[Polling for Interest] Cheap DIY mmWave Radar Sensors Using ESPhome - Buy or Build

💬 Lounge
1
  • I'd like instructions to build my own and set them up in HE
  • I'd like to buy a pre-built sensor that I would add to HE myself
  • Get lost Ryan
0 voters

tl;dr - I'm gauging interest before putting in the work to write a detailed guide to build an ESPhome-based mmWave radar for pretty cheap. If there's enough interest in it, I'm also toying with the idea of selling pre-built sensors that would just need to be connected to WiFi, added to HE, and configured.

One of the biggest doors that having Home Assistant has opened is the ability to use ESPhome. For me, the biggest win here has been the ability to build mmWave radars, with extremely granular configurations for sensitivity, for extremely cheap. For example, I have several built using:

ESP32-S2 Mini - $1.89
LD2410C radar - $2.71
3D printed case - $0.50
6' USB cable - $3

Add a few dollars for shipping and it's right around $13. Since the shipping costs do not scale when bought in bulk, 100 of these would cost ~$8/each.

The other benefit is that they're small. The configuration above is about 1" tall and 1.5" wide (38.85mm x 29.95 mm). The depth depends on which mounting back is used (I have different ones for mounting the sensor flat, at an angle, in the ceiling corner, and tilted down), but ranges from 0.5" to 1.25" (14mm to 32 mm). There are slight more expensive ESP32 options (~$2.20) that bring this size down even more.

image
image1638×1233 139 KB

The LD2410C has a built-in Bluetooth module allowing it to be connected to a mobile device and using an app to configure everything with it. The dashboard below is from Home Assistant, but the graphing in the app is extremely similar.

image
image1258×490 31.9 KB

Each number on the x-axis represents a certain distance (0.2m or 0.75m; configurable in the app) with independent threshold settings for detecting movement and still beings (can at least detect 45 lb dogs).

Why do I bring this up:
I still have some HE hubs sitting around. I was toying with the idea of working on getting these sensors integrated into HE and documenting the whole process of building the sensor to integration.

Also, realizing there are folks out there that haven't an appetite for soldering and/or access to a 3D printer, I thought I could bulk order the parts, assemble and complete the initial setup process (installing the ESPhome firmware on the ESP32 board), and sell these for ~$20-25 including shipping ($5 flat rate in the US). This would massively depend on interest. I value my time (like I hope everyone does), so unless I could get enough to drive the parts cost down, then it wouldn't be worth my time to solder, print cases, and setup. A smaller pool would result in a higher price which would negate the big upsell for these (aside from size).

4 Likes
2

Instructions definitely, would be a fun project, thanks for considering doing that.

Purchase maybe... Main reason is I'm not sure I really need any more mmWave sensors, my primary interest in doing the project is just for the fun of doing something new. :slight_smile:

Glad you still hang out around here, you always share good info and ideas. :+1::grin:

2 Likes
4

Sorry, I thought this was for other sensors hooked up with an ESP32, more than just radar.

5

Usually a post needs multiple flags to be temporarily hidden.

2 Likes
6

Just to clarify, do you 3d print the whole case including individual mounting backs? Or does the sensor case attach to different backing pieces?

I’d probably be interested in ordering a few, but haven’t really taken the plunge with mmWave sensors yet, so would probably have to experiment with placement.

1 Like
7

Just as an FYI, there is already someone doing this, although it's been sold out for awhile:

I'd be interested in building one (or more) myself, but I'd also be interested in supporting your work. Maybe an option to "tip" you for the instructions? That wasn't an option, so I voted for both build and buy. '-)

Would this support zones? That's an important option for me to meet my use case.

8

Why limit to only radar sensors?

9

OP seems to have a pretty specific objective in mind, with parts and pricing etc.

Personally, I appreciate that. Particularly as one of the people that wouldn’t be able to put the sensors together myself.

If everyone’s favorite sensor starts getting added on, the whole thing becomes a discussion of everyone’s favorite sensors, instead of the actual objective of the thread, which was to see how many people are interested in this specific sensor project.

2 Likes
10

I'm thinking the OP's thread title could've been more specific as well.

11

I wouldn't bother, and would just buy the pre-made Athom mmWave ESPhome capable sensor. I have one, and it works very well (better than my Aqara FP2 in my specific installation...).

I'm convinced that using only PIR to initiate motion detection windows, and then using mmWave to maintain detection of motion/presence is the way to go.

Disclaimer: for MY uses/installation locations. YMMV

(they often go on sale too, if one is patient)

3 Likes
12

The back is replaceable with the back itself providing the different mounting options. From left to right: corner mount that points the sensor downward (45° angle); multi-option mount that can be used to angle left, right, or down (30° angle); flat mount.

image
image867×277 27.1 KB

No. The one you linked is using a different radar that has that capability that's a bit more expensive, a little more tedious to setup, and (if I'm not mistaken) does not have sensitivity adjustments.

Because I'm not trying to start a business. My whole reason for doing any of this is more to help the community than anything else. Even selling just the radar sensor for $20 would be at a loss once labor is factored in. If I start adding in more components, that quickly turns into a negative number for me (and I do legitmately mean me losing money out of pocket) without jacking the price up which then, essentially, means I'm "competing" with the market.

Spot on.

That's definitely an option for folks to take. My thought with providing pre-built is that there may be folks wanting to dip their toes in, but don't currently have the know-how/tools to do so. So, instead of buying a $30 sensor, this could get them a prebuilt, one trick pony for about half that price to try out, take apart, test out flashing firmware, etc.

3 Likes
13

Makes sense. I guess I would say that without a PIR sensor to use in conjunction with the mmWave, I wouldn't use it.

But that's just me. Many others will be fine with just mmWave, and you are right that the price point/cost to make is great.

14

I wonder if that's a radar problem or perception or some combination. The LD2410C is pretty quick. Maybe a half second. I can't find what radar the Athom setup uses, but I can tell it's a different one.

1 Like
15

Other new designs are using PIR in conjunction with mmWave for false motion rejection. Here is another example (that coincidentally also uses a LD2410):

I think many people (like me) have converged on the idea that mmWave by itself takes too much tuning and/or has too many false positives on initial detection (especially with pets), and augment it with PIR.

Anyway, sorry to derail the thread. You have a neat design!

16

Not derailing at all. There's no issue highlighting some of the potential caveats.

The one you linked is using the OG LD2410B sensor. They have a number of different sensors available. Some folks have started building a comparison chart:

I could've sworn I saw a feature breakdown somewhere on HiLink's website, but I can't find it now.

2 Likes
17

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.