Microwave Radar

tmpBA87As soon as one mentions “microwave radar” I’ll bet you’re thinking “expensive” – well how does under 2 quid sound…


These little devils are movement detectors, bought from Ali Express on a whim and they use the PCB as the antenna (3Ghz) – hence the ridiculously low price. So just a quick one – I ordered 2 and they arrived the other day – I’ve hooked them up to 5v (they work from 3v3 to 20vDC) and use less than 3ma. 3 pin connector gives a positive output on movement – and the output stays high for around 30 seconds.

Updated April 2017 - see also http://tech.scargill.net/microwave-for-the-weekend/

On the back there is a pair of connection points (soldering iron needed) and if you put a 10k resistor across them (I found this out by trial and error as there is very little info on these units – at least in English) the pulse goes down to around a second. VERY sensitive and although I’ve only done the briefest of tests I can confirm they can detect hand movement through 0.5” of chipboard !!

According to the blurb they pick up movement 360 degrees and have a detection of 6-9 metres. Applications include security, intelligent lighting, toys etc.

What I DID find when testing this and a later model (see above) is sensitivity to power supply - it is probably no good sticking it on the same supply as an ESP8266, the noise spikes will make the board unreliable. My solution is a 10r resistor in series with the 5v supply - and a 330u 6v cap (from the microwave board end of that resistor) - to ground - works a TREAT.

Both sensors work well, let's have your thoughts  (no theories about getting cancer from radiation please).


66 thoughts on “Microwave Radar

  1. interesting... which are the point where to solder the resistor? the 2 big holes near the corner marked as 2? why exactly need it? thanks 🙂

      1. thanks, i've seen them, thanks
        why exactly put that resistor? which improvement you get?
        and how does this sensor compare to a standard PIR one?

  2. I've looked at these a couple of times but hesitated to buy any, I'm wondering how good they would be outside (humidity etc.) to detect someone walking through a passageway at night, I suppose I should give one a try, only a couple of quid but all the 2 quids add up !....

      1. I was planning to put one in a sealed box with a esp-12 to post a MQTT status message and use Node red to check the time of day and send a push notification to my phone if someone walks through at night or if nobody is at home (need to work on the latter though )

        1. That seems extremely reasonable. You could even do that with an old ESP-01 - though as the ESP-12s are the same price.... yes, not a bad idea. ACTUALLY it's a very good idea as I've been pondering something easy to use for when I'm in my office (often referred to as a shed) and pop into the house...... hmmm...

  3. Hey Scargill,

    Have you tested the assertions of these modules working upto 6-9 meters.
    Please post whether it really works that far or just a marketing strategy.

    1. I think having started the ball rolling (there's nothing out there about these) I may leave location testing to others - right now mine is being awkward and staying on - which makes it difficult to test. More later when I have time to get the soldering iron out.

  4. I'm no expert in microwave sensors but they operate differently than an infrared, the microwave constantly pings out a signal and measures the time for the echo, like Ultrasonic, my concern is that if there is a wall too close to the sensor it may permanently trigger it due to the echo time being to small, if the software is good it should only trigger when the echo time changes as you move towards it or away.

    Also I wonder if you can put a metal shield below it to mask out cats and dogs.

    Have to order one to play with I suppose

        1. Well - like those other esp boards you've been playing with it does include a 240V power supply and a relay as well as the box... and some "controls" .

          I'm sure you could find a corner to stick an esp in as well.

          1. Oh, I missed that - in that case that's a GOOD deal!! Well, I'm stuck right now as in 2 weeks we're heading off to Spain for the summer. If I order now, it's a little dicey - so I'll wait and order one for Spain (so that the Spanish postal service can lose it) 🙂

  5. These just output the raw signal. You need to build an instrumentation amplifier and a bandpass filter to detect movement. These just output a few mV at the difference frequency that represents the speed of the object. You would be better off getting one of the pre made might control modules from ali at only a few bucks more and pulling them apart. (The are pretty badly made so they are easy to hack).

      1. Good find. I got the wrong one a while back. A lot more trouble. I'm ordering one of these to give it a try.

  6. Can the be made directional, (ie put them in a metal box with a small aperture), to scan full 360.
    this way you can build 'real radar'.

    1. That was the idea - but I can get a movement detection - but not any kind of delay..... I thought it might be nice to stick it on a servo and do the whole "voyage to the bottom of the sea" radar thing.

    1. Oh that helps a lot - thanks Venkatesh. It does make you wonder when someone brings out a board like this - which must cost a lot to setup - why they don't make proper instructions available in a couple of languages.

      1. Peter : These are mainly made for some product companies in China like motion detection electric bulbs or some sort of other equipment and hence the information mostly in chinese. I have imported other sensors which doesn't come with a manual and I had a hard time searching manual using google and after finding it I have another problem, all are chinese manuals. There is no single website that converts a chinese PDF to english docx/pdf.

        1. Google Translate does a fair job.
          I received source code for a USB/Digital IO board, in Visual C, but with all the comments and text labels in Chinese! Copy/Paste/Translate and it's now all in English...
          (and extensively modified!)

    2. Thank you for that!!
      With Google Translate under Android, I got some useful, some odd, info:

      1) Sensitivity: "Sensitivity adjustment resistance is inversely proportional to the size of the distance range: 10K- 100K (default 20K) Delay time adjustment Vatican." (Vatican?!?)

      2) Time delay: "The delay time of the delay is proportional to the size of the resistance adjustment range: 1K- 250K (delay of one second resistor resistance without delay of 30 seconds)"

      3) Connector: "Terminal from the bottom is - + - 0 + - 0 as input to output TTL high negative public" (Too many symbols in a row there; I'm lost!)

      4) Photo sensitivity?: "This welding hole photosensitive resistance welding position" (I have no idea)


      1. Since these were probably designed to be associated with light fixtures, there's an input to connect an LDR, and presumably it uses this input to decide on something else to do depending on the light level.

  7. A search for microwave radar on the Banggood web site, reveals a range of units in different enclosures. There are 12/24v versions as well as 220/240v.

  8. It still looks like early days with these units. There are a few home security companies selling them, but it doesn't seem like there's too much hands-on experience in the hobby market yet. It rather looks as though they might be too sensitive for a lot of applications. I saw one guy who had installed one in a bathroom light fitting (because the PIR sensor wouldn't sense presence behind glass panels -- the shower stall?) and discovered that it triggered whenever anyone moved in adjacent rooms. Reducing the sensitivity to prevent those unwanted triggers also meant that it wouldn't trigger for movement inside the bathroom either. He also tried tinfoil shielding, to no avail. I guess his next move is a Faraday cage for the whole bathroom. 🙂
    On a slightly different tack, the PCBs in the Bangood photos are just about the messiest I've ever seen. How is the quality on the units you got? Has Aiden had a chance to play with the ones he ordered, yet? Do all of the units function?

    1. Hi there Pucebaboon

      Quality? Well, for something that costs a couple of pounds, ok. Neither good nor bad. As for reliability... I'm finding them way too sensitive right now, not really had a chance to test properly.

  9. I use them for about a year for my lighting. They all work perfectly. The aperture is easily tunable with a piece of self adhesive alufoil.

    1. Solipso : So with alufoil we can focus the radar scanning in a particular direction ?

      1. Solipso : So with alufoil can we focus the radar scanning in a particular direction ?

      2. Exactly. The alufoil creates "shadow" - a direction, where the sensor is blind. It is usefull as it also senses through the wall and switches the light on whenever you move in a neighbouring room.

    1. So - Venkatesh... I'm really not following this.. firstly - effects of using microwave motion sensors.... have you looked at the amount of power these modules DON'T use? The microwaves coming out of them are utterly trivial. Secondly, that video - what on EARTH is he on about? "microwaves will go right through you and trigger a photoelectric sensor" ??? So someone had better tell the microwave oven people who seem to think the microwaves will only penetrate a very short distance into food like an inch or two - and that is a powerful heating unit, not a £2 milliamp-powered toy as we have here - and since when do photoelectric sensors pick up microwaves? I thought they were for reacting to light levels.

      It is early in the morning so maybe I'm missing something.

      1. Peter : I just ordered couple of motion sensors same as yours and also one from dfrobot which has more range. I know PIR sensors are harmless, but microwave I was kind of thinking if there is an issue as the microwave sensor will be ON for 24/7 and if there are any ill-effects on health. These are really not certified and hence I got this doubt.
        I have limited understanding of electronics, communications and physics and hence I posted the question so that master's like you can shed some light on the issue. Thank you.

        Working Voltage: 5V + 0.25V
        Working Current(CW): 60mA max., 37mA typical
        Size: 48.5x63mm
        Detection Distance: 2-16M continuously adjustable
        Emission Frequency: 10.525 GHz
        Precision Frequency Setting: 3MHz
        Output Power (Minimum): 13dBm EIRP
        Harmonic Emission: < -10dBm
        Average Current (5%DC): 2mA typ.
        Pulse Width (Min.): 5uSec
        Load Cycle (Min.): 1%
        Sensitivity (10dB S/N ratio) 3Hz to 80Hz bandwidth: -86dBm
        3Hz to 80Hz Bandwidth Clutter: 10uV
        Antenna Gain: 8dBi
        Vertical 3dB Beam Width: 36 degrees
        Level 3dB Beam Width: 72 degrees

  10. Is there any way to direct the beam? I'm looking at a movement detector but only along a narrow(ish) beam with a range of at least 50cm. I was thinking this maybe ideal but seems too sensitive.

  11. Well, I have a few of these from IC Station with a white pcb but it appears to be different than the blog describes ...there are 2 sets of dual through holes one for a resistor the other for a capacitor. I suspect it is for a RC type charge discharge time delay. These units are Doppler so it is velocity they measure and there appear to be two kinds 5GHz and others are at 10Ghz. 10ghz is what autos are using for collision avoidance. I hace some phased array auto collision sensors that do an electronic sweep. The units I have are 5 Ghz and that we are discussing here do trigger upon motion but the response is variable.
    I'm using a esp8266 12e to monitor the sensor, There is a lot of false triggering. We desperately need a schematic they use a BISS0001 IC as do some PIR sensors. Most are aware they have to use Doppler as light speed is about 1 ft per nanosecond ..echo times are too short for radio ranging

    1. I have bought 5 different Chinese microwave motion detectors altogether. For the one you mentioned, white PCB with resistor and capacitor slots, I had spent hours tinkering with it, because I had the same random triggering issue as you. I was testing it by powering directly from my Arduino, until I managed to get it to function properly by using an external power supply. After powering it directly from my bench power supply at 5v, the signal became extremely stable. Despite the listing stated the detection angle is 180 degrees, my tests indicated it's not. It appears to be 360, just like the sensor stated on the page.

      Both sensors are also highly sensitive, a lot more than PIRs. It could detect even the most minute movement of my arm. Also to add from my tests, it has no issue detecting movement about 50-100cm behind a 10cm concrete wall.

      I also have the one which is the size of a penny, it has 5 pins. I can never get that one to work properly. The signal is always high. The pinout for this one could be found on taobao and those Chinese eBay-like sites.

      1. The white pcb Chinese radar switch is 360 degrees and I don't think you can place a metal reflector on one side without continually triggering it. It may be these are designed to be placed in the center of a space possibly on the ceiling to turn off lights when no movement is detected. Running at 3.3v is much less reliable than 5v. The BISS0001 has triggering delays... when I can get to it I'll trace back to the BISS pins and get the purpose of the Cap and resistor through holes. The description is 180 deg but either side will trigger it.
        Thanks for the info Lio and the chance to confirm it.

  12. As to microwaves and risks. Most are aware that the power of these devices is very very low. Most of all the frequency is in the several gigahertz range. Quantum mechanics decrees that the energy of a photon is proportional to frequency. Gigahertz photons lack the energy to break DNA bonds or ionize atoms within cells but they can jiggle free electrons found in metal conductors and other electrons which when bound to atoms will jiggle the atoms. Atoms jiggling about is commonly called heat. Microwaves induce heat but at the power these small devices emit it is very very inconsequential. Just as well since we are bombarded 24/7 by the cosmic microwave radiation from the primordial universe.

    1. Anything known about the risk of these with pacemakers? (I don't mean walking past them some yards away, I mean the guy tinkering with them on the workbench about 6 inches away...)

  13. So I have broken out the LDR port to header wires, the R6 time delay pads to header wires, and the R9_1 distance pads to wire headers also removing the 20k resistor there. That makes a 9 pin header now for this device instead of 3! why is this good? LOGIC! lots more logic to play with. I have run the LDR port to a 222 transistor with a 1 mega ohm resistor in series with the emitter. this acts as an on off switch for the radar. pull arduino pin high, transistor pulls and turns off radar. nice! then i hooked up a MPC41100 digital potentiometer (100K) to the header wires for R9_1. this acts as an arduino controllable variable ranging system. Now I am on to programming radar ranging ping logic that will grow and shrink the radar field and sense when it is 'pinged'. this will act as a super fast, collision avoidance early warning system for various robots i am building. not quite directional but ranging and thats all i need for now. pics, schematics, and arduino code soon to come...

  14. There are misunderstandings here about how these work. Measuring the time for a reflection (like a radar) would be impractically difficult and expensive on such a cheap unit - especially at short ranges where the round-trip time is only a few nanoseconds. The way these work is quite simple and clever - they send out a constant microwave carrier this is reflected back by all sorts of things (walls, floors, etc). Whatever the reflection, it is normally at the same wavelength as the original signal (although each reflection will have a different phase shift depending on how far away it is). The unit is looking for a reflected signal that is at a slightly different wavelength to the transmitted signal, which would be caused by a reflector moving towards or away from it (causing the reflected wave to be slightly compressed or expanded, i.e. a doppler shift). The unit detects the hetrodyning (=frequency difference, found by simply mixing them) between the outgoing and returned signal, and this is what's used to detect motion. In short - something continuously moving towards/away from the unit is detected. Anything not moving in that plane is not detected. If you put a scope on some of the circuit points you will be able to find this hetrodyne difference signal - it will be a voltage that moves up and down in a sinewave pattern as you move towards/away from the receiver. The amplitude depends (to some degree) on how close you are. The frequency depends on how fast you're moving towards or away. There is no real way to tell absolute distance; for that look at ultrasonic sensors (that do use time-of-reflection; hugely easier because sound is so slow)

  15. Lots of comments, sorry, not going to read through them all.. just skimmed.
    Looks like these gizmos use a BISS0001 PIR sensor chip and a single transistor microwave oscillator and receiver using the same trace as an antenna. must be some sort of beat requency thing going on: when something moves fast enough, the reflected signal is out of sync enough to give a beat frequency out that the PIR chip can use as pulses as if it had a PIR sensor attached; giving motion sensing without the PIR.
    The timing resistor that y'all have been fiddling with is attached to the output timing timer on the PIR chip.
    the data sheet is available at seedstudio ( just the PIR chip, not the board) http://www.seeedstudio.com/wiki/images/2/2f/Twig_-_BISS0001.pdf
    I popped the daughter board off ( a bit of time with an iron and a bit of leverage at each pin equally.. round and round I went) and the circuit is pretty much a PIR sensor board that someone did some very clever microwave work on the daughter board.. 'specially clever, as it probably needs no precision components or tuning, as it is receiving it's own signal.
    Hope this helps someone..

    1. Update:
      pin 4 of the daughter board has the beat frequency / doppler signal on it.. It is only a few mV peak to peak, but that must be enough for the PIR chip. Amplitude is not affected by rate of movement, and target size and reflectivity does not seem to make much difference. I even tried a corner cube radar reflector, and that did not make an appreciable change in signal: if it sees you, it sees you.

      With a bit of work, you could probably turn this into a speed gun ( amplifier, zero crossing detector or schmidt trigger, and a counter.. scale it, and you would have feet per minute or miles per hour..)

      1. You can feed the doppler signal straight into a sound card mic input btw, it's almost ideal, at least for simple testing. The doppler signal isn't terribly useful for distance (although you do get some amplitude changes) but yes, for speed gun it's great, that's how most speed detectors work.

  16. Has anybody confirmed that the modules Peter linked to do in fact work at 3.3V? I have ordered something similar (XYC-WB-D1 modules) and they only work at 5V, which is an inconvenience...

    1. I've got XYC-WB-D1 modules as well, and they seem to work at 3.3V. I'm not sure which resistors to modify to change the behaviour though.

  17. Yeah, there is amost no data to be found on the net about the XYZ-WB-D1, unfortunately. I seems to be a decendant of the original design, using just one single board (which probably reduces manufacturing complexity). Unfortunately it lacks the convenient multiple pads of the WB-DC, featuring only one single pair of pads labeled "PT" (photo transistor??).
    Got it yesterday, along with some other tiny module which is a round sandwich design with a diameter of only about 1 inch, named "XYZ-WB-12" which features undocumented pad pairs named "PT", "Time" and "R15" (guess: sensivity).

    1. Is it possible to detect "Relative Motion"? I am trying to detect the positions of the objects in the room. If I make it directional by using alufoil, and put it on a servo or a stepper motor and scan the room; will I be able to get positive response at some angles? No response at all, or response at all angles?

      1. So, I bought a few of these units, and tested myself. Yes you can direct the beam by masking most of the antenna with a piece of alufoil leaving a small gap. Yes, I get response at SOME angles, not all, not none, but they do not seem to be consistent. I placed the unit on a stepper motor and scanned 275 degrees by 7.5 degrees increments. I did both digitalRead and analogRead on Arduino. I plotted analog echoes on XY. Broadly speaking, some echoes are overlapping but mostly they are inconsistent. Hard to say that we get echoes for the same object consistently.

  18. I've been experimenting to used it as presence detector in combination with an ESP-01 in the same housing. Although I tried to shield it from each other, the WiFi signal was influencing the radar signal and false detection's where made.



Comments are closed.