Setup: Pretty easy - biggest choice Thread or Zigbee. It’s configured as Thread out of the box, but I choose to switch to Zigbee to get all the bells and whistles. Replaced an FP2 with the FP400. The walk round the room process was a great way to begin setting up the room, but was fairly easy to fine tune that space in the app and add doors, furniture etc.
Once setup the zones worked brilliantly. Some needed a little fine tuning but can now detect very accurately, down to individual seats on the sofas and chairs.
Setup and tried a few automations using presence at triggers. Detection of the dogs vs people is a little hit and miss but again, a little tuning and and the spatial learning tool seems to that resolved 99%.
Seems faster and more accurate than the FP2. Still experimenting with light sensing and building those out.
Very happy with the FP400. I’m using mostly via HomeKit and honestly it’s not really missed a beat.
Will now be replacing my old Ikea PIR in the kitchen with the FP2 as there’s no need to be using the extra people sensing capabilities of the FP400 which is currently in a large lounge/dining area in 3 zones.
Really pleased with the upgrade overall
9 Likes
@mrplowuk Thanks for sharing your detailed experience with the FP400! It’s great to hear how you’re putting it through its paces.
Your setup approach really resonates — going with Zigbee to unlock the full feature set is a solid call, especially for a multi-zone space like your lounge/dining area. The fact that you’ve dialed it in to detect individual seats is impressive; that level of granularity is where these sensors really shine compared to traditional PIR.
A couple of things stood out from your post:
On the Thread vs Zigbee choice — you’re spot on that Zigbee mode opens up Aqara’s richer customization options, while Thread gives you direct Matter platform access without a hub. The FP300 (which shares this dual-protocol architecture) works the same way, so your experience aligns well with what we’ve seen from that product line.
On fine-tuning accuracy — your mention of using spatial learning and interference configuration to sort out the dog-vs-human detection hits on exactly the right levers. For anyone else reading, the key steps are: careful install height/angle, marking edges and entrances properly, setting interference zones for moving objects, running space intelligent learning when the room’s empty, and enabling AI person detection if needed.
The FP2 hand-me-down plan is smart — the FP2’s still a capable sensor for simpler spaces like a kitchen where you just need basic presence triggering. No point over-serving the use case.
Curious how the light sensing automations are shaping up as you build those out? That’s an area more users are starting to explore, and real-world setups like yours help the rest of us figure out what’s actually workable versus just theoretically cool.
Keep the updates coming if you hit any interesting edge cases
2 Likes