How I built the world's first 3D ultrasonic scanner(alextoussaint.com)
alextoussaint.com
How I built the world's first 3D ultrasonic scanner
https://www.alextoussaint.com/2021-04-28_How-I-built-an-ultrasonic-3d-scanner.html
44 comments
Very interesting! Do you have any links for me to learn more?
There are a few limits on what you can scan with only 3 mics. I detail everything in my article. Really suggest you read it.
There are a few limits on what you can scan with only 3 mics. I detail everything in my article. Really suggest you read it.
I only quickly checked and found this:
https://rusneb.ru/catalog/000224_000128_0000488231_19751015_...
-- click on "Skachat' PDF" (second from the top on the left). Unfortunately it is in Russian, not sure whether English translation exists.
This patent seems only for (X,Y) coordinate capture, but the same authors later worked on 3D setups as well.
EDIT: let me know if you need help with translation
https://rusneb.ru/catalog/000224_000128_0000488231_19751015_...
-- click on "Skachat' PDF" (second from the top on the left). Unfortunately it is in Russian, not sure whether English translation exists.
This patent seems only for (X,Y) coordinate capture, but the same authors later worked on 3D setups as well.
EDIT: let me know if you need help with translation
I definitely would like an English translation. Though, from what I understand (the diagram and the translation of the title) it's a graphic input device that can work out a few points at the time, and not in the same sphere (as echos will overlap, a point I develop in my article).
So it's not properly able to scan an object like I did?
Apart from that, how have you found that untranslated 1975 Russian patent? Google Scholar?
It's impressive what they were able to build with such low-tech.
So it's not properly able to scan an object like I did?
Apart from that, how have you found that untranslated 1975 Russian patent? Google Scholar?
It's impressive what they were able to build with such low-tech.
Replied to your personal address.
Phased array ultrasonics[0] and Multibeam echo sounding[1] has been around for decades[2]. This project in particular is an example of what's called an air coupled ultrasonic phased array. It's an interesting project, but it's not first. Congrats on making it happen though, and thanks for the write up.
[0] https://en.wikipedia.org/wiki/Phased_array_ultrasonics
[1] https://en.wikipedia.org/wiki/Multibeam_echosounder
[2] https://patents.google.com/patent/US5675552
[0] https://en.wikipedia.org/wiki/Phased_array_ultrasonics
[1] https://en.wikipedia.org/wiki/Multibeam_echosounder
[2] https://patents.google.com/patent/US5675552
Author here, the title is a bit misleading. In the article, I specify that it's the first 3d ultrasonic scanner (of my knowledge) that works in the air.
Have you seen the way lasers are used to do this? NDT or non destructive testing has use dual laser systems to create ultrasound for several decades. Check it out if you are interested.
I can try to find that papers if you would like.
I can try to find that papers if you would like.
I think one of the subjects of my exam for entering engineering schools (a thing we do in France) was about that.
Would really like to see papers on the subject, because I don't really understand how you can use lasers in opaque materials. Will also look at my test to see if it's really about that.
Would really like to see papers on the subject, because I don't really understand how you can use lasers in opaque materials. Will also look at my test to see if it's really about that.
> I think one of the subjects of my exam for entering engineering schools (a thing we do in France) was about that.
Cool. For which school was it? And Which year? (I'd like to get the subject and read it actually)
Cool. For which school was it? And Which year? (I'd like to get the subject and read it actually)
It is this year's entrance exam (PSI) for the École Normale Supérieure and the Polytechnique school.
The subjects have not been officially published. But here's a scanned copy:
https://www.cpge-paradise.com/Concours2021/PhysPSI.pdf
It's in French though!
The subjects have not been officially published. But here's a scanned copy:
https://www.cpge-paradise.com/Concours2021/PhysPSI.pdf
It's in French though!
I don't speak French at all, but it sounds funny to me that the École is both Normal and Supérieure. Like, make up your mind, school!
I don’t speak French either but I like the etymological sense that Superior school means Post-Secondary/College better than the English version. It seems like a logical and efficient way to get the meaning across.
That's probably what's happening, but I think I will keep the funny version in my head canon.
Not at all! There are many "Ecole Normale" which are just schools for teachers. The "Superieure" refers to the fact that you have to pass a hard exam to get in. And once there you can do research or become a teacher in a preparatory school (which prepares you to do the ENS's exams, in a pyramidal way)
Thanks! X-ENS subjects are usually really well done and insightful, I'll have a look!
Fascinating! Awesome project. Interference of waves seems to be such a fertile ground for all sorts of unexpected powerful possibilities.
There have been some projects to make “sonic cameras” using MEMS mic arrays that this reminds me of. https://hackaday.com/2020/08/10/acoustic-camera-uses-many-ma...
The maths is exactly the same, except that the beamforming is at the reception rather than at the emission. But it works based on the same principle.
> If it's so easy, why hasn't anyone done it before?
Medical 3D ultrasounds are a thing:
https://en.wikipedia.org/wiki/3D_ultrasound
That doesn't take away anything from how cool this project is though!
Medical 3D ultrasounds are a thing:
https://en.wikipedia.org/wiki/3D_ultrasound
That doesn't take away anything from how cool this project is though!
It’s a lot easier when the waves are propagating in mostly liquid. Like a sonar.
I guess that the main option is the second one. Lidars, once they are solid-state and cheaper will be a better alternative (smaller, more precise and much faster).
The first option isn't the good one. Ultrasounds are already used in car parking assists and people trying to improve them would have gone the same route as I did.
The first option isn't the good one. Ultrasounds are already used in car parking assists and people trying to improve them would have gone the same route as I did.
Not sure about the first, but it is really cool regardless. I tried doing a similar thing once upon a time with RFID (SAR), but I couldn’t figure out how it would work.
This seems even lower level, since I was basically using off the shelf components, and you built the whole thing yourself!
Really cool.
This seems even lower level, since I was basically using off the shelf components, and you built the whole thing yourself!
Really cool.
well, you sir, are the first.
except maybe those lamers: https://en.wikipedia.org/wiki/Medical_ultrasound
:)
:)
The title in the submission is inaccurate & not actually the title of the article. The missing bit:
> that works in the air, using off-the-shelf components (< $100).
> that works in the air, using off-the-shelf components (< $100).
Amazing. I was literally working in a factory at that age
Reminds me of phased array radars. Where they form an array of transmitters into a moveable “beam”
https://en.m.wikipedia.org/wiki/Phased_array
Radar is at its simplistic sensing a pulse out and listening for a return. But then the engineers wanted more.. of course home radar would not be really safe.
Cool project by the way.
https://en.m.wikipedia.org/wiki/Phased_array
Radar is at its simplistic sensing a pulse out and listening for a return. But then the engineers wanted more.. of course home radar would not be really safe.
Cool project by the way.
It is exactly what it is, except I used ultrasounds (40khz, easy to work with) instead of electromagnetic waves.
Hah! Back in my undergrad days I’ve thought about doing exactly this, but I couldn’t figure out the beamforming part, so I just measured distances in (microcontroller) software by correlating transmitted x received waveforms.
Awesome project! I’m happy the author figured it out! I feel vindicated :)
Awesome project! I’m happy the author figured it out! I feel vindicated :)
Thank you!I've been helped a lot by how technology changed for the past few years (lots of good resources online, was able to do simulations easily, MCU programming a lot easier, etc).
> I hated using a Chinese company for a job that could be done in France, but no local companies were offering the same service.
They exist, except they are more pricey.
They exist, except they are more pricey.
Would you have any links please? I didn't see any that did small development batches for a price lower than 200 EUR.
https://aisler.net/
they definitely aren’t as cheap as Chinese offers but I always found them to be reasonably priced. They manufacture in Germany.
they definitely aren’t as cheap as Chinese offers but I always found them to be reasonably priced. They manufacture in Germany.
I'm in France ;-)
You can enter your country in the "Ship to" field.
Check e.g. Eurocircuits
Check e.g. Eurocircuits
Interesting, but yeah, he fought way too hard with the analog part driving the transducers
I get it, it's a hobby project, but I feel the knowledge of analog electronics is getting lost in a way (and yes the analog people are the worse at explaining things - I think I saw YT videos more informative than whole university courses)
I get it, it's a hobby project, but I feel the knowledge of analog electronics is getting lost in a way (and yes the analog people are the worse at explaining things - I think I saw YT videos more informative than whole university courses)
Author here. It's all digital (square waves!).
There were definitely well-known circuits for driving transducers, but I had to improvise to make a circuit small and simple enough to build 100 of them.
Building new circuits requires tinkering and many tries, regardless of the builder's knowledge.
There were definitely well-known circuits for driving transducers, but I had to improvise to make a circuit small and simple enough to build 100 of them.
Building new circuits requires tinkering and many tries, regardless of the builder's knowledge.
> well-known circuits for driving transducers, but I had to improvise to make a circuit small and simple enough to build 100 of them
Yes, maybe an Opamp that works at 40V could have worked (with an appropriate feedback config), but your end result looks ok. I'm just saying that putting the transducer in the emitter was a non-starter (as you correctly pointed out)
> Building new circuits requires tinkering and many tries, regardless of the builder's knowledge.
Absolutely, I'm not denying it :) It's a great job nonetheless
Yes, maybe an Opamp that works at 40V could have worked (with an appropriate feedback config), but your end result looks ok. I'm just saying that putting the transducer in the emitter was a non-starter (as you correctly pointed out)
> Building new circuits requires tinkering and many tries, regardless of the builder's knowledge.
Absolutely, I'm not denying it :) It's a great job nonetheless
You need to be more verbose as you say analog tech is pretty unusual knowledge so even though I've fought through the 101 course you still make no sense to me. It's really sad and I think it's a provlem I share with many.
> I'm just saying that putting the transducer in the emitter was a non-starter
I don't get it. What do you mean by that?
I don't get it. What do you mean by that?
There are two reasons why you don't put it there, one of which you explained "The NPN capacitor here only cuts out the alimentation of the capacitor, but there's nothing to discharge it.". Yes, the transducer has (theoretical) infinite resistive load.
But another reason (which would be valid for a resistive load as well) is that anything plugged into the emitter will have its impedance "reflected" to the base.
Even if you're talking about a digital circuit you need to think about transistor biasing. A pure capacitive load on the emitter means no biasing.
One of the ways you could work around this is to have an inductor in parallel with the transducer so it would resonate at the frequency you want https://en.wikipedia.org/wiki/LC_circuit (though not very practical to your case) - but ideally this would go on the collector, not on the emitter
Your 3 transistor solution is ok for the most part. Maybe it's doable with only one transistor but probably would need an inductor and/or wouldn't be too efficient.
(Not the best explanation, but this would require some trial and error and actually thinking a bit about the analog aspects of the circuit)
But another reason (which would be valid for a resistive load as well) is that anything plugged into the emitter will have its impedance "reflected" to the base.
Even if you're talking about a digital circuit you need to think about transistor biasing. A pure capacitive load on the emitter means no biasing.
One of the ways you could work around this is to have an inductor in parallel with the transducer so it would resonate at the frequency you want https://en.wikipedia.org/wiki/LC_circuit (though not very practical to your case) - but ideally this would go on the collector, not on the emitter
Your 3 transistor solution is ok for the most part. Maybe it's doable with only one transistor but probably would need an inductor and/or wouldn't be too efficient.
(Not the best explanation, but this would require some trial and error and actually thinking a bit about the analog aspects of the circuit)
That's smart. Didn't think about it!
Though, the emitter is pulsed and I also rely on the fact that it stops resonating after I cut out the power. Unless my LC oscillator has a low Q, it will keep oscillating after I cut out the supply, which is something I'd prefer to avoid.
Though, the emitter is pulsed and I also rely on the fact that it stops resonating after I cut out the power. Unless my LC oscillator has a low Q, it will keep oscillating after I cut out the supply, which is something I'd prefer to avoid.
> Unless my LC oscillator has a low Q, it will keep oscillating
Correct! You can add a resistor in parallel to reduce Q (but it will naturally "not be perfect" since the transducer is putting energy out as sound waves and it's not really a capacitor.
Correct! You can add a resistor in parallel to reduce Q (but it will naturally "not be perfect" since the transducer is putting energy out as sound waves and it's not really a capacitor.
[didn't read your article, your claim has put me off, sorry]
-- use an electric spark gap as the wideband acoustic source
-- use 3 simple electret mikes on 3 axes. Mikes don't need to be high quality (spark intensity can always be raised as necessary), so they were hand-assembled using metallised mylar film (used nowadays to pack some food, or ESD-sensitive electronics components).
The rest was data processing.