Electron microscope image capture via microcontroller (with drill bit animation)
Vložit
- čas přidán 29. 08. 2015
- This video shows a microcontroller-based image capture system that makes it easier for me to record animations with my scanning electron microscope.
Support Applied Science videos: / appliedscience
The diagram in the video is missing equal-valued resistors for the second opamp in the video processing chain. Gain=-1
Teensy firmware: drive.google.com/file/d/0B4YX...
Processing script for PC image viewing ( not finished!):
drive.google.com/file/d/0B4YX...
The Teensy microcontroller: www.pjrc.com/store/teensylc.html - Věda a technologie
this man is a genius
+MrPolymath0 I just do not understand where he takes the money and motivation.
+0MoTheG 840$ per video from patreon, ad revenue from youtube aka google, plus he works at google X as far as i know. and motivation isnt a problem if you are really interested in something...
+Felix Dietz He should work for DARPA
h
+John Ecstasy
I have none of those at home. I would like to have 1/5 of one.
If this is the SEM you got from Sweden, I helped the previous owner move it from his apartment. =D
Man I barely understand any of this. I do appreciate these videos though. Everytime I watch one, I at least learn a little more.
+Nugenrules This video was a little more technical than usual. Stay tuned for more general science demos!
+Applied Science I totally dug the technical side. please post technical videos in the future. I understand the general science angle but you have a good capacity to belay the info. great work!
+Nugenrules yep...Ben is a clever dude.... Hey Ben how come you not still in VR?
+Nugenrules im with you hre. no idea what hes talking about but still watch.
+. This . gets it. ^^^^^
Fantastic project as always Ben. Awesome to see the MDO being put to such good use! You know, there is a Search feature as well, that would allow you to search and mark for specific events in the USB data, making it easy to navigate to them. Also, when you scroll in the Event Table, that centers the zoom window in that spot.
+w2aew Thanks! I still haven't forgotten about the ICs that you sent me! They'll make an appearance soon. The search function on the MDO is definitely handy, but in this case, triggering on specific USB data was my goal. I couldn't get the scope to trigger on any sort of USB data packet, even though they were definitely present in the stream. Let me know if you have any tips regarding this, or let me know who to talk to regarding USB triggering.
+Applied Science It should be straightforward to do that. You may have to put triggering into "Normal" instead of "Auto" to prevent a self-trigger, especially if long memory depth is used. I'll have to give this a try myself and let you know...
FYI, I believe the metallization on those chips is gold, so you may not have to sputter them.
Applied Science rtxasz xx
I must have watched a few dozen of your videos in the past week, and I can't get enough! You give very detailed and informative explanations, and I'm impressed with how self-sufficient you are to do all the design, hardware, and firmware, and see these projects through. Keep up the great work!
Followed you pretty much all the way through, having a background in Electronic Engineering (but I'd never dare define myself one).
I must congratulate you on the exceptionally clear reasoning and logical steps taken throughout your project. Not only that, but you made me reconsider how something as seemingly daunting in my eyes as an SEM can be confronted so calmly if looked at in a methodical fashion.
This has definitely spurred me to try something similar with some vintage analog data acquisition devices of my own [even though with far less sophisticated testing equipment than yours ;) ]
As little value as it may have to you, you have all my respect. With your polyedric nature and inquisitive mind, you're a true scientist and a scholar.
Stay curious and keep on churning videos (if you wish to, of course).
c(_) Cheers
Pretty sure he travels back in time every couple of years when he's improved, just to live in a different country, with a different name to start over.
@@hardwareful Hmmm, I read a book about people who do that...
This is so incredibly technical it just blows my mind, you sir are clearly some kind of technomancer.
Amazingly well done. Some of us can imagine such things, but not many can go the extra mile (or two thousand) to figure it out and actually make it happen. I enjoyed your no bs straight forward explanation of your work. Bravo!
I love how you put the culmination of all your hard work at the very beginning. It makes for a great video. Thank you.
This video has been really helpful in aiding me to accomplish something similar in the lab I work in. Thanks for sharing!
I'm a software dev and understood about 90% of what you mentioned. In a nutshell, you created a custom analog to digital converter and wrote your own software to pull the digital image data over the USB bus and display it on a modern computer. Just wanted to say you are someone I admire and are probably up there with the best of the best when it comes to engineering. Unfortunately I don't have the opportunity to work with the hardware side of things but it's something that interests me probably more than the software side. Keep up the awesome videos.
It doesn't get more interesting than this!
Thank you!
I only understand small amounts of what goes into these types of videos but it's fascinating none the less! Keep up the great work and please don't ever stop being awesome =)
Really awesome that you have a Vintage SEM. I have always been fascinated in these ever since I was a kid.
I gained about 20 IQ points just watching this video. And more importantly it inspires me to get back to messing with my Arduino.
I recently got back into EE and understood all this stuff. This is awesome!
the data transmission part is really helpful thanks for sharing!
Not being an EE, I understood about 5% of what you said. Humbling and very interesting, this is the shortest 17 minute video I've ever watched. Thanks for setting all of this up and making this video.
Very cool and innovative way to capture the analog image data.
I truly hope that you are using your obvious gift to teach young people!
Thanks for the video!
Wow, wow, wow! The quality of your content is outstanding! Many thanks for sharing this
I know this is an old post but it’s great work! What is more, your “ping pong buffer” is perfectly elegant it’s effectively the ur-example of “page-flipping”. If you are processing in real-time and have no control over the speed or period of acquisition, it’s the best option if you have the resources. If resources are limited you can sometimes implement a page-shifting method using less memory but its algorithmically more complex and can be difficult to implement on some hardware. Synchronous access often ends up even more complex in such scenarios; even if you have the speed, depending on the hardware it might not be reliable.
Instruction Clear Enough. Successfully build SEM with 0,5 nm resolution. Thank you.
+Clear Instructions You managed to build a SEM with only him giving a good overview on how he converted the analog video signal out of a already functional SEM?.. that's impressive
Awesome job man! I'm drooling over all your awesome gear! Super nice oscilloscope!
Really enjoyed this one Ben. Great solution. Looking forward to the scanning control.
Fantasitc to see it all leading to this. I tried to build one myself with an electron gun from a crt HD TV (there were a few), I feel for your efforts, I had the 2 stage oil and diff stack, still have the edwards pirani glass guages, outgassing and special greases, berilium copper ? photomultiplier. Had to call it a day before I got anything working so Its great to see you go all the way. Just had a demo of a JEOL sem this week, sample in hand to image was ~ 5 mins, no sputtering,no gasses,only a 2 stage pump. wow. Just found your channel. Love it. Ive done DMA on arduino, alternativley the pic32MZ has a fast ADC and runs at 200MHz. Great vid, nice scope.
Very thorough walk-through. Could build one from your explanation if needed.
You're a natural teacher.
The Ping-Pong buffer is basically a ring-buffer with two entries. Synchronizing on whole buffers is much more efficient than synchronizing byte-accurate access. Adding a few more buffers to your ring might improve throughput (at the expense of latency) or help mask transmit jitter.
Thanks for making these videos. I love shit like this and it makes my days better every time I see one.
This is amazing. Like most of the viewers (it would seem), I too didn't comprehend most of the things you presented (I lack the proper technical background) to us. But, because of your clear explanations, I was able to take some of it with me. Looking forward to future videos. Thank you, Ben!
-Martch
+Martch Zagorski That's what is so fantastic about these videos and his work. He very clearly explains the method behind everything.
Man I'd love to be as handy as you are. This is so darn exciting. Great stuff man!
One of the higher quality CZcams channels right here.
What a nice result! I'd be proud to have made this! Spin a small PCB, USB B on the side! The only 1980s SEM with USB out! 👌
That was awesome, your videos are the best. Keep up the good work.
This Ben, It s the most intelligent person I ve ever seen on the screen and in real live. Keep going Ben! There are fans out there!
That's cute putting the duty of the routines on the scope. On the C64, we would change the border color in routines allowing us to see the amount of time spent processing, which we tried to do in the vertical blanking period. The start of the vertical blanking (actually the border) was a NMI we used for our 'trigger'.
You should make a video just about the electron microscope I don't think im alone in thinking that it would be great to see & learn more about it. Also thank you for showing me the coolest thing I have seen all day :)
Holy fuck, I didn’t understand a single thing that this man was saying but it was beautiful to hear his enthusiasm. Guy obviously loves what he does.
way cool project! I hope you make another video using the setup, I would love to see that. thanks Ben.
Using pins to debug the microcontroller is a pretty friggin genius idea!
you are my favorite type of human! Awesome video.
Wow - a sophisticated retrofit!
I do some simple electronic stuff and want a scope, but I can't really justify paying for a good one. You seem to use yours all the time.
Ben,
Cool little project. That SEM has got to be a lot of fun to play with.
Craig
God level Genius !!
More than information I was motivated and inspired . Thanks a lot.
Amazing work you do! As i was watching the video I was thinking wow that scope is really nice. That Tektronix MDO3104 cost 10K on ebay! No wonder it's so impressive.
I really enjoy your videos! You are very motivating - thank you so much!
Wow! kind of rocket science and you could explain it - that's great!
Well...this is very inspiring to me.
I mean.. the processing app is simple but really, really cool indeed!
Thank you!
This man is dedicated to things many humans don't dedicate themselves to. (Hi Ben)
I would love to spend a day with you in your workshop. The things you do are so interesting, i came across a video you did a little over a year ago. I think i watched every video until the early hours of the morning.
Pure engineering. Well done, loved it.
Splitting sync and info... very clever and efficient solution!
Amazing work, as always!
This is the best Science Channel
can't wait to try this at home!
Great work on this project... this is not something I would have ever tried, just because it seems just so crazy hard! Nice.... :D
That is very impressive. Thanks for sharing it.
Ben, you're so awesome, yours is in my top 3 channels, you remind me of Chris Hadfield :)
Love the vids thanks for doing them some of it way over my head but that's the only way to learn from UK
Ben: The newer SEM machines that succeeded yours probably improved by means of a signal/data out. So it's quite the accomplishment that you built something that extends that in such a short time. I learned a few things from your video including processing.org and a lot about the tektronix too. Thank you
You made me realize that I'm not an "electronics enthusiast".. I'm just a consumer that solders some stuff from time to time.
I have watched these 9 videos several times but I must have missed why there are spark plugs in the vacuum chamber. I love your channel!
Can't find words to fit my appreciation
Thanks for the unique video👍📗📚📊
This is fantastic! Thanks for sharing.
Pure brilliance!
I learned a lot from this video! Thanks!
You sir are a wizard.
As usual, great stuff!
I'm not sure it matters much here, but using an opamp as a comparator does have its drawbacks. The saturation of the opamp might influence timing, whereas a comparator is built to avoid this.
And thanks for another super video... Keep up the good work
Seconded.
Lucid and concise - an *excellent* video.
Great one. Very informative. Thanks
Awesome! Looking forward to scans of all sorts of interesting moving things.
Awesome Report Thank-you!
This is amazing!
Nice! I am sooo in love with your oscilloscope... *_*
This is just so cool
This is so cool.
Wait, I think I remember the initial video you did when you got this beast.
Nice job
Really neat way of capturing images! :)
Have you compared some sample images with the ones coming from your scope?
Very impressive bit of work. I was thinking about how I would have tackled the problem and decided I'd probably have just stuck a digital camera to that camera adaptor box with a really long shutter. The shutter could be controlled by noticing the bright dot move back up to the top left corner. Not as elegant as your solution but perhaps a bit quicker.
+saturdayocean Yes, that was the other solution I'd pondered. I thought it would probably be slightly more work than just controlling the shutter. Either way I think I'd probably have sucked all the raw data into a PC and attacked with it code. As you can probably guess I'm in software development, when the only tool you have is a hammer...
+Wobblycogs Workshop ...you become extremely skilled at using that hammer and learn how to be very creative with its application?
+TheCammerhammer ...every problem looks like a nail. I prefer your end to the saying though.
+Wobblycogs Workshop
Sure that would work, but it's not really much better than making polaroids and putting them in the scanner afterwards. But it will only work in a perfect world.
Taking an image of the screen will never give you as high of a resolution and dynamic range. You don't know if the brightness of the screen is uniform or even linearly related to the signal. The electronics for the screen will introduce some brightness "distortion" and maybe the physical construction of the screen (electron gun or phosphor) means what the signal says and what you see on the screen isn't proportional. The image could even be distorted (pincushion/barrel) and you have distortion due to the screen shape and camera lenses. Then you're taking a picture with a device, that's much better at seeing colors than raw brightness levels (in that case a black&white polaroid may even have more dynamic range than a digital camera). White balance and other mojo the camera is doing will further decrease the dynamic range.
After you've figured all that out, you have to do it all again when you change anything on the microscope, starting with just the screen brightness but also other more important stuff like sweep time (I'm pretty sure the 1 image / 11s isn't the only setting).
Back in the 1990’s I had a serial video image capture device for Macintosh that worked a lot like this. It took so long to scan a single frame of video that I bought a 4-head VCR so I could pause still frames and capture them, otherwise it was useless except for capturing images from a live video camera with a still subject.
You are a very smart dude.
To make hid work faster on video you can use 4 bit delta compression. All bits set can be an escape code. When the escape code is hit you can use the next 4 bits to cover sync or numbers bigger than 4 bit. This pretty much doubles the bandwidth but still may not be enough for your project.
"I realised I could fix it and so then I had to" - This is the mindset of an engineer, and the mindset that consumes about 90% of my time.
Wonderful verbiage....
I thought for sure he is going to use an FPGA for this one but no he is smart enough to make a MCU do the job ..great video thanks
This is awesome. I really admire the stuff you do. I want to start making my own video portfolio like this, but I always feel like I don't have something worth making a video about. Do you have any tips for how to generate project ideas and motivation?
+JoeJoeTater Thank you! I very often feel like I don't have a video-worthy project. Once a problem is solved, it sometimes seems too simple to talk about. The trick is to just start the camera rolling, and talk about your work. You can always edit or re-shoot some bits later to make a better video. You'll find that when you start talking, you have to go through the thought process of your viewers, and this will help bring out the details that make the project interesting in the first place. Sometimes people do not want to post their code for a project because it's too hacky, or unfinished, or they're worried the masses will ridicule them. The overwhelming advice from the community is to just post the code! Don't worry about hacky/unfinished. There is certainly someone who will benefit from partially-working code, and if everyone posted their projects, the total volume of knowledge on the internet will grow and have compounding benefits. Motivation itself is a more tricky subject.
wow, I worked with that SEM model 20 years ago and it was old even then :-)
Very cool!
You are a wizard.
Once you're controlling X&Y with a micro controller you can digitally limit the range to trade in some real estate for frame rate. If you manage a higher resolution with a slow and accurate digitally controlled scan you can even use it to "zoom in" (slow down the scan rate for higher resolution, limit the spatial range, end up with the same frame rate and pixel-resolution)
By the way, would it be possible to use another (electro) magnet as a "magnifying lens" for the electron beam?
That duty cycle trick for measuring time spent in interrupts is really neat, I wish I'd thought of that during a project a few years ago...
Ninja! way outta my league!
Code's looking good. Double buffering is the right choice here.
If you have enough memory for 2 buffers AND your acquisition is by definition "quantized" into lines, then i'd prefer this over a ring buffer.
Only thing I'd improve here would be to remove the duplication of code with bufferA and bufferB.
Get a read and a write pointer and just alternate them between the buffers.
Sure, you get an extra indirection, but code gets shorter and more streamlined.
Keep it up man :>
Cool gadgets bro...
Ben, do you have a gallery of most, if not all of your experimental SEM photos? I'm interested in seeing them.
Very nice.
Thanks . Beautiful
Incredible work, and thanks for sharing the source code :) I was a bit thrown when you said "half of each of these nibbles is a sample" - was that a slip of the tongue (a nibble is 4 bits)?
Love your videos. I'm hoping to get a SEM of some sort along with a micro or nano scale 3D printer within the next couple of years so I find this video very exciting! I'm curious how much you got your SEM for, and whether you think the quality and resolution (or other features) you're getting compares to what a newer model SEM would deliver?
He got this one as a university discard from Sweden for shipping cost. Probably still several thousand dollars to ship from SE to the Bay Area.
holy crow you are a genius!
One suggestion, please re-label this video something more consistent with the bulk of the content, such as "homebrew circuit design and application to modernize/hack vintage electron microscope"