That is wonderful and I'm super interested in seeing more. I loved seeing the wheel bearing design stuff. You say you made it too complicated but I love seeing it :)
Very cool! You could probably detect the position of a train with infrared lights and receivers (or LEDs and LDRs or similar) and adding another set of magnets to the last car means you effectively increase the launch track by the length of the train. Really hope to see a part 2!!
Yo this is amazing. First off, you can import SVGs or DXF's into kicad as a footprint lolol. So you can draw up your coil in inkscape or illustrator and then import into kicad. Thats how a lot of art pcbs or just logos get onto boards, obvi usually the silk layer but you can do it as a copper layer. In the Knex roller coaster world, the creme de la creme for bearings is all ceramic bearings. They're somewhat pricey at like ~$3 a piece if you sleuth aliexpress, but man my knex coaster rips with ceramic bearings now. Alsoooo for thermals on the PCB, you should be able to do a ground flood of the entire PCB, essentially adding a copper flood to the rest of the board that isn't being utilized for the coil, thats attached to whatever trace that is attached to the coil, and that will help dissipate heat, but I actually don't know that much about magnetism/RF so that might mess with the coils. I considered doing this once for my knex coaster, and just went with a quadcopter motor/ESC and Arduino connected to a chain. Its just sooo much simpler and has lasted thousands of cycles with almost no maintenance: ua-cam.com/video/uckMLsE-Tts/v-deo.htmlsi=Mb8IQ5Yay7Uztjbo I'm subbed and hope to see a full layout!! Sick work man I know how much time all this shit can take, and then filming and editing is so wild.
Dude that is so cool. I'm working on mini LSMs for my model roller coaster as well, and some of your ideas I've never thought of. the PCB coils are super cool.
Since the launcher is magnetic, then it can also be used to catch the cars. Voila. Anywhere you want brakes or a booster you can also have booster and brakes! Such a deal!
Hold up. You launch with LIM, you slow down with eddy current breaking. 1) all you need for eddy current breaking is an aluminum or copper plate that will fit on both sides of your magnets. Cheaper and won’t give you headaches 2) Eddy current breaks don’t get as hot. 3) you can be lazy and not do as much math to get it set up. And 4) most modern roller coasters use eddy current breaking. They don’t require any power to work so if a coaster looses all power the train can still slow down. Eddy currents can’t stop your train though. They need motion to work. Which is why they will also have friction breaks to bring the train to a full stop.
This is a pretty sweet project, modified railgun projects are so awesome. If you do another video on this it would be cool to do a more typical railgun that way the whole track could be powered, but I know that would be difficult to implement
interesting, I was thinking about building a miniature version of the Tykrupp Multi elevator. It uses linear motor to operate and can go sideway. Could be another uses for your pcb coil :p
Key to a good linear motor is position tracking and field regulation. Some form of low-latency feedback instead of assumption by physical modeling of the train during acceleration could improve its performance greatly.
Put a hall effect sensor or an IR led/phototransistor pair before each coil right on the PCB and use that to control when each coil switches. No control board required, and you can chain as many as you want.
What an awesome project! I'd love to see a follow-up. I'm also wondering if a thicker copper on the PCB would reduce the resistance. Any chance you could share the 3D and KiCad files?
what kicad version are you using? kicad 6 added support for curved traces through the "fillet traces" command, later versions added new route modes that let you draw curves while routing, and kicad 8 allows graphic items (including arcs) to be associated with nets, allowing them to be used as traces while retaining the features of the graphic item. EDIT: pre-8, you could create your geometry as graphic items and convert it to traces with the "create from selection" tool
Well I wish v8 was out when I started the project! Filets may have still been a thing but it was just easier for me to generatively design the coils because of the constantly changing radii
In my opinion, one way to increase the effectiveness of the design is to maximise the coils self inductance : put a piece of steel betweenthe coils. anealed iron seemed to be a soft type ferromagnetic material, and that would boost the magnetic fields' strength. You could do the same on the magnets, the steel would redirect the magnetic field towards the coils. You will need some tests, but putting steel will help, which configuration is the best ? I don't know
It would be substantially easier to get feedback if your magnets were on the track and electronics on the cart. Two hall sensors interleaved like a quadrature encoder. Granted then you'd have to deal with making the coils more robust. But you wouldn't have to populate a track with PCBs, just magnets.. easier for curves. As it is you'd likely need to put hall sensors all over your boards, to figure out where the cart is in relation to phasing positions of its magnets.
I just watched a video about a launch, were the UA-camr switched to coils and the blade... So he attached a coil to the train and than added two springs to the sides of the train which touched the metal plates which he connected to negative and positive power, he just switched the polls every centimeter so the could got automatically powered according to how it speeds up. Hard to explain and can't find it anymore... The coil on the train would give it more mass too.
Yeah Tom Stanton’s? A very simple and elegant solution for sure. Has its own advantages and downsides. Having a proper coil definitely helps with power and efficiency, but also is a much larger setup and the only control you have is go or no go haha!
Yes you can! That does make controls a lot simpler but the problem then becomes getting that much power to the trains. Batteries add unnecessary moving mass and you’d have to still charge them somehow. Brushes would work but that impacts reliability and adds maintenance. Also it’s a lot of power to pass through brushes! This is all a big reason why rollercoasters have the coils on the track like this.
@@3DprintedLife and I just remembered that each coil can only be powered so long before it overheats so having a few coils that are constantly working won't work
Better performance yes, reliability no. The goal here was accurately recreating roller coaster tech tho so railguns are definitely out of the question haha!
Very nice. I can't help but think there must be something fundamentally wrong with the wheels though - probably the bearings as you mention - it kinda looks like they're gripping the track.
@@3DprintedLife Could do a kind of preproduction run with the intent to have others test and mess around with the design to help figure out different ways to optimize it
Epic job! Have you thought of simplifying everything by putting the magnets on the track and coils on the coasters? it will make the electronics much easier, power comes from the tracks or contact points, and alternating the polarity of the track/contact points causes the change of polarity in the coils.
I don't think all 6 wheels should be. touching the track at once. You need a tiny gap so that 4 at most are ever touching. Ceramic bearings would help a ton too. They don't need to be clamped, just rigid. gravity always keeps a set of wheels touching the rail.
Need to be, no. But I want a functional but accurate model, so all 6 should be touching. The added friction does slow it down a lot so I may try to find some sneaky middle ground, like having the wheels still be compliant but hard-stop them so they nominally don't all touch.
I mean...you joke but I designed the system electronics and coils to run at around 1kW peak output (500W per segment, 2 segments at most active). With a few minutes between launches to cool the coils off of course.
Don't trust LIDAR for something so precise. It might have a repeatability of 1cm in ideal conditions with enough time, but as the train moves that repeatability will go out of the window. Your original idea of having sensors at each segment would be much more reliable.
Yeah I definitely can't use this for real-time feedback. A laser displacement sensor was what I wanted to use but that was way too pricy. My goal will be to get those proximity sensors working and perhaps add a second on each segment so I can get constant train position feedback.
KiCad really is great and it's free and open source. Also their newer versions which I wasn't using at the time do apparently support curves as well as converting dxfs into copper fill/traces.
@@3DprintedLife What did you learn? How to work around a specific issue on an outdated version of kicad doesn't seem like a skill that will ever be used again. Is it the Python you got better at?
oh I'd absolutely watch a whole series of these mini roller coaster videos
I second this. I loved the knex sets as a kid and this video is really nice to see, I love seeing these complex machines in small scales
The project farm reference at 3:00 is quite genius
LED's were a nice touch! Also loved the PCB manufacturing + the build process. Thanks for some awesome work!
That is wonderful and I'm super interested in seeing more.
I loved seeing the wheel bearing design stuff. You say you made it too complicated but I love seeing it :)
0:50 love the ksp music
Username checks out
Very cool! You could probably detect the position of a train with infrared lights and receivers (or LEDs and LDRs or similar) and adding another set of magnets to the last car means you effectively increase the launch track by the length of the train. Really hope to see a part 2!!
Great video! Yes we want to see a full length video please!!
This is really awesome! I am thinking about building a 1:87 scale "Wilde Maus" and seeing your wheel assembly helped me a lot, thank you!
9:35 As an engineer, I can confirm that the neopixel leds are absolutely essential to both rollercoasters as well as linear motors 😎
The lights are the most important part! They add +10% speed!
This is incredible, a series of mini coasters would be great to see
lmao i loved the "we're going to test that" reference 😂
i'd love a full series! cool project and great work!
3:00 Project Farm reference! truly a man of culture
Yo this is amazing. First off, you can import SVGs or DXF's into kicad as a footprint lolol. So you can draw up your coil in inkscape or illustrator and then import into kicad. Thats how a lot of art pcbs or just logos get onto boards, obvi usually the silk layer but you can do it as a copper layer.
In the Knex roller coaster world, the creme de la creme for bearings is all ceramic bearings. They're somewhat pricey at like ~$3 a piece if you sleuth aliexpress, but man my knex coaster rips with ceramic bearings now.
Alsoooo for thermals on the PCB, you should be able to do a ground flood of the entire PCB, essentially adding a copper flood to the rest of the board that isn't being utilized for the coil, thats attached to whatever trace that is attached to the coil, and that will help dissipate heat, but I actually don't know that much about magnetism/RF so that might mess with the coils.
I considered doing this once for my knex coaster, and just went with a quadcopter motor/ESC and Arduino connected to a chain. Its just sooo much simpler and has lasted thousands of cycles with almost no maintenance:
ua-cam.com/video/uckMLsE-Tts/v-deo.htmlsi=Mb8IQ5Yay7Uztjbo
I'm subbed and hope to see a full layout!! Sick work man I know how much time all this shit can take, and then filming and editing is so wild.
Dude that is so cool. I'm working on mini LSMs for my model roller coaster as well, and some of your ideas I've never thought of. the PCB coils are super cool.
K'Nex coasters about to get crazy
I actually designed it to use the same size as the knex version…so you could 😅
Could you mention how your LSMs can move trains from standstill? iirc coasters have to be in motion for them to operate?
9:35 beautiful sound choice
I don't think I've ever seen a video more to my interests, lol
I would indeed love to see more on this project!
This is sick
Since the launcher is magnetic, then it can also be used to catch the cars. Voila. Anywhere you want brakes or a booster you can also have booster and brakes! Such a deal!
Hold up. You launch with LIM, you slow down with eddy current breaking. 1) all you need for eddy current breaking is an aluminum or copper plate that will fit on both sides of your magnets. Cheaper and won’t give you headaches 2) Eddy current breaks don’t get as hot. 3) you can be lazy and not do as much math to get it set up. And 4) most modern roller coasters use eddy current breaking. They don’t require any power to work so if a coaster looses all power the train can still slow down. Eddy currents can’t stop your train though. They need motion to work. Which is why they will also have friction breaks to bring the train to a full stop.
that's super cool, I love to see more linear electric motors from makers
In the early days, the Lego guys were reluctant to visit the theme park.
This is a pretty sweet project, modified railgun projects are so awesome.
If you do another video on this it would be cool to do a more typical railgun that way the whole track could be powered, but I know that would be difficult to implement
omg thats so cool!!!! I want it for when I build redforce!
I'd love to see more of this!
Awesome video as usual. Can't wait to see the next one!
This is really funny to watch. I'm doing my own PCB motor design right now and have been running into similar issues. Nice work!
pls do more of this this is so so cool!
Very cool, hope to see more 3d printrd rollercoaster videos
I'd love to see more videos about this!
Awesome project!
The king is back! nice build
Man this is so cool! I am inspired to copy you and try something on my own
is part 2 up?
Roller coaster tycoon was awesome
That's awesome! Nice work.
interesting, I was thinking about building a miniature version of the Tykrupp Multi elevator. It uses linear motor to operate and can go sideway.
Could be another uses for your pcb coil :p
You are a smart dude. And hard working. That’s a lot of cool stuff you did.
Thanks for making such a great video❤❤🎉🎉😊
finally a new 3dpl video
Key to a good linear motor is position tracking and field regulation. Some form of low-latency feedback instead of assumption by physical modeling of the train during acceleration could improve its performance greatly.
Put a hall effect sensor or an IR led/phototransistor pair before each coil right on the PCB and use that to control when each coil switches. No control board required, and you can chain as many as you want.
Waiting impatiently for when this kind of kit is available as an add-on for Coaster Dynamix sets.
that was such a cool build. Subbed!
Thanks for the sub!
Pls more Videos of the roller coaster!!!!
Great video
Very cool
It is a super cool project and I hope you will continue it! If I wasn't already subscribed, I'd consider doing it just to watch the follow-up :)
What an awesome project! I'd love to see a follow-up. I'm also wondering if a thicker copper on the PCB would reduce the resistance. Any chance you could share the 3D and KiCad files?
what kicad version are you using? kicad 6 added support for curved traces through the "fillet traces" command, later versions added new route modes that let you draw curves while routing, and kicad 8 allows graphic items (including arcs) to be associated with nets, allowing them to be used as traces while retaining the features of the graphic item. EDIT: pre-8, you could create your geometry as graphic items and convert it to traces with the "create from selection" tool
Well I wish v8 was out when I started the project! Filets may have still been a thing but it was just easier for me to generatively design the coils because of the constantly changing radii
Really cool
ohhh this is so cool!!! was the led launch pattern inspired by tron lightcycle run? in any case, super cool project and id love to see more!
Maybe unintentionally haha. The train itself was inspired by Velocicoaster
The video was so neat, and I was wondering, do you have a link to the power supply you are using?
In my opinion, one way to increase the effectiveness of the design is to maximise the coils self inductance : put a piece of steel betweenthe coils. anealed iron seemed to be a soft type ferromagnetic material, and that would boost the magnetic fields' strength.
You could do the same on the magnets, the steel would redirect the magnetic field towards the coils.
You will need some tests, but putting steel will help, which configuration is the best ? I don't know
Consider me obsessed 😍
1:08 ...or use an already existing plugin for it?
ksp music goes crazy
Super cool project. Have you tried Teflon as a low friction glide material?
Would be so cool if you were to make a desk version of the "shuttle loop" from the OG Roller coaster Tycoon
Now that is definitely achievable!
we need a serie
It would be substantially easier to get feedback if your magnets were on the track and electronics on the cart. Two hall sensors interleaved like a quadrature encoder. Granted then you'd have to deal with making the coils more robust. But you wouldn't have to populate a track with PCBs, just magnets.. easier for curves. As it is you'd likely need to put hall sensors all over your boards, to figure out where the cart is in relation to phasing positions of its magnets.
I feel like @ryantheridemechanic needs to weigh in on the restraint failures at 0:05... 🤣
Nice job!
Kicad has a curved trace add-on btw
👀 is this recent? I started this project a long time ago and couldn’t find anything then hah
I just watched a video about a launch, were the UA-camr switched to coils and the blade... So he attached a coil to the train and than added two springs to the sides of the train which touched the metal plates which he connected to negative and positive power, he just switched the polls every centimeter so the could got automatically powered according to how it speeds up. Hard to explain and can't find it anymore... The coil on the train would give it more mass too.
Yeah Tom Stanton’s? A very simple and elegant solution for sure. Has its own advantages and downsides. Having a proper coil definitely helps with power and efficiency, but also is a much larger setup and the only control you have is go or no go haha!
2 sponsors is crazy
would it work if you were to put the coils on the train and magnets along the track?
Yes you can! That does make controls a lot simpler but the problem then becomes getting that much power to the trains. Batteries add unnecessary moving mass and you’d have to still charge them somehow. Brushes would work but that impacts reliability and adds maintenance. Also it’s a lot of power to pass through brushes! This is all a big reason why rollercoasters have the coils on the track like this.
@@3DprintedLife and I just remembered that each coil can only be powered so long before it overheats so having a few coils that are constantly working won't work
Yes, just yes.
how is the plastic fixed to the brass tube?
superduperglue
hella cool.....
are you gonna make a coaster with this?
Nice, I got it
Nice project! My important question is what kind of iced coffee was that?
Asking the real questions. Salted caramel iced latte 😋
@@3DprintedLife Delicious
Storm runner mentioned??
You bet! One of my all time faves
Feels like a railgun launcher would have been better but hey! Good work is still a good work!
Better performance yes, reliability no. The goal here was accurately recreating roller coaster tech tho so railguns are definitely out of the question haha!
hi,
The spools, what is the mm³ of these coils and amount of windings? max voltage and amps??
Epic!
Very nice. I can't help but think there must be something fundamentally wrong with the wheels though - probably the bearings as you mention - it kinda looks like they're gripping the track.
Time to add a plc system and swing launch
What Ratchet & Clank Future song are you using at 0:51 and 3:37 ?
Frost Waltz!
While this is cool, the coolest model coaster on yt has drive motors that mimic the motion of a full scale coaster.
I can't be the only one who instantly heard the KSP orbit music near the start right?? (Is this just some royalty free track?)
Good catch ksp is great. Yeah it is royalty free music lol, frost waltz
Man i would buy this as a kit... hope you will start selling it
It was and still is a consideration…but would need a lot of optimizations first
@@3DprintedLife Could do a kind of preproduction run with the intent to have others test and mess around with the design to help figure out different ways to optimize it
Yeah, go on with a full series!
And hey, it's not like you can't launch nerfs in it, don't you think?, Ok at least put few bullets in a seats
You watch Project Farm, don't you? 😂
Eyyy glad someone got the reference!
Epic job! Have you thought of simplifying everything by putting the magnets on the track and coils on the coasters? it will make the electronics much easier, power comes from the tracks or contact points, and alternating the polarity of the track/contact points causes the change of polarity in the coils.
Yes definitely makes electronics easier, but that isn't how rollercoasters do it!
Looks like you may have found your project for 2025
I don't think all 6 wheels should be. touching the track at once. You need a tiny gap so that 4 at most are ever touching. Ceramic bearings would help a ton too. They don't need to be clamped, just rigid. gravity always keeps a set of wheels touching the rail.
Need to be, no. But I want a functional but accurate model, so all 6 should be touching. The added friction does slow it down a lot so I may try to find some sneaky middle ground, like having the wheels still be compliant but hard-stop them so they nominally don't all touch.
You can't not continue this to fruition.
Just hook up a microwave transformer lol
I mean...you joke but I designed the system electronics and coils to run at around 1kW peak output (500W per segment, 2 segments at most active). With a few minutes between launches to cool the coils off of course.
Don't trust LIDAR for something so precise. It might have a repeatability of 1cm in ideal conditions with enough time, but as the train moves that repeatability will go out of the window. Your original idea of having sensors at each segment would be much more reliable.
Yeah I definitely can't use this for real-time feedback. A laser displacement sensor was what I wanted to use but that was way too pricy. My goal will be to get those proximity sensors working and perhaps add a second on each segment so I can get constant train position feedback.
100 car per second centrifugal magazine launched coaster when?
your unnecesarily nice comment
WOW kicad sucks. Welcome to 1975 where CAD software doesn't do curves. Unbelievable.
KiCad really is great and it's free and open source. Also their newer versions which I wasn't using at the time do apparently support curves as well as converting dxfs into copper fill/traces.
@@3DprintedLife That's a lot of hours lost as a consequence of not updating software. Wow!
That version wasn’t out yet when I started this project a year and a half ago. And not lost hours if I learned something new from it
@@3DprintedLife What did you learn? How to work around a specific issue on an outdated version of kicad doesn't seem like a skill that will ever be used again. Is it the Python you got better at?