Have a look at Hi-Link from China (on Aliexpress).
They offer 50GHz, 10GHz, 24GHz and 60GHz FMCW radar module boards for 10-30€ which are easy to reverse engineer on firmware and PCB level.
A lot of them are CW radars, a few FMCW and they also use the Infineon 60GHz radar chips. Very unusual for Infineon: with all tools and datasheets available without signing a NDA. Down to the register level.
There's an interesting Chinese company around (ICLink with their ICL1122/ICL1112) which offer highly integrated radar ICs. They can spit out raw ADC measurements of their downmixed baseband using Quad SPI at 20 or 40MHz when switching on their debug mode. Price range: ~10€. But datasheets are difficult to find. Example board: LD2410S.
The analog ones are easy to play with. You just need a DAC to drive their VCO and then can sample the I/Q pins with an ADC. That how a lot of the 24GHz modules (like the LD1125H) work.
> The analog ones are easy to play with. You just need a DAC to drive their VCO and then can sample the I/Q pins with an ADC
Do you have any reference or notes on how to access the IQ pins on one of these devices (ideally one of the FMCW ones)? I've been wanting to play around with one of these 24 GHz or 60 GHz units for coherent radar but it seems like most of the boards only report on distances over serial links. If there's an easy way to tap into the analog IF signal after down conversion I'd love to see how to do that!
What's cute about this is how far they went to make hobbyists happy. There's a way to connect it to Bluetooth and WiFi, and an Android app. There's compatibility with some Adafruit products. That's unusual for IC data sheets.
This is a phased array device. Angular resolution 20 degrees, range resolution 1 meter.
It's not a Doppler radar, so it can detect fixed objects. So if you're using it for people detection, you have to tell it where the fixed objects are. A ceiling mounted unit will see the floor. OK for people counting and such. Range is only 10 meters.
If you just want a motion detector to turn on a light, and IR isn't working for you, there are cheaper microwave detectors.
> What's cute about this is how far they went to make hobbyists happy.
Professionals are using hobbyist tools more and more, since these tools are just so much easier to use and don't require wading through datasheets which is like filling out tax forms.
Also IoT usecases require flexibility in connectivity.
Plus, rapid prototyping makes it easier / more likely to bring a volume product to market, and selling the same parts to hobbyists and manufacturers amortizes R&D costs more widely. It’s just makes a ton of sense all the way around.
As far as seeing and ignoring fixed objects, you can also remove any returns that have a near-zero velocity in radar and focus only on those objects that are moving.
Of course, indoor settings have a lot of non-stationary objects as well that might not be targets of interest to you, like fans, curtains blowing in the breeze, etc. So you can also develop algorithms to remove those signatures too.
Seeing fixed objects can be beneficial as well, for example, if you have a sensor deployed in a room but you don't know a priori what the room looks like. Longitudinal results and long range statistics can take you pretty far in seeing the room extents and layout and furniture, etc. Though a lidar sweep is better if you can get it
I’ve played around with this (the KIT-CSK-BGT60TR13C). It is a very interesting gadget and the kit doesn’t need much configuration. In an afternoon I made a car detector for my bike that detects the presence of an approaching car from behind. I thought I had a winning startup idea then found out this idea has existed for years. :(
They don’t suck. The Garmin one is also a rear light. It integrates with the bike computer to show detections.
When a car approaches from behind you get a visualisation on the bike computer showing you how far away the car is and an audio alert. If the car is approaching fast you get a different audio alert and the light flashes differently to warn the approaching car. It also detects up to three cars.
There might be improvements to be had but the implementation is pretty solid.
Ah. The Garmin Varia looks like it’s $200-300. Would be hard to go up against a brand like that, but wonder if you could compete on cost. The margins look great.
There is tons of cheaper alternatives now. Some chinese brand. The implementation is particularly hard to get right (how to avoid object that enters the radar field of view like passing next to a fence among other tricky scenario like a car that follow at the same speed as you but suddenly accelerate to pass you, etc). A lot of competiting radars will have false positives or negatives. That’s why Garmin is still the favorite product on the market. Wahoo just released one as well, and the reviews are quite positive so far. Bryton makes one, Magene, Giant, Trek, igpsport, etc.
Not even that. I think I had mine for under 150 (on some sale). Also, there are a few cheaper alternatives, but some tests have shown them to be less reliable (either false positives, or worse - false "all clear").
There's one feature they all miss. They only measure relative velocity to You. That's definitely better than nothing, but in my experience the most dangerous drivers aren't necessarily the fastest ones, but those who pass too closely. So, something like "angle of approach" measurement would be a feature I'd personally be willing to switch my Varia for something else.
There's some cheaper knockoffs but they tend to struggle with false positives, poor integration with our device, or weird noncompliant USB ports. (Most versions of the Garmin are still micro-usb so they're not perfect either)
Pleasantly surprised to see electronics parts manufacturers on front page HN. The advancement of sensors nowadays is staggering - I'm reminded of the recent airborne dust sensor that's literally orders of magnitude smaller because it uses a new backscatter detection mechanism. All this technology available at very reasonable prices too.
It is fascinating to me how 'hackers' went from being people who operated at the margins of established technology (2600, Capn Crunch anyone?), to being the literal establishment, in the sense that probably a good chunk of the 'hackers' here work at one of the big 5 or plausibly could.
> Regulatory issues were likely a major factor that led to the demise of Soli and Motion Sense on future Pixel models. Soli operates in the 60GHz frequency, which is reserved for military and government use in India.. Many of the Google Pixel 4's Motion Sense gestures are available.. Nest Hub's Soli radar extends far enough to detect when you're sleeping, and to track your breathing.
I got one of those cheap waveshare 60GHz modules with heart rate detection a few years ago. My goal was to build a sleep tracker - stick it under the bed pointing up, detect getting in and out of bed and falling asleep.
Unfortunately even just pointed straight at me with no obstructions and nobody else in the room the data was more or less random noise...
I wonder if now (after few more years of development, more reputable manufacturer and more money) it is worth trying again?
Thinking of starting a project to detect rainfall/snow/mist - has anybody used those mmWave sensors in a similar application? I've researched some of those and seems many are build for specific application, like a car rain sensor, which hardly provides broader output to apply further modeling.
Tie that chip to a beamformer (silicon labs have a few) and you have a phased array radar, which is a radar that does not move at all (pretty cool in my opinion)
Also, 15usd is not cheap for this kind of chip. You can buy a full wifi 7 rf/modem or a 4 core arm64 soc with this kind of money.
You can't use an external beamformer with this chip; it has the antenna built into the package itself. The chip doesn't have pins for RF input/output to bypass the built-in antenna.
60GHz radar is very different from WiFi. 15USD actually seems about right for the functionality this chip offers.
60GHz is actually poor for through-wall detection as this frequency is heavily attenuated by building materials - lower frequencies like 24GHz or UWB radar would be more suitable for both applications you mentioned.
An antenna large enough to work at 100 mhz would be impractical. That's a wavelength of 3 meters, and incapable of reliable detection of objects less than about 0.7 meters in all dimensions (1/4) wavelength, not to mention the required size of the antenna array
Update: chatgpt told me that the wavelength (3m) is too large to detect humans, antenna size would be impractical, more power would be needed for transmission, and reflections are more of a problem than with higher frequencies.
That band is often chosen _because_ of the absorption band of oxygen. Significant attenuation at those frequencies means it limits range, allowing for higher frequency reuse and less interference over a greater area.
If you’re worried about things exploding or something, don’t worry. Unless you were exciting a fuel/air mixture up to the combustion point you have nothing to worry about. So that would take a deliberate change to the chemical makeup of the air and much much much higher transmit power levels.
But I guess you should be careful if you are trying to modify this into an RF pulse welder and someone left their cutting torch gas cylinders venting nearby. </s>
obstacle avoidance for rc airplanes would be nice, maybe combine it with one of the autopilots out there. Learning to fly rc airplanes is pretty easy these days, the flight controllers can hold your hand through just about everything.
I built a FT Explorer with their FT Aura gyro system and it's so much more fun flying in a gentle breeze when not having to worry about keeping the plane upright and level. Radar ground detection and obstacle avoidance would be a crazy cool addition.
Amazon with all its power and incentive still can’t figure out there’s a limited number of toilet seats I might need. Linkedin is still offering me (50ish male, 30+ years in it, 15+ years of teaching in universities, board member etc) jobs as a kindergarten teacher. If anything, I’m amazed they are still so bad at this.
Heh, people detection, gaze detection and people hotspots detection (i.e. where people mostly stand, say at an intersection) have been a thing for 15-20 years now, I consulted on such a project in the 2000s, using simple mobile cameras that were placed in a van, the van moved, and they'd make "maps" and sell that data to advertisers who chose where to place fixed/changeable ads banners. They eventually got that data from cell providers, and didn't need the cameras anymore.
So much data collected for targeting ads, only to finally conclude that the only product I'm interested in buying is a handful of cool mmWave sensors from Infineon.
Have a look at Hi-Link from China (on Aliexpress).
They offer 50GHz, 10GHz, 24GHz and 60GHz FMCW radar module boards for 10-30€ which are easy to reverse engineer on firmware and PCB level.
A lot of them are CW radars, a few FMCW and they also use the Infineon 60GHz radar chips. Very unusual for Infineon: with all tools and datasheets available without signing a NDA. Down to the register level.
There's an interesting Chinese company around (ICLink with their ICL1122/ICL1112) which offer highly integrated radar ICs. They can spit out raw ADC measurements of their downmixed baseband using Quad SPI at 20 or 40MHz when switching on their debug mode. Price range: ~10€. But datasheets are difficult to find. Example board: LD2410S.
The analog ones are easy to play with. You just need a DAC to drive their VCO and then can sample the I/Q pins with an ADC. That how a lot of the 24GHz modules (like the LD1125H) work.
> The analog ones are easy to play with. You just need a DAC to drive their VCO and then can sample the I/Q pins with an ADC
Do you have any reference or notes on how to access the IQ pins on one of these devices (ideally one of the FMCW ones)? I've been wanting to play around with one of these 24 GHz or 60 GHz units for coherent radar but it seems like most of the boards only report on distances over serial links. If there's an easy way to tap into the analog IF signal after down conversion I'd love to see how to do that!
https://github.com/dm5tt/LD1125H_ReverseEngineering here.
What's cute about this is how far they went to make hobbyists happy. There's a way to connect it to Bluetooth and WiFi, and an Android app. There's compatibility with some Adafruit products. That's unusual for IC data sheets.
This is a phased array device. Angular resolution 20 degrees, range resolution 1 meter. It's not a Doppler radar, so it can detect fixed objects. So if you're using it for people detection, you have to tell it where the fixed objects are. A ceiling mounted unit will see the floor. OK for people counting and such. Range is only 10 meters.
If you just want a motion detector to turn on a light, and IR isn't working for you, there are cheaper microwave detectors.
> What's cute about this is how far they went to make hobbyists happy.
Professionals are using hobbyist tools more and more, since these tools are just so much easier to use and don't require wading through datasheets which is like filling out tax forms.
Also IoT usecases require flexibility in connectivity.
Plus, rapid prototyping makes it easier / more likely to bring a volume product to market, and selling the same parts to hobbyists and manufacturers amortizes R&D costs more widely. It’s just makes a ton of sense all the way around.
As far as seeing and ignoring fixed objects, you can also remove any returns that have a near-zero velocity in radar and focus only on those objects that are moving.
Of course, indoor settings have a lot of non-stationary objects as well that might not be targets of interest to you, like fans, curtains blowing in the breeze, etc. So you can also develop algorithms to remove those signatures too.
Seeing fixed objects can be beneficial as well, for example, if you have a sensor deployed in a room but you don't know a priori what the room looks like. Longitudinal results and long range statistics can take you pretty far in seeing the room extents and layout and furniture, etc. Though a lidar sweep is better if you can get it
What about for things like moving objects? I’ve used the AD pluto system but looking for something more affordable
I’ve played around with this (the KIT-CSK-BGT60TR13C). It is a very interesting gadget and the kit doesn’t need much configuration. In an afternoon I made a car detector for my bike that detects the presence of an approaching car from behind. I thought I had a winning startup idea then found out this idea has existed for years. :(
I'd like to know about this product, got more info? Not a bot, just a cyclist who's fed up of drivers who don't care about my safety.
Garmin do one called Varia
Can it detect the distance of the car? If so what's the range/accuracy like?
Could you share how you did it? Interested in this as a hobby
Those implementations might still suck.
They don’t suck. The Garmin one is also a rear light. It integrates with the bike computer to show detections.
When a car approaches from behind you get a visualisation on the bike computer showing you how far away the car is and an audio alert. If the car is approaching fast you get a different audio alert and the light flashes differently to warn the approaching car. It also detects up to three cars.
There might be improvements to be had but the implementation is pretty solid.
Ah. The Garmin Varia looks like it’s $200-300. Would be hard to go up against a brand like that, but wonder if you could compete on cost. The margins look great.
There is tons of cheaper alternatives now. Some chinese brand. The implementation is particularly hard to get right (how to avoid object that enters the radar field of view like passing next to a fence among other tricky scenario like a car that follow at the same speed as you but suddenly accelerate to pass you, etc). A lot of competiting radars will have false positives or negatives. That’s why Garmin is still the favorite product on the market. Wahoo just released one as well, and the reviews are quite positive so far. Bryton makes one, Magene, Giant, Trek, igpsport, etc.
Not even that. I think I had mine for under 150 (on some sale). Also, there are a few cheaper alternatives, but some tests have shown them to be less reliable (either false positives, or worse - false "all clear").
There's one feature they all miss. They only measure relative velocity to You. That's definitely better than nothing, but in my experience the most dangerous drivers aren't necessarily the fastest ones, but those who pass too closely. So, something like "angle of approach" measurement would be a feature I'd personally be willing to switch my Varia for something else.
There's some cheaper knockoffs but they tend to struggle with false positives, poor integration with our device, or weird noncompliant USB ports. (Most versions of the Garmin are still micro-usb so they're not perfect either)
Pleasantly surprised to see electronics parts manufacturers on front page HN. The advancement of sensors nowadays is staggering - I'm reminded of the recent airborne dust sensor that's literally orders of magnitude smaller because it uses a new backscatter detection mechanism. All this technology available at very reasonable prices too.
It is nice to see more technically oriented things on the main page. It is frustrating to see very high level more businessy topics.
It is fascinating to me how 'hackers' went from being people who operated at the margins of established technology (2600, Capn Crunch anyone?), to being the literal establishment, in the sense that probably a good chunk of the 'hackers' here work at one of the big 5 or plausibly could.
What's the best "Margins News" site today?
Slashdot is HN minus good behavior and a touch of 4chan. Their site is such shit with the ads though.
Slashdot was certainly dead after taco left
And webtech topics are equally boring and/or frustrating.
[dead]
"What Is mmWave Radar?: Everything You Need to Know About FMCW" (2022), 30 comments, https://news.ycombinator.com/item?id=35312351
"Inside a $1 radar motion sensor" (2024), 100 comments, https://news.ycombinator.com/item?id=40834349
"WhoFi: Deep Person Re-Identification via Wi-Fi Channel Signal Encoding" (2025), https://news.ycombinator.com/item?id=44685869
Feels like various companies have tried sticking this in smartphones and no one has yet found a use for it. At least in the smartphone form factor.
https://en.wikipedia.org/wiki/Pixel_4#Motion_Sense
https://www.androidpolice.com/google-pixel-4-killed-motion-s...
> Regulatory issues were likely a major factor that led to the demise of Soli and Motion Sense on future Pixel models. Soli operates in the 60GHz frequency, which is reserved for military and government use in India.. Many of the Google Pixel 4's Motion Sense gestures are available.. Nest Hub's Soli radar extends far enough to detect when you're sleeping, and to track your breathing.
Apple went full LIDAR instead. The extra resolution means you can go way beyond presence/motion/gesture detection.
I got one of those cheap waveshare 60GHz modules with heart rate detection a few years ago. My goal was to build a sleep tracker - stick it under the bed pointing up, detect getting in and out of bed and falling asleep.
Unfortunately even just pointed straight at me with no obstructions and nobody else in the room the data was more or less random noise...
I wonder if now (after few more years of development, more reputable manufacturer and more money) it is worth trying again?
Unfortunately even just pointed straight at me with no obstructions and nobody else in the room the data was more or less random noise...
What steps did you take to limit the detection bandwidth? SNR is often a bandwidth problem, not a sensitivity problem.
Thinking of starting a project to detect rainfall/snow/mist - has anybody used those mmWave sensors in a similar application? I've researched some of those and seems many are build for specific application, like a car rain sensor, which hardly provides broader output to apply further modeling.
Tie that chip to a beamformer (silicon labs have a few) and you have a phased array radar, which is a radar that does not move at all (pretty cool in my opinion)
Also, 15usd is not cheap for this kind of chip. You can buy a full wifi 7 rf/modem or a 4 core arm64 soc with this kind of money.
You can't use an external beamformer with this chip; it has the antenna built into the package itself. The chip doesn't have pins for RF input/output to bypass the built-in antenna.
60GHz radar is very different from WiFi. 15USD actually seems about right for the functionality this chip offers.
It looks great but the range is not, I already have some project idea in mind but with that range it won’t work.
Can this be used to monitor people through walls?
Could it be used as a "life" detector in collapsed structures after earthquakes?
60GHz is actually poor for through-wall detection as this frequency is heavily attenuated by building materials - lower frequencies like 24GHz or UWB radar would be more suitable for both applications you mentioned.
Why do you say 24GHz and not, e.g., 100MHz? (Apart perhaps from it being within a reserved band).
An antenna large enough to work at 100 mhz would be impractical. That's a wavelength of 3 meters, and incapable of reliable detection of objects less than about 0.7 meters in all dimensions (1/4) wavelength, not to mention the required size of the antenna array
Update: chatgpt told me that the wavelength (3m) is too large to detect humans, antenna size would be impractical, more power would be needed for transmission, and reflections are more of a problem than with higher frequencies.
60 GHz is pretty much Line of Sight. If you can't see it - it will not work.
Interesting that it happens to be the harmonic frequency of oxygen. For these power levels it is probably fine but something to keep in mind.
That band is often chosen _because_ of the absorption band of oxygen. Significant attenuation at those frequencies means it limits range, allowing for higher frequency reuse and less interference over a greater area.
Wouldn't it be the same if you just decreased the transmission power?
If you’re worried about things exploding or something, don’t worry. Unless you were exciting a fuel/air mixture up to the combustion point you have nothing to worry about. So that would take a deliberate change to the chemical makeup of the air and much much much higher transmit power levels.
But I guess you should be careful if you are trying to modify this into an RF pulse welder and someone left their cutting torch gas cylinders venting nearby. </s>
Pretty cool, I imagine it's only a matter of time until they're plug and play with Home Assistant
I don’t know how similar is to this board, but I’ve already had this 60GHz gadget running with esphome connected to homeAssistant for a few months! :)
https://www.seeedstudio.com/MR60BHA2-60GHz-mmWave-Sensor-Bre...
obstacle avoidance for rc airplanes would be nice, maybe combine it with one of the autopilots out there. Learning to fly rc airplanes is pretty easy these days, the flight controllers can hold your hand through just about everything.
I built a FT Explorer with their FT Aura gyro system and it's so much more fun flying in a gentle breeze when not having to worry about keeping the plane upright and level. Radar ground detection and obstacle avoidance would be a crazy cool addition.
“Quick! How can we use this to gather data for ad targeting?”
Amazon with all its power and incentive still can’t figure out there’s a limited number of toilet seats I might need. Linkedin is still offering me (50ish male, 30+ years in it, 15+ years of teaching in universities, board member etc) jobs as a kindergarten teacher. If anything, I’m amazed they are still so bad at this.
Don't worry, with the new AI tools they're going to get a lot worse
> Linkedin is still offering me (50ish male, 30+ years in it, 15+ years of teaching in universities, board member etc) jobs as a kindergarten teacher.
What's wrong with that? It sounds like you've got plenty of relevant experience!
Heh, people detection, gaze detection and people hotspots detection (i.e. where people mostly stand, say at an intersection) have been a thing for 15-20 years now, I consulted on such a project in the 2000s, using simple mobile cameras that were placed in a van, the van moved, and they'd make "maps" and sell that data to advertisers who chose where to place fixed/changeable ads banners. They eventually got that data from cell providers, and didn't need the cameras anymore.
So much data collected for targeting ads, only to finally conclude that the only product I'm interested in buying is a handful of cool mmWave sensors from Infineon.