Mehrdad Majzoobi Releases an Open-Hardware Adapter for M.2 A/E-Key Modules on the Raspberry Pi 5

Maker Mehrdad Majzoobi has released an open source design for an M.2 M-key to A/E-key adapter board — intended to make it easier to chop-and-change between Non-Volatile Memory Express (NVMe) and machine learning/artificial intelligence accelerators on the Raspberry Pi 5.

"It is frustrating that we need to buy a separate PCIe HAT/HAB [Hardware Attached on Top/Hardware Attached on Bottom Raspberry Pi 5 add-on] for using AI accelerators (Google Coral, Hailo, etc.) or WiFi cards that do not use the same M.2 M-key connector as NVMe drives," Majzoobi explains. "This also makes swapping out the PCIe HAT to use [Raspberry] Pi 5 with non-storage devices painful. So I designed an open source simple adapter that lets you connect any A/E-key cards to most Raspberry Pi PCIe HATs/HABs made for storage."

The board itself is pretty straightforward: there are no active components, merely an edge connector for insertion into a Raspberry Pi 5's M.2 HAT or similar accessory exposing an M.2 M-key slot and an A/E-key slot for the device to be adapted — with machine learning and artificial accelerator modules top of the list, but other M.2-format PCI Express modules are also compatible.

"Saving money was not the reason I designed this," Majzoobi says, in comparison to off-the-shelf equivalents costing as little as $2 each, "but you can make this for less than $1 [per unit] at 100pcs. Generally I trust things that I design myself more than something super cheap on AliExpress, specially if takes me a short time to design. I ensure I follow all routing guidelines for PCIe high speed lines and stress test my designs afterwards."

More information is available in Majzoobi's Reddit post, while the design files are available in the Ubo Pod hardware repository on GitHub under the reciprocal GNU General Public License 3 — where it's used to provide easy swapping between NVMe storage and AI acceleration on the underside of an Ubo Pod unit.