lattice iCE40UP5K FPGA on a single board with PMOD-compatible headers; and the UPduino v3.1 ($36), a standalone Lattice iCE40UP5K FPGA development board with open-source KiCad schematics used in university digital logic and embedded systems courses worldwide.

What is the pico2-ice?

The pico2-ice is a low-cost open-source development board combining the Raspberry Pi RP2350B microcontroller with the Lattice iCE40UP5K FPGA on a single compact board. It supports MicroPython, includes 4MB SPI Flash, 8MB qSPI SRAM, RGB LEDs, and PMOD-compatible 0.1” headers exposing all GPIO. It is the successor to the pico-ice and is used in university courses and professional rapid prototyping. Open-source KiCad schematic and layout, OSHW certified.

What is the UPduino?

The UPduino v3.1 is a $36 open-source Lattice iCE40UP5K FPGA development board widely used in university digital logic and embedded systems courses. It features 5.3k LUTs, 1Mb SPRAM, 120Kb DPRAM, 8 multipliers, and is PMOD-compatible. Full open-source toolchain support including Yosys, nextpnr, and IceStorm.

Engineering Services

What engineering services does tinyVision.ai offer?

tinyVision.ai offers: MIPI to USB system design and sensor bridging; FPGA RTL development on Lattice iCE40 and CrossLinkU-NX devices; Zephyr RTOS firmware and camera driver development; custom PCB and carrier board design; BLE, WiFi, GPS, LTE, USB, and GigE connectivity; AI/ML edge integration; and full prototype-to-production engineering support. The team includes engineers across the US, France, South Africa, and India.

What industries does tinyVision.ai serve?

tinyVision.ai serves medical device manufacturers, industrial automation companies, robotics developers, and UAV/drone manufacturers with MIPI-to-USB hardware and embedded vision engineering services. For education, they supply FPGA development boards to universities worldwide including UCLA and the University of Washington.

How do I get started with tinyVision.ai engineering services?

Email sales@tinyvision.ai with your camera, use case, target platform, and timeline. Most engineers receive a same-business-day response. You can also use the contact form.

Ordering and Shipping

How long does shipping take?

Most in-stock orders ship within one week, often sooner. Orders ship from Carlsbad, CA, USA. International shipping is available.

Where can I buy tinyVision.ai products?

All products are available directly at tinyvision.ai/collections/all. For volume orders or production quantities, contact sales@tinyvision.ai for pricing.