There used to be an old hardware store in Atlanta that had a loose hardware section. They also had a large trash can full of all the fasteners that strayed from their boxes. It would have been a challenge to buy and sort that.
Honestly, the best thing about these videos and your designs is how you're not willing to accept such high error rates or deficiencies in design even when the model "does the job". So many model creators out there wouldn't balk at an 8% error rate let alone doing error rate analysis altogether. But when your model is going to be used by hundreds or thousands, the extra 6% means you're wasting hundreds of third party man hours. Good job
@@EmGi_ I thought of similar a few years ago. Idea being a simple connector that could be mass produces by injection molding and then used in 3d-printed parts.
I truly enjoy your problem solving approach. It's what we do as engineers. To continue with the fun, you could sort by weight, instead of shape. Your current approach is based on the principle of a sieve. Since you are sorting scaled shapes of various lengths, made of the same material, one metric that is reliable and shape-agnostic is the weight. To this end, you can use one of several mechanisms but for a zero electronics system centrifugal force separation is a common practice. Once you have separated the bulk (in the pharma industry this is done with spinning grooved shallow cones) you can then proceed to run your item on a channel that has trap doors calibrated per weight. You can run the same bulk through the circuit until all the items are sorted. Another approach used in the pharma industry is to run the bulk in a single layer (easy to do) on a "sorting disc" that has little cavities that can only be occupied by a specific pill size and shape. It works very reliably. In your case because you have a bolts head and the item shape is anisotropic, you would need a more permissive cavity width-wise, but it would still work.
Interesting approach. Havent thought about a Weight activated trap door. might be worth a try. Curious if its possible to get such a high accuracy with a 3d printed trap door
To solve the standing up screw problem you could make the shield a ramp widening towards the top so that the screws tilt back falling into the reservoir again. This should not affect the other screws.
Really enjoy your series on bolt sorting. You have had way better results than me. Like others have said, rotating vibrating tray is the first step, just like the industry.
I just discovered your channel. This is such an amazing project. Your unstoppable attitude is so encouraging. So much is going through my mind right now. Thank you for sharing this idea. Your connector system with the tool is also so clever.
Hi EmGi, Love the video and your designs. I have two suggestions which may reduce the error rate while sorting; 1) introducing a saw tooth pattern on the higher wider section of the red bends, this would induce a bounce toward the narrow lower part which would reduce the chances of a smaller bolt sliding on the lip 2) induce vibration on the red filters this could reduce the odds of smaller bolts riding to the end and keep the hex head bolts moving
I was going to post these suggestions as well. I think a slight speed bump might help. Small enough for it to pass over naturally, but big enough for it to "jump" and cause some movement.
So you could add a angle on the inside of those sorting blocks so instead of flat ledges they are angled like this / but at about 45 degrees so the the ones small enough get led guided through and dot have flat ledges to worry about and then for the bigger screws it will still be small enough to catch the screw you want
Random thought: instead of a smooth down ramp, embrace the turbulence by switching to sloped steps forcing the bolts to drop and giving more chances to align or drop through. As for the standing bolt extending the overhang down at the same location and angle of the top lift hill may work to drop them back down into the starting basin provided the bolts aren't so short to get under it already
you could try using the smooth print plate to take away the layer lines and drop the flat part into a slot with the shiny side facing where the hex bolt was getting stuck
maybe for the curved ramp you could remove the right angles, so instead of a shelf it is a tapered channel. this should prevent bolts from being able to "ride" the flat surface...
Great stuff! a bit efficient space wise would be even greater! for example snap lock of "filter blocks" to containers or some kind of hinges to make it smaller while not used
I notice that top faces of the sorting ramp appear to be flat. What happens if you give that a V profile? the shorter bolts should still swivel into a vertical orientation.
I’m just curious what’s the difference between the three levels of support? I do not mind supporting you, but I just can’t tell which level I should choose.
On youtube there is no big diffrence between the three levels. All of them will get you a mention in my youtube videos. Its mainly for people who want to support me and the work im doing
I started Modeling with Fusion360 but then switched to Blender (as im already experienced with blender and found my workflow im faster in blender). The animations and simulations are done in Blender as well
@@EmGi_ oh woah that is super cool! I just got a 3d printer and i chose fusion 360 (personal use) as it seems helpful to learn. but your visuals are SO GOOD!!! and the simulation too... i wonder if i should look into blender as well thanks for the response!
Your feeding mechanism is your problem. I thought about it for a min or 2 and I would use the gravity by spinning the screws around, forcing them to be side ways... The solution was obvious, make a rotary feeder to align all the bolts in the same direction. By searching "rotary feeder" I found a short of an already existing factory machine doing it (near the end of it) th-cam.com/users/shortsNnifJspPjWk . You could do multiple stairs for all the diferent kind of bolts. From there sorting them out by length is a breeze
I think your initial approach could have worked if you didn’t make it ramp and just let the pusher advance them along so they moved at a constant speed instead of accelerating. Just some armchair engineering. Great video.
“Now that I’ve created the problem, it’s time to solve it using 3D printing.”
Absolutely legendary XD
Don't forget, it's always for 10x the cost and 2x the time than otherwise.
The sheer dread I felt when you mixed all the bolts xD
yeah that hurt lol
It hurt my physically
Same
There used to be an old hardware store in Atlanta that had a loose hardware section. They also had a large trash can full of all the fasteners that strayed from their boxes. It would have been a challenge to buy and sort that.
Hmm, the 10 second trivial fix isn’t feasible, might as well do a thesis level engineering project. You sir, are a true engineer.
Honestly, the best thing about these videos and your designs is how you're not willing to accept such high error rates or deficiencies in design even when the model "does the job". So many model creators out there wouldn't balk at an 8% error rate let alone doing error rate analysis altogether. But when your model is going to be used by hundreds or thousands, the extra 6% means you're wasting hundreds of third party man hours.
Good job
I really like the circular lock thing.
im planning to improve it that concept and maybe create a universal part connection system out of it
@@EmGi_ I thought of similar a few years ago. Idea being a simple connector that could be mass produces by injection molding and then used in 3d-printed parts.
Ever thought about building a system that uses vibration. It seems like most industrial sorters use that. Nevertheless, really like your videos!
i watched a few videos to research how they work. but i dont remember why i didnt go further on that route
That was my thought as well, then I realized that maybe this device doesn't need an even louder feed mechanism 😅
It's not just vibration. It's micro throwing and that's what makes the system quite complex. 😊
I truly enjoy your problem solving approach. It's what we do as engineers. To continue with the fun, you could sort by weight, instead of shape. Your current approach is based on the principle of a sieve. Since you are sorting scaled shapes of various lengths, made of the same material, one metric that is reliable and shape-agnostic is the weight. To this end, you can use one of several mechanisms but for a zero electronics system centrifugal force separation is a common practice. Once you have separated the bulk (in the pharma industry this is done with spinning grooved shallow cones) you can then proceed to run your item on a channel that has trap doors calibrated per weight. You can run the same bulk through the circuit until all the items are sorted.
Another approach used in the pharma industry is to run the bulk in a single layer (easy to do) on a "sorting disc" that has little cavities that can only be occupied by a specific pill size and shape. It works very reliably. In your case because you have a bolts head and the item shape is anisotropic, you would need a more permissive cavity width-wise, but it would still work.
Interesting approach. Havent thought about a Weight activated trap door. might be worth a try. Curious if its possible to get such a high accuracy with a 3d printed trap door
"Now I have created the problem that I am going to solve"
Damn, yoinking ideas from Apple now! XD
At first my anxiety of seeing mixed bolts filled my soul, then the wonder of your solution healed me!
To solve the standing up screw problem you could make the shield a ramp widening towards the top so that the screws tilt back falling into the reservoir again. This should not affect the other screws.
Yes!. And print it sideways. So the print lines go with the direction movement of the screw
I love the mechanism to extend the pusher reach!
Really enjoy your series on bolt sorting. You have had way better results than me. Like others have said, rotating vibrating tray is the first step, just like the industry.
I just discovered your channel. This is such an amazing project. Your unstoppable attitude is so encouraging. So much is going through my mind right now. Thank you for sharing this idea. Your connector system with the tool is also so clever.
Thank you! appreciate the kind words. Im planning to further improve the connector system and publish it so everyone can use it for his own projects
Could you add a little foot to kick the device every 3rd cycle or more or less frequent
Great project and result. I'm intrigued by that 3D printed joining device.
You've got me in tears man. 2:45
Hi EmGi, Love the video and your designs.
I have two suggestions which may reduce the error rate while sorting;
1) introducing a saw tooth pattern on the higher wider section of the red bends, this would induce a bounce toward the narrow lower part which would reduce the chances of a smaller bolt sliding on the lip
2) induce vibration on the red filters this could reduce the odds of smaller bolts riding to the end and keep the hex head bolts moving
I was going to post these suggestions as well. I think a slight speed bump might help. Small enough for it to pass over naturally, but big enough for it to "jump" and cause some movement.
So you could add a angle on the inside of those sorting blocks so instead of flat ledges they are angled like this / but at about 45 degrees so the the ones small enough get led guided through and dot have flat ledges to worry about and then for the bigger screws it will still be small enough to catch the screw you want
Great engineering skills. Really enjoy watching your videos.
Random thought: instead of a smooth down ramp, embrace the turbulence by switching to sloped steps forcing the bolts to drop and giving more chances to align or drop through. As for the standing bolt extending the overhang down at the same location and angle of the top lift hill may work to drop them back down into the starting basin provided the bolts aren't so short to get under it already
you could try using the smooth print plate to take away the layer lines and drop the flat part into a slot with the shiny side facing where the hex bolt was getting stuck
So cool! Would love to see an improved countersunk version and a button head version. I’ve got lots of drawers waiting to sorted… 😄
I wonder if the ramp at the end had a few steps if it would help in jarring the smaller bolts loose/ allow them to fall out better
Hey, Very nice video, good problem solving.
What I recommend is to use pitch angled gears to smoothen the process, reduce noise.
That you even come to the point that you have to create problems to solve says a lot about you.
If you chained multiple modules, you could sort all the sizes at the same time. You'd need multiple motors though, or just one with more torque
Hey at school I had the chance to make one with an industrial camera and a robotic arm and even with that it was a huge pain in the a** 🤣 good job !
Man, that first minute was hard to watch. 😅
Glad you have a solution to fix the problem now👍
OK, I joined your Patreon. I don't even need any of your things but they are all very cool and I love your attitude.
Thank you! Really appreciate the Support
Now that you've solved one problem, you've created another in that people will just mix up their bolts to watch this thing sort them 😂😂😂
maybe for the curved ramp you could remove the right angles, so instead of a shelf it is a tapered channel. this should prevent bolts from being able to "ride" the flat surface...
1:19 this upsets me in places that no person should be upset
Subscribe or i will do it again
You are a genius, thank you very much, great video editing.
Great stuff! a bit efficient space wise would be even greater!
for example snap lock of "filter blocks" to containers or some kind of hinges to make it smaller while not used
1:01 broke my heart
1:00 that scared me and broke my heart 😱😲
Now just a countersunk head screwsorter and I can clear out my other box of random screws! That one contains a mix of all kind of screws
12:25 I’d be interested in the fastener you use there being a separate file we could download and use
Soon. I want to improve that concept before i Publish a standalone Version
Please DON´T TRY MIXING SCREWS AT HOME!!! ....is so dangerous...EmGi is a PRO!!
Certified screw Operator
What made you decide not to use a vibrating bowl feeder like has been the way for decades?
super belle réalisation comme toujours
How do you do that neat simulation?
All done in blender
You should make a tutorial on how to design those circular locks, as long as you don’t dont consider them a company secret of course.
I want to improve them a bit before publishing them as a standalone version
what if you split the device into stages? first stage separates by thickness, next stage - by length?
very impressive , what is the SW you use to perform the simulation ?
I used Blender 4.2
@@EmGi_ 我喜欢你的作品。你提到的Blender 4.2我将在完成手上项目后学习它。
Somebody show this guy the fillet and chamfer tool....
LoL
I notice that top faces of the sorting ramp appear to be flat. What happens if you give that a V profile? the shorter bolts should still swivel into a vertical orientation.
Excellent work
Thank you!
This is amazing, did you go to school for engineering or is this just a hobby?
What program did you use for the physics simulations?
Blender 4.2
What software did you use for the simulation?
Blender 4.2
Great project and awesome video! New subscriber ;)
What was used for the physics simulation?
Blender 4.2
Add a champfer (60° fase) to the edges. This should guide the bolts
How many times did you need to sort the bolts while building the bolt sorting machines? :D
The circular part joiner feels backwards. With screws and bolts you turn it clockwise to tighten, but your joiner you turn it counterclockwise?
Good point 🤔
Very interesting video, thanks 👍
1:16
My brain was screaming
I’m just curious what’s the difference between the three levels of support? I do not mind supporting you, but I just can’t tell which level I should choose.
On youtube there is no big diffrence between the three levels. All of them will get you a mention in my youtube videos. Its mainly for people who want to support me and the work im doing
5:25 I have 3 related questions! what software do you use for your 3d modeling? your (amazing) visuals? And finally, the physics simulation?
I started Modeling with Fusion360 but then switched to Blender (as im already experienced with blender and found my workflow im faster in blender). The animations and simulations are done in Blender as well
@@EmGi_ oh woah that is super cool! I just got a 3d printer and i chose fusion 360 (personal use) as it seems helpful to learn. but your visuals are SO GOOD!!! and the simulation too... i wonder if i should look into blender as well
thanks for the response!
@@EmGi_ Holly, that animation is wild. How much time it took? Probably months. Thanks you for video.
World you try wood screws please?
The universal concept of entropy disliked this video 😂😂😂😂
1:08 this hurts my feelings
3:19 just make a car engine like piston long enough and stick it out the side and a few gears
1:01 NOOOOOOOOOOOO
Your feeding mechanism is your problem. I thought about it for a min or 2 and I would use the gravity by spinning the screws around, forcing them to be side ways... The solution was obvious, make a rotary feeder to align all the bolts in the same direction. By searching "rotary feeder" I found a short of an already existing factory machine doing it (near the end of it) th-cam.com/users/shortsNnifJspPjWk . You could do multiple stairs for all the diferent kind of bolts. From there sorting them out by length is a breeze
I think your initial approach could have worked if you didn’t make it ramp and just let the pusher advance them along so they moved at a constant speed instead of accelerating. Just some armchair engineering. Great video.
Here's another headache for you to think about - sort bolts by head type. Cap Head, Countersunk, Button head etc....
This is the perfect task for AI image recognition.