How does this motion sensor work?

[TheBigT]

Hi Everyone,

I was having trouble with a car alarm that was installed in my car by the previous owner. It turned out that the root cause was an oversensitive motion sensor that was installed under the roof liner (see attached pics). The maker claims that it can detect movement inside the car (https://www.defa.com/product/motion-sensor-25m/), but I've never seen a design like this, and I'm puzzled as to how it works. Can anyone shed some light on its principle of operation?

Thanks,
  TheBigT

[SiliconWizard]

Looks like some kind of simplified doppler radar to me.

[DaJMasta]

My first guess was using that bar as some kind of deflection measurement, but it's not that flexible.  I know you can use piezo elements for this, maybe the plastic piece under the bar is the key?

I agree the thing on the end of it sort of looks like an RF stub.

[glarsson]

Probably a radar implemented using one transistor. See this video: https://youtu.be/9WiJJgIi3W0

[Benta]

The brass strip is a vibration sensor (sensor itself must be under the strip) for when the car is being moved/towed.
For the RF part I don't know.

[TheBigT]

Thanks everyone for the tips and ideas! The video with the cheap radar sensors was very interesting, too bad that the reviewer couldn't figure out how they worked either The construction of my unit is quite different though, it doesn't have the serpentine strip that all units in the video have, and has that mysterious brass bar instead.
There is nothing under the brass strip, it is just a strip, raised approximately 4 mm above the PCB. The brass is quite sturdy, it doesn't move or vibrate without applying quite a bit of force. The plastic piece at the end is a plastic screw, with a head for a flat head screwdriver at the bottom. Is it possible that they use a screw instead of a single post or standoff to be able to precisely tune the distance between the PCB and the brass?
There must be a good reason for that bar, since it is the only hand-soldered component on the PCB, and inserting the plastic screw must also take time and effort.

[TheBigT]

Taking a closer look at the RCWL-0516 module (https://engineersportal.com/blog/2018/7/5/arduino-wall-penetrating-motion-sensor-using-the-rcwl-0516-microwave-radar-module), the RF stage construction shows a lot of similarities between my module and that one. In both cases there is a high frequency transistor with base, emitter and collector connected to larger metal structures. The biggest difference is that the base on my module is connected to the brass instead of the serpentine trace. The small square trace for the emitter looks almost identical though.  The other difference is that my unit uses off-the-shelf components for the signal processing, instead of the the Single Magical IC. So I'm ready to accept the hypothesis that this is the same fundamental circuit, but I'm still super-curious as to how it actually all works.

[glarsson]

You are not alone. We are all curious about this alien design. A doppler radar built with one transistor and some pcb traces is just pure magic.

Do you have an instruction manual? Does the plastic screw do anything like adjusting the distance between the pcb and the strip? Tune for car cabin size or sensitivity?

[mikerj]

Here is a schematic of the RCWL-0516 that also shows how it works.  This is not a doppler radar as such, it seems like it depends on the tuning of the microwave transmitter being peaked when a reflecting surface (e.g. person) is moves through multiples of a 1/2 wavelength from the antenna and the circuit is simply a bunch of filtering and a comparator to count these peaks.

[glarsson]

Same BISS0001 as in most motion detectors.
Still magic in this single transistor circuit.

[Zero999]

Single transistor? There are two transistors near the antenna.

The transistor is just an amplifier in the oscillator. All the clever stuff is done by the integrated circuits.

[glarsson]

The RCWL-0516 page shows only one transistor, Q1. There's only one connection from the 3.2GHz part (Q1) to pin 14 (opamp input) on the IC. Everything RF is done by Q1.

[Zero999]

The RCWL-0516 page shows only one transistor, Q1. There's only one connection from the 3.2GHz part (Q1) to pin 14 (opamp input) on the IC. Everything RF is done by Q1.

I was referring to the picture attached to the original post, which clearly shows two transistors, although one of those three pin SMT packages could be a voltage regulator. In any case, the transistors are not doing anything remarkable. The one near the antenna is just an RF oscillator.

[IanMacdonald]

Probably microwave, and those things give more false alarms than an Orkney social worker.

[glarsson]

Probably microwave, and those things give more false alarms than an Orkney social worker.

It surely can't be that bad?

[TheBigT]

Turns out BigClive also has a video on the topic: