Hi Clive, We called that area by B a tuning stub when I worked at Motorola. You design the capacitor values as needed, but because of capacitor and high frequency part tolerances you need to match the capacitance to the high frequency part. We used a laser to remove a small amount of metal to 'tune' the circuit to the exact frequency. The squiggly track is also to match the impedance of the circuit to the antenna (VSWR) to deliver maximum power to the antenna (or from, in this case), and it is a simple formula of frequency to a single sign wave (2 kHz = a wavelength of 35 cm). It reduces the tract length (from straight, as in an collapsible FM radio antenna) by zig-zagging it on the substrate.
Yep. That stub in combination with the ground plane next to it *and* the ground plane under it form a tuned circuit making the basis of a single-transistor RF oscillator. The squiggly bit is the antenna + z matcher in one piece. It’s not really that simple of course- the resonator and the antenna/matcher interact too kind of like a tuned-plate tuned- grid tube oscillator works but the coupling isn’t as strong, until someone walks by and it is. Anyway, Clive is right, it’s a Doppler detector. Reflected RF adds to or subtracts from the current in the antenna and the difference gets passively filtered and sent to the active low pass filter about which he is also correct.
Wow, even trimming it off with LASERS! That's finicky! I'm impressed. I knew RF analog was a bit of a trick but I hadn't realized it was that much of a trick.
It's true; alot of rf guys I know 'cheat' by watching a scope while adding components rather than just designing it so it should work in the first place.
I really love the longer/complex ones. Its fascinating to see how those boards work, especially with how well you explain all part but keep it simple at the same time ❤.
Don’t doubt yourself, Clive! You had it right the first time. It clearly says 0.2 milli seconds per div on the scope image, which makes the frequency 2kHz ish. If it was 0.2 nano seconds, the frequency would have been 2 gigahertz. By the way, doppler frequency is about 20Hz per m/s, so a waving hand might give 50 to 150Hz, and a swaying Clive closer to 30Hz. Not sure if this makes a case for or against it using doppler shift, but considering it only uses an audio-frequency low pass filter, I’m leaning towards yes - I can’t imagine how an LM358 could respond to any kind of reflected power ratio measurement, cos it would have to be operating at RF frequency to do that, or at least have an RF detector (diode and cap integrator?) as the front end, rather than the low pass filter. Then again, if we guess that it’s operating around 5GHz or so, the reflected wave would interfere with the transmitted one in a varying way as objects moved in its vicinity. Nah… that would have no effect on the oscillator - it would be way too tiny. I’m sticking with doppler, via mixing of the transmitted and received waves generating a doppler difference frequency.
I've played around with standalone versions of these "doppler radar motion detectors" and I believe they are doppler based. That one transistor oscillator is working pretty hard, and also serves as a mixer for the signals that are bounced back from nearby targets. Their shifted frequency based on the motion of the object causes the low-frequency mixing product to be generated and detected in that "audio" region.
It definitely is Doppler. What you’re missing is that the outgoing and incoming signals are mixed in the antenna (the squiggly trace) and the low frequency sum and difference are what the passive LC filter sees, and its output goes to the active low pass filter. A swaying Clive should be more in the .3-3Hz neighborhood…
err, 3 1ohm resistors in parallel seems to be 0.33 ohm? @2:10 . Also, around @11:22 you have "0.5 nanoseconds is about 2 kilohertz". I think you misread the "ms" as "ns"?
The question-mark cap may be coupled to the antenna to give a feedback to the base of the transistor to get the circuit to oscillate. Reflected RF will then be mixed by the same transistor. The doppler frequency is then filtered by the first op-amp. This is a typical bandpass, It will be tuned to pick out the typical frequency's made by objects moving at human typical speed. To filter out flying bugs.
What an adventure! Going through this circuit has been totally captivating. Thank you for being the guide 😊. You clearly had quite a challenge putting this description together 👍
Very interesting, I enjoyed the video. Note that these will go deaf around wifi routers as the router emissions do drown out the doppler signals. A few meters of separation is normally good enought o fix that. Appreciate the effort that went into this one.
Yeah, good old fashioned front end capture. It’s a consequence of the 5GHz band having to be shared by devices that do different things, and (in the US) one of them having to accept interference by the other.
I'm sorry Clive, I fell asleep (repeatedly), every time you got to 'nano seconds'... I just can't help it. You are clearly brilliant though, I wish we could meet, so that I could learn more... Keep up the good work, regards, Mike
This module was inspired by the RCWL-0516 module. That one operates at 3.2 GHz. Didn't think they could make it cheaper? They did. They eliminated the need for that chip, with clever filtering and amp section.
@@hubertnnn Good point. LM358 is cheaper than dirt. So are resistors and caps. I don't know the costs, but given the way Chinese operate, I wouldn't be surprised if they came across tons of these Op Amps on the cheap, and designed around them. They do that for all through hole components which have no market left since SMD.
Quite a complex functioning circuit! It’s like a solid state version of the proximity fuzes used in WW2, plus a microcontroller to select different modes. I was surprised to see several of these online claiming 5 or 10 meter sensing range. Pretty significant for such a small, low cost and low powered unit. Well done with the reverse engineering!
Interesting little circuit! My experience with "radar" detectors is a bit unusual - I installed one on my balcony light, outside the room I have the PC, so that I could note when people passed by at night. I found out that it was able to sense across the wall and trigger on any movement I made, inside the house! This was easy to fix with a grounded metal plate, but I didn't bother. It's still there, but I disconnected its power - that's how lazy I became :-)
I have read all the comments after watching the video twice & now realise i am an Expert! I verified this status by a simple process of rating my grasp of the technology at each point in the design. Zero across the board and an overall average of nil on my scorecard is worthy of note. Nonetheless i enjoyed it all thanks Clive & all.
You should start a podcast, I love having your videos on in the background just for the background noise. And of course they’re informative and interesting too
There was a live stream channel on RUclips, but I recently moved to Twitch due to problems with RUclips. The twitch channel is also called bigclivedotcom.
The circuit does use Doppler shift. The oscillator transmits a signal. It is reflected, if the object it reflects off is moving, a Doppler shift is created. The reflected signal is received on the same antenna as it was transmitted from and mixes with the oscillator signal to produce a beat frequency equal to the frequency difference between the two signals (i.e. the Doppler frequency. Its this signal that is then being filtered and detected I think. This is exactly the way the first proximity fuses for shells worked. They had a single 'peanut' valve packed in wax to allow it to withstand the shock of launch which did the work of the transistor in your circuit. Basically one active component functions as transmitter, receiver and mixer. I know about this because the same principle is still in use today - although we are at least using transistors now!
Thanks for the detail, Clive. I have one of those solar street lights with LiFePO₄ cells in it that's been running for years. It's surprising to see that it will occasionally respond to insects or bats chasing them, strong winds, especially when accompanied by rain. Fortunately, the thing has a "full on" time of about 30 seconds, so it's not really an issue.
I believe you have it right on Doppler. The microwave oscillator is so fast that even a slow reflected movement will change the frequency down to an audio signal not just a different RF signal. The opamp can pick this up and filter it to a DC level and trigger the light.
Re that op amp feedback circuit, the main element is that cap going directly from out to -in. That drops the gain to unity for higher frequencies. The whole thing looks like an active low-pass filter. The input cap blocks DC and the voltage divider recenters the waveform at 2.5V.
Multi-path interference. Basically heterodyning itself but with slightly different phases, the result of which is a low frequency signal. The difference in signal path is doing the work, not the speed of movement (doppler).
It likely is just constructive/destructive interference with the reflected wave. For mixing (heterodyne) you need multiplication. Such nonlinearity is not given in this simple circuit.
Really fascinating subject matter, I understand the principles of the build, but would not attempt to build it myself. I've e purchased plans to build a Bluetooth speaker with 10 watt speakers. I bought all the components and have started to build. You are a bad influence on some people Clive as this is the fourth project that I've begun since I started watching you years ago. Believe me when I say it is not just electronic projects but all kinds that stop the decay of the grey matter. Good post, thank you👍
the rf section reminds me a lot of a regenerative receiver, where you monitor the current draw to pick off the received signal. I guess this is not much different than this- it is just receiving its own signal after it bounces off something.
2:06 Are the resistors in parallel or series? Your photo looks like parallel but could easily be deceiving me... Irrespective, many thanks for your continued work... It's all greatly appreciated. [edit] Ah watched a bit further... 4:13 they are in parallel, thanks!
8:07 looks like that 150 Ohm resistor serves double duty as both a shunt and as a means for setting the DC bias for the transistor. The capacitors across it serve as bypass to ground for the RF. Anyway, the designers seem to be interested only on the average current across the "shunt", thus the 1st order low-pass filter formed by the adjacent 1K5 and 47K resistors and capacitor (the 47K resistor would have negligible attenuating effect, and could be omitted with no major impact on performance). In your diagram, there are a few essential microstrip elements missing. I've simulated it as is, and the oscillator performs poorly.
The doppler we had sensed movement thru the wall, floors and ceiling. I was very confused for some time before I understood why the light was on so much.
The wiggly track is actually the antenna. If I'm reading it correctly (I'm by no means an expert in this...) the antenna transmits a frequency, and when it receives the reflected signal, the interference on the antenna gets amplified by the op-amp circuit. The scope screen does look like it shows 0.2ms, but then, I didn't do the measuring, so you likely know best ;) But if the 2kHz is correct, it's 0.5ms, not ns. The filtering caps in the op-amp section, especially the one between the 1M resistors, might be used to vary amplification depending on the frequency.
Thanks Clive you rock!! We appreciate the work you put into your videos and reverse engineering, making it easier (but still very complicated) to understand. I find myself nodding like ya that makes total sense...but I am still totally clueless! Ha ha ha...Love your videos and I will always return for more!
The "splooge" capacitor is not a normal place for one. With some opamps, it could cause oscillations. The LM358 happens to not be normal under the hood and not very fast either, so all bets are off for why it's there.
In my garage I have a RaspberryPi server with a camera and Python code to send me, when image pixels change, photos to my email. I was overloaded with photos in my email. I put one of those RF detectors on the GPI and modified the code to only send email if the RF detected motion. It works just fine. I keep the extra images/movies on a thumb drive that are less than two weeks old. Each night a crontab schedule deletes old images. It's the best of my hobby projects yet. RF detectors are much better than PIR.
There's an amazing animation of a solid state beam steering antenna. Can be made similar to the traces there, but is a weird asterisk shape: * . Slow mode field animation shows if you send signal/power to one side not the other it "forms" the beam to point in the other direction, and likewise can fill each finger with more or less power/pulsed waves to shape it's direction. I wonder if those through holes help shape the direction?
Thanks :) 8:32 is an economy version of a twin T filter in the negative feedback loop of the first opamp. So the first opamp is a narrow pass filter. I shudder to thing what the bandwidth is More gain less bandwidth and all that
I don't think that 2.5kHz signal is meant to be there at all, it's a spurious oscillation of the band pass filter (the left hand op amp). The output of this op amp is supposed to be a dc level at half the supply voltage when there's no movement. The threshold of the comparator (the right hand op amp) is about 0.12V below that dc level. So any low frequency Doppler frequency that gets through the band pass filter larger than 0.12V will be detected and turn the light on. The presence of the large 2.5kHz oscillation will produce continuous output and prevent it from working. By attenuating the spurious signal you have prevented this from happening and allowed movement to be detected. The bodge capacitor was probably and attempt to stop the spurious oscillation.
It's doppler, Andreas tested some, but what's the nighttime standby drain like? I have some modules, but they greedy, I forget the drain, but not suitable for battery powered home automation, which is a shame. There are some quite advanced programmable ones now, I believe they can detect your chest raising and lowering as you breathe at night while asleep, which makes a perfect presence detector, finally.
Ain't no LM358 on this planet that can output a signal in the GHz region! It's 0.5 mS, or about 2 kHz, which the LM358 can certainly handle. Having said that, one well-known problem with an actual LM358 chip (not an updated LM358) was its awful totem pole output. If you bias the op amp in its mid range, as is the case here, then the output will have odd crossover distortion artifacts as the output tries to alternate between sink (PNP) and source (NPN) of currents to some light load. The fix is to add something like a 1 kohm resistor to ground on each output, which effectively biases the output so that the NPN output transistor in conjunction with the pulldown resistor are always supplying the output current. The PNP output transistor is then effectively cut off. Any newer updated version of the LM358 (which will never include "LM" in its name) fixes this known flaw in the opamp. Whenever I see any kind of design that uses an actual LM358 in it, it immediately tells me that the designer was not very experienced in linear design. ruclips.net/video/VgodYtiD_F0/видео.htmlsi=wK2MK04FJWuOPijM
I have a complaint, you didn't put a link to tell us where to buy one of these this time. I believe that the "doppler" works in a similar way to a theramin, moving close to it will change the frequency of the oscillator and the change is detected by the first op-amp. ( I could be wrong, but that is how it looks to me )
These radar detectors rather than PIR are a real pain . I fitted a light with one built in , it sees you through the door in the hall , and even through the floor to the landing upstairs . The light is nearly always on because the cloakroom is dark and waiting for a trigger , which it almost constantly gets .
I'm having trouble undering how the RF section at 7:14 will oscillate. I don't see a feedback path back to the base of the transistor. Could it be from the physical placement of the tracks which is causing capacitive or inductive coupling to the large pad at the base?
At RF EVERY part of the circuit is inductive and capacitive. And they all resonate. Even changing a track width, or placing it 0.1mm closer to another track changes the frequency.
I work at a place that sells fitness equiptment and we have a TM that is notorious for blowing the lower control board because of friction. I wonder if big clive would be interested in finding out why the boards are so junky straight from the factory. I wish i knew how to get ahold of him but im not great with computers. Its from a BH S5TIB. They used to be good TMs. We are stuck with a whole container of those and they are nightmarish for us
If it is doppler, there'd be some sort of "mixer" on the pcb track. So imagine one part of the PCB oscillates generating an RF output (transmit beam if you like), and then the receive part of the same PCB has to interfere with the reflection signal. Or "mix". So the result would be a low frequency. Perhaps that 2kHz signal means the receive part of the PCB was "out". Instead of generating frequency bursts, it is probably meant to generate very low frequency pulses. Like on or off blips. Think of police radar detectors which use horns and precision cavity resonators. School "Young's slits wave experiments". What's the sensitivity range like?
The PIR is ineffective with ambient temperatures around 48c, common in tropical regions. Doppler however is succeptible to false trigerring around water bodies and plants and trees.
omg...doppler radar for solar lights...... so can like detect an incomming Russian Su 25 .....Su34 or Mig31... and light up my garden to warn me..... you soooo remind me of old electronics teacher... apart he had a white board... Top videooooo N x
As long as it provides a Homer Simpson "Light goes on, light goes off, light goes on, light goes off!" level of entertainment, then that's what matters... :P
It's milliseconds 😆 Clive and not Doppler radar. For this you would need two parch antennas. The RF circuit is just an oscillator. Someone analyzed it years ago. Guess you need to Google for it. I believe what you saw was interference with the reflected wave.
The Dobbler Radar circuitry is very interesting. Some half century ago I tried to build emergency morse radio for standard SOS-frequency, now defunct. Suitable transistors were quite expensive, I recall.
Yeah, but does it come Pink?!😆 RF is easier than AC. Clive, you would pick it up in about 20hrs. Rule of thumb. 1. All RF circuits are either, a filter, or an amplifier, or a mix of both.
Interesting fault 2x👍 The PCB antenna working are something I never really wanted to know about, (as long as they work). We had microwave antennas for speed trap detection and Im just happy they all worked out the box. It was interesting to see the difference between the UK and USA versions.
Curious Marc calls microwave circuitry "the blackest of black arts in analogue electronics". If anyone who enjoys Big Clive's teardowns and explanations of electronics ISN'T subscribed to Curious Marc, I suggest they look him up. There is a magnificent stage-by-stage analysis of how data, voice, telemetry, television and more was encoded and decoded into a single signal sent back from the Apollo capsules to Earth during the Moon missions.
Just sharing a sneaky VS1838 Infrared Receiver Transistor failure, here, to get the word out... I was repairing a solar light, which was controlled by an IR remote control. I discovered that its 1838 was faulty, which was causing the MCU to keep the light switched off. Here's the sneaky part... I tested the transistor function of the 1838 with a multimeter and the transistor tested as working. The problem was that the 1838 looks for a 38KHz "modulation" signal from the IR transmitter (inside the remote control.) The 1838 had lost its ability to read that modulation signal. As said, sneaky, because the transistor function tested as working. Hope that helps someone. Cheers.
2:07 - “… three one ohm resistors in parallel to give about 3.3 ohms …” Please try that one again: three one ohm resistors in parallel will give about 0.33 ohms, whereas three one ohm resistors in series will give about 3 ohms.
When an inductor is used for switching on and off, as it switches off it makes a reverse current for a short time and the resistor and capacitor could reduce that effect.
There are plastics that are IR transparent - for all that extra complexity (and need for power), I sure hope radar actually adds something useful... Black magic is also a good place to hide back doors ;-)
i am too tight to pay for a TV license, and i never watch YT ads. but i'll gladly contribute to your work.
Thanks.
Hi Clive,
We called that area by B a tuning stub when I worked at Motorola. You design the capacitor values as needed, but because of capacitor and high frequency part tolerances you need to match the capacitance to the high frequency part.
We used a laser to remove a small amount of metal to 'tune' the circuit to the exact frequency.
The squiggly track is also to match the impedance of the circuit to the antenna (VSWR) to deliver maximum power to the antenna (or from, in this case), and it is a simple formula of frequency to a single sign wave (2 kHz = a wavelength of 35 cm). It reduces the tract length (from straight, as in an collapsible FM radio antenna) by zig-zagging it on the substrate.
Yep. That stub in combination with the ground plane next to it *and* the ground plane under it form a tuned circuit making the basis of a single-transistor RF oscillator. The squiggly bit is the antenna + z matcher in one piece.
It’s not really that simple of course- the resonator and the antenna/matcher interact too kind of like a tuned-plate tuned- grid tube oscillator works but the coupling isn’t as strong, until someone walks by and it is.
Anyway, Clive is right, it’s a Doppler detector. Reflected RF adds to or subtracts from the current in the antenna and the difference gets passively filtered and sent to the active low pass filter about which he is also correct.
@HytelGrp
"Tract" vs Track.
Clever Bugger, I'm impressed, honestly. I keep learning things that I don't quite understand... er, completely.
Wow, even trimming it off with LASERS! That's finicky! I'm impressed. I knew RF analog was a bit of a trick but I hadn't realized it was that much of a trick.
@HytelGrp
The Wavelength of 2kHz =
149,.896 Kilometres.
What is VSWR.
RF is a black art - even to RF experts.
It's true; alot of rf guys I know 'cheat' by watching a scope while adding components rather than just designing it so it should work in the first place.
@@sometimesleela5947 Are you saying RF stuff can actually be designed instead of just stumbling upon a working configuration? :)
@@QwarzzYou can kinda design qualified guesses xD
We called it PFM…. Pure f&$cking magic!
All geometry with unknown parasitics....
Thank you Clive.
We all are hoping that hand is healing up well.
Burns can be problematical especially on busy hands.
I really love the longer/complex ones. Its fascinating to see how those boards work, especially with how well you explain all part but keep it simple at the same time ❤.
Very bright guy our clive. Well impressed with his knowledge.
Meh, nanoseconds vs. milliseconds. What's six orders of magnitude between friends?
It depends if you are measuring from the root or from the asshole. 🤣 OR maybe that is meters rather than seconds?
But that statement is incorrect.
Have a little think about it.
20 Aspro a day takes a man's pain away
micro.
I have a million friends who could answer the question, but he's out of town.
Don’t doubt yourself, Clive! You had it right the first time. It clearly says 0.2 milli seconds per div on the scope image, which makes the frequency 2kHz ish. If it was 0.2 nano seconds, the frequency would have been 2 gigahertz.
By the way, doppler frequency is about 20Hz per m/s, so a waving hand might give 50 to 150Hz, and a swaying Clive closer to 30Hz.
Not sure if this makes a case for or against it using doppler shift, but considering it only uses an audio-frequency low pass filter, I’m leaning towards yes - I can’t imagine how an LM358 could respond to any kind of reflected power ratio measurement, cos it would have to be operating at RF frequency to do that, or at least have an RF detector (diode and cap integrator?) as the front end, rather than the low pass filter.
Then again, if we guess that it’s operating around 5GHz or so, the reflected wave would interfere with the transmitted one in a varying way as objects moved in its vicinity. Nah… that would have no effect on the oscillator - it would be way too tiny.
I’m sticking with doppler, via mixing of the transmitted and received waves generating a doppler difference frequency.
I've played around with standalone versions of these "doppler radar motion detectors" and I believe they are doppler based. That one transistor oscillator is working pretty hard, and also serves as a mixer for the signals that are bounced back from nearby targets. Their shifted frequency based on the motion of the object causes the low-frequency mixing product to be generated and detected in that "audio" region.
It definitely is Doppler. What you’re missing is that the outgoing and incoming signals are mixed in the antenna (the squiggly trace) and the low frequency sum and difference are what the passive LC filter sees, and its output goes to the active low pass filter.
A swaying Clive should be more in the .3-3Hz neighborhood…
err, 3 1ohm resistors in parallel seems to be 0.33 ohm? @2:10 . Also, around @11:22 you have "0.5 nanoseconds is about 2 kilohertz". I think you misread the "ms" as "ns"?
He's reviewed so many chinese electronics, he's began to use their accuracy.
The question-mark cap may be coupled to the antenna to give a feedback to the base of the transistor to get the circuit to oscillate. Reflected RF will then be mixed by the same transistor. The doppler frequency is then filtered by the first op-amp. This is a typical bandpass, It will be tuned to pick out the typical frequency's made by objects moving at human typical speed. To filter out flying bugs.
What an adventure! Going through this circuit has been totally captivating. Thank you for being the guide 😊. You clearly had quite a challenge putting this description together 👍
Very interesting, I enjoyed the video.
Note that these will go deaf around wifi routers as the router emissions do drown out the doppler signals.
A few meters of separation is normally good enought o fix that.
Appreciate the effort that went into this one.
Yeah, good old fashioned front end capture.
It’s a consequence of the 5GHz band having to be shared by devices that do different things, and (in the US) one of them having to accept interference by the other.
I'm sorry Clive, I fell asleep (repeatedly), every time you got to 'nano seconds'... I just can't help it. You are clearly brilliant though, I wish we could meet, so that I could learn more... Keep up the good work, regards, Mike
This module was inspired by the RCWL-0516 module. That one operates at 3.2 GHz.
Didn't think they could make it cheaper? They did. They eliminated the need for that chip, with clever filtering and amp section.
I concur..I use the "cheap like borscht" RCWL-0516 units with a FET to switch on LEDS, alarms, etc.
Wouldn't a single chip be cheaper than 5 layers of filtering?
@@hubertnnn Good point.
LM358 is cheaper than dirt. So are resistors and caps.
I don't know the costs, but given the way Chinese operate, I wouldn't be surprised if they came across tons of these Op Amps on the cheap, and designed around them.
They do that for all through hole components which have no market left since SMD.
@@hubertnnn LM358 can be had for 1.5c each, 10 passives cost maybe 1c. So you are competing with a cost of ~3c.
Analog RF posse reporting Sir 🫡
It is Doppler shift sensing a low frequency RF signal ~2kHz.
The length and shape of the antenna gives it away. 🙂
Quite a complex functioning circuit! It’s like a solid state version of the proximity fuzes used in WW2, plus a microcontroller to select different modes.
I was surprised to see several of these online claiming 5 or 10 meter sensing range. Pretty significant for such a small, low cost and low powered unit.
Well done with the reverse engineering!
Interesting little circuit! My experience with "radar" detectors is a bit unusual - I installed one on my balcony light, outside the room I have the PC, so that I could note when people passed by at night. I found out that it was able to sense across the wall and trigger on any movement I made, inside the house! This was easy to fix with a grounded metal plate, but I didn't bother. It's still there, but I disconnected its power - that's how lazy I became :-)
I have read all the comments after watching the video twice & now realise i am an Expert! I verified this status by a simple process of rating my grasp of the technology at each point in the design.
Zero across the board and an overall average of nil on my scorecard is worthy of note.
Nonetheless i enjoyed it all thanks Clive & all.
You should start a podcast, I love having your videos on in the background just for the background noise. And of course they’re informative and interesting too
There was a live stream channel on RUclips, but I recently moved to Twitch due to problems with RUclips. The twitch channel is also called bigclivedotcom.
The circuit does use Doppler shift. The oscillator transmits a signal. It is reflected, if the object it reflects off is moving, a Doppler shift is created. The reflected signal is received on the same antenna as it was transmitted from and mixes with the oscillator signal to produce a beat frequency equal to the frequency difference between the two signals (i.e. the Doppler frequency. Its this signal that is then being filtered and detected I think. This is exactly the way the first proximity fuses for shells worked. They had a single 'peanut' valve packed in wax to allow it to withstand the shock of launch which did the work of the transistor in your circuit. Basically one active component functions as transmitter, receiver and mixer. I know about this because the same principle is still in use today - although we are at least using transistors now!
Thanks for the detail, Clive. I have one of those solar street lights with LiFePO₄ cells in it that's been running for years. It's surprising to see that it will occasionally respond to insects or bats chasing them, strong winds, especially when accompanied by rain. Fortunately, the thing has a "full on" time of about 30 seconds, so it's not really an issue.
I believe you have it right on Doppler. The microwave oscillator is so fast that even a slow reflected movement will change the frequency down to an audio signal not just a different RF signal. The opamp can pick this up and filter it to a DC level and trigger the light.
Re that op amp feedback circuit, the main element is that cap going directly from out to -in. That drops the gain to unity for higher frequencies. The whole thing looks like an active low-pass filter. The input cap blocks DC and the voltage divider recenters the waveform at 2.5V.
Wow...this is an extremely complicated board...very cool!
Thank you for explaining it so well :)
Multi-path interference. Basically heterodyning itself but with slightly different phases, the result of which is a low frequency signal. The difference in signal path is doing the work, not the speed of movement (doppler).
It likely is just constructive/destructive interference with the reflected wave. For mixing (heterodyne) you need multiplication. Such nonlinearity is not given in this simple circuit.
Good job Clive, thank you for your service Sir
Thanks Clive, sounds like it took a long time to get this one reverse engineered
Ya think! Yay BC
Really fascinating subject matter, I understand the principles of the build, but would not attempt to build it myself. I've e purchased plans to build a Bluetooth speaker with 10 watt speakers. I bought all the components and have started to build. You are a bad influence on some people Clive as this is the fourth project that I've begun since I started watching you years ago. Believe me when I say it is not just electronic projects but all kinds that stop the decay of the grey matter. Good post, thank you👍
Nice workaround for fixing a bad module :)
You continue to demonstrate your brilliance with our appreciation!
👍👍 Good job Clive, thank you.
the rf section reminds me a lot of a regenerative receiver, where you monitor the current draw to pick off the received signal. I guess this is not much different than this- it is just receiving its own signal after it bounces off something.
11:25 i assume it's 0.2ms on the photo. so 2k5kHz is right (the device will probably not get that resolution on 2,5GHz ;-)
The opamp with RC network shown is a typical band pass filter
very cool Clive. Thanks for sharing ❤️
I agree! RF is black magic. Good job sorting it out.
2:06 Are the resistors in parallel or series? Your photo looks like parallel but could easily be deceiving me...
Irrespective, many thanks for your continued work... It's all greatly appreciated.
[edit] Ah watched a bit further... 4:13 they are in parallel, thanks!
8:07 looks like that 150 Ohm resistor serves double duty as both a shunt and as a means for setting the DC bias for the transistor. The capacitors across it serve as bypass to ground for the RF.
Anyway, the designers seem to be interested only on the average current across the "shunt", thus the 1st order low-pass filter formed by the adjacent 1K5 and 47K resistors and capacitor (the 47K resistor would have negligible attenuating effect, and could be omitted with no major impact on performance).
In your diagram, there are a few essential microstrip elements missing. I've simulated it as is, and the oscillator performs poorly.
Oh this would be an interesting follow up video!
Good intervention sine wave Doppling Clive
The doppler we had sensed movement thru the wall, floors and ceiling. I was very confused for some time before I understood why the light was on so much.
And often stuff like trees moving outside too.
The wiggly track is actually the antenna. If I'm reading it correctly (I'm by no means an expert in this...) the antenna transmits a frequency, and when it receives the reflected signal, the interference on the antenna gets amplified by the op-amp circuit. The scope screen does look like it shows 0.2ms, but then, I didn't do the measuring, so you likely know best ;) But if the 2kHz is correct, it's 0.5ms, not ns.
The filtering caps in the op-amp section, especially the one between the 1M resistors, might be used to vary amplification depending on the frequency.
I believe the track your referring to is an inductor, if you mean the one on the emmiter of Y2
Ooh I see the track you mean my bad, your absolutely correct, apologies
Don't think it's doppler, I think it's a bad self feedback, which your waveform suggests also, being in a constant state
Wow and Wow! Yo Da Man! I'm going to have to watch this many times to even begin to get an idea of a grasp on this black art. So Interesting
RF design is indistinguishable from magic.
Thanks Clive you rock!! We appreciate the work you put into your videos and reverse engineering, making it easier (but still very complicated) to understand. I find myself nodding like ya that makes total sense...but I am still totally clueless! Ha ha ha...Love your videos and I will always return for more!
The "splooge" capacitor is not a normal place for one. With some opamps, it could cause oscillations. The LM358 happens to not be normal under the hood and not very fast either, so all bets are off for why it's there.
“ Built By You ,” Clive Would be Bespoke Not Far From Perfect Like This One.”❤
In my garage I have a RaspberryPi server with a camera and Python code to send me, when image pixels change, photos to my email. I was overloaded with photos in my email. I put one of those RF detectors on the GPI and modified the code to only send email if the RF detected motion. It works just fine. I keep the extra images/movies on a thumb drive that are less than two weeks old. Each night a crontab schedule deletes old images. It's the best of my hobby projects yet. RF detectors are much better than PIR.
Waa. Neato
There's an amazing animation of a solid state beam steering antenna. Can be made similar to the traces there, but is a weird asterisk shape: * . Slow mode field animation shows if you send signal/power to one side not the other it "forms" the beam to point in the other direction, and likewise can fill each finger with more or less power/pulsed waves to shape it's direction. I wonder if those through holes help shape the direction?
Thanks :)
8:32 is an economy version of a twin T filter in the negative feedback loop of the first opamp. So the first opamp is a narrow pass filter.
I shudder to thing what the bandwidth is More gain less bandwidth and all that
I'd say that copper pad is an inductor, not capacitor as it is a 'short' possibly forming a feedback circuit in an oscillating transmitter?
I don't think that 2.5kHz signal is meant to be there at all, it's a spurious oscillation of the band pass filter (the left hand op amp). The output of this op amp is supposed to be a dc level at half the supply voltage when there's no movement. The threshold of the comparator (the right hand op amp) is about 0.12V below that dc level. So any low frequency Doppler frequency that gets through the band pass filter larger than 0.12V will be detected and turn the light on. The presence of the large 2.5kHz oscillation will produce continuous output and prevent it from working. By attenuating the spurious signal you have prevented this from happening and allowed movement to be detected. The bodge capacitor was probably and attempt to stop the spurious oscillation.
incredible what technology we can get for a small buck nowadays.
Good tear-down. Never knew you could dopler rf, sure you dont mean phase shift detection?
It's doppler, Andreas tested some, but what's the nighttime standby drain like? I have some modules, but they greedy, I forget the drain, but not suitable for battery powered home automation, which is a shame. There are some quite advanced programmable ones now, I believe they can detect your chest raising and lowering as you breathe at night while asleep, which makes a perfect presence detector, finally.
The standby current is definitely significant compared to PIR.
2:08 in 3 one ohm in parallel?
"with schematic"
Aw yiss, the good stuff!
"and fix"
That's the luxury!
Ain't no LM358 on this planet that can output a signal in the GHz region! It's 0.5 mS, or about 2 kHz, which the LM358 can certainly handle. Having said that, one well-known problem with an actual LM358 chip (not an updated LM358) was its awful totem pole output. If you bias the op amp in its mid range, as is the case here, then the output will have odd crossover distortion artifacts as the output tries to alternate between sink (PNP) and source (NPN) of currents to some light load. The fix is to add something like a 1 kohm resistor to ground on each output, which effectively biases the output so that the NPN output transistor in conjunction with the pulldown resistor are always supplying the output current. The PNP output transistor is then effectively cut off. Any newer updated version of the LM358 (which will never include "LM" in its name) fixes this known flaw in the opamp. Whenever I see any kind of design that uses an actual LM358 in it, it immediately tells me that the designer was not very experienced in linear design. ruclips.net/video/VgodYtiD_F0/видео.htmlsi=wK2MK04FJWuOPijM
I have a complaint, you didn't put a link to tell us where to buy one of these this time.
I believe that the "doppler" works in a similar way to a theramin, moving close to it will change the frequency of the oscillator and the change is detected by the first op-amp.
( I could be wrong, but that is how it looks to me )
These radar detectors rather than PIR are a real pain .
I fitted a light with one built in , it sees you through the door in the hall , and even through the floor to the landing upstairs .
The light is nearly always on because the cloakroom is dark and waiting for a trigger , which it almost constantly gets .
I'm having trouble undering how the RF section at 7:14 will oscillate. I don't see a feedback path back to the base of the transistor. Could it be from the physical placement of the tracks which is causing capacitive or inductive coupling to the large pad at the base?
It's hard to say when working in such high frequencies. Everything behaves differently.
At RF EVERY part of the circuit is inductive and capacitive. And they all resonate. Even changing a track width, or placing it 0.1mm closer to another track changes the frequency.
That was milliseconds, not nanoseconds. If you had .5ns wavelength of that thing, it'd be at 2GHz
"There are advantages on the doppler, especially if it's working"
I work at a place that sells fitness equiptment and we have a TM that is notorious for blowing the lower control board because of friction. I wonder if big clive would be interested in finding out why the boards are so junky straight from the factory. I wish i knew how to get ahold of him but im not great with computers. Its from a BH S5TIB. They used to be good TMs. We are stuck with a whole container of those and they are nightmarish for us
Thank you, keep working.
Why did you leave a blank page between the pages of the schematic
What is Doppler
To stop the image coming through under the bright lights.
If it is doppler, there'd be some sort of "mixer" on the pcb track.
So imagine one part of the PCB oscillates generating an RF output (transmit beam if you like), and then the receive part of the same PCB has to interfere with the reflection signal. Or "mix".
So the result would be a low frequency. Perhaps that 2kHz signal means the receive part of the PCB was "out".
Instead of generating frequency bursts, it is probably meant to generate very low frequency pulses. Like on or off blips.
Think of police radar detectors which use horns and precision cavity resonators. School "Young's slits wave experiments".
What's the sensitivity range like?
The PIR is ineffective with ambient temperatures around 48c, common in tropical regions. Doppler however is succeptible to false trigerring around water bodies and plants and trees.
Excuse me, the filter on my filter's filter is not filtering.
Could you add an extra filter on that filter?
@2:09 Aren't three 1 Ohm resistors in parallel .33 Ohms?
Maybe I misunderstood you. I thought you said 3.3 Ohms.
🙂 OK, @4:15 you said .33. 🥰🤩
omg...doppler radar for solar lights......
so can like detect an incomming Russian Su 25 .....Su34 or Mig31...
and light up my garden to warn me.....
you soooo remind me of old electronics teacher...
apart he had a white board...
Top videooooo
N x
As long as it provides a Homer Simpson "Light goes on, light goes off, light goes on, light goes off!" level of entertainment, then that's what matters... :P
Very nice
It's milliseconds 😆 Clive and not Doppler radar. For this you would need two parch antennas. The RF circuit is just an oscillator. Someone analyzed it years ago. Guess you need to Google for it. I believe what you saw was interference with the reflected wave.
The Dobbler Radar circuitry is very interesting. Some half century ago I tried to build emergency morse radio for standard SOS-frequency, now defunct. Suitable transistors were quite expensive, I recall.
Yeah, but does it come Pink?!😆
RF is easier than AC. Clive, you would pick it up in about 20hrs.
Rule of thumb. 1. All RF circuits are either, a filter, or an amplifier, or a mix of both.
Interesting fault 2x👍
The PCB antenna working are something I never really wanted to know about, (as long as they work). We had microwave antennas for speed trap detection and Im just happy they all worked out the box. It was interesting to see the difference between the UK and USA versions.
Curious Marc calls microwave circuitry "the blackest of black arts in analogue electronics". If anyone who enjoys Big Clive's teardowns and explanations of electronics ISN'T subscribed to Curious Marc, I suggest they look him up. There is a magnificent stage-by-stage analysis of how data, voice, telemetry, television and more was encoded and decoded into a single signal sent back from the Apollo capsules to Earth during the Moon missions.
He would, it doesn't take a genius to work it out
Years ago I went for a job interview at RF Micro. The lead engineer called 100MHz "practically DC."
Just sharing a sneaky VS1838 Infrared Receiver Transistor failure, here, to get the word out... I was repairing a solar light, which was controlled by an IR remote control. I discovered that its 1838 was faulty, which was causing the MCU to keep the light switched off. Here's the sneaky part... I tested the transistor function of the 1838 with a multimeter and the transistor tested as working. The problem was that the 1838 looks for a 38KHz "modulation" signal from the IR transmitter (inside the remote control.) The 1838 had lost its ability to read that modulation signal. As said, sneaky, because the transistor function tested as working. Hope that helps someone. Cheers.
358 circuit looks like a bandpass and was maybe oscillating.
Meanwhile, somewhere in China... "Ah Big Clive he fix our module, we need put capacitor back"
Pretty sure it's 5ms on a DSO "scope" scare quotes. You are not seeing microwaves.
Thank you
2:07 - “… three one ohm resistors in parallel to give about 3.3 ohms …”
Please try that one again: three one ohm resistors in parallel will give about 0.33 ohms, whereas three one ohm resistors in series will give about 3 ohms.
‘Trigged’ on the scope… is that like triggered with extra trigonometric math? There’s even spare characters 😂
When an inductor is used for switching on and off, as it switches off it makes a reverse current for a short time and the resistor and capacitor could reduce that effect.
Next step - download and disassemble the code in the microprocessor!
Unknown processor and probably locked.
Wow, that's not a BC108 then!
I wonder if you could use it for a proximity fuse for your ground to air missile. Maybe your next video. 😂😂😂
its a puzzle and a maze
If it's not true Doppler, would it be appropriate to call it a Dopplerganger? 🤣
Luckily not your right hand 🖐 😂
You need a few jewels on the glove, hope your hand is healing.
There are plastics that are IR transparent - for all that extra complexity (and need for power), I sure hope radar actually adds something useful... Black magic is also a good place to hide back doors ;-)
Sonst aber meisterhaft!
T.y. The remotes.. power reduction feature, I encountered on solar banks for fairy lights The long dim rice lights.. self dimming...
hello world
Are you allowed to say 'Black Magic'? Will youtube demonetise you?
If you tube don't, Nestlé might.