DIY sonar scanner (practical experiments)

[@stevejohnson1685]

I worked for a medical ultrasound company in the late 1980s, where we did phased array beam steering with very elaborate FPGAs, DSPs, and only the very highest level processing with general-purpose microprocessors. Staff was dozens and dozens of very talented electrical engineers and programmers. What you've done by yourself with just a contemporary microcontroller is extremely impressive!

[@MrScooter46290]

Great comment. This is incredible.

[@mphRagnarok]

Well he's standing on the shoulders of hundreds of thousands or millions of engineers and corporations that made all the ready to use tools he's using...

[@tessierdsl]

​@@mphRagnarokWell, you use the same engineers to write condescending comments... 🤷‍♂️

[@lokeshchandak3660]

​@@tessierdslIt does not read as condescending to me.

[@Out_of_order]

You should make some videos about your experience what you have done in the past

[@JohnBysinger]

I don't comment often, but man, after two decades working in cellular wireless data network engineering, this is the best phased array explanation I've ever seen. The visualization you did, the solenoids in water, the pick and place plotting, and the overlay at the end each by themselves were awesome. But showing all three was just amazing. I've tried to explain this to peers on numerous occasions and it's a difficult concept to grasp without seeing it. Very well done.

[@KhaladFaraday]

Yes.

[@sentiparole1165]

​@@KhaladFaradayIndeed

[@danoneill8751]

Funny thing: having only ever seen this one explanation, my reaction to your comment is: nah, you're clearly not so bright, its a very simple idea. I feel robbed of the experience of reading about it in a physics or engineering textbook, being flummoxed and then finding this superb, absolutely excellent video.

[@JohnBysinger]

@@danoneill8751 I've been working with it in practical application long before it made it into any textbooks. Frankly if it is in them now, I'm impressed, usually engineering texts lag a good decade or two behind current tech. We were testing 5G phased array hardware in 2014-15 and planning for it around 2010. So yes, a video like this would have been immensely useful a decade ago. Oh and often the simplest of ideas can be both completely non-obvious and even difficult to explain to an audience that's never been exposed to it.

[@ryaneGT]

@@danoneill8751 such an interesting point! I have also spent a lot of academic and professional effort as an engineer trying to understand phased arrays and agreed this video does it better than anything I've encountered so far! Would love a better explanation from a signal isolation/receiving perspective as much as the beam forming.

[@AppliedScience]

Really really nice work!

[@joseparedesalbuja8293]

You are like twin brothers. Thank you for the many amazing vídeos. Greetings from Ecuador

[@Timsturbs]

@@joseparedesalbuja8293 Escuchame!!!

[@PilotPlater]

I knew I'd see applied science in the comments!

[@Scrogan]

Inb4 phased X-ray array

[@Reth_Hard]

Yeah! I already watched a video about Starlink's phased array but this one really made me understand how it works!

[@quetzalcoatl-pl]

As a professional programmer, I really love your 'coding' intermissions xD the one with drill and glove was hilarious!

[@pskcalvo]

After almost 40 years coding I found my lifetime goal... I want to type like him !! 🤣

[@markgoogolplex2572]

I've tried all of them but much prefer the roller pin method. It seems slightly faster than the glove-drill with less strain on the hands.

[@Sylfa]

I'm just glad he didn't resort to cookie cutters, that would really take things too far…

[@Albtraum_TDDC]

@@pskcalvo chatGPT got you covered now.

[@denizcancgsar2810]

Roller is my favorite.. xD

[@antonmilosevic]

I have to say it. THIS IS YOUR BEST VIDEO EVER. I was kinda getting sad that you didn't upload that frequently anymore, but the wait was well worth it! Hut ab mein Freund :)

[@magge7910]

Can only underline that. So good. Well done.

[@Francois_Dupont]

i agree

[@mayankraichura]

Totally agree 👍🏻

[@Mr_Ponzo]

Me too

[@maicod]

I guess the low frequency lies in the fact there's so much pcb waiting time and video segments filming and testing etc. etc. time needed

[@janlinzmaier2930]

Finally someone with a local pcb manufacturer as sponsor instead of a chinese one.

[@Maxjoker98]

Wow, you visualizations are top-notch! The best visualization I've seen on the subject, and mostly practical too!

[@kennethhicks2113]

I agree, was introduced to pesa in 80's and the water demo is the best I've ever seem to demo.

[@subminuentisch]

I never comment on KZbin but that reminded me of a little visualizationproject i made a while back i never published it though. When i saw 3:13 i thought how the hell did he get his hands on my private videos. If you are interested i just unprivated all the related videos i have. Includes PAR dipole and singlepoles and some other stuff.

[@subminuentisch]

Also when i realized back then that em waves pretty much behave like water waves a lot of things became way more intuitive so that water experiment was pretty good

[@sumduma55]

@@subminuentisch back in the 90s, I was experimenting with much similar techniques but my goal was to isolate sounds within a crowd of noises. Of course I was severely limited on processing power and it didn't go far outsideof some limited success with tones. I hear that in the mid 2000s, somebody created an algorithm that was way more efficient. Anyways, I'm pretty sure the same effects can work backwards where an aray of sensors can be used to pick the same wave anomaly from natural sounds or even a focused beam like this returning. I used recordings to vary the timing in playback of three to six microphones placed a few millimeters apart to isolate waves from specific targets. I think this can be done in real time today to discriminate the sonar response to just the target area increasing the resolution somewhat significantly. I haven't played in this arena in a long time. My memory could be rusty on it. I'm just tossing it out there in case anyone else might be interested in taking things further.

[@jamesbaker7112]

I helped remove the sonar transducers from USS Fox CG 39 during an overhaul in '89. It took two men to lift them. They were covered in sharp cooling fins and there was no good way to hold them. We had to carry them up six decks of ladders, the length of the ship to the quarterdeck and down the gangplank to a truck.

[@thethoughtemporium]

That was absolutely outstanding. Incredibly good job! Very cool project

[@mrbane2000]

You two guys have so much knowledge and put up effort, i subbed you both. This guy put effort even in analyzing beam and having simulation. Props to you

[@djdurban3294]

If you have taught yourself about phased arrays and gone on to develop this experiment without help from anyone else then I truly admire you. This is brilliant work. 👏🏽

[@bardenegri21]

Holy production quality batman! Good job and it was real fun watching you toil over this project on livestream.

[@ted_van_loon]

now just wait untill he makes a headband with frequency controllable vibration actuators so he can feel what the scanner scans. so he can literally see like badman(or feel), and even see/feel behind objects.

[@paulatreides6779]

Your effort in making this, from the device design itself to the video filming/editing is OUTSTANDING!

[@muniswamy100]

bro, you nailed it, synthetic aperture sonar theory and working in under 15 mins. great apprecciation for all the deligence put for making this amazing video.

[@TheRationalPi]

Did I miss the part where he used a synthetic aperture? He certainly *could* have done it with CNC, but I don't think he did.

[@muniswamy100]

@@TheRationalPi, well, he wasn't specific about synthetic aperture, but i felt the video basically conceptualized it. However, for Synthetic aperture, the array needs to be moved.

[@TheRationalPi]

@@muniswamy100 I think he stayed pretty specifically on phased arrays. It would be really cool to see him go back and do some synthetic aperture stuff with that CNC though. He could hook the rig from the end of the video to it and get some really great resolution out of it!

[@muniswamy100]

@@TheRationalPi i'd love to see that. Above all, video was highly educative..

[@Nets-nutsBr]

13:48 - "Let me display the sonar as an overlay": Blew my mind. I understood how it works. Tks

[@MrTegidTathal]

Very cool! You can actually use alternating scan rates to see further than your sample rate. Assuming you are scanning/listening constantly (so you transmit a pulse, wait to receive for t time then transmit again exactly when that time is over) you are actually seeing distant items in your near field as their echoes are still returning to your receiver (2t period). The only time you cannot hear them is during the transmit period. If you alter the scan rate, you can cohere them together (only the coherent signals will add up) over time and extend your physical scan distance. The different transmit rates are effectively a distance phase.

[@amazing763]

Same technology used at the lab where I worked 50 years ago to steer seismic arrays, spread out over hundreds of miles, to locate earthquakes. Obviously can't pick up seismometers and move them about, but phased array math does it just fine. Good collaboration of geophysicists and electrical engineers.

[@bobrowland6936]

Great demonstration! You did an outstanding job using water waves to show wave interference patterns. Back in 1985 I was using ultrasonic transducers similar to the ones you are using to create a binocular sonar system with a single transmitter and two receivers. Unfortunately, I did not realize how large the amplitude of the sonar signal was and over a single weekend I damaged my hearing. That was a real bummer and I stopped working with ultrasonic transducers. But about ten years ago I started working again on a binocular sonar system for my robot. This time my transmit level is much lower. It took me a while to create the algorithms to correlate the two echoes, but after a couple years I now have a working binocular system which can map object ( X & Y) six feet in front of the robot within a +/-15° cone. The accuracy is better than I expected, about +/-3.5cm.

[@markb1911]

This just blew my mind. Everything about it. The Presentation, examples, explanation, shots, editing, and method used to impart understanding on this topic are just… It’s just art at this level. also… The rolling pin got me.

[@pronouncedoiler7193]

I'm a medical ultrasound researcher, and this is a really good primer on beamformation theory without digging into the math. Though in medical ultrasound "phased array" means an array with at most half wavelength spacing, I understand radar guys are a bit more liberal with the usage. Still, it's important to know that this is just the tip of the iceberg. This is just fixed transmit steering for narrowband signals in the far field with fixed weights. Dynamic nearfield focusing of broadband signals in both transmission and reception is standard operating procedure these days, and there is a ton of nuance in optimizing the parameters. Also, your main sidelobe bottleneck here is your grating lobes, which are the result of the transducer spacing. Those cannot be easily tuned on the fly, and thus are critical hardware considerations for the array designer. The other side lobes can be tuned by adjusting the weights before summation, which we US folks call apodization. Still this is pretty impressive for doing this on the cheap side of things. Would love to build something like this myself.

[@tymofiichashurin1602]

This is precisely the density of information, which nearly every course should have. It is a lot of work has been done and that is why it is such a cool and concise explanation. Thank you for that! :)

[@HennerZeller]

Really well done, in particular the effort that went into explaining phase arrays on a level better than most science communicators would do. Happy beamforming!

[@robiniuteee]

13 years in telco career, this was the first time actually and visually it got clear what happens in the real world. I practically was on my toe how fast i can give a compliment after watching 7 mins of the video. Great job, really appreciate

[@kierancampbell3322]

Apart from being a great explanation of phased arrays, the production quality is awesome in its own right. The visuals, including overlaying the dolls, are so good! And I too enjoyed the coding 'methodologies' :) And the reversing skit!

[@DaveTheYellowDart]

Dude, nice job! So cool. About 15 years ago or so I built a very crude audible beamformer to play with for fun. I just used an Arduino and 10 piezo transducers that have a resonant frequency around 3 kHz IIRC, held in a laser-cut acrylic frame. I would read in the value from a pot and convert that into a desired azimuth angle, and then calculate the appropriate offset between pulses. It was a good simple project because I could drive the transducers directly from the GPIOs of the Arduino so the circuit was simple, just an Arduino on a breadboard, a pot, and then the 10 piezo transducers. (I also drove the transducers differentially, so each transducer would connect to 2 GPIOs, and I would set them to 0,1 to drive them forward and 1,0 to drive them back, doubling the amplitude compared to just on/off.) It wasn't particularly practical, it just made this annoying drone sound, but you could definitely tell that it was directional, by moving your head, or sweeping the signal. Not bad for an 8-bit processor running at a few MHz. Really awesome video, well done.

[@peterbonham5540]

As an aircraft avoincs apprentice in the 70's I worked on weather radar systems that had multi KW magnetrons, physical wave guides and mechanical dishes that swept an arc. At that time the new tech stuff was coming in with phased slot arrays, gunn diodes and digital signal processing. Absolutely mind blowing shift in tech. Got a lot of redundant really strong magnets out of it though.

[@robwilton9539]

The Sonar 195M wasn't new then Pete, it was invented in 1935 if my memory serves me correctly. It was probably the first phased array device ever.

[@markTheWoodlands]

Excellent combination of theory and practice. This is the best introductory tutorial on beamforming that I have seen to date.

[@rammss]

I'm amazed by your roller coding skills xD

[@CarlMahnke]

The work you put into this is insane.

[@Ry_TSG]

this is legitimately the coolest DIY project I've ever seen. I've seen the videos of the little Arduino sonar projects that rotate back and forth like old fashioned radars, but to see a demonstration of a modern phased array radar using the same method is absolutely amazing.

[@sam42j]

The amount of work that went into this video... Just wow.

[@MattAtHazmat]

Long time ago, I tried to build something similar- the biggest issue with doing a phased array like this is the sidelobes. You really need some small transducers- if memory serves (without my handy copy of Skolnik's Radar handbook nearby), best spacing is about 0.7 of the wavelength. without that kind of spacing, the sidelobes (things outside the 60 degree sweep) have nearly as big a return as the objects within the sweep. Your simulation seems to show this? What may get you even better resolution is a Synthetic aperture setup (SAR, or sidescan sonar) Cool thing about SAR is that you can do it with an FFT. At 40KHz, you need about 5mm spacing. Way smaller than any transducers I could source at the time (25ish years ago). Mind you, I did this with a PIC16C71 (UV Eraseable!) that has 288 BYTES of RAM and an 8 bit A/D converter, and all in assembly! You can do a variant with a "zig-zag" of the transducers so at least in the far field, the spacing is pretty close and you can reduce the sidelobes. Worked on it for a bit but... LOOK A SQUIRREL!

[@firstmkb]

Amazing work! I am burdened by so many herds of SHINY squirrels, so I have never gotten that far. This does reawaken my desire for an open source sonar for pleasure craft. I don’t want to spend $1,000 for a commercial system OR run aground!

[@upendrahatiya5686]

well side lobes can be reduced by using amplitude control as well with phase control in a binomial distribution. this way one can significantly reduce side lobes but with a cost of directvity.

[@MattAtHazmat]

@@upendrahatiya5686 Huh? modulating the phase is exactly what is happening here- and the directivity is reduced exactly because of the sidelobes. Directivity is exactly the parameter that this setup lacks. I believe what you're talking about is the cross range apodization in a focusing sense. Still, you need to get the individual elements much closer together- shorter than a wavelength.

[@MattAtHazmat]

@@firstmkb thanks. A side-scan sonar is *relatively* straightforward math. You make a 2D array of your data- for example columns are individual readings, increasing in time as you (radar speak: "range bins"). Take time domain measurements at a regular (distance/cross-range) interval. Do an FFT across the rows ( the same "range" bins, but differing in cross range): BAM. the output of your FFT is your synthetic aperture at that range. It took me a while to catch on, with the help of some very smart radar engineers, but the "focusing" is done with well established algorithms.

[@jtjames79]

@@firstmkb I remember reading an old story about how maybe seals use an array of electrodes on their tongues connected to sonar. Back then the hardware was ridiculously expensive (particularly processing power), bulky, and power hungry. I want to use this technology to create "sonar" for the blind. So I would also mix in radar, IR, visual spectrum, etc. So you could switch around like predator vision. So even non-disabled could use it like let's say an electrician, who could then walk into a room and sense any shorts. Or IR so fireman can sense the strength of flames that they can't see. Etc.

[@dsuess]

Mind blown!! This is truely amazing to see how you broke down a complex topic into something more simple. You sir, are an amazing engineer. Thank you for sharing!

[@taterbits]

This is an awesomely made and put together (edited) video. Very well done mate. Two thumbs up!

[@Juzujka]

Very nice work! I guess, high level of side lobes is due to - Wide spectrum of signal. If you use short pulse with sharp edges then you have wide spectrum. Different parts of spectrum gives different lobes. Band filter would help. - Different power level transmitted by individual transducers. Calibration would help.

[@PabloMayrgundter]

Great work! An idea: since you have a closed loop testing rig, maybe you could tune the emitters with a matrix of weights. For each sensing location record the decibel level, with perhaps a bit of modulation of each emitter's output level. let's say you expect to receive X dB at that point, and you vary emitter A +- 5 dB, but only record -4.8 to +4.7, you can then correct the power output on each to reach the desired output. Repeat this for each emitter, so you have a vector of tuning weights for the target location. Repeat that yet again for each target location (could be 1, 2 or 3 dimensions). So you end up with a calibration matrix and can use it to precisely tune the emitted waveforms to have more consistent power densities where your beam intersects the target, and so a more consistent return to analyze and then a higher resolution display result. Maybe :)

[@joshielexable]

Genius...

[@JaredNorton]

literally said "dude, come on!" out loud at 7:36. So freakin cool, what a video. As a first year EE student, super inspiring stuff.

[@PrashantKarade85]

This visualising of phased array is amazing. I've worked with antennas but never thought it can be implemented with ultrasonic sensors. Great work 👏

[@ClaymateDesigner]

That's the best explanation of a phased array I have ever seen. Well done...

[@Nicolas_dG]

The quality of this video is amazing! A real pleasure to watch :)

[@ntomata0002]

Excellent build, good explanation, a lot of solid work to set everything right. I really liked it. But, why you did phased array (only) in transmitter? The same principles that apply on the transmitter, can be used to the receiver too. So by sending one single pulse and getting the return by several receivers, using the same techniques for beam forming you can combine the same received signals with different phase differences to get focused reception from every angle you like. The obvious advantage is that you can get your scan in a single transmission/reception cycle by simply combining the same received signals with different phases. Even better you can use arrays in both transmission and reception to get even less noise. You can also use different spacing in transmitters and receivers to get the sidelobes in different angles so the transmission and reception lobes not to overlap, making noise even less. Also you can use different frequencies for each direction in order to choose the better combinations of sidelobes. Finally, you can get a second array of transmitters above the current array offset by half transmitter to get even more transmitters in the same space focusing the beam even more (the height difference will even make the beam more focused on one plane). Ofcourse you can make the array 2D, making it possible to 3D scan without physical movement but I think that at this point it starts do get too far. Anyway, you can make a lot of cool stuff with these ideas.

[@NorroTaku]

The water was just killer production effort Loved it Very satisfying video Super informative Also the overlays and editing was top

[@RealKingOfTheLab]

Amazingly well presented. Too many times I've just heard people toss around "beam forming" and "phased array", knowing they and I probably didn't have a firm grasp on how it was actually executed in the hardware level. Thank you. :)

[@maicod]

ofcourse it triggers my curiosity. You are a real master making and especially clearly explaining it all ! Thanks Bitluni, this video deserves a large audience !

[@RADARTechie]

I am an IFF technician and use phased array every day. I love this technology.

[@geoffscott265]

I had always put phased arrays in the "too difficult for me" camp. Thank you for a wonderfully straightforward and easy to understand explanation. You can't put a price on this kind of teaching skill ...well you probably can but I'm not paying it 😀 Seriously though, great video (I did LOL a couple of times) and your project was brilliant.

[@nighthawk4028]

IIT professors will not be able to explain and demonstrate as simply as you have done it. Wonderful.

[@rayh594]

You can probably improve this a lot by transmitting using CDMA, such as using Gold sequences that are mostly orthogonal. Then you don't have to wait for the echo of a pulse before doing the next one. You can just step through Gold sequences and decorrelate them in the receiver. You need good linearity on the receiver + rx amplifier to do this. The echo delay does not have to limit your update rate at all.

[@linhthai75]

I second motion this comment, and hope your comment would reach the author of this video. But, after a little thought I think this quite a tall order. As for such develoment attempt CDMA+PhaseModulation would required RF and digital engineering skills (ie. VHDL/Verilog for the FPGA). CDMA is one of the coolest tech this century, so cool that everyone ran off and bought a +$USD300 piece of CDMA eleclectronic and stuff it in our pant pocket, every other year.

[@rayh594]

@@linhthai75 There is no need to go the lengths you are talking about. The transmit waveform could be generated using an SPI port and a shift register (74HC595 is around 30c for one), so that is very easy. Shift rate would be 40kHz times 8, but PN sequence data rate is still 40 kbits/sec. The ESP32-S3 has two CPUs, so one could be used for transmit, but honestly it could be easily done using 1% of a CPU, or the low power IO CPU in the ESP32. Then, suppose that one measurement normally takes 100ms, to speed it up by 8, we'd use 8 Gold sequences, and transmit them in sequence, for 12 ms each. The, we'd collect data at the normal rate, but correlate it in software with the gold sequences. The sequences would overlap in flight, but would not overlap during transmission, so the TX circuit would be unchanged. The receive side would receive overlapping gold sequences from the last 100ms, and would figure out the time of flight using correlation. This would be CPU intensive, but I think could still be done completely in software. If one sequence is 12ms, then it has to be around 512 bits long. Correlation would be around 16 million operations, per 100ms, which I think is quite doable on the two 240 MHz LX-6 CPU cores in the ESP32. That's doing it the "dumb" brute force way. I've seen correlation sped up by an order of magnitude by using bitwise operations in a CPU followed by a bit count, at the cost of a some loss in sensitivity.

[@roba2440]

@@rayh594 the only thing you have to watch for is that the transmitter can withstand the ~50% transmit duty cycle and doesn’t melt! :)

[@zaccrisp9988]

​@Linh Thai Your level of understanding is great, what do you do for work?

[@stevejohnson1685]

Or look at the technique that HP patented for laser rangefinders - a swept frequency transmitted, beat against the delayed "chirped" echoes. If the frequency of the sweep and the carrier frequency have the right relationship, out pops the distance.

[@MikeDQB]

I had a basic understanding of how this worked before this video... Your visualizations really helped it sink in and I feel like I have a much greater understanding now. Thank you!

[@Audio_Simon]

As someone who plays in acoustics and electronics this video impressed me so much! The content is great but the production is amazing too! 👏

[@hoofheartedicemelted296]

Imagine if these could be miniaturised and put on a 360 headband with tiny pressure pads that implement a gentle pulse on the skin opposite the transducers. You would literally be giving blind people their sight back. Kind of like the new anti hero Mobius soon to be released in the cinema. Such is the innovation of creators. Thank you sir.

[@Farming-Technology]

This is amazing. Your proof of concept at 7:00 was more effort than some put into a whole video. Subbed!

[@viniciusfriasaleite8016]

What a good video! To make something similar, I would use an array of receivers and measure the direction of arrival like an automotive radar. It's probably faster, but needs more signal processing and signal modulation. Take a look at FMCW radars!

[@IndependentNewsMedia]

A method of improving the the resolution is to place a cone in front of the receiving sensor, as the transmitters can cause interference due to their close proximity, a second method is to have a transmit, receive, sleep cycles, transmit short burst e.g. 100ms and during transmit period the receiver is switched off, and then receiver switched on period 50ms and the transmitter is switched off, and and third period both transmitter and receiver switched off 20ms (sleep cycle save battery), varying these parameters can set the distance of detection. Please also be a ware the weather effects detection distance and resolution, on damp days speed of sound varies, in fog the transmissions are dispersed, and in heavy rain or snow false triggering. This method is utilised in the Railway industry to detect a train has stop in a platform. Great video, keep up the good work. 👍

[@JBrierley02]

This is fantastic. Well explained, extremely well illustrated, with excellent working examples and perfect pacing. I wish all such video’s were as good!

[@NunoCunhaSRW]

Nice video... i work with this stuff, and just to let you know, you were experimenting with directing the beam, left to right, its great but you can do that with a moving arm. Now, instead of manipulating the direction you can use the delays to focus the beam, concentrating the energy in a specific depth... assuming you have 8 transducers, directing the beam left to right you apply delays like 1 2 3 4 5 6 7 8 ms, or 8 7 6 5 4 3 2 1 ms, depending to each direction you want to direct the beam left or right... but if you use a delay like 8 6 4 2 2 4 6 8 ms or 2 4 6 8 8 6 4 2 ms you can concentrate the front wave in a specific depth making the image clearer. :D any way great project👍

[@TorbjrnViemNess]

Really cool project! I made a single transmitter ultrasonic ranging system once, and there's one more trick you might want to consider implementing (if you haven't already) for better sensitivity along the sensing range: make an adjustable gain for the receive amplifier and program that with the controller (I used a JFET as one of the resistors in an opamp circuit and had a filtered PWM signal to drive the gate, but other solutions might exist). That way you can start out with a low amplification (to avoid saturation and picking up unnecessary noise for close ranges) and gradually increase it as the sound waves travel to increase range.

[@douglasengle2704]

Seeing how just a few sonar transmitters and one receiver were able to provide a vague image, besides being amazing from a relatively simple clever configurations, made me think how some means of vision may have quickly evolved in nature for very primitive life. The beam shaping ability was significant, but not as sharp with a narrow beam changeable angle compared to a highly focused beam aimed by mechanical rotation. This interference wave multi wave source generator gives an entire image the width of the beam all at once or rather in one cycle of the source signal. The receiver is likely not taking a snap shot of the entire image in once source signal cycle or there about. This is the best demonstration of the physics of phased array I've seen. Thank You.

[@ClokworkGremlin]

The fact that the phased array allows you to move the beam without needing any moving parts (besides the transducer diaphragms, of course) is really cool. I'll have to keep that in mind for future design ideas. [edit]You could use a low-pass RC filter to smooth out the hard edges and feed the transducers with a pure sine wave. Might improve your scan quality. Might not.

[@stevejohnson1685]

And further refinement by "fading in" and "fading out" the train of sine waves ("Chebyshev windowing") ...

[@philliprhyner850]

Excellent work! If you switch the transmit array to a receive array and perform an FFT across the array (not time domain but the array domain at every point in time). This is closer how underwater acoustics, and likely Starlink, is done for receive 2D arrays (a 2D FFT). Your example is actually similar in hardware to most side scan sonar systems for underwater, just with piezo-ceramic transducers. The other difference is that the array is used only for receive while the single transducer is only used in transmit. The transmit transducer is meant to have as perfectly spherical beam pattern as possible. If you have the results of a beam pattern of a single transducer, you can use "window" function of the array to minimize the side lobes for both transmit and receive. One common window for acoustics is the "Hamming Window". If the transmit transducer does not have a perfect spherical beam pattern, you can modify the window to increase or decrease sensitivities at specific angle. For your inverse application, the window algorithms would improve the transmit signal. Also for the underwater acoustics realm , the shape of the piezo has direct affect on the beam pattern. Transducers used in arrays are specifically shaped like narrow rectangles. An array would look something like this, IIIIIIIIIIIII. The beam pattern coming off in the narrow width is very wide while the beam pattern coming off the wide dimension is very narrow (~ 1 degree). This is how the array can digitally process along one dimension (out lateral from the center line of a vessel), while repeat transmits and receives along the tow track at specific intervals have the high resolution along the tow track, due to the tight physical beam pattern. Having examples performed on the table create multi-path. Hanging the test in open air would improve the result, along with using a receive array normal to the transmit array. Apply a hamming window to the transmit array will create a line transmit window that would minimize multipath. Subscribed and impressed!

[@mpja]

This video really impressed me! It was a great demonstration of applied physics to solve a problem. The visualizations were also great to watch.

[@m_riatik]

i learned about phased arrays in my undergrad and was just visualising it in my head the whole time, which worked for me but it's not as exciting. i've never seen a visualisation as good as this!! really makes you appreciate the physics behind the tools we use everyday.

[@DantalionNl]

Intuitive, easy and effective. This could have many applications such as passive non-intrusive presence detection and crowd estimations. In this field I believe it to be highly important to perform the detection without using cameras for obvious privact reasons.!

[@kelvinchow3663]

Awesome, I had a final year project for ultrasonic detection using DSP around 1996. Although I had exposure in beam forming in the later study, I stopped doing practical hardware experience after the project. I am really appreciate of what you did. Indeed this is what I want to start again. Your inputs and materials should be valuables resources that I can start again.

[@tommysecondspace6977]

Great video! The amount of work that went into the projekt and the video really shows. It is fascinating to me how all the little parts and projects, the 3d printing, the cnc, the micoprozessors, the programming come together and we are able to create something like this on our own.

[@stevegreen65]

A very cool project and a fun video. I was a SONAR Technician in the US Navy and I never dreamed I see a video of someone building a phased array SONAR in their home. 😁👍

[@73Ferret]

Really interesting - I used to work with underwater survey systems and when multi beam echo sounder came along they changed what was possible on a massive scale. An imaging sonar system I worked with later had 56000 beam elements…

[@mickeyfilmer5551]

I knew about phased arrays - but couldn't find ANY explanation as to how they worked-your video did that. Subscribed forever and looking forward to learning much more from you - even in my advanced age! Thank you Young Man : Sir! big respect to you!

[@NathanOKane]

Absolutely incredible explanation, experiment, and testing. I'm floored by this video wow!

[@DanielMaidment]

I studied this at uni, and this is by far the best explanation I've seen.

[@alexkram]

This is amazing! Super impressed with the engineering and especially the visual displays. I have never known about your channel before but I am an instant subscriber. The automated coding with rolling pin, etc is freaking hilarious too.

[@88njtrigg88]

The visuals in you're videos, make them incredibly intuitive. Thank you very much kind Sir.

[@TheHookUp]

Saw you previewing this in the last Makercast, but this turned out better than I could have imagined. Awesome project and awesome video.

[@amitsci]

What’s the makercast? Sounds like I should be attending!

[@claytonhutton3695]

@@amitsci It's on youtube. Just search there. But it's a cast of a few different people/makers going over different projects, ideas, etc. It's pretty cool.

[@jimmythepowerful]

The genius mad scientist! Ppl Overlook how many skills this guy has, coding, electronic soldering, 3d printing, math, physics, engineering… a lot more.. wow

[@nenioc187]

Excellent video! And as a programmer, I can tell: yes, that's exactly how we program stuff! ;-)

[@quetzalcoatl-pl]

Yup. Totally. And when we "go for a lunch break" we really are just taking our time to recharge drills' batteries.

[@mattstaab6399]

i suprised he decided to show all his time spent programming, most people skip it

[@marsvandeplaneet7786]

Currently I;m working on placing muliple type of sensors on my rc truck.Your 'array story' gives me a whole mew dimension. Tnx a lot!

[@samarpan.das.]

This video is too good. Well produced and scientific. 🔥🇮🇳

[@danriches7328]

Could not have said it better myself!

[@stanleydsouza6723]

This same steps , by adding more mimics as beacons placed or spread around can generate clear image, mimics with different wave pitch, signals, Thanks to this valuable researchers in ultra Sound, Sonic sound wave ..to real image output..

[@erygion]

This is a great video! I was laughing at the "hacking" (a rolling pin!? 🤣) and besides that the information was very interesting and well explained with visuals. I've never seen ultrasonic sensors set-up like this, definitely new to me and I just made a "radar" using two servos but THIS is 5 levels next level, I had no idea you could have control like this... and how it's done, fascinating! Very well done but that's no surprise coming from you.

[@jafinch78]

Will be interesting to see if bitluni advances the development into a cost effective larger array like the transducer array boards used in the Soundlazer. There is more that can be done with these for directing ultrasonic beams as well as detecting also. I actually thought he was going to make a synthetic stone skipping effect on water at first... of course only with the sound beam and not a stone. Surprised actually he didn't focus an ultrasonic beam on water... then I thought about and this is more for the specific application. Was thinking he can use an array for detecting and emitting.

[@gmag11]

The way you visually explain this topic is wonderful. My congratulations

[@roba2440]

If you had an array of receivers, instead of transmitters, and used a single wide angle transmitter, you could form your beams on receive, _ in parallel_ , in software. This would allow your frame rate to be much faster, as you would be scanning all the angles with one transmitted pulse. Your frame rate would be limited to the echo return time. You would need to simultaneously capture the signal from each receiver separately.

[@transient0126]

Yes but you won’t have directional control of the pulse…this is necessary to focus the energy on all objects in front of your system. A single pulse will produce a stronger response directly in front. Not ideal.

[@roba2440]

@@transient0126 the transmitted beam pattern is the product of the element factor and the array factor. The element factor, ie the angular response of the projector must be broad enough to insonify the sector in front of you, or it wouldn’t have worked as well as it did. The array factor can only restrict the insonified sector - it can’t make sound go where the element can’t transmit but it can focus it as you said. The lack of focus from a single projector allows you to fully insonify the whole sector in front of the projector all at once - good thing, but the lack of focus also reduces the intensity on the target, bad thing - made up for by the improved receive gain reducing the background noise level. You need to think about both the signal and the noise, because it is signal to noise ratio that you need and you are detecting signal against a background noise level. It would be useful for you to measure the beam pattern of a single projector and then of your full array.

[@DjWojtas1]

Visualizations are great but sound effect and music is just sooo perfect

[@ArktikosAdventures]

I'm so glad there are intelligent passionate people out there like you who do the things that us mere mortals could only dream of doing. You guys advance our civilization in so many ways. Thanks for a great video.

[@bobf9749]

As a novice in this, I suddenly have a basic understanding of the mysterious term “phased array”. Thank you.

[@EatRawGarlic]

Very cool! I noticed Matlab has a phased array toolbox and I have been wanting to play around with that for a while myself. Looking forward to you adding the third dimension to your setup, and maybe switching to RF in time? If you get that to work, mount an array in front of a VR headset, to scan your environment and then show a 3D representation of it inside the headset.

[@lloydrmc]

Subscribed. I appreciate the high quality, thematically representative music, and the zany keyboard manipulation cutaways.

[@rockyhill3]

That was amazing! I wish youtube would allow me to give you way more likes! You've inspired me to make a coding bot with a drill and a glove! Great video!

[@CoADD-51]

This is probably the most well done video I have seen on KZbin recently. I am not an engineer, physicist, or mathematician, but I feel like I could at least have a logical conversation with someone regarding how a phased array actually works. Excellent video!

[@edward17829991]

so coooooooool~ how about 2d array? I've learned the theory, but your execution is top-notch. It's a pleasure to watch you doing these!

[@sharonbraselton4302]

giid údeàr

[@DerSolinski]

Hands on the best explanation on phase arrays I have seen so far. Well done, I enjoyed it.

[@zachcrawford5]

Very cool, I wonder if having more "ears" and only one transmitter would be a better approach. Biological sonar seem to take this approach (and it seems to be very fast and accurate). Phased arrays are good for "digitally" moving a beam around, which is great for speed, if you can take advantage of it but it seems to come at the cost of more noise in the returning data since you now have to deal with the interfering information from multiple wave fronts.

[@basecius]

Yes indeed. The receiver part is a bit more complicated than the transmitter, so over all it would be a more complicated setup. But with the benefit that you could send a pulse, receive the response (keep the data from the multiple receivers separate), and then decide in what directions you want to look. I.e. many directions for each pulse. And then of course, you could have multiple Rx AND Tx. Lets say a vertical Tx array, and a horizontal Rx array. That way, you could do a full 3D scan.

[@kevin42]

One key advantage of the multi transmitter approach is that it blasts stuff with a lot higher power density, thus leading to a stronger response.

[@Breeegz]

@@basecius If you are investing in the multiple Tx & Rx, then you could decide at time of pulse which array is the Tx and which is the Rx. You could switch back and forth like the single ultrasonic sensor does, but with more channels.

[@Vega3gx]

@@basecius rx phased array is complicated, but not as complicated as the beam steering algorithm needed to do tx effectively. I'm not sure but I strongly suspect much of the noise and side lobes we see is due to the methods he used for active scanning. RX can all be done in post

[@Arrogan28]

Wow, i had no idea what the deal was with phased arrays, but in one short video you opened my eyes to so much possibilities. Thank-you!!!

[@steve0steel]

It would be interesting to see if a sine wave would create a clearer signal since the square wave includes multiple frequency components.

[@maxhouseman3129]

Sure. But a true sine is difficult to create digitally. Theoretically in a square wave there are infinite frequencies, practically its depending on the slew rate.

[@MikePorterII]

I can't say thank you enough. This was such a hard concept to grasp without being able to visualize it. Thank you so much for what had to be quite the labor of love to make this!

[@peter.stimpel]

Haha, I know what your neighbors are saying when you walk through the forest carrying all this gear: Look, he is now as well into finding strange energy, poor boy ... LOL. That was a fun journey.

[@appidydafoo]

Amazing. The glove on the drill and the rolling pin slayed me.

[@ronidaffan5904]

Wow !!! The most comprehensive video about phase array by far !!! Weldone !!!

[@miltonlai4850]

I see a major commitment of money and time behind this 14 minutes video. Very good work!

[@Reach41]

Best demonstration of a complex physics phenomenon I’ve ever seen. Thank you!

[@FatherMartin]

I have never seen, or read, a better explanation of phased arrays. Thank you!

[@BrunoHenriqueMattos]

How can this be a work of only one person? Incredible 😮 👏🏻