Microwave Black Magic Guru Needed

[tecman]

I am sure some of you have seen, or read about, the RCWL-0516 proximity sensor.  A surprisingly effective device that you can snag from ebay for under 60 cents.  I am very knowledgeable in many EE areas, but not so much on microwave details.  Having searched the net for more details, you mostly come up empty.  Several have described the single transistor in the circuit as an oscillator, mixer, demodulator and amplifier.  Must be quite a transistor.

So, for the gurus, explain how this works.  Is it a "radar" as has been described with the return signal and corresponding doppler shift being detected ?  Or is it loading by a body causing increase transistor current or something similar ?  It's sensitivity seems omni-directional, so how does that occur with a planar, and seemingly directional pattern ?  And what is the purpose of the ring trace on the PCB ?  Is it a director of some sort ?

Curious minds want to know.

paul

[ogden]

Is it a "radar" as has been described with the return signal and corresponding doppler shift being detected ?  Or is it loading by a body causing increase transistor current or something similar ?  It's sensitivity seems omni-directional, so how does that occur with a planar, and seemingly directional pattern ?  And what is the purpose of the ring trace on the PCB ?  Is it a director of some sort ?

It is FMCW radar which mixes transmitted signal with reflected and on output you get doppler shift frequency. Ring trace likely is not rat-race coupler - microwave signal combiner/splitter which is present in other CW radar modules.

p.s. I'm not microwave guru 

[Mechatrommer]

Ring trace is rat-race coupler - microwave signal combiner/splitter.
p.s. I'm not microwave guru 

doesnt look like so. maybe somesort of a circular patch antenna
p.s. I'm not microwave guru 

[Bud]

i think it is a resonator.

[SiliconWizard]

Seems like it's based on the Doppler effect and is more or less described in the following patent: https://patents.google.com/patent/EP3091605A1/en

[SaabFAN]

I'm wondering myself how they work.
The Chip is apparently designed for Ultrasonic or Infrared motion detectors and the transistor is doing something that creates a simliar input-signal to the chip.

From what I learned so far, it is an oscillator that puts out a signal which is radiated via the Patch-Antenna and also mixes returning waves. If the waves have some doppler-shift on them, they appear at the input of the chip.
There's also a different theory out there by RogerClark (one of the main guys in the STM32Duino-Project). I don't remember exactly what his theory was, but it basically stated that this thing is no radar.
EDIT: Found it! http://www.rogerclark.net/investigating-a-rcwl-9196-rcwl-0516-radar-motion-detector-modules/ - His hypothesis is, that it uses "wave propagation interference" instead of mixing a doppler-shifted signal.

I think, the best way would be to petition The Signal Path on Youtube to make a video. He certainly has the Knowledge to tell us about the microstrip-structures and how the device works, and he has the Measuring-Equipment to show analyzer each part of the circuit

[tecman]

The op-amp of the BISS0001 chip is just looking at the emitter current, with some RC low pass filtering, AC coupled to another high gain stage op-amp.  This could see doppler frequency due to movement, or it could also be changes to the emitter current due to loading caused by absorption of the RF. 

Still not sure

Paul

[TheUnnamedNewbie]

Just a stab in the dark, I am not familiar with RF at these frequencies.
The two rings remind me of a few omnidirectional patch antenna designs I've seen. Not a rat-race coupler, on account of not having the necessary feed lines or ground planes (For a microstrip rat-race).

My wild hypothesis is that this is somewhat similar to backscatter communications. Reflections are complex. As the room changes, the phase sum of the reflections changes too (all the reflections together are in phase if there is not too much movement. As a result, the "antenna" appears different in the eyes of the oscillator. Thus, as things move around, the antenna seems to be "changing" and this results in the voltage at that node changing too (because the antenna impedance changes, the current flowing changes too). This can then be measured.

I don't think it is doppler-effect, based on the measurements by Roger Clark showing speed didin't change much in terms of voltage swing, and the patents claiming doppler effect have an additional pair of diodes to form a diode-mixer - not present on this board. I don't see how the transistor itself would double as a mixer *and* have the result of said mixing appear on the input of the IC.

I'm still not completely confident at how the feedback works because there seem to be some feedback. I will investigate further.

[Bud]

I think, the best way would be to petition The Signal Path on Youtube to make a video. He certainly has the Knowledge to tell us about the microstrip-structures and how the device works, and he has the Measuring-Equipment to show analyzer each part of the circuit

There are reviews w analysis of a similar type of device on youtube. i am pretty sure Big Clive did one. Try seaching youtube Big Clive channel.

[Howardlong]

I don’t have one of these, but my strong suspicion is that it’s simply reflected wave interference mixing with the carrier, just like how the old Gunnplexer Gunn diode cavitiy radar trap setups worked of old, but at a lower frequency. The difference is that radar traps had horns to provide some directionality, whereas this looks like there’s not much directionality, so there’ll be plenty of multipath interference and the associated low frequency component when mixed with the carrier when there is any movement of a reflective or absorbent medium.

The distances being discussed are well into the far field for this device bearing in mind its size and frequency. If it were detecting impedance perterbations you’d need to be in the near field, which for the size and estimated frequency of this device is going to be under six inches.

[Zero999]

Does it detect static objects or do they have to be moving?

The Doppler effect is only present when the object is moving, so it can't be using that if it's detecting a stationary object. If the object has to be moving for it to detect it, then it might be using the Doppler effect, but it's not a certainty.

I suspect it's just monitoring the loading on the antenna. Reflected radiation would change the amplitude of the signal at the antenna and it should be fairly easy to detect that.

[tecman]

It does not detect static objects.  That would point to doppler.  However if you are looking at oscillator loading, it too would also not detect stationary objects, since the detection is AC coupled (low frequency) and only a change in time is detected.

So being AC coupled, either a LF doppler signal or a varying change in oscillator loading could produce the same result !

paul

[Zero999]

It does not detect static objects.  That would point to doppler.  However if you are looking at oscillator loading, it too would also not detect stationary objects, since the detection is AC coupled (low frequency) and only a change in time is detected.

So being AC coupled, either a LF doppler signal or a varying change in oscillator loading could produce the same result !

paul

Loading could theoretically be used to detect stationary objects. If the RF level is known, when it's unloaded, then a threshold detector could be used to detect whether or not anything is reflecting the radiation back. If there's a reflective object, in front of the antenna, then peak antenna voltage would increase, due to the reflected radiation, even if there's some destructive interference.

Yes, it's true that either loading or Doppler could be used to detect moving objects and I suspect it's the former, rather than the latter, due to the lack of circuitry required for Doppler.

[ogden]

It does not detect static objects.  That would point to doppler.

Most likely it is doppler. Relevant patent: https://patents.google.com/patent/US4672379

Truly magic MW design where transistor is both oscillator and mixer, stripline is part of both antenna and oscillator. And yes, I have no idea what actually ring trace on opposite side to curved stripline is - could be element of antenna to change radiation pattern or just antenna itself. As correctly noted, it is not rat-race coupler because have just single tap. We shall seek for proper microwave guru for comments.

I am not sure how legal such radars are because of frequency (in)tolerance.

[Howardlong]

It does not detect static objects.  That would point to doppler.

You have to have relative movement for Doppler detection.

I think we’re thinking too hard, it mixes its own signal with all the multipath reflections. If something moves, there will be a change in the mixer output level due to the phase change compared to the carrier, that is what’s being detected.

The only difference between that and Doppler is that typically Doppler is directional between two objects, and is designed to avoid or mitigate against multipath.

[0xfede]

It is a CW (unmodulated) doppler antenna. The circuit is a super regenerative one that uses microstrip lines for both oscillating and mixing the doppler result using Q1 as the only semiconductor. The IC does not take any role in RF.

Best,
0xfede

[Howardlong]

I’m not sure I’d call it super regenerative, but I fully agree that the transistor both generates the carrier and is used as the receive mixer, and that the IC is only seeing baseband, or low IF at best.

[0xfede]

Yes it is a super regenerative as it use the feedback thru the anular ring microstrip antenna (that is basically a parallel RLC circuit) to give a positive and phase shifted feedback from the collector to the base at the center frequency. In the PDF the 'antenna' is an inductor with a central tap connected by a microstrip cap to the collector. Furthermore there is an LC element on the base towards ground and another cap towards the emitter. That's all. The math for each element is fairly simple since the frequency is not really high and knowing the name of each element is easy to find. I don't have this circuit so I cannot take the mechanical measures needed to make the simulation.

Best,
0xfede

[rfeecs]

I think language may be getting kind of confused here.

Every oscillator has positive feedback.  Super regenerative is a more specific term.

My understanding of super regenerative would be an oscillator that repetitively builds up and then quenches.  The bias point (say the emitter voltage) shifts during this process, following the amplitude of the oscillation.  The presence of a received signal would alter the timing of the build up / quench cycle, effectively sampling it.  Low pass filtering the bias voltage will detect the amplitude of the received signal.  So the super regenerative receiver has an LO oscillator, mixer and detector in one.

It's not obvious if this is a super regenerative receiver or just a CW oscillator.  Although the spectrum in the pdf file from the first post seems to indicate that the roughly 3150 MHz carrier is being modulated at about 20MHz, so that could be quenching at 20MHz.

To nitpick, microstrip has a ground plane.  This circuit doesn't, so it's not microstrip.  If you stick it in a box you might call it suspended microstrip or suspended stripline.  Close enough.

A Doppler radar would measure the speed of the target by measuring the frequency shift of the reflected signal.  This doesn't appear to measure the frequency shift, but just measures the amplitude of the combined incident and reflected signals and compares that to some thresholds.

[0xfede]

I think language may be getting kind of confused here.

I know that my English is awful and I'm trying to get better. Sorry for that.

It's not obvious if this is a super regenerative receiver or just a CW oscillator.  Although the spectrum in the pdf file from the first post seems to indicate that the roughly 3150 MHz carrier is being modulated at about 20MHz, so that could be quenching at 20MHz.

As you also noted the signal quenches (40db on the SA at the center frequency). That's why I'm pretty sure it's super regenerative. And was exactly the term I was looking for.

A Doppler radar would measure the speed of the target by measuring the frequency shift of the reflected signal.  This doesn't appear to measure the frequency shift, but just measures the amplitude of the combined incident and reflected signals and compares that to some thresholds.

I just doubt that this approach could work especially in a circuit without any trimming and produced at the lowest possible cost due to its criticity. I'm sticking with doppler.

Best,
0xfede

[Mechatrommer]

To nitpick, microstrip has a ground plane.  This circuit doesn't, so it's not microstrip.

to my understanding, ground plane is only to comply with EMC emmision (or probably to achieve some characterized impedance required), without it every tracks will become an antenna. the fact the OP circuit is not fully grounded (plane) suggesting its an antenna and the fact that the tracks and rings/circular got coupled with other tracks/nodes suggesting they are lumped elements or "intentional" stray elements. their dimensions and geometry should be intentional too to get required elements values and effects, hence an oscillator/resonator as others suggesting.

A Doppler radar would measure the speed of the target by measuring the frequency shift of the reflected signal.  This doesn't appear to measure the frequency shift, but just measures the amplitude of the combined incident and reflected signals and compares that to some thresholds.

or probably when there is disturbance/shift/movement in the reflected signal received by the antenna/resonator will change the fixed biased transistor's emitter current and hence voltage and thats what the IC is detecting, just maybe.

p.s. I'm not microwave guru 

[0xfede]

To nitpick, microstrip has a ground plane.  This circuit doesn't, so it's not microstrip.

to my understanding, ground plane is only to comply with EMC emmision (or probably to achieve some characterized impedance required), ......

In a microstrip circuit the ground is there to provide a reference to the signal and is exactly the same for a DC net: without a reference and a return path there is nothing happening. What he said here is true and he is just nitpicking (his words).

I design MW for this purpose (alarm / motion detection) usually at 10.54GHZ and 24.125GHZ for job. I tried many many years ago the  "measures the amplitude of the combined incident and reflected signals" approach for an ultra low cost @2.4GHZ and the result was that is impossible to have a reliable detector in this way. This is why the majority of the commercial products in this area are tipically DRO based, CW doppler effect with a simple diode mixer.

Best,
0xfede

[Zero999]

I think language may be getting kind of confused here.

I know that my English is awful and I'm trying to get better. Sorry for that.

It's not obvious if this is a super regenerative receiver or just a CW oscillator.  Although the spectrum in the pdf file from the first post seems to indicate that the roughly 3150 MHz carrier is being modulated at about 20MHz, so that could be quenching at 20MHz.

As you also noted the signal quenches (40db on the SA at the center frequency). That's why I'm pretty sure it's super regenerative. And was exactly the term I was looking for.

A Doppler radar would measure the speed of the target by measuring the frequency shift of the reflected signal.  This doesn't appear to measure the frequency shift, but just measures the amplitude of the combined incident and reflected signals and compares that to some thresholds.

I just doubt that this approach could work especially in a circuit without any trimming and produced at the lowest possible cost due to its criticity. I'm sticking with doppler.

Best,
0xfede

How stable is the oscillator?

How much phase noise does it have?

How fast does the object need to be moving, in order for it to detect it?

How far away does it detect the object?

A stable frequency is required to detect the speed of slowly moving objects.

An object travelling at 10m/s would create a Doppler shift of just 105Hz at 3150MHz, which is only 33.36 parts per billion. I doubt such a crude oscillator is stable enough for Doppler radar, unless it's used to detect supersonic objects.

I think language may be getting kind of confused here.

Every oscillator has positive feedback.  Super regenerative is a more specific term.

My understanding of super regenerative would be an oscillator that repetitively builds up and then quenches.  The bias point (say the emitter voltage) shifts during this process, following the amplitude of the oscillation.  The presence of a received signal would alter the timing of the build up / quench cycle, effectively sampling it.  Low pass filtering the bias voltage will detect the amplitude of the received signal.  So the super regenerative receiver has an LO oscillator, mixer and detector in one.

It's not obvious if this is a super regenerative receiver or just a CW oscillator.  Although the spectrum in the pdf file from the first post seems to indicate that the roughly 3150 MHz carrier is being modulated at about 20MHz, so that could be quenching at 20MHz.

To nitpick, microstrip has a ground plane.  This circuit doesn't, so it's not microstrip.  If you stick it in a box you might call it suspended microstrip or suspended stripline.  Close enough.

A Doppler radar would measure the speed of the target by measuring the frequency shift of the reflected signal.  This doesn't appear to measure the frequency shift, but just measures the amplitude of the combined incident and reflected signals and compares that to some thresholds.

Yes, that makes much more sense to me.

[Howardlong]

How stable is the oscillator?

How much phase noise does it have?

How fast does the object need to be moving, in order for it to detect it?

How far away does it detect the object?

A stable frequency is required to detect the speed of slowly moving objects.

The Gunn diode cavities in Doppler radar are also notoriously unstable, but as the mixer uses exactly the same unstable but coherent reference, it is of negligible consequence.

While I agree this isn't technically Doppler radar, many of the underlying phenomena that make Doppler radar work are the same with this device. You can indeed do very much the same thing with a Gunnplexer. This device is measuring phase differences, whereas Doppler traditionally measures frequency differences, but keep in mind that frequency is simply the derivative of phase wrt time.

In addition to detecting phase, this device may also be detecting amplitude, a not uncommon side effect both in detectors, and as a result of the constructive and destructive interference.

I am wondering what the purpose is, if any, of it FMing at ~20MHz, or is that just a feature of the design, possibly noise from the power supply?

Anyway, I ordered some to keep me occupied.

[tecman]

I am wondering what the purpose is, if any, of it FMing at ~20MHz, or is that just a feature of the design, possibly noise from the power supply?

Anyway, I ordered some to keep me occupied.

I am running mine off a 9 v battery, so PSU noise is not a factor.  On the spectrum analyzer I saw a much narrower spread of frequency compared to the plot in the attached documentation.

paul