Shadowboard Build from the Basement Shop 

Thanks Christoph! Just finished watching your PrintNC video. Awesome job, really like both the content and presentation of your videos. (I need to start using audio the way you do) I think you might have convinced me to start designing my next CNC... It's sort of a disease. Very jealous of your space and equipment. Someday I hope to catch up... Keep in touch! -Scott

@@scottprints Hi Scott, Thanks for the comment! yeah the CNC is like a disease, once you're infected, your other projects will suffer from a lack of time :D

Hi, Is there any chance I could get a diagram or a step by step "how to" for the software and what else is needed for this set up? I have a raspberry pi and camera, id ike to make one of these. Thanks @@scottprints

@@johnbarrera7998 hi John, I really don't have more instructions past what I put in the video, sorry.

Thanks! Probably under-promoted because I haven't released much content since this video. I have a ton of ideas, but time... Not so much.

Guess I'd better start designing that!

@@steveallwin1056 Thanks for the feedback, nice to see that this video is still getting a few views. I'm working towards being able to film again. Since I made this one, I've moved. Hobbies have had to take a backseat to other parts of my life, but things should start settling down soon.

Thanks Chuck, glad you remember the old office benches.

Thanks for watching and for the comment!!

More to come!

Yeah, you're going to need to include something in the photo to establish scale (say, a coin or something else of known size). But it can be done with a cell phone camera.

Thanks Benjamin, I appreciate the encouragement!

Hey, thanks! I feel honored to be the recipient of your first RU-vid comment!

Cool hack, thx for sharing!

It was a lot of fun to make! Actually, I'll be working on it more today. I upgraded to a raspberry pi 5; looking forward to seeing how that changes the performance. Plus I want to make some code changes.

Good luck, and thx for the kind words! Concerning Python: your local public library will probably have some good "getting started with Python" books. That got me started and I was able to hobble through enough code to make the shadow board box from there. Please keep me posted!

Thank you, I used to be a tooling manager at several aircraft manufacturing facilities. I got really frustrated with typical shadowboard design & fabrication processes so I came up with what you saw me demonstrate in the video.

it definitely peaked my interest! I went through the whole "cell phone, inkscape" process, and I gotta say - I'd much rather build the shadowbox and use that process. Genius!

Thanks for noticing that. This is one of those projects that builds on lots of other skills, and it would have been really easy to go off on tangents in any number of places while documenting this process. Appreciate the comment!

Thx!

Yep.... Not like working on cars!

100%

Thanks Riggsron, I appreciate the feedback. I really enjoyed making that video, I have another topic in mind fory next one. Just need to find the time to actually make it.

Welll, feedback seems to be fairly positive. Guess that's my signal to start working on Part 2. Thanks!

Thx!

Thx!

Thanks for the comment! Truthfully, I don't have any training in this area either. I watched a lot of RU-vid videos and played with Python a bit. I never went to school for any sort of CNC or programming education.

Yeah, I enjoyed working for Vycom. I left almost a year ago, went back into aerospace. Though I do very much miss access to scrap HDPE...

Thx! It's a work in progress. Got a few ideas for improvements, but I have no idea when I'll have enough time to play with it again.

Thanks!!

You're welcome. Glad it was helpful.

I should charge Vycom for my marketing!! Thanks for the feedback!

"Stuff Made Here" (I think his name is Shane) just did a video on a robot that measures puzzle pieces. It also uses a camera and a backlit light. Shane's significantly better at this stuff than I am, so his video might give you some ideas as well.

Thanks Pat! Great to hear from you again!!

I'm thinking that I'll do another video on a different method. One that doesn't require Python scripts or a CNC. Standby for that!

Good luck! Post some pics if you complete the effort and share the link!

Ha! I haven't even finished my own toolbox yet!

Why thank you!

Thanks, it was a fun video to make.

Thx!

Thanks!

Good luck! Don't forget to post back here and let me know how your version of this project worked out!

Thanks! I appreciate the feedback!

I've been told that before! Thanks!

Thanks! It's though to decide what to do with that workflow. On one hand, people seem interested. On the other, I'm absolutely a novice programmer and I don't want to present bad practices (and I know that I my code could be vastly improved)

@@scottprints Very cool video! I'm also very interested in the build of the image processor.

I would love to see that as well.

​@@scottprintsHow about someone that works in Python every day? I'm getting close but I'm missing something. I could help clean up the script you're using. :)

@@StellarEclipseStudio I think I'll take you up on that! How do I contact you outside of RU-vid?

Agreed! That would be awesome for all kinds of stuff!!

Came here to say the same thing, I need one of those light boxes In my life

Thank you, it's always fun to see the boards that you put up on LinkedIn as well!

Thanks Josh, I'm planning on doing a video about a less technical version of this technique. I think that learning Python, CAD/CAM, and access to a CNC is a fairly formidable barrier for most people that watch this video. But there are some other methods that can create the same effect. Hope to be able to work on that video soon. Thanks for the comment!

Thanks! I plan to as time allows.

I wish I had the time! Would love to make a video like that, but I have had 10 minutes to call my own in months. Thanks for the suggestion though!

@@scottprintsthere are definitely people who would value that information and would enjoy watching! I subscribed in hopes I see that video pop up on your channel soon.

@@tomaschacon4275 Hi Tom! I'd love to make a detailed video about the camera system, I really would! Unfortunately, I just don't see a lot of free time in my schedule anytime soon. Thanks for the note; if a few dozen hours worth of free time ever falls into my lap, I'll make a camera system video for you. Take care!

@@scottprints No worries, I understand how busy life gets! If you get a free chance, you should post a few pictures of the internals especially wiring and post them on your community page. I get a full video may not be in the cards but some photo references would mean a lot. Thanks again, awesome video!

Thx!

Thanks for watching! I appreciate your input, Josh!

You're quite welcome, thanks for the feedback!

Thanks for the comment! Yep, we cut this stuff at work, and it isn't terribly hard on the tooling. Hey.... I've been wanting to ask about doing this on a flatbed scanner. How does the profile capture work with tool edges that are a couple inches away from the glass? I wanted to re-create this contraption with a scanner, but I'm worried that thick tools will be out of focus if parts of them aren't right on the glass.

@@scottprints I dont think I can say that I have done anything 2" thick. If it had straight sides I think it would not be a problem since as you know th scanner bar is always looking straight up iff you will as it moves along the bed. I have done lots of cardboard patterns for things that I want to profile and cut in more expensive materials like carbon fiber and such where you really want to get it right the first time. It has always worked perfectly. Lately I have been tracing around wrenches, pliers and the like for foam inserts. That is what brought me to your video. What my method lacks in refinement however is the fact that I have to bring a PDF of each tool into autocad and then manually trace around the object that i want to shadow with a polyline drawn manually. It is a tedious process and it can take a few hours to trace around a whole tray of wrenches. Sometimes I profile cut reliefs when I want it to go all the way thru the 13mm EVA foam, and other times I cut relief pockets in the foam by laser engraving partially thru the foam for shallower stuff. I then stack up 1 to 4 layers of the foam with the corresponding cutouts and glue them all together. Again it is not a a fast process, but it sure does look amazing when it is done. I use 24" square EVA foam gym floor tiles from home depot or Amazon as it cuts quickly and beautifully with no residual stink on the CO2 laser. Again thanks for the inspiration!

Thanks!

Good to hear! What sort of projects are you working on?

@@scottprints I'm currently in tech school during my highschool school day. Me and my other fellow classmates are working on a shelf with 6 casters on it 4 of them being locking and the shelf's have rubber on them. We are building it for Aurora flight sciences. They have cool cnc machines, fixtures, robot machine tending, paint booths it was really cool to tour there facility they even have a computer inventory system that is like a vending machine but for tools and PPE... We finished the shelf but still have to mount the shelves on the beams which at are a 45 degree angle. There will be 4 shelves on it to hold airplane rotors that Aurora flight sciences manufactures. I'm still learning manual and cnc machines at my school and I find it very interesting. I like making things and I learn something everyday in this industry I love it.

Hello Mike, If I were you, I'd back into the process from the end. So start with your laser and cut a piece of foam with it. Can you make a shadow board? If the answer is "yes, but it takes forever" then you're ready to start improving the other parts of your process. That's how I did it, I started cutting foam with a scroll saw and constantly looked for ways to take the pain out of the overall process. If you have a CNC of some sort that can cut a shadow board, then start looking for ways to automate. A next logical step might be to map out your tools in some CAD software (semi manually), then after your team is comfortable with that process, build a rig like the one I demonstrated so that the silhouette generation process gets easier. Hope that helps, let me know if you have more questions.

Thanks! It was fun to build.

Somebody finally noticed the humor!! Congratulations Bob, you're the first! And shops are fun; highly recommend that every person should build their own 'me' space someday. They're great!

Thanks!

I'm impressed that you recognized the Pro-Lift mat! I've cut up a few of them. Too bad they're so small, I wish they came in bigger sizes.

@@scottprints haha yeah -- I've been down this road a few times! I actually ended up stacking gardening knee pad/mats from the dollar tree for my shadow boards -- as it cuts perfectly with a razor. The pro-lift it's very economical for what it is -- but it is so difficult to get a clean cut -- --- I even chucked a nice endmill in my hand drill and it cut fine -- but it was tedious and hard to get it to look right :) It's actually pretty nice for rolling around on the ground tho :) It's a shame they don't make them wider -- finding bulk foam like that is $$$ -- just goes to show economics of scale matter!!

@@IroncloudMetalworks I have a theory on why bulk foam is hard to obtain. I could be wrong, but I think that shipping costs destroy the economics of good shadow board foam. It can't be terribly expensive to produce; but the shipping costs are probably why it's so hard to find.

It's all Raspberry Pi hardware. You can find separate tutorials for hooking a pi to a touch screen, input buttons (switches), camera, etc. The scanner is just following all those separate instructions while making the wires long enough to fit into a box. That's pretty much it.

I wish I could! But I have no time right now. That CNC you saw me operating in the last video is in about 200 pieces now. I need to find a new workshop and get my equipment set up again, and that all has to happen in the scraps of time that aren't taken by other priorities. I'd love to make more videos, but I just don't see any room in my schedule for the effort. Thanks for your interest though, it's always nice to get positive feedback about my builds.

You're welcome!

Hello Nils, Thanks for your note. Unfortunately, I'd rather not distribute my code. Couple reasons for that: I'm not a great programmer myself, and I know that if I distributed my code I would be inundated with requests for assistance from people that were having trouble adapting my poorly-written code to their own use cases. Second reason is that Potrace requires a license if it's going to be used inside of a proprietary program. While my script is for hobbyist use, I'd hate for someone else to use my script inappropriately as a basis for a commercial program that might come back and bite me later.

Classic Games, circa 1960! Great chess set!

Thanks!

Good luck, keep me posted on your progress. No, I don't have my scripts posted. Mostly because I'm a novice programmer and this was a learning experience for me. Consequently, my script is poorly commented. While it works, it could be MUCH better and I know that I'll have to re-do it soon. Cost of the material is always a challenge. As I said in the video, I'm lucky enough to work in the factory that makes Playboard. I was allowed to take home some scrap for this project. If I didn't have that resource, I'm not sure I would have been able to make the video. Thanks for your comment!

​@scottprints , that's an amazing video! Your scanning device is a brilliant idea! I was wondering what are the dimensions of the back light surface (what's to biggest tool you can fit in)? Speaking of dimensions, how do you define the distance between the camera and the tool? What's this distance value in your case?

Ending scripts, starting scripts, and looking cool. They came in a 5 pack, so I decided to wore them all in.

Thanks! I have used my CNC router to cut foam before. It works, but dense foams work much better than soft foams. I think Kaizen foam would be OK to cut, but might be pushing the limit on "too soft to cut with a router". Dunno... only way to know for sure is to give it a shot. Appreciate the subscribe! Hope I get some time to make another video someday. Life's been pretty busy lately.

@@scottprints "Have you been busy?" that is a joke from a band named 12 foot Ninja on a song called One Hand Killing and they do a bit of comedy at the intro of the video... pretty funny but if you don't know them it may not be... Good song also if you like modern progressive rock music. I also have been busy, got a new job a year ago and turns out this squad works massive amounts of overtime and Saturdays and right through some of the holidays!!! It sucked but I learned a lot and got a different position where we only work a bunch of OT! But better than not having a job. But I did literally noting other than work for the entire year so it was rough. Anyway thanks and I'm looking forward to seeing more of your videos. Cheers

My masters is in 3D printing

Who's Dustin Hoffman? Anyhow, thx for the kind words! I made that video more than a year ago and I'll bet that those two drawer inserts have saved me hours of "where did I put that (insert tool)". Probably hasn't been a compete wash yet, but the drawer inserts are still really helpful.

Hi. Sorry, all that I have documented is what I presented in the video. A few other people have asked for more info, but I'm not sure how to pull it together. Just the video you saw was a lot of work; I'm not sure I could cover everything in depth without neglect other parts of my life.

At the moment, those scripts are so cobbled-together that I think they'd be more trouble than they're worth. Maybe someday if I re-write them, I'll put them up on GitHub or something.

@@scottprints Even documenting the steps a bit more would be valuable. I can write my own Python, just could use a head start on the process.

@@StevePrior I agree. It's better to have a jumbled toolbag of code to start out with and then polish than to have nothing to start with.

It's the way to go! Do me a favor: post back here after you've had the conversation with your boss. I'm curious to know how it turns out. Are you in aviation?

@@scottprints will do! No I'm in industrial automation. Ever heard of Festo?

@@jpwd1999 no, but I'll head to Google right now. I'm all about industrial automation. Stated differently, I don't know anything about industrial automation, but I think some Python and a few sensors placed strategically around my factory might be super useful.

Do you have a more specific question? I did everything that I could (short of posting my script) to convey the process in my video.

@@scottprints Short of posting the script, it would be most useful to me to understand the libraries you used to automate the workflow of photo to usb stick at the push of a button, so I could have a go a coding something similar. I've taken PiCamera and PyPotrace from the video, but have little understanding how you isolated the tool from the background. Although I get the theory, not sure how you produced that in code.

@@__stuart__ sure! Use PiCamera (or PiCamera2 nowadays) to make a 3 dimensional array in NumPy. Then you can use some match functions on each pixel to determine its distance from either the black or white corner. Once you have that distance, each point in the array can be categorized as either "tool" or "background" That ought to get you started! Good luck!

@@scottprints Just wanted to loop back and close this off. I've got some code that I'm happy with for now (these things are never "finished", right?) to extract the tool from the background using PiCamera2 and OpenCV, now just need to find some time to have a play with Fusion360 and the CNC. Thanks for the pointers - really useful starting point.

@@__stuart__ Code is never finished. Good luck, and keep me posted!

Thanks for the feedback! I really don't have any additional info past what I put in the video. If you have a specific question, might be able to point you in the right direction. But I don't have a detailed "how-to" guide for this project.

Really love the photobooth, and the video!

If I were you, I'd start by working through the software as 1's and 0's are free. Get started with your language and hardware of choice (I used Python and Raspberry Pi products) and build out a working model on the bench. If you can get that to work, then go bigger and get a box to put it all together with a backlit stage. Good luck!!

Thanks for the feedback; but I'd rather not share my code anymore.

Well, the trick is in how the cube is broken up. I find that I get better results if the split is closer to the white corner of the cube. That way, the pi has a better chance of labeling shiny tools as "tool pixels" -otherwise shiny pixels can get mistaken for background pixels.

But yes, the black lining is a good idea. I've been toying with the idea of painting the inside of the box black.

Hi Sam, Scale on a tool of known size and remove the barrel distortion will get you most of the way there. Parallax error is still an issue, but keeping the camera as far from the tool as possible minimizes that error.

⁠@@scottprintshow to counter the barrel distortion using a phone? Taking overhead pictures using a white background with lights to minimize shadows? Thanks for the video, info and your time

@@CisnerosShop well, I think "how" you take the photo (in your case, with a phone) is irrelevant. What's important is knowing how much barrel distortion is inside of your images and then removing it. You may want to do a search for "openCV barrel distortion correction" for some examples that show Open CV being used with Python for this function. Then you can take that knowledge and apply it to your project.

Hello, Yes, that's pretty much the Python workflow. I prefer to have Potrace output .dxf files, but the .svg or .eps formats you mentioned would work just as well.

@@scottprints 1 question.. do you have all the pictures on 1.dxf or are those multible

@@hrykne6134 I have the script spot it out both ways, generating each .dxf individually and one master .dxf. that way I can use whatever output I need and send the rest to the recycle bin. Gives me more flexibility.

are they useful tuts for the opencv part, I want to first rotate ä distortion operation but just finding those chess board tuts with cam adjusting...

In dense toolboxes, it can make sense to stand tools on their sides. That usually means putting them into deeper foam or getting support gizmos to hold them in place. There's also a lot of influence from the sort of environment the tools are in. I've had a few guys with automotive backgrounds mock my "overkill" because this just isn't needed in an automotive environment. But guys that come from aviation tell me that I haven't gone far enough.

Could be an SVG if you want. But I use a DXF. Both would be suitable for this sort of application.

Another question, about the voltage converters. You have two, what are they used for?

@@ammonsequiptment9877 pi runs off of 5 volts, the other changes the voltage going to the LED's, which is how I change their brightness.

I built it for about $3000 in materials.

Hmm... I think that looping over every pixel in your image is causing your slow code. Try setting up your image in an array (say, with numpy) and do operations on the array instead of individual pixels. Hope that helps. I'm not a terribly accomplished programmer myself.

this I could try... especially since picamera2 (libcamera) does provide an output in an array 🎉... and did you automate the scaling? I also did object recignition with a ref and scale it by converting with potrace but this also gets sticky😅

@@hrykne6134 all my scaling variables are fixed as my setup has no distance variables. So I don't need auto-scaling.

so, I nearly finished..😅 did you do any optimizing on the BMP files? what threshold number to decide wheter pixel go white or black in the BMP file? I am asking because I not already happy with the lines of the tool in the bmp file.. they not that smooth...

@@hrykne6134 depends on the intensity of the backlight. I made it a variable and change it if it seems to be giving me bad outlines. You're going to have to set up strategies: shiny tools will need a different set of variables than flay ones, etc.

Yup. Need to design that drawer next.

I think that exists actually. I can't remember what it's called, but some company has a business model that's similar to what you just described.