If you’ve ever used a Raspberry Pi 5 with PCIe add-ons, you probably know the struggle — NVMe drives use an M.2 M-key connector, while many AI accelerators (like Google Coral or Hailo) and WiFi cards rely on A/E-key connectors. That usually means juggling different HATs or swapping hardware around just to test different modules.
Maker Mehrdad Majzoobi decided that was a hassle worth fixing — so he designed an open-source M.2 M-key to A/E-key adapter board that makes switching between NVMe storage and AI accelerator modules a breeze.
Majzoobi explains the frustration well:
“It is frustrating that we need to buy a separate PCIe HAT for using AI accelerators or WiFi cards that do not use the same connector as NVMe drives. This also makes swapping out the PCIe HAT to use Pi 5 with non-storage devices painful.”
The adapter is a simple passive board — no active components, no fancy chips. It just bridges the M.2 M-key slot on your Pi 5’s PCIe HAT to an A/E-key slot, letting you use AI accelerators, WiFi modules, or other PCIe devices without extra hardware headaches.
Sure, you can find similar adapters online for a couple of dollars, but Majzoobi had another reason for making his own:
“Saving money was not the reason I designed this… I trust things I design myself more than something super cheap on AliExpress, especially if it 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.”
He also points out that if you build in bulk, these can be made for under $1 per unit at 100pcs. Not bad for a tool that can make your Pi 5 setup way more flexible.
The design is fully open-source under the GNU General Public License v3. You can grab the files from the Ubo Pod hardware repository on GitHub, and Majzoobi shared more details in his Reddit post.
If you’re working on a Pi 5-based project with AI accelerators or swapping between NVMe and other PCIe devices, this little adapter could save you a lot of time (and frustration).
This project is a perfect example of the DIY spirit in the Raspberry Pi community: identify a real pain point, build a simple but clever solution, and share it with everyone.
It might not be flashy, but for anyone experimenting with edge AI on the Raspberry Pi 5, Majzoobi’s adapter could be a game-changer.
0 Comments