About
This shield is compatible with the OpenMV N6 and OpenMV Cam RT1062.
The Wi-Fi HaLow Shield connects your OpenMV Cam to the world using the Morse Micro MM8108, a second-generation Wi-Fi HaLow (802.11ah) SoC operating in the sub-GHz bands from 850MHz to 950MHz. Unlike 2.4GHz Wi-Fi, sub-GHz signals travel much farther and pass through walls, foliage, and machinery far more effectively, giving you over a kilometer of range in open air.
The MM8108 supports 1/2/4/8MHz channel widths and MCS0-10, including sub-GHz 256-QAM modulation for up to 43.3Mbps at 8MHz. That's enough bandwidth to send images, not just sensor readings, and perform firmware updates. The integrated power amplifier delivers up to 26dBm (400mW), and receive sensitivity reaches -108dBm, so you can trade throughput for distance by dropping to a narrower channel and a lower MCS when you need to reach further. Full RF specifications are in the MM8108 datasheet.
Because Wi-Fi HaLow is still Wi-Fi, it's native IP. Your existing socket code, MQTT, and HTTP all work unchanged. There's no gateway translation layer and no proprietary protocol stack to learn. Security is WPA3 with Protected Management Frames (PMF) and Opportunistic Wireless Encryption (OWE), with hardware AES and SHA-2 acceleration. For a primer on the technology itself, see the Wi-Fi Alliance's Wi-Fi HaLow resources.
The MM8108 offers four low-power modes: Hibernate (<1uA), Deep Sleep (<2uA), Snooze (<20uA), and standby with SDIO from 29uA. Combined with the shield's power inputs below, this lets you deploy a camera that sleeps, wakes to capture and run inference, transmits only the result, and goes back to sleep, all on batteries and a kilometer from the nearest access point.
The Wi-Fi HaLow Shield also powers your OpenMV Cam from a 1.8V to 5.5V source connected using a standard 2.1mm DC barrel jack, attached via the 2-pin JST headers. This is perfect for powering your OpenMV Cam from three alkaline batteries in series using a standard battery holder with a 2.1mm DC plug. Two alkaline batteries in series also work, although the full capacity of both batteries cannot be used completely before the shield shuts down.
You may attempt to use four alkaline batteries in series; however, the fully charged voltage of four alkaline batteries in series may slightly exceed the onboard buck/boost converter's maximum allowed voltage. Proceed with caution.
The shield also features a wide voltage range 6-36V input, attached via the 2-pin JST headers, which is robust to reverse voltage and features robust transient surge protection. Use this input to power your OpenMV Cam from 12/24V power supplies without having to worry about damaging your camera. The wide range voltage supply outputs 5.6V on VIN via an ideal OR'ing diode so that you can power your OpenMV Cam from multiple shields with OR'ing diodes (e.g. all of our dual header column shields). This input can be used at the same time as the 1.8V to 5.5V input.
Finally, the shield has a robust 0-6V ADC input (AIN) connected to the 1.8V to 5.5V DC input, which outputs 0-2.8V to your OpenMV Cam on P6 for measuring the attached battery voltage. The input is robust to reverse voltage. If you don't need the ADC input, you can disconnect it from P6, too.
Note that the 1.8V to 5.5V input directly powers your OpenMV Cam via its RAW pin, which bypasses the onboard battery charger. The shield is designed to draw microamps of current in operation, allowing you to deploy an OpenMV Cam to run for years on alkaline batteries while your OpenMV Cam is in deep sleep (e.g. see the OpenMV Cam RT1062, which draws ~30uA in deep sleep).
Applications
This shield is perfect for customers who need to deploy OpenMV Cams far beyond the reach of regular Wi-Fi, across a factory floor, a farm, a job site, or a parking lot, without trenching cable or paying a monthly cellular bill. Regular 2.4GHz Wi-Fi gives up at the far end of a building, and LoRa has the range but nowhere near the bandwidth to move an image or update your firmware. Wi-Fi HaLow sits right in between, which makes it a natural fit for an OpenMV Cam that runs inference on-device and transmits detections instead of video.
