Not strange at all. I was taught back in the 70's that all that "Doppler detection" wasn't how things really worked. I've played with the horn out of a old microwave motion detector and have seen it for myself. Here's how it works: You have a free running oscillator that transmits a signal. (In the case of my unit, a Gunn diode.) You also have a receiver that picks up the transmitter AND the reflected signals. AM demodulate the signal (in the case of my unit, it was just a simple microwave diode halfway down the horn. The PC board was the filter capacitor.). Then look for low frequency AC. (Ignore the DC value) That's it. Why does that work? The combination of reflected signals will either add or subtract from the signal received from the transmitter. If something moves, that will disturb that "balance" and cause a change in the DC value (aka a AC signal). AC detection can be quite sensitive and find small changes in the DC signal. Outside of two semiconductors on the horn assembly, everything else was power supply or low frequency audio. No narrow-band filters or precision measurements. (Cheap, cheap, cheap.) I challenge you to find something that really measures the frequency. I doubt if you'll find one, because measuring frequency is complex. The closest you'll come is that you can claim that it's measuring the beat frequency between the transmitted and reflected signals - but that circuit will still be substantially the same as I've described. In that case, the diode detector will be called a "mixer". But you end up with the same results - a AC signal the represents movement.
Russel has it right. The wavelength at 5.8 ghz is 51 millimeters. Half that is a complete phase reversal. So a movement of 25 millimeters, one inch, is enough to completely reverse the phase of the reflection and when mixed with other non-moving reflections in the room will change the amplitude on the receiver. A ceiling fan would probably drive it nuts but it may be filtered to ignore rapid and continuous changes (low pass filter).
Russel you just described a downconverter and doppler phase shift detection! All you need is usually a diode, almost always biased IME, to supply a nonlinearity. I've also commonly seen a transistor used as the mixer in place of the diode.
These modules use a method that was briefly popular in the 80s in the electronic hobby magazine. IIRC, ETI did one, for example. They work like this: The microwave oscillator runs continuously at a resonant frequency set by that very obvious worm-like track and probably a surface mount capacitor. This oscillator emits a few milliwatts of energy as best it can. The additional stub antenna (that white wire) helps radiate a bit more RF energy. In the simple version of these things, a simple diode is connected between that resonant circuit (on one side) and the input of a really high gain low frequency (typically audio bandwidth) amplifier on the other. The oscillator therefore not only emits (low level) RF energy into the general area, but therefore also acts as the local oscillator for the diode, just like a "direct conversion receiver". OK, so nothing much happens if there is no movement, but the instant there is movement, the returning reflected RF signal will now contain the Doppler-shifted oscillator frequency. That happens to mostly be at audio/low frequencies, and they get amplified, filtered and detected, in this case in the BIS0001 sensor. It's a chip specifically designed for this sort of task, but normally used in that fashion with the PIR sensor, amplifying its low level low frequency signals. The old ETI (and other 80s) designs typically used an LM324 quad op amp to boost and filter the diode detected signal. Nice and neat. And another idea that we've seen before, but now nicely manufactured in vast numbers at ultra cheap prices. Thanks for the teardown, Clive
@@krecikowi a couple milliwatts at most. If on the same frequency as wifi (these little oscillators with a room as part of the tank drift around like crazy) it would lower your reception a bit. No health effects
I keep watching his videos hoping that I will one day have an 'ah-ha' moment and electrical circuits will one day make sense to me. I'm a programmer, and I understand how computers work all the way down to the logic gate level. But the logic gates are made of transistors in electrical circuits. And every time I have tried to learn that level... it just doesn't compute. I can tell I am missing something tremendously fundamental, something about the way I am actually trying to think about it is incorrect or incomplete, which prevents anything else from making sense. I can skip over that level and go into how the transistors themselves work, with their NPN layers and migration of 'holes' on the molecular electromagnetic level even. But the circuits... nope. Maybe I'll reach a critical mass of hearing various explanations of various things from Clive and one day it will all 'click'... that's my hope.
I have to thank your for kindling my interests in electronics once more. I've had a nerd gun that I was planning to turn into a sort of laser rifle prop with lighting and soundfx since last year, but your videos got me thinking about it again :) Just came back from the shop with the sound trigger module and the little audio amp for the speaker
See, all the magic ends outside the topmost board with an antenna and weird PCB trace. That board contains a simple one-transistor RF oscillator, which doubles as an RF mixer (it has much in common with super-regenerative receiver topology). So it generates and radiates the RF energy, and at the same time is mixing generated signal with the reflected signal. The result contains harmonic with a frequency equal to the DIFFERENCE of the forward and reflected signal frequencies, and this is something about 5 - 10 Hz. This is fed to the second low-speed board. So it works like this: no motion - no difference tone, something moves - there is a difference tone on the output. The frequency of the difference tone is proportional to the object speed, and amplitude depends on reflected signal strength. The weird trace on the RF board is most probably an RF microstrip transformer/resonator (and then this is probably a Hartley-type oscillator). A bunch of capacitors is there for supply decoupling. So no RF on the second board, it deals only with the difference tone.
It's always amazing to me how the human mind ( including myself) naturally wants to improvise and use one thing for something else! Would still be living in the dark without it ! 👀
Isn't that the way of things? Here I was, the poster child of "ignorance is bliss," gradually figuring out and working with RF stuff, and it turns out I've been practicing the Black Arts the whole time. ;) lol jk I really liked this tear-down, Clive. You're right about the track layout being important to HF designs. The 'troubles' begin when signal wavelengths begin to coincide with track lengths, in the radio wave to microwave frequencies; radio is 3 Hz to 300 GHz ( λ = 100 km to 1 mm), microwave is 300 MHz to 300 GHz ( λ = 1 m to 1 mm), but I usually say 'beware' around & above the GHz region. Basic components start acting weird, or at least not as anticipated. Those non-ideal values such as ESR, ESL, and indeed, track (and ground plane!) layout, become critical.
Q1 is probably a NPN BJT. The strangely shaped PCB traces make resonant circuits at 5.8 GHz, and putting them at the right spot in the transistor circuit makes an oscillator. Movement changes the tuning of the circuit and the voltage changes a little bit. The other chip just amplifies the crap out of that voltage change and triggers when it gets to a threshold. The lots of tiny holes are plated thru via holes connecting the outside of the top of the board to ground around the edge, or making the output line bigger. You can design impedance matched lines on PC board knowing the dielectric constant and thickness, and doing math on it (or looking it up on a nomograph). I've used standard FR-4 up to 5.6 GHz, amplifiers and antennas and whatnot. I first saw this circuit around 1964 where they used a tube oscillator at around 300 MHz. Around 1971 I built one at 10 GHz using a reflex klystron that I found laying around. Ran it into my guitar amp.
Can confirm that Q1 is indeed an NPN BJT, a high frequency one. Though I don't fully understand radio myself so this didn't bring me closer to understanding how the signal was transmitted, received and mixed.
Think of it as using the whole room as part of the oscillator circuit. A change in the resonance of the room makes a change in the DC current supply in the oscillator. Couple that into an amplifier, trip a comparator, and you've got a burglar alarm.
Thanks for explaining! So could I use the output from the amplifier directly on an MCUs ADC input to find out which way the voltage shifted, up or down, and get the direction of travel of the moving body?
No, all you get is a perturbation. It's a kind of self-mixing where you get a baseband detected signal: near DC. The problem is that your detected signal has no negative or positive shifted frequencies that could tell you whether the doppler shift was positive or negative, approaching or receding. It wraps around the zero (DC) point and both appear positive. You can count the frequency and figure out the speed, however. You can also tell things about the size or distance to your intruder, too, for what that's worth.
James Van Damme If I understand, a moving body causes an oscillation across 0V perhaps due to a phase change or the like from the resonating frequency, however you cannot determine the movement direction. You say speed and distance can be determined, do you mind elaborating? Thank you very much for taking the time to reply.
Ok. I've not read every single comment but I can see that most people are missing the point! What we want to know is: Are those LEDs warm white or bright white?
I saw something similar in a 1970s hobby book of schematics. They called it something like "UHF proximity detector," and it was a 1-transistor UHF oscillator with a little antenna, then other transistors on the DC side of things. I think they were just measuring the current draw, and amplifying any AC changes. I didn't build one, so I don't know how many meters distant it would still detect motions.
The microwave sensors have been around since the 90's, they were an optional extra in the car alarms we used to fit at Road Radio. We used to fit one under the drivers seat, for the car alarm purposes it just used to create a voltage drop to trigger the alarm. The ultra sonic detectors were pretty much standard, but a window left open a little bit would set them off.
This is a possible method of operation that will turn your brain inside out: There may be no oscillator on the module what so ever for the 5.8GHz. I suspect what you have is a couple of narrowband tank circuits (think coil and variable cap in a crystal set), that use far field capacitance (IE a human body) as one of it's finite elements to couple them (or possibly the only element as the polarities are at 90 degrees IE vertical and horizontal). If you charge one to say a couple of volts it will then cause oscillation in the other (and it's self) every time the far field capacitance changes (IE your body moves). You only need miniscule changes in capacitance to activate the circuit as at 5.8GHz even 1pF is only about 27 Ohms impedance. If you are only looking for a micro amp current flow from 2.7V you would only need 1/100,000 pF capacitance change in the far field capacitance. When Idle this would draw a current in the kappo-amps range on FR4 or less on a ceramic substrate if it is clean. The only detector you would need is a Gunn diode or similar (think cats whisker in a crystal set). The KGB used a very similar method to this with a remote exciter to listen in on a bug that nobody figured out for years as it just looked like a metal clip.
This whole thing was making me think of the alarms around about the climax point in the 1992 movie Sneakers, I think those were ultrasonic, but when ya mentioned standing still and posh alarms as well as those ones that use both IR and microwave...
Units I used back in the day fooled me in a similar way. It used separate transmit and receive horns but the receive horn just had an audio amp connected directly to it! The two horns were cast into one piece of metal with a small notch between them. What the notch did was mix a small amount of the transmit signal back into the receiver. What happened was the reflected signal mixed back into the horn with the sample of the transmitted signal and any frequency difference mixed in the detector and audio came out. Effectively, the receive circuit could only see the difference signal of the heterodyned result. This means the microwave signal was not basically detected at all in the receiver -- only the difference product between the transmitted and received signals. I believe the "detector" was nothing more than a slightly forward biased diode if I recall. This was about 30 years ago. Bottom line: stupidly simple circuitry that boggled my mind as well. The nice thing is since you're just checking a difference frequency there is no real need for tuned circuits or transmitter frequency stability!!! I modified one into a "doppler radar gun" which I could use to check speeds although I never got it totally accurate with the circuits I built.
It definitely looks like a single transistor oscillator for the RF generation. What I can't tell is if the output is split, one going to the antenna and the other going to the comparator. I can guess it's using something similar do a hetrodyne radio receiver, but the local oscillator is also the transmitter oscillator. When you mix the local oscillator and the reflected signal together, the remaining signal is in the low hertz range. Even though you're using really high frequency RF, the detected signal isn't.
I'd say it's a simple, free-running heterodyne oscillator where the antenna forms the primary load; the reflected signal is so strong that it causes the oscillator to either 'lock on' to some frequency dictated by the reflection pattern or start fluctuating -- and those fluctuations are detected by the rest of the circuit and used to light the lamp up.
Mora Fermi If the "load" (person moving) is modulating the main oscillator, how can you tell if they are moving? You don't have a stable reference to compare the reflected signal to. Instead of measuring the pull of the oscillator, you mix the oscillator (who's frequency isn't stable) with the reflected signal. The difference frequency is the movement of the person, and that can be detected with a simple AM conversion.
It is unnecessary to use the original signal as a comparator. What will happen is the room is filled with reflections and they reinforce or cancel to varying degree. At 5.8 gigahertz the wavelength is 51 millimeters so it takes only 25 mm (one inch) to obtain a complete phase reversal and produce a detectable signal strength change.
Any sufficiently advanced technology is indistinguishable from magic. - Clarke It looks like you stole that lamp out of Rick Deckard's apartment in Bladerunner.
Great summary as always, Clive! If you want more "Radar" based detectors (cheap, seeing as you're a Scot :-) ), Bang Good has heaps for "a good price". I use them in combination with PIR to reduce false triggering in our outdoor lights.
The detector send hf and receive hf but with a little frequency. Then you multiply what is sent by what is received. Then you remember that sin(a) sin(b) = [cos(a - b) - cos(a + b)]/2 so mixing the 2 signals gives you a sum of 2 sinus. One with the frequency offset and one with almost twice the original frequency that you discard with a low pass filter. At the end you end up with a sinusoidal signal with a frequency proportional to the speed of the object in your direction and amplitude corresponding to the size and proximity of the object. This simple signal is easy to detect.
Sounds like a pretty good security light: someone comes up to peek in your windows at night while you're away, lights come on. Unfortunately my neighbourhood is hoaching with cats and I expect the ones that get up on the roof would set it off constantly.
This has been spoken about on other electronics channels of late. I am more convinced that the module just transmits and measures the SWR, uses the BIS0001 to compare the signal. The signal frequency on these units seem to be more around the 3.8Ghz. As pointed out by someone on another channel the single transistor is non linear and can have two voltages present at the same time. The holes are just via's for the ground plane. The SWR theory makes sense when you find that they don't disturb each other.
Draws two cars for illustration purposes, then sais "A motorbike is a good example because of the noise they make" and forgets about the two cars ... But seriously Clive, I love your video's :D
Here's a link that discusses the functionality of this device and the type of transistor used (Bipolar High Frequency Transistor): ua-cam.com/video/jAeFQEHWLZU/v-deo.html (credit is given there to this video for sparking interest as to how it works) A potential candidate for the transistor used: www.nxp.com/files/product/doc/MMBR941.pdf
On second thought at microwave frequencies, the doppler change will be in the audio range instead of sub-audio. My radar lamps at home seem to not go through mirrors. I can wave my hand all over a large mirror on a dresser but move away from it and the light picks up.
It's really not that complicated! it's just an RF oscillator that operates at a high enough level to radiate a useful level of signal. The reflected signal will be picked up on the same antenna, and through non linear action in the same transistor mixes with the oscillation. The beat (the frequency difference due to Doppler) between the transmitted and received signals becomes impressed on the current flowing in the device, and is simply a very low amplitude, low frequency audio signal developed across a resistor in series with the oscillator's collector or emitter. Since it is very similar in amplitude and frequency to the output of a PIR sensor, they've simply used the BISS001 to do the high gain low noise amplifying and level detection. I first saw this very sensitive single transistor RF stage design in a high-end car alarm back in the 80's and thought it very clever then. That would detect movement outside the car and give you a verbal warning to back off!
I was just about to say that. With a two inch wavelength the beat not going to be much over a couple of Hz. I wonder how two of them work in close proximity. Would they lock together or just stay on all the time?
That makes sense. Yes I remember one of those car alarms actually telling me and everyone else on the pavement who walked past it to "back off". Very irritating.
What is really crazy, even with this bizarre technology, this microwave/Doppler lamp at 5,43 cost from eBay is less expensive than a standard LED lamp from a hardware store or department store in Canada.
Yes! Finally you got one. Don't you just love the cheesegrater heatsink fins? Mine got more and more erratic over time, pretending not to notice me sometimes. I have taken it down to try and fix it but haven't gotten around to it. I don't think this works with the doppler effect. I think this is just a single transistor oscillator and it detects changes in overall reflection/absorbtion in the room by monitoring the supply current to the oscillator. The PIR IC is able to detect very tiny changes. 7Y MMBR941L Mot SOT23 npn RF fT 8GHz MRF941 So just a NPN transistor. I've even taken this into the radiated EMC room and it makes a spectrum of narrow lines at around 3Ghz or something. The other "radar" detectors on Ebay (with the white dome case) do really use the doppler effect. You can tap into the signal and make a speed detector, but I think this one is different. It is strange because it really was so incredibly sensitive.
You should make a camera system that hooks into this lightbulb and the activators are in parralel. so you could have lightbulbs in seperate ends of a building and hook up cameras to go with it. If nothing happens, cameras save power and storage. If motion is detected, camera activates and catches movement. Could be your pet or could be a thief. Who knows.
Some microwave doppler modules use a mixer to mix the transmit output with the received reflection... with a static target the output is a dc level... movement of the target produces a beat frequency which is dependent on speed of relative movement....
I believe that transistor is a high speed JFET used to amplify the transmitter signal. If not then it is the oscillator at which point I would expect the capacitor to be of very small value. Either way, doppler detection modules are just frequency comparators used to trigger other functions.
Standard operation for cheap microwave radar modules. Rf is generated by a Dro, dieelectric resonance oscillator, just a fet oscillator with a ceramic 'puck' as the tuned element. Rf is generated then split 2 ways, one goes to the transmit aerial and gets radiated out into the room, the other goes to a mixer, this mixer mixes this signal with the received signal from the receive antenna. If an object is stationary theres no diffrence in frequency between the transmitted signal and the received signal, as soon as an object in range moves theres is a difference in freq, this freq difference depnds on the speed of the moving object, because people move very slowly compared to the speed of light the output freq of the mixer is only a few Hz. The 'voodoo' tracks on the board are microstrips, and these are often used for filtering at microwave frequencies.
And it is also presumably capable of detecting anybody flying an R/C drone or a WiFi signal in the area! (we pilots often use 5.8GHz analog video feeds for our FPV video singal in our drones.)
It's a homodyne transciever, basically measuring the SWR of the antenna as the people (or other stuff) move. They used the same idea at 915 MHz back in the 70's for security systems.
I would go with that, I used to buy Doppler modules from RS to make alarms for long coaches, they contained a Gunn diode and a receive antenna in the same "wave guide" a small portion of the signal was sent straight to the receive antenna inside the module to act as a reference, the rest sent out, the "bounced" back signal (off of a moving object) is then added/subtracted from the original and these changes were used to trigger the alarm circuit.
I'm thinking what is going on here might be as simple as measuring the power consumed by a microwave oscillator that's coupled to an antenna. One of the ebay listings linked in another comment has a decent picture of what appears to be a very similar design. Looks like the chip on there is most likely a single microwave transistor. One pin is connected to power with a bunch of bypass capacitors. Another pin seems to be set up for biasing the transistor. Then the output is connected to that long curved trace which acts as the antenna. Could be set up as a slot antenna with the hole in the ground plane on the back side. The far end of that trace appears to be connected to ground with some capacitors and a resistor. The capacitors provide the RF return while the resistor provides the DC return. Looks like the voltage across the resistor gets picked off and routed to an output pin. My guess is the BISS0001 chip is just being used to detect changes in that output signal. The idea is that the antenna acts as both a transmit and a receive antenna at the same time. Since it's part of the feedback network for the transistor, the transmitted signal that gets bounced around in the environment with interfere with the operation of the oscillator, changing the effective load impedance and therefore the performance of the oscillator (frequency and/or amplitude). Changing frequency and amplitude will change the power consumption of the oscillator, which is relatively easy to measure. Now, what would be interesting would be to set up a spectrum analyzer to see what happens to the oscillator frequency. Might also be interesting to see how susceptible it is to external interference - i.e. 5 GHz band wi-fi.
If you have ever played with regenerative receivers, you know that they can interfere with other receivers, since they transmit as well as receive. Since they are only a weak oscillator, they can be easily disturbed by stray capacitance and conductors placed near them. I think this is how this circuit works. Think of a metal detector that detects anything that conducts. A grid dip meter is similar a tuned circuit upsets it's grid bias and shows a "grid dip" when the meter is tuned to the circuit under test.
actually the frequency shift is in the nanometers for RF. what they actually do is to mix it in with the original frequency and you get a very low, easy to sample frequency.
These do use the doppler method as you say, but they don't do clever frequency filterring or anything, if something is moving, there is the incident rradiation, and the reflected frequency sets up a beat frequency with it, so as you move, you get an audible beat that is higher pitched the faster you are moving. Way more simple than you were even saying. Presumably that PIR chip would detect even small amplitude beats and trigger. It's probably using a gun diode eather in a wresonant cavity or a tuned bit of trackwork! I'd bet it has one set of tx/rx antennas, and the beat is detected from it's current draw.
you really need to send that one to dave jones for a two minute teardown. i'd like to see what he thinks of it. i'm sure it will be totally obvious caveman level stuff to him.
AFAIK Dave is no RF "black magic"-person. I'd think Mikeselectricstuff has much more of an idea of how it works. Another option is to have W2AEW look at it, he's a genius with RF-circuits.
I add the Signal Path Blog for consideration. I would put my money on Shahriar if I needed someone to quickly figure out the workings of an RF circuit.
Some of the commentators to this video look to have worked out everything or almost everything based upon the video alone if you are interested in the workings
Clive the 7Y transistor seems to cross references as an MMBR941LT3 NPN high frequency transistor, able to operate up to 8Ghz which would seem to go along with the application. Here's its datasheet.. pdf.datasheetcatalog.com/datasheet/motorola/MMBR941BLT1.pdf
I'm pretty sure (could be wrong) the piece of wire is the transmitter and the wavy track is the receiver. The weather radar receiving antenna on aircraft are built in s similar manner but on a larger scale. they also transmit the signal at the same time witch it witchery to me but all i need to know is how to test it and replace what needs to be. :)
*Off Topic Mystery:* Clive, you seem to be a man who has come across a lot of weirdness in his time -- and I am not referring to Alex Salmond's past girl friends here -- and therefore you may have noticed this effect or heard about it some other way. You know those self gluing envelopes they sell in the post office? The ones you just have to press together and they are supposed to stick securely (Do they buggery!). Well, I had occasion to pull one open and I did so by pulling on the flaps. I happened to be standing in a darkish room at the time and to my astonishment I noticed what appeared to be tiny blue sparks appearing along the glue line. I pushed the two parts of the flap back together to see if I could repeat the effect and as the glue came apart, there were the sparks again. They looked like the kind of sparks you get from a low voltage current jump between a hot wire and a metal surface. Ever seen it or know why it is happening? PS. I thought it may be some kind of static discharge but I was amazed to see it happening on glue.
That's an interesting effect. Then again, dry glue has to work in some way after all, and relying on electrostatics (on a molecular level, anyway) does not sound so outlandish. Long molecules giving off charge as they are stretched are not unheard of either. A (bio)chemist should have more to say on this, I imagine.
Scotch tape does that as well. Static charge generated by separating two surfaces. Some time ago someone detected x-rays generated by unraveling scotch tape. AFAIR this effect was published in Phys.org.
I wonder if water thru PVC pipe or ceiling fans will trigger this one, that is a problem with RADAR sensors, some have algorithms to ignore those things, pet immunity under a certain weight etc... Of course a dual tech PIR and RADAR like security sensors would be false free. You may want to partially obscure the cds cell so it functions with some ambient light.
An intriguing little device indeed, makes me wonder what frequency it operates at too, and what kind of interference it probably puts out that annoys other electronics (especially digithell TV & radio receivers)...
Hey all. Just to confirm it is indeed a phase shift detector. +5v ground and a floating ground to trigger. unlike the road sensors (that have now been replaced with pole mound dopler units) this does not work on capacitance or like the road sensors that used EM detection aka a metal detector. Dopler detection units or at least the many I have seen don't have a transmitter rod like this lamp. Dopler only needs a finely tuned pcb antenna that does both jobs. As Clive stated they are found in dualtek (as they are known in the trade) alarm sensors and have a range ajustment via a small pot usualy. If the 5.8Ghz is correct that's a little to close to the wifi frequencies and on a board like like that intermodulation and drift may well create a problem as I doubt the build quality is going to be top notch. I am going to order one myself as I have a SDR that will show the entire spectrum. Very interesting lamp and a great video as allways.... P.S. surely the correct value resistor in place of the LDR will solve the light sensing problem, at those power levels who cares if it comes on during the day!
It appears that most of the similar modules don't have the extra antenna stub. I wonder if this had one to change the range. If i was using this lamp for room use I'd probably add a dot of black marker ink on the front of the LDR to make it light at higher lighting levels.
+bigclivedotcom That may be the case, there are 360 degree dualtec detectors that would need to detect just the room area. I have seen at least 10 different brands of alarm ceiling sensors and they all have the waveguide pcb etch but not the vertical antenna. This popular model for example..... g01.a.alicdn.com/kf/HTB1DAesKXXXXXcKXpXXq6xXFXXXy/Wired-PIR-MicroWave-Ceiling-font-b-Sensor-b-font-with-font-b-360-b-font-font.jpg has a single flat pcb with a single etched antenna. I am wondering if the type of sensor pcb you have in the lamp is geared towards presence detection for room lighting, given the LDR on the board I think this is the case. I have yet to see the insides of that type of sensor. BTW ultrasonic sensors are still in use but only employed in very stable area's. I have seen them in safety deposit centres for example.
I wonder if putting nail polish or a bit of paint on the photo cell to cover some of it would make the light come on sooner or in a brighter room where light is wanted.
I did some research a few weeks ago, and it turns out that the 'proper' pronunciation of the metric prefix "giga" is to say it like "jiga". Almost no one actually says it that way, but if you look it up in a dictionary or on wikipedia or wherever, all references are consistent. Doc Brown pronounced it correctly, and we're the ones all saying it wrong. This disturbed me greatly. My whole life is a lie.
Greetings from LA: Just checked eBay for these lamps and found what look like what you "took to bits" (not a side by side comparison) but I do not recall your mentioning these were AC85-265Volt. Your expose showed a SMPS (or was that a BUCK regulator) inside. I'm a retired elect. tech but miss the diagnosing part of the job. Your videos fill that need (including the occasional accidental overloads ;^) Never seen a capacitive dropper circuit here. Jim
Clive said it puts out 2mW. 5.8GHz Wi-Fi can use 50-800mW so I think it blows it away. On the other hand, maybe the bulb could be used to detect network activity. :P
No, it cannot. The bulb is looking for the difference between two signals. The signals themselves just happen to be high frequency. Imagine overtaking someone, you walk 6 km/h and they walk 3 km/h. Now, imagine you were driving instead of walking: one going 120 km/h and the other one going 123 km/h. The duration of the overtake would be roughly the same even though the speed of travel is really fast. Same sort of thing here, the signal difference you are trying to resolve has more to do with how fast *you* are walking towards/away from the light bulb, not the signal emitted. The signal frequency just allows for more accurate spacial and temporal resolution.
Is this not based on a very simple Gunn Diode transmitter? Maplin did a doppler alarm in the 1990s based on the Gunn effect. No oscillator needed. The Diode does it. There is a mixer and a detector circuit. Transmitted and received waves are mixed. The logic compares the transmitted frequency with the received frequency by looking for (AF)beat frequencies in the mixed signal. Google Gunnplexer.
Can you substitute the 'Baby in the bath' drawing for all future stick figures please. That video was your 'Finest Hour' Mate. I think about it every time you pick up a pen.
Hi bigc, consider buying a 3 in 1 universal lens kit for mobile. The macro lens in it is really great and it'd be useful for you (and in effect the viewers) because with that attached to your iPad, you can show us those tiny components really close. It's worth buying for the macro lens, though the other two in that kit are not that great in image quality. :)
Maybe something to do with the beat frequency? That is the original gets mixed with the doppler shifted reflection and you get the difference between the two processed in some way?
That's exactly the way it works. It mixes the received reflected signal with the transmitted signal. If something moves towards or away from it then there will be slightly different frequency reflected, it mixes that with the original transmitted signal giving the small difference in frequencies of just a few Hz, and that is what's detected and turns the light on, and Bob's your uncle.
I sometimes see the automatic doors at supermarkets here in the US still use ultrasonic detectors. I think the industry had since switched to RF, but the old tech is still out there buzzing/ticking away. I don't know if ultrasonic was also used for the application in the UK...
yes somwthing like that. radio waves have a magnetic wave that suts at 90 degrees to the propergation orentation of the radio wave (trying to remember a course I did 44 odd years ago so forgive me if wrong) the Theromin "listened" to the changes in the magnetic field I seem to remember. a bit like a metal detector the field coil also acts as a detector. hold it still and you just get a constant tone. move it and you get a change in tone if something magnetic is close by. Have you noticed a DAB radio if its on a weak signal will get louder or quieter as you move around the room near it? it is possible (just a suggestion) to make a totally passive sensor that works on the frequancies of DAB broadcasts 2 sensors close by would be able to cancil out external effects and detect local movement??
Why couldn't the chip work in the same fashion as it does when hooked up to a PIR sensor? It could simply detect a change in the pattern of returned energy with antenna acting as illumination and an antenna array acting as "eye".
These lamps are nice and handy but the build in power supply has a short livetime. Two of these died after serveral month. You can get new power supply for about 1,-€ on ebay.
P.P.S. A request for you Clive. It would be great if you could get hold of some X10 bits to play with. A great system to have a play with though I have mixed feelings about sendind data over domestic wiring!
you can get a lot of mircowave setups that basically just need to supply power to them and they self oscillate, and as you said, its all based on controlled impedance circuit board design.. but yeah with a single antenna you can actually do rf-mixing in the ocscilator itself, and the output signal frm it is a signal that is a signwave in the Hz range that is the difference between the transmitted and received frequency. and that can be converted pretty easy into a movement speed.. not sure how that device uses it, but with opamps and capaciters could easily setup something that basically is "if Hz > x then trigger" but yeah... power, ground, and signal. Anyhow the signal output is going to be sub kilohertz ... at 5.8 GHz, the output will be about 39 Hz for every m/s of the motion... ie if the person is walking 1 m/s the output will be 39 hz, if 2 m/s then about 77 Hz.
This would be a neat hack with a makeshift house alarm that is very very discrete. If you were to combine this with an arduino with an Ethernet shield, and a loud piezo or speaker, you could have the arduino send out an email notification, and also fire off that speaker. It would be pretty neat.
Clive, I'm wondering if the transistor at the base of the transmit antenna (vertical wire) is an RF amplifier, with the zig-zag track being the receive antenna.
Not strange at all. I was taught back in the 70's that all that "Doppler detection" wasn't how things really worked. I've played with the horn out of a old microwave motion detector and have seen it for myself.
Here's how it works: You have a free running oscillator that transmits a signal. (In the case of my unit, a Gunn diode.) You also have a receiver that picks up the transmitter AND the reflected signals. AM demodulate the signal (in the case of my unit, it was just a simple microwave diode halfway down the horn. The PC board was the filter capacitor.). Then look for low frequency AC. (Ignore the DC value) That's it. Why does that work? The combination of reflected signals will either add or subtract from the signal received from the transmitter. If something moves, that will disturb that "balance" and cause a change in the DC value (aka a AC signal). AC detection can be quite sensitive and find small changes in the DC signal. Outside of two semiconductors on the horn assembly, everything else was power supply or low frequency audio. No narrow-band filters or precision measurements. (Cheap, cheap, cheap.)
I challenge you to find something that really measures the frequency. I doubt if you'll find one, because measuring frequency is complex. The closest you'll come is that you can claim that it's measuring the beat frequency between the transmitted and reflected signals - but that circuit will still be substantially the same as I've described. In that case, the diode detector will be called a "mixer". But you end up with the same results - a AC signal the represents movement.
Russel has it right. The wavelength at 5.8 ghz is 51 millimeters. Half that is a complete phase reversal. So a movement of 25 millimeters, one inch, is enough to completely reverse the phase of the reflection and when mixed with other non-moving reflections in the room will change the amplitude on the receiver. A ceiling fan would probably drive it nuts but it may be filtered to ignore rapid and continuous changes (low pass filter).
Russel you just described a downconverter and doppler phase shift detection! All you need is usually a diode, almost always biased IME, to supply a nonlinearity. I've also commonly seen a transistor used as the mixer in place of the diode.
These modules use a method that was briefly popular in the 80s in the electronic hobby magazine. IIRC, ETI did one, for example. They work like this:
The microwave oscillator runs continuously at a resonant frequency set by that very obvious worm-like track and probably a surface mount capacitor. This oscillator emits a few milliwatts of energy as best it can. The additional stub antenna (that white wire) helps radiate a bit more RF energy.
In the simple version of these things, a simple diode is connected between that resonant circuit (on one side) and the input of a really high gain low frequency (typically audio bandwidth) amplifier on the other. The oscillator therefore not only emits (low level) RF energy into the general area, but therefore also acts as the local oscillator for the diode, just like a "direct conversion receiver".
OK, so nothing much happens if there is no movement, but the instant there is movement, the returning reflected RF signal will now contain the Doppler-shifted oscillator frequency. That happens to mostly be at audio/low frequencies, and they get amplified, filtered and detected, in this case in the BIS0001 sensor. It's a chip specifically designed for this sort of task, but normally used in that fashion with the PIR sensor, amplifying its low level low frequency signals. The old ETI (and other 80s) designs typically used an LM324 quad op amp to boost and filter the diode detected signal.
Nice and neat. And another idea that we've seen before, but now nicely manufactured in vast numbers at ultra cheap prices.
Thanks for the teardown, Clive
Do you know how powerful those are? any interference with WiFi etc.? Health concerns?
@@krecikowi a couple milliwatts at most. If on the same frequency as wifi (these little oscillators with a room as part of the tank drift around like crazy) it would lower your reception a bit.
No health effects
Sometimes I just put a vid of Clive on running on the background. It soothing. Yes its weird..... I know.
you're one of us. we are legion.
asmr
I keep watching his videos hoping that I will one day have an 'ah-ha' moment and electrical circuits will one day make sense to me. I'm a programmer, and I understand how computers work all the way down to the logic gate level. But the logic gates are made of transistors in electrical circuits. And every time I have tried to learn that level... it just doesn't compute. I can tell I am missing something tremendously fundamental, something about the way I am actually trying to think about it is incorrect or incomplete, which prevents anything else from making sense. I can skip over that level and go into how the transistors themselves work, with their NPN layers and migration of 'holes' on the molecular electromagnetic level even. But the circuits... nope. Maybe I'll reach a critical mass of hearing various explanations of various things from Clive and one day it will all 'click'... that's my hope.
I'd like a Big Clive voice for my GPS... "Turn left in a wee bit".
His voice would be good on a Sat Nav until he says "I'm not 100% sure" which he says frequently!
Then you would have to revert to a paper map :-)
Reading the title fast I thought you'd put that lamp into the microwave
lol me too xD
Haha, me too 😂
I have to thank your for kindling my interests in electronics once more. I've had a nerd gun that I was planning to turn into a sort of laser rifle prop with lighting and soundfx since last year, but your videos got me thinking about it again :)
Just came back from the shop with the sound trigger module and the little audio amp for the speaker
Hey thanks for making this videos! I just ordered one of those "radar modules". Maybe I can use them for a home security system.
See, all the magic ends outside the topmost board with an antenna and weird PCB trace. That board contains a simple one-transistor RF oscillator, which doubles as an RF mixer (it has much in common with super-regenerative receiver topology). So it generates and radiates the RF energy, and at the same time is mixing generated signal with the reflected signal. The result contains harmonic with a frequency equal to the DIFFERENCE of the forward and reflected signal frequencies, and this is something about 5 - 10 Hz. This is fed to the second low-speed board.
So it works like this: no motion - no difference tone, something moves - there is a difference tone on the output. The frequency of the difference tone is proportional to the object speed, and amplitude depends on reflected signal strength.
The weird trace on the RF board is most probably an RF microstrip transformer/resonator (and then this is probably a Hartley-type oscillator). A bunch of capacitors is there for supply decoupling.
So no RF on the second board, it deals only with the difference tone.
I could listen the Clive waffling on about electronics for days on end. Such pleasant background noise, and educational to boot. :)
It's always amazing to me how the human mind ( including myself) naturally wants to improvise and use one thing for something else! Would still be living in the dark without it ! 👀
One of my favourite tear downs Clive. Keep up the good work.
Isn't that the way of things? Here I was, the poster child of "ignorance is bliss," gradually figuring out and working with RF stuff, and it turns out I've been practicing the Black Arts the whole time. ;) lol jk
I really liked this tear-down, Clive. You're right about the track layout being important to HF designs. The 'troubles' begin when signal wavelengths begin to coincide with track lengths, in the radio wave to microwave frequencies; radio is 3 Hz to 300 GHz ( λ = 100 km to 1 mm), microwave is 300 MHz to 300 GHz ( λ = 1 m to 1 mm), but I usually say 'beware' around & above the GHz region. Basic components start acting weird, or at least not as anticipated. Those non-ideal values such as ESR, ESL, and indeed, track (and ground plane!) layout, become critical.
Q1 is probably a NPN BJT. The strangely shaped PCB traces make resonant circuits at 5.8 GHz, and putting them at the right spot in the transistor circuit makes an oscillator. Movement changes the tuning of the circuit and the voltage changes a little bit. The other chip just amplifies the crap out of that voltage change and triggers when it gets to a threshold.
The lots of tiny holes are plated thru via holes connecting the outside of the top of the board to ground around the edge, or making the output line bigger.
You can design impedance matched lines on PC board knowing the dielectric constant and thickness, and doing math on it (or looking it up on a nomograph). I've used standard FR-4 up to 5.6 GHz, amplifiers and antennas and whatnot.
I first saw this circuit around 1964 where they used a tube oscillator at around 300 MHz. Around 1971 I built one at 10 GHz using a reflex klystron that I found laying around. Ran it into my guitar amp.
Can confirm that Q1 is indeed an NPN BJT, a high frequency one. Though I don't fully understand radio myself so this didn't bring me closer to understanding how the signal was transmitted, received and mixed.
Think of it as using the whole room as part of the oscillator circuit. A change in the resonance of the room makes a change in the DC current supply in the oscillator. Couple that into an amplifier, trip a comparator, and you've got a burglar alarm.
Thanks for explaining! So could I use the output from the amplifier directly on an MCUs ADC input to find out which way the voltage shifted, up or down, and get the direction of travel of the moving body?
No, all you get is a perturbation. It's a kind of self-mixing where you get a baseband detected signal: near DC. The problem is that your detected signal has no negative or positive shifted frequencies that could tell you whether the doppler shift was positive or negative, approaching or receding. It wraps around the zero (DC) point and both appear positive. You can count the frequency and figure out the speed, however. You can also tell things about the size or distance to your intruder, too, for what that's worth.
James Van Damme If I understand, a moving body causes an oscillation across 0V perhaps due to a phase change or the like from the resonating frequency, however you cannot determine the movement direction. You say speed and distance can be determined, do you mind elaborating? Thank you very much for taking the time to reply.
Ok. I've not read every single comment but I can see that most people are missing the point!
What we want to know is: Are those LEDs warm white or bright white?
Cold white. :(
I saw something similar in a 1970s hobby book of schematics. They called it something like "UHF proximity detector," and it was a 1-transistor UHF oscillator with a little antenna, then other transistors on the DC side of things.
I think they were just measuring the current draw, and amplifying any AC changes. I didn't build one, so I don't know how many meters distant it would still detect motions.
Just because I said last night I was tired of LED light videos you show me the most high tech 5$ light bulb I have ever seen
The microwave sensors have been around since the 90's, they were an optional extra in the car alarms we used to fit at Road Radio.
We used to fit one under the drivers seat, for the car alarm purposes it just used to create a voltage drop to trigger the alarm.
The ultra sonic detectors were pretty much standard, but a window left open a little bit would set them off.
You have taken something apart and have been left with more questions than when you started, I know that feeling lad :)
On a plus note it has now been explained by the RF gurus in the comments here. So now it makes sense. And it's very clever indeed.
This is a possible method of operation that will turn your brain inside out: There may be no oscillator on the module what so ever for the 5.8GHz. I suspect what you have is a couple of narrowband tank circuits (think coil and variable cap in a crystal set), that use far field capacitance (IE a human body) as one of it's finite elements to couple them (or possibly the only element as the polarities are at 90 degrees IE vertical and horizontal). If you charge one to say a couple of volts it will then cause oscillation in the other (and it's self) every time the far field capacitance changes (IE your body moves). You only need miniscule changes in capacitance to activate the circuit as at 5.8GHz even 1pF is only about 27 Ohms impedance. If you are only looking for a micro amp current flow from 2.7V you would only need 1/100,000 pF capacitance change in the far field capacitance. When Idle this would draw a current in the kappo-amps range on FR4 or less on a ceramic substrate if it is clean. The only detector you would need is a Gunn diode or similar (think cats whisker in a crystal set). The KGB used a very similar method to this with a remote exciter to listen in on a bug that nobody figured out for years as it just looked like a metal clip.
This whole thing was making me think of the alarms around about the climax point in the 1992 movie Sneakers, I think those were ultrasonic, but when ya mentioned standing still and posh alarms as well as those ones that use both IR and microwave...
Units I used back in the day fooled me in a similar way. It used separate transmit and receive horns but the receive horn just had an audio amp connected directly to it! The two horns were cast into one piece of metal with a small notch between them. What the notch did was mix a small amount of the transmit signal back into the receiver.
What happened was the reflected signal mixed back into the horn with the sample of the transmitted signal and any frequency difference mixed in the detector and audio came out. Effectively, the receive circuit could only see the difference signal of the heterodyned result. This means the microwave signal was not basically detected at all in the receiver -- only the difference product between the transmitted and received signals. I believe the "detector" was nothing more than a slightly forward biased diode if I recall. This was about 30 years ago.
Bottom line: stupidly simple circuitry that boggled my mind as well. The nice thing is since you're just checking a difference frequency there is no real need for tuned circuits or transmitter frequency stability!!!
I modified one into a "doppler radar gun" which I could use to check speeds although I never got it totally accurate with the circuits I built.
It definitely looks like a single transistor oscillator for the RF generation. What I can't tell is if the output is split, one going to the antenna and the other going to the comparator. I can guess it's using something similar do a hetrodyne radio receiver, but the local oscillator is also the transmitter oscillator. When you mix the local oscillator and the reflected signal together, the remaining signal is in the low hertz range. Even though you're using really high frequency RF, the detected signal isn't.
I'd say it's a simple, free-running heterodyne oscillator where the antenna forms the primary load; the reflected signal is so strong that it causes the oscillator to either 'lock on' to some frequency dictated by the reflection pattern or start fluctuating -- and those fluctuations are detected by the rest of the circuit and used to light the lamp up.
Mora Fermi
If the "load" (person moving) is modulating the main oscillator, how can you tell if they are moving? You don't have a stable reference to compare the reflected signal to. Instead of measuring the pull of the oscillator, you mix the oscillator (who's frequency isn't stable) with the reflected signal. The difference frequency is the movement of the person, and that can be detected with a simple AM conversion.
It is unnecessary to use the original signal as a comparator. What will happen is the room is filled with reflections and they reinforce or cancel to varying degree. At 5.8 gigahertz the wavelength is 51 millimeters so it takes only 25 mm (one inch) to obtain a complete phase reversal and produce a detectable signal strength change.
Any sufficiently advanced technology is indistinguishable from magic. - Clarke
It looks like you stole that lamp out of Rick Deckard's apartment in Bladerunner.
Great summary as always, Clive! If you want more "Radar" based detectors (cheap, seeing as you're a Scot :-) ), Bang Good has heaps for "a good price". I use them in combination with PIR to reduce false triggering in our outdoor lights.
I have to admit, when you showed off what you found my first that was "hahaha, that's AWESOME!"
Got to admit I kinda like finding unexpected things!
No more noisy clicking of the cord pull switch, when you nip to the loo in the middle of the night. Good as a security light for the garage too.
The detector send hf and receive hf but with a little frequency. Then you multiply what is sent by what is received. Then you remember that sin(a) sin(b) = [cos(a - b) - cos(a + b)]/2 so mixing the 2 signals gives you a sum of 2 sinus. One with the frequency offset and one with almost twice the original frequency that you discard with a low pass filter. At the end you end up with a sinusoidal signal with a frequency proportional to the speed of the object in your direction and amplitude corresponding to the size and proximity of the object. This simple signal is easy to detect.
1st sentence, read ... little frequency OFFSET.
Sounds like a pretty good security light: someone comes up to peek in your windows at night while you're away, lights come on.
Unfortunately my neighbourhood is hoaching with cats and I expect the ones that get up on the roof would set it off constantly.
This has been spoken about on other electronics channels of late. I am more convinced that the module just transmits and measures the SWR, uses the BIS0001 to compare the signal. The signal frequency on these units seem to be more around the 3.8Ghz. As pointed out by someone on another channel the single transistor is non linear and can have two voltages present at the same time. The holes are just via's for the ground plane.
The SWR theory makes sense when you find that they don't disturb each other.
The Terminator of sensor lamps?
Draws two cars for illustration purposes,
then sais "A motorbike is a good example because of the noise they make" and forgets about the two cars ...
But seriously Clive, I love your video's :D
Here's a link that discusses the functionality of this device and the type of transistor used (Bipolar High Frequency Transistor):
ua-cam.com/video/jAeFQEHWLZU/v-deo.html
(credit is given there to this video for sparking interest as to how it works)
A potential candidate for the transistor used:
www.nxp.com/files/product/doc/MMBR941.pdf
On second thought at microwave frequencies, the doppler change will be in the audio range instead of sub-audio. My radar lamps at home seem to not go through mirrors. I can wave my hand all over a large mirror on a dresser but move away from it and the light picks up.
It's really not that complicated! it's just an RF oscillator that operates at a high enough level to radiate a useful level of signal. The reflected signal will be picked up on the same antenna, and through non linear action in the same transistor mixes with the oscillation. The beat (the frequency difference due to Doppler) between the transmitted and received signals becomes impressed on the current flowing in the device, and is simply a very low amplitude, low frequency audio signal developed across a resistor in series with the oscillator's collector or emitter. Since it is very similar in amplitude and frequency to the output of a PIR sensor, they've simply used the BISS001 to do the high gain low noise amplifying and level detection. I first saw this very sensitive single transistor RF stage design in a high-end car alarm back in the 80's and thought it very clever then. That would detect movement outside the car and give you a verbal warning to back off!
Those alarms were great fun, see how many times you could get the warning before the battery died :op
I was just about to say that. With a two inch wavelength the beat not going to be much over a couple of Hz. I wonder how two of them work in close proximity. Would they lock together or just stay on all the time?
Roughly what I was thinking, Doppler shifted echo mixed with the transmit CW signal produces a low frequency signal that the PIR chip can detect.
They've been around awhile. en.wikipedia.org/wiki/Proximity_fuze
That makes sense. Yes I remember one of those car alarms actually telling me and everyone else on the pavement who walked past it to "back off". Very irritating.
What is really crazy, even with this bizarre technology, this microwave/Doppler lamp at 5,43 cost from eBay is less expensive than a standard LED lamp from a hardware store or department store in Canada.
Yes! Finally you got one.
Don't you just love the cheesegrater heatsink fins?
Mine got more and more erratic over time, pretending not to notice me sometimes. I have taken it down to try and fix it but haven't gotten around to it.
I don't think this works with the doppler effect. I think this is just a single transistor oscillator and it detects changes in overall reflection/absorbtion in the room by monitoring the supply current to the oscillator. The PIR IC is able to detect very tiny changes.
7Y MMBR941L Mot SOT23 npn RF fT 8GHz MRF941 So just a NPN transistor.
I've even taken this into the radiated EMC room and it makes a spectrum of narrow lines at around 3Ghz or something.
The other "radar" detectors on Ebay (with the white dome case) do really use the doppler effect. You can tap into the signal and make a speed detector, but I think this one is different. It is strange because it really was so incredibly sensitive.
You should make a camera system that hooks into this lightbulb and the activators are in parralel. so you could have lightbulbs in seperate ends of a building and hook up cameras to go with it. If nothing happens, cameras save power and storage. If motion is detected, camera activates and catches movement. Could be your pet or could be a thief. Who knows.
Some microwave doppler modules use a mixer to mix the transmit output with the received reflection... with a static target the output is a dc level... movement of the target produces a beat frequency which is dependent on speed of relative movement....
I believe that transistor is a high speed JFET used to amplify the transmitter signal. If not then it is the oscillator at which point I would expect the capacitor to be of very small value. Either way, doppler detection modules are just frequency comparators used to trigger other functions.
Standard operation for cheap microwave radar modules.
Rf is generated by a Dro, dieelectric resonance oscillator, just a fet oscillator with a ceramic 'puck' as the tuned element.
Rf is generated then split 2 ways, one goes to the transmit aerial and gets radiated out into the room, the other goes to a mixer, this mixer mixes this signal with the received signal from the receive antenna.
If an object is stationary theres no diffrence in frequency between the transmitted signal and the received signal, as soon as an object in range moves theres is a difference in freq, this freq difference depnds on the speed of the moving object, because people move very slowly compared to the speed of light the output freq of the mixer is only a few Hz.
The 'voodoo' tracks on the board are microstrips, and these are often used for filtering at microwave frequencies.
And it is also presumably capable of detecting anybody flying an R/C drone or a WiFi signal in the area! (we pilots often use 5.8GHz analog video feeds for our FPV video singal in our drones.)
I wonder if it's triggered by rain or passing cars
if theres rain and passing cars inside your living room, you have bigger problems than a light occasionally coming on. ;)
The author reports passing cars can trigger it; probably not rain as many tiny signals will tend to cancel each other out.
Jim Steinbrecher: LOL LOL LOL That was too good !!!
It's a homodyne transciever, basically measuring the SWR of the antenna as the people (or other stuff) move. They used the same idea at 915 MHz back in the 70's for security systems.
Possably a Gunn diode (TED) oscillator connected to load detecting circuit.
I would go with that, I used to buy Doppler modules from RS to make alarms for long coaches, they contained a Gunn diode and a receive antenna in the same "wave guide" a small portion of the signal was sent straight to the receive antenna inside the module to act as a reference, the rest sent out, the "bounced" back signal (off of a moving object) is then added/subtracted from the original and these changes were used to trigger the alarm circuit.
I'm thinking what is going on here might be as simple as measuring the power consumed by a microwave oscillator that's coupled to an antenna. One of the ebay listings linked in another comment has a decent picture of what appears to be a very similar design. Looks like the chip on there is most likely a single microwave transistor. One pin is connected to power with a bunch of bypass capacitors. Another pin seems to be set up for biasing the transistor. Then the output is connected to that long curved trace which acts as the antenna. Could be set up as a slot antenna with the hole in the ground plane on the back side. The far end of that trace appears to be connected to ground with some capacitors and a resistor. The capacitors provide the RF return while the resistor provides the DC return. Looks like the voltage across the resistor gets picked off and routed to an output pin. My guess is the BISS0001 chip is just being used to detect changes in that output signal. The idea is that the antenna acts as both a transmit and a receive antenna at the same time. Since it's part of the feedback network for the transistor, the transmitted signal that gets bounced around in the environment with interfere with the operation of the oscillator, changing the effective load impedance and therefore the performance of the oscillator (frequency and/or amplitude). Changing frequency and amplitude will change the power consumption of the oscillator, which is relatively easy to measure. Now, what would be interesting would be to set up a spectrum analyzer to see what happens to the oscillator frequency. Might also be interesting to see how susceptible it is to external interference - i.e. 5 GHz band wi-fi.
Ok, Now Hack that into a security system, and then I'll be very VERY impressed.
the chip 1am - straight away thought of your song where you can't sleep watching youtube
If you have ever played with regenerative receivers, you know that they can interfere with other receivers, since they transmit as well as receive. Since they are only a weak oscillator, they can be easily disturbed by stray capacitance and conductors placed near them. I think this is how this circuit works. Think of a metal detector that detects anything that conducts. A grid dip meter is similar a tuned circuit upsets it's grid bias and shows a "grid dip" when the meter is tuned to the circuit under test.
Finally someone who also think that Gigahertz pronounced as Jigahertz sounds much better!
actually the frequency shift is in the nanometers for RF. what they actually do is to mix it in with the original frequency and you get a very low, easy to sample frequency.
These do use the doppler method as you say, but they don't do clever frequency filterring or anything, if something is moving, there is the incident rradiation, and the reflected frequency sets up a beat frequency with it, so as you move, you get an audible beat that is higher pitched the faster you are moving.
Way more simple than you were even saying. Presumably that PIR chip would detect even small amplitude beats and trigger.
It's probably using a gun diode eather in a wresonant cavity or a tuned bit of trackwork!
I'd bet it has one set of tx/rx antennas, and the beat is detected from it's current draw.
you really need to send that one to dave jones for a two minute teardown. i'd like to see what he thinks of it. i'm sure it will be totally obvious caveman level stuff to him.
AFAIK Dave is no RF "black magic"-person. I'd think Mikeselectricstuff has much more of an idea of how it works. Another option is to have W2AEW look at it, he's a genius with RF-circuits.
The Signlent is not ideal, too low frequency for the fundamental in this application.
wow looks like i hit a nerve i didn't even know existed.
I add the Signal Path Blog for consideration. I would put my money on Shahriar if I needed someone to quickly figure out the workings of an RF circuit.
Some of the commentators to this video look to have worked out everything or almost everything based upon the video alone if you are interested in the workings
AAAgh why do i love these kind of video's so much just opening and searching and well just looking ;D
We call this a Hyperfrequency detector in France as in alarm systems ;)
"Radio circuitry is just a black art" so true
Clive the 7Y transistor seems to cross references as an MMBR941LT3 NPN high frequency transistor, able to operate up to 8Ghz which would seem to go along with the application. Here's its datasheet..
pdf.datasheetcatalog.com/datasheet/motorola/MMBR941BLT1.pdf
"Neeeeeeeeeeeyooooooooooooooo"
- Clive, 2016
Sam Delaney lol
Also as the car approaches you it looks a little bluer, and as it goes away it looks a little redder.
My grandma's terrified to use a microwave thinking it's going to cook her organs or something, so this thing would scare the hell out of her.
Perfect for in the living room when everyone is asleep. Movement outside the house it turn on
I'm pretty sure (could be wrong) the piece of wire is the transmitter and the wavy track is the receiver. The weather radar receiving antenna on aircraft are built in s similar manner but on a larger scale. they also transmit the signal at the same time witch it witchery to me but all i need to know is how to test it and replace what needs to be. :)
*Off Topic Mystery:* Clive, you seem to be a man who has come across a lot of weirdness in his time -- and I am not referring to Alex Salmond's past girl friends here -- and therefore you may have noticed this effect or heard about it some other way.
You know those self gluing envelopes they sell in the post office? The ones you just have to press together and they are supposed to stick securely (Do they buggery!). Well, I had occasion to pull one open and I did so by pulling on the flaps. I happened to be standing in a darkish room at the time and to my astonishment I noticed what appeared to be tiny blue sparks appearing along the glue line. I pushed the two parts of the flap back together to see if I could repeat the effect and as the glue came apart, there were the sparks again. They looked like the kind of sparks you get from a low voltage current jump between a hot wire and a metal surface.
Ever seen it or know why it is happening?
PS. I thought it may be some kind of static discharge but I was amazed to see it happening on glue.
That's an interesting effect. Then again, dry glue has to work in some way after all, and relying on electrostatics (on a molecular level, anyway) does not sound so outlandish. Long molecules giving off charge as they are stretched are not unheard of either. A (bio)chemist should have more to say on this, I imagine.
yeoldeengineer
It was startling to see it. I think you could be right though.
Scotch tape does that as well. Static charge generated by separating two surfaces. Some time ago someone detected x-rays generated by unraveling scotch tape. AFAIR this effect was published in Phys.org.
Milan Trcka
Thanks.
I wish you could show close ups sometimes. Even in HD it's hard to see the little circuit boards you're showing
I wonder if water thru PVC pipe or ceiling fans will trigger this one, that is a problem with RADAR sensors, some have algorithms to ignore those things, pet immunity under a certain weight etc... Of course a dual tech PIR and RADAR like security sensors would be false free. You may want to partially obscure the cds cell so it functions with some ambient light.
i could hear Clives gears turning in his head when he said the little module could be use for other things... :)
It could simply be a Tunnei diode based oscillator. Very common and cheap to build too.
Thank you for the headphones warning!
last time I came this early ,9 months later you were born
first of all that jokes old second of all you look like your 10
An intriguing little device indeed, makes me wonder what frequency it operates at too, and what kind of interference it probably puts out that annoys other electronics (especially digithell TV & radio receivers)...
5.8GHz at extremely low power of about 2mW. So not going to do much harm to anything.
I say ol' chap, what's your fascination with lights? What am I missing?
Hey all. Just to confirm it is indeed a phase shift detector. +5v ground and a floating ground to trigger. unlike the road sensors (that have now been replaced with pole mound dopler units) this does not work on capacitance or like the road sensors that used EM detection aka a metal detector. Dopler detection units or at least the many I have seen don't have a transmitter rod like this lamp. Dopler only needs a finely tuned pcb antenna that does both jobs. As Clive stated they are found in dualtek (as they are known in the trade) alarm sensors and have a range ajustment via a small pot usualy. If the 5.8Ghz is correct that's a little to close to the wifi frequencies and on a board like like that intermodulation and drift may well create a problem as I doubt the build quality is going to be top notch. I am going to order one myself as I have a SDR that will show the entire spectrum. Very interesting lamp and a great video as allways.... P.S. surely the correct value resistor in place of the LDR will solve the light sensing problem, at those power levels who cares if it comes on during the day!
It appears that most of the similar modules don't have the extra antenna stub. I wonder if this had one to change the range. If i was using this lamp for room use I'd probably add a dot of black marker ink on the front of the LDR to make it light at higher lighting levels.
+bigclivedotcom That may be the case, there are 360 degree dualtec detectors that would need to detect just the room area. I have seen at least 10 different brands of alarm ceiling sensors and they all have the waveguide pcb etch but not the vertical antenna. This popular model for example..... g01.a.alicdn.com/kf/HTB1DAesKXXXXXcKXpXXq6xXFXXXy/Wired-PIR-MicroWave-Ceiling-font-b-Sensor-b-font-with-font-b-360-b-font-font.jpg has a single flat pcb with a single etched antenna. I am wondering if the type of sensor pcb you have in the lamp is geared towards presence detection for room lighting, given the LDR on the board I think this is the case. I have yet to see the insides of that type of sensor. BTW ultrasonic sensors are still in use but only employed in very stable area's. I have seen them in safety deposit centres for example.
I wonder if putting nail polish or a bit of paint on the photo cell to cover some of it would make the light come on sooner or in a brighter room where light is wanted.
I did some research a few weeks ago, and it turns out that the 'proper' pronunciation of the metric prefix "giga" is to say it like "jiga". Almost no one actually says it that way, but if you look it up in a dictionary or on wikipedia or wherever, all references are consistent. Doc Brown pronounced it correctly, and we're the ones all saying it wrong. This disturbed me greatly. My whole life is a lie.
Well jiga is the french way of pronouncing Giga which is fair enough considering the SI was created in France.
Oxford Dictionary says it is pronounced /ˈɡɪɡə/, while Wikipedia and Wiktionary show both /ˈɡɪɡə/ and /ˈdʒɪɡə/
this is like when I thought back to Spanish class and realized that Arnold schwarzenegger is the one pronouncing California correctly.
Oh good. Because Jiga sounds better.
Greetings from LA: Just checked eBay for these lamps and found what look like what you "took to bits" (not a side by side comparison) but I do not recall your mentioning these were AC85-265Volt. Your expose showed a SMPS (or was that a BUCK regulator) inside. I'm a retired elect. tech but miss the diagnosing part of the job. Your videos fill that need (including the occasional accidental overloads ;^) Never seen a capacitive dropper circuit here. Jim
I'll bet that thing plays havoc with 5.8 GHz Wifi.
Clive said it puts out 2mW. 5.8GHz Wi-Fi can use 50-800mW so I think it blows it away.
On the other hand, maybe the bulb could be used to detect network activity. :P
No, it cannot. The bulb is looking for the difference between two signals. The signals themselves just happen to be high frequency. Imagine overtaking someone, you walk 6 km/h and they walk 3 km/h. Now, imagine you were driving instead of walking: one going 120 km/h and the other one going 123 km/h. The duration of the overtake would be roughly the same even though the speed of travel is really fast. Same sort of thing here, the signal difference you are trying to resolve has more to do with how fast *you* are walking towards/away from the light bulb, not the signal emitted. The signal frequency just allows for more accurate spacial and temporal resolution.
Klikkitse
Exactly. It's such a small signal, and it's only being used to detect changes in the area around it.
There's a bunch of similar breakout modules that are popping up lately; being marketed as 'radar modules'.
Might be worth getting a few to hack up?
I wonder if the oscillator is dependant on the feedback of the reflected pulse, it has to have reflected feedback to keep it running.
Is this not based on a very simple Gunn Diode transmitter? Maplin did a doppler alarm in the 1990s based on the Gunn effect. No oscillator needed. The Diode does it. There is a mixer and a detector circuit. Transmitted and received waves are mixed. The logic compares the transmitted frequency with the received frequency by looking for (AF)beat frequencies in the mixed signal. Google Gunnplexer.
Not weird I do it too, but also listen to him, clever fellow.!!!
Can you substitute the 'Baby in the bath' drawing for all future stick figures please. That video was your 'Finest Hour' Mate.
I think about it every time you pick up a pen.
Hi bigc, consider buying a 3 in 1 universal lens kit for mobile. The macro lens in it is really great and it'd be useful for you (and in effect the viewers) because with that attached to your iPad, you can show us those tiny components really close. It's worth buying for the macro lens, though the other two in that kit are not that great in image quality. :)
Maybe something to do with the beat frequency? That is the original gets mixed with the doppler shifted reflection and you get the difference between the two processed in some way?
That's exactly the way it works. It mixes the received reflected signal with the transmitted signal. If something moves towards or away from it then there will be slightly different frequency reflected, it mixes that with the original transmitted signal giving the small difference in frequencies of just a few Hz, and that is what's detected and turns the light on, and Bob's your uncle.
I am an RF tech and wish I was here to comment when this came out. Nobody is around now for it to have much worth
+Brandon Cornett There will be more of the tiny RF sensors featured in future videos.
bigclivedotcom great. Awesome and ever change because your content is just perfect
RF magical angry pixies
That noise makes me want to chew my fingers off
I sometimes see the automatic doors at supermarkets here in the US still use ultrasonic detectors. I think the industry had since switched to RF, but the old tech is still out there buzzing/ticking away. I don't know if ultrasonic was also used for the application in the UK...
yes somwthing like that. radio waves have a magnetic wave that suts at 90 degrees to the propergation orentation of the radio wave (trying to remember a course I did 44 odd years ago so forgive me if wrong) the Theromin "listened" to the changes in the magnetic field I seem to remember. a bit like a metal detector the field coil also acts as a detector. hold it still and you just get a constant tone. move it and you get a change in tone if something magnetic is close by. Have you noticed a DAB radio if its on a weak signal will get louder or quieter as you move around the room near it?
it is possible (just a suggestion) to make a totally passive sensor that works on the frequancies of DAB broadcasts 2 sensors close by would be able to cancil out external effects and detect local movement??
Isn't it more practical to make inter-connector between lamp and socket with this detector?
Why couldn't the chip work in the same fashion as it does when hooked up to a PIR sensor? It could simply detect a change in the pattern of returned energy with antenna acting as illumination and an antenna array acting as "eye".
i cant wait for the results for ''the supergayrainbow exploding powersupply'' thing.
These lamps are nice and handy but the build in power supply has a short livetime. Two of these died after serveral month. You can get new power supply for about 1,-€ on ebay.
P.P.S. A request for you Clive. It would be great if you could get hold of some X10 bits to play with. A great system to have a play with though I have mixed feelings about sendind data over domestic wiring!
you can get a lot of mircowave setups that basically just need to supply power to them and they self oscillate, and as you said, its all based on controlled impedance circuit board design.. but yeah with a single antenna you can actually do rf-mixing in the ocscilator itself, and the output signal frm it is a signal that is a signwave in the Hz range that is the difference between the transmitted and received frequency. and that can be converted pretty easy into a movement speed.. not sure how that device uses it, but with opamps and capaciters could easily setup something that basically is "if Hz > x then trigger" but yeah... power, ground, and signal. Anyhow the signal output is going to be sub kilohertz ... at 5.8 GHz, the output will be about 39 Hz for every m/s of the motion... ie if the person is walking 1 m/s the output will be 39 hz, if 2 m/s then about 77 Hz.
Those cars you drew must be a Lada or some other Russian car.
A volvo 240 does make for a great radar target ;)
As for desoldering those pins, I wonder if the modded yihua tips can do it. Mine can solder onto 2 mm thick sheet metal
I like it better when you say "millamps".
I wonder if it interferes with 5.8Ghz wifi/wireless stuff.
Would a bulb like this make a Easy Bake oven into a microwave oven?? Now THAT would be something...
At 2mW output power it would not really make a good job of baking cakes.
So you would need A LOT of them to make it effective. I'm not hearing a "no" here!
;-)
You could probably get a tiny magnetron and stuff it and an inverter into a dead LED lamp housing. :P
This would be a neat hack with a makeshift house alarm that is very very discrete. If you were to combine this with an arduino with an Ethernet shield, and a loud piezo or speaker, you could have the arduino send out an email notification, and also fire off that speaker. It would be pretty neat.
Clive, I'm wondering if the transistor at the base of the transmit antenna (vertical wire) is an RF amplifier, with the zig-zag track being the receive antenna.
Fun fact: Light also has doppler shifts. Though your eyes wont detect it.
Maybe there are a gun diode / pin diode like on radar oscillators.