Это видео недоступно.
Сожалеем об этом.
Making a Jumping PCB Robot
HTML-код
- Опубликовано: 17 авг 2024
- My attempt at making a flexible PCB robot frog 🙃🐸 Help me make more projects by supporting me on Patreon 🠊 / carlbugeja
PCB Manufacturer 🠊 www.pcbway.com
PCB Design Software 🠊 carlbugeja.com...
Free Version 🠊 www.altium.com...
My Flexible PCB Actuators 🠊 flexar.io
This project is Open-Source and licensed with Creative Commons Attribution Share Alike 4.0 International License
Cameras/Tools I use (affiliate links):
Main Camera 🠊 amzn.to/30tRgY7
Second Camera 🠊 amzn.to/2ZLm2wy
Light Setup 🠊 amzn.to/3jtqckA
Magnifying Lamp 🠊 amzn.to/3jrcaQv
Vernier Caliper 🠊 amzn.to/3eTju3I
Multi-meter 🠊 amzn.to/3hq6B2R
Digital Milligram Scale 🠊 amzn.to/30HIw1X
Website 🠊 carlbugeja.com
RUclips Channel 🠊 / carlbugeja
Twitter 🠊 / bugejacarl
Instagram 🠊 / bugeja.carl
Facebook 🠊 / bugeja.carl
Linkedin 🠊 / carl-bugeja-0b922a135
Music:
Carmelized - RUclips Library
AI 2 - RUclips Library
Ella Vater - RUclips Library
Not Impossible - RUclips Library
This "failures" are more interesting and usefull than a hundred success videos. Thank you!
True
the beauty of engineering comes from the failures 🙃
Carl Bugeja pain and despair also come from failures :) Persistence and grit defy them.
I was bricking DIY micro RC receivers for a year and couldn’t figure out why. Until I found a workaround for what seems like a hardware bug in my laptop.
@@CarlBugeja give to me your wtsup number
True I mean another good example is electroboom he is also super deluxe funny
That's fantastic Carl! Have you thought of using opposing coils instead of permanent magnets? That might improve the power to weight ratio.
The mass of the magnet is what makes it jump
Never mind, I see what you mean. The battery mass is high enough that using just coils might work
thank you so much :) opposing two coils alone does not create a magnetic field strong enough to repel/attract them.. i have some tests coming up soon (next month or two) related to electromagnets that might give some more answers to this question
Or maybe the coils can load a spring with a ratcheting action and then it can jump like a flea, maybe using the coil movements to steer while in the air.
You might want to try putting a small platform in the centre of each coil (on the bottom) so that air can enter the space under the coil easier. Perhaps the small vacuum is holding it down?
Hi Carl,i'd like to give you an advice about jumping, idk if you tried that before , but i noticed that one of the major problem of your design (still awesome, I was never able to do something similar and I'm following your journey with a lot of enthusiasm) is the very low potential elastic energy that the robot can accumulate during compression, this due the low elastic constant of the pcb material. I know that a spring effect would require bigger mag fields and so bigger coils and mags, but why not trying to exploit a capacitor discharge in combination with some lp boost converter in order to feed the coil with short pulses of relatively high voltage electricity. In this way, enhancing the elastic constant of the pcb using , for example, aluminium planar springs, you should be able to achieve better results without changing too much the size of your robot. Hope to be useful :)
does it work if you have stiffer metal springs that are _also_ coils? Aka they are coiled into a roughly conical shape and when a current is passed through them they compress themselves against the magnet.
Easy and free good test you can try is to add a second layer of coils to increase potential energy. Use also the arms to be more stiff - you already have many PCB's. so, why not?
terrible advice
you are right actually i was thinking the same way but the problem is the design itself even though we add springs to it still this design at current stage cannot work harsh but true. just some magnets pulling and pushing against each other wont do. first thing is we need a lot of force to jump which means we need potential energy storage like proper springs or rubber bands will also do. next we need a way to pump energy into them like the PCB linear actuator carl designed in his previous videos. next we need a lot of design improvement. i have a suggestion for carl he can use the PCB LA design to pull the springs from all sides towards the centre straining the springs and suddenly releasing them that should create enough force
So if you think about physics of jumping, you can't get more energy out of external energy harvesting.
If you wan't a real jump, you need to meet the "sweet spot". In your case the moment when you can enforce the anti gravitational force in upward direction, when the inertia of the lifting motion is enough to pull the mass that is still on the ground.
If you think of when you jump, you push yourself up. Right in the moment, when you feel gravitation less, you pull your legs to yourself. This order is a case for control algorithms. The problem you disrespect with fixed/variable frequencys for proof of concept, is that the actual momentum is not linear. It is by the power of 2 depending on time, its linear but not straight.
I would love to see a version 2 (3)! Keep it up, you make excellent videos and you inspire me to graduate college (studying computer engineering, but EE was my first love)!
Underrated engineering channel
🙏
Do you considered air suction forces between coils and table ? Maybe try to run it on some kind mesh table (maybe speaker grill)
I would suggest to add some legs to reduce vacuum under the coils.
that could be effecting yes and it would be an interesting test to make.. but i still would like it work on flat surfaces to be more "usefull"
@@CarlBugeja stick something below very light that at least lets air rush in. Like 4 small polystyrene balls. For next version, make a hole at the center for air inrush. You can't fight against one atm.
I had similar thoughts to Norbert SadeXo.
curious if you could take some inspiration from origami frogs? folded tabs off the main body could help solve this, plus give some extra lift?
Nice project. Do you ever tryed to pump negative signal thru the coils. In this case you can try to increase the reliasing force on the magnets. Force will push further the magnets.
This would benefit form having Some feet as this will reduce the energy needed to jump due to less losses due to suction
Use superglue and baking soda to build some nubs for feet?
This so much, it seems to be stuck to the ground?
Or at least holes in the bottom PCB to allow some airflow.
This is so cool! The thing I would do is to double, triple, etc. stack the coils - the primary source of the filed/force. Also, don't think you need wide coil, but stacking as high as possible with low/similar to magnets' diameter. Thanks for the videos!
Yes please.
Also, I'd love to see a video on your smt soldering technique. I noticed you used some tape on your stencil. Is that just to affix it to the ocb, or was it for clearance adjustment when you spread the solder paste?
I did use tape for this video but its not the best way to do it.. maybe i will do a video on that topic someday :)
You are going to succeed Carl, your ideas and projects are absolutely awe inspiring and innovative as hell
How about using more of a centipede like set up with smaller multi layer pads, and a tiny bar magnet instead of the leapfrog and button magnets? A sequential firing of the pads can be used to mimic the insects walking pattern...
That sounds awesome, and more doable than jumping.
Aside from jumping, the fact that this robot was drifting can still be used to your advantage. A robot that can move around with no mechanical or moving parts can still be very useful in certain scenarios.
It's amazing seeing your robots evolve and improve with time. Keep up your good work!
I would love to see a working frogbot! 🐸 I really enjoyed seeing how you work with trial and error.
What you need imo are longer legs that are way less stiff, this allows the magnet board to accelerate longer before pulling the coils
People in 2000 : we will have flying cars in 2020
Meanwhile 2020 : jumping frog go brrrrrrrrr
That's pretty interesting so far. Maybe putting the heavy part _down_ instead of _up_? Would that allow it to get more travel (and thus accelerate) before it has to pull the whole mass along? As it is, it starts under maximum load, so it doesn't build up much momentum. Just making this up as I go along, so take it with a grain (or a shaker) of salt. Just a gut feeling. Or, could you make it tumble instead of hop? Make it a little more spherical in shape and then it could selectively shift the center of gravity by pulling the magnets to one side until it rolled over? Your ability to design and build tiny little intricate things like this always amazes me, and your obviously enthusiasm for the subject matter make your videos a real treat to watch. Take care, and good luck!
As they say, there is no success without failure, however, I like to think failures, are learning curves, If anyone stands a chance at making a digital jumping frog, it is you!
What happens when you make holes in the center? I think the foot area is so big that the outside air pressure makes it "stick" on the surface. You could also try a non flat surface where air can flow more freely, like a sponge or a wire mesh.
You could also give the robot skids to stay off the ground.
i'm not sure if making holes in the middle will resolve this issue but i agree that having a smaller foot area will yield better results
@@CarlBugeja It's all about better air flow underneath. It acts like a suction cup at the moment. I'm sure you'll come up with somethign great.
9:27 is full on terror mode.
This Frog is so cooool!!!!!! You are such a smart and cool guy!
I'm interested in seeing where this goes but man I don't expect much more of a jump with these magnets. My thoughts
1) The resistance to flipping magnetic fields is significant at this small a scale
2) Diminishing returns after 2 or 3 magnets/coils
3) You're losing some inertia to the dampening effect of the flexible PCB
4) The only way you're going to get the weight down with these materials is with fewer components and less soldier
5) Your idea might have some broader application at a larger scale. Build a monster sized one might give you some ideas :D
Most complex promotion of flexible PCBs manufacturer ever)
Hey! I love these projects of yours! Don't give up! As an aspiring robotics engineer myself, I'd love to see you making more versions and more bots!
i'd love to see a part 2
I think if you increase the number of coils instead of decreasing them, it has a higher chance to succeed. Don't forget that you have a first prototype that can jump more or less as you wanted(at least in height) with only one coil and no battery, so you know that one coil can compensate for the electronics that make it work, if you want to also have the battery onboard you need to increase the number of coils (maybe like 8 coils in two coils per leg configuration for example) so that it can compensate for the weight of the battery. Keep trying it is a fun and teaching journey for all of us :D Great work!
Good job man, i rly enjoy your videos. i'm in my first year as an electrical engineering student and i find your videos as being rly inspirational
Thank you so much!! I'm glad you like them! :) good luck with your studies
Yes pls try version 2!
Also, why use magnets?
Why not use more coils on top and run the current in an opposite direction?
You could reduce your overall weight this way.
Granted you'd need more current to run all your coils, for which you could design some sort of capacitive circuit that can give short bursts of high current.
This was recommended by RUclips and I’m so glad it did. This is so cool! Pls make a version 2
You Working Hard You Deserve More than RUclips
I like your persistence. Consider introducing a texture to the underside of the pads so it doesn’t have to overcome a vacuum when leaving the surface. Keep trying.
I think a big problem are the 'massive' and solid coils on the bottom, it is actually a solid plate, no air can go thru it. When it tries to jump it moves air that wants to push it down, a resistance. When released, same story, it pushes away the air between the plate and the table surface, again a resistance. It is now acting like a speaker cone. Maybe you can add some vent holes, air leaking points, that reduces the amount of air to move. You can test personally by taking a solid plate and try to jump with the plate. Almost impossible. After this take a plate with many holes in it and try to jump again. Much easier! Notice: I am not responsible for any injury performing these tests. :-P ;-)
Hi, you should make 4 flexible arms more rigid so when it releases from magnetic will take a high jump.
I think that maybe changing the design to have some springs between the body and the feet and then using the coils and the magnets to compress those springs. Then release the magnets and the velocity gained from the springs could help it jump.
I figured it would only bounce based on appearance alone; our ability to jump is amazing and takes a feat of evolution. It's one of our great advantages as human beings. Short of adding a spring, that thing is hardly going to be taking any substantial leaps. It's still fun to watch and see how adjusting frequency can produce such reactions. This video has definitely earned a like and save. keep doing what you do Carl!
You really are one heck of an expert in designing PCB's.I really want to learn to design PCB's like you
Remember that not all actuators/motors need permanent magnets. For this weight critical application, replace the magnets with top side coils. I also recommend investing time to find a higher energy density (watts per kg) energy storage system... supercapacitors have 2-20 times higher energy density than batteries
Hi Carl! Just stumbled to your channel and saw this frog project. You are very skilled and you have really grate and creative ideas! Keep going! Though there is a drawback in your designing process. You have to pay more attention in the physics of your projects. For example, here, you have a really large flat surface (quad coil system) pushing on your desk. This works as a suction cup and your system struggles to detach from the desk. Try to attach 4 polystyrene foam hemispheres under each coil. Also, try when you design the PCBs to take in to account the moment of inertia on the primary axis you will be using or at least try to spread symmetrically your components in order not to have to re-calibrate the whole system. Finally, for the specific project try to determine the "force" or the weight a single coil can elevate and how far (hang your magnet from a string right above a single coil and play with the coil strength and extra weight on the magnet). I hope that helps a bit. Nice videos! ;-)
You've got a few things going on there. I would place some air holes through your coil designs for starters, there is too much flat area needing a big air inrush for a snappy jump response. Also, I would find the jumping frame resonant frequency, run it at that frequency for movement and then slide the coils on intended motion axis oppositely out of phase from the other coils, each other for effective movement. Think more kangaroo than frog.
Thanks a lot for show the whole process...it really gives motivation
I think you could achieve directional control using just one centrally located magnet and three “steering coils” positioned a little off axis. Because the steering coils are not directly in line with the magnet, they will exert a lateral force upon the magnet and a corresponding movement in the opposite direction while jumping.
To reduce size and save weight, the coils can be made into wedges that meet in the center of the PCB instead of being circular.
As a further refinement, it might be possible to design a hybrid system consisting of one large outer coil to provide most of the loft for jumping wrapped around three smaller inner coils that add an asymmetric steering force as previously described.
Have you taken into account the airflow? Lifting the large footprint vertically up from a flat surface creates a pressure difference and a sort of vaccum underneath keeping the coils stuck to the ground. I would suggest drilling some air holes in the centre of the coils so that as the coils lift air can enter through the centre and not just from the sides.
I’ve done a terrible job explaining this but I’ll give another example. A piece of paper flat on a table... it’s difficult to lift it vertically up, but easier to peel it up wards from one side to allow the air to fill the space you are creating.
Either way try drilling an air hole in the bottom panel:)
Glue on some thin plastic stiffener on the legs. Trace out the bottom 4 circles and the legs. This shouldn't add much weight, but it'll give some rigidity and high spring tensions. You know those hard plastic packages where they melt the borders shut, the ones notorious for being hard to open. I forget what they are called. Or grab some thin vacuum forming plastic sheets and cut that out. Maybe some projector plastic sheets? Idk if they are still around :D
the 4 cables has a spring effect try changing material for harder springer effect, then the frequency will change too. It's working as a mechanical oscillator, you might get higher more powerful jumps.
Have you tried to increase the magnetic flux capacity of the coils themselves? Try using a ferrite paint on one side of a coil and see if that helps.
314/5000
a great project.
I think you have three serious problems:
1. the conductor track width, this limits the current flow and thus the magnetic field strength
2. the large base, perforate it so that the air can flow faster between the base and the board
3. As already mentioned, the spring force
Altium and Keil - good choise! Very interesting
Maybe the problem is the big flat surface of the coils. It air can t flow fast enough under the coils. So it creates suction. It is like you try to pick up something flat (like a sheet of glas) from a flat surface. The faster you try, the harder it gets. It may be an idea to make little dots under the coils or to fold them a little bit. This will reduce the amount of the planar area of the coils. I hope this helps
Problem ( At least what I think ): Motion is caused by an external force. The action and reaction of your system are contained in itself. Action -> PCB Coil is pushing the magnet upward:-> Simultaneously:-> Reaction -> The magnet is pushing the coil downward.......Both coil and magnet are part of a single system. It's like pushing your car from inside.
I think you should try with a larger disk magnet
One problem I see is that there is a "giant suction cup" on the bottom. Could be causing a huge loss due to the viscosity of air, the relative flatness of both the PCB and the surface it is on. You might try reversing the setup, magnets on bottom coils on top or to add standoffs (maybe headder pins) on the base.
yes please make version 2
increasing the number of coils may be beneficial. While the battery may also increase, the control circuit portion in the whole device will become smaller.
You could try if increasing the magnetic force helps. Add some metal to the coils. You could try to paint it with magnetic paint. It contains small metal particles. It will increase the weight of the feet, but metals are known to increase the field strength in the order of 1000.
The principles of this technology is very important, You could investigate how you can apply in the microrobotics. of medical applications.
Very interesting… my first idea if you do a version 2 of it would be to place a hole in the middle of the coils… at least in the space between the 4 coils. Reduce the air pressure on the base plate.
I think you need the legs to be stiffer rather than softer! If I understand correctly then the legs are a spring, the magnets tension the spring and then release to get a jump. If the springs are stiffer you'll get more jump from the same compression. The only issue would be that you'd also need more force from the magnets to pull them together. I don't know if it's possible but I think the best solution would be to match the force of the spring to the magnet so it can just about pull them together.
as the force of the magnets is non linear with distance you might be able to optimise the springs by making those non linear as well, for this I think you'd need the spring to be thick at one and and thin at another. You might find that is theoretically a good idea but practically too much effort
it looks like you could get height by controlling the coils more finely. fire one set, then the other set allowing it to lever forward a bit. ie kicks up on hind legs, then fires off with the front legs.
use a velcro pad at the base. This will help the vibration advance and slide on the road in one direction. you would only have to control the direction
Hey carl, here are the things that I would do differently for this particular design:
1. Use a piece of iron at the center of the coils, which would not only help in increasing the magnetic field but also will help with the weight distribution.
2. Make the length of the connector even longer to give more headspace for the magnets to move.
3. Use a higher grade of neodymium magnets.
I think you should try to change 2 things
1. you should harden a bit the legs to gain more elastic energy accumulated.
2. make the magnetic field switch reverse and add pushing on jump instead of just letting it go.
I'm sure these changes will give a better result . and of course you should do the next step
i think you should combine this with something like shape memory alloys for example nitinol or maybe design a actuator that can store potential energy like springs. just replace those 4 ends of the pcb with springs and this combined with your mechanism or the SMA mechanism should provide many times more jumping force than you have now.
Amazingly creative work, I would definitely like to see a version 2. I suspect you may be able to cleverly use the pcb material as a sort of spring to assist the hop. Perhaps reducing to 2 "legs" and putting 2 coils on each with a stiffer joint - the attraction of the first coil could pull a magent that was out of range into the attractive field of the second coil, kind of acting as a 2 stage arm of a lever allowing you to use the mechanical rigidity as a spring when you release/reverse the electromagnetic polarity.
do make a part 2 of this robot try adding few more layers in coil to reduce the overall resistance and to increase the magnetic pulling strength and also make the frog arms more stiffer
Got no recommendations but I just wanna say your projects are awesome
Make the pcb helmholz chambers so leg vibration adds additional acoustic resonant lift
Space frog 🐸🚀
Mini compressed air system spinning a mini turbine powering mini van de graff through a micro super capacitor to power the legs instead of batteries
That would be awesome
Keep on tinkering 🤖
Haha even if we say no i am pretty sure you have already started designing the next version. You are determined enough to beat all the issues and this is why you will be successful bro
Magnet at the bottom coil on top. Like a pogo stick the rubber feet would be the magnet the outer tube would be the coil. The coil would need a Center hole for a stabiliser rod that connects to the magnet at the bottom. The whole unit will be able to operate without flexible pcb I think and will jump with the weight of the pcb and the battery.
Cover the springs with slightly stiffer layers of PP packing tape to increase how much energy they can store per bounce?
I see you already mentioned what I basically just said
You can integrate the bluetooth chip directly on the pcb right the extra weight of the the bluetooth pcb will also affect the jump or you can make a small spy radio module which will eliminate the chip completely
we need to talk about microfluidics Carl, that's a nice pump you've got there.
Yes please. Make a version 2, version 3, version 23.. Then crowdfund the mass production of them.
Imo it needs more current. Say add a 50-75C smaller lipo baterry and up the current but with a shorter duty cycle. Or have rwo settings, low current for compression and a high current pulse for decompression.
Also flatter magnets would also help as you dont even have a flux ring around it so after it separates from te coil the flux loses are huge, thats why im sugesting doing a say, 15-20A pulse, to concentrate all the power while the magnets are close to the coils
Love this project and really interested to see a v2! Another idea, maybe you could ditch the battery and use your existing coils as a wireless charging receiver for a capacitor instead
Add stiffer arms or more springy, and make the attraction force stronger ,that way you only have to realse the magnetic hold and it will automatically spring up without the need of the magnets pushing it upwards, i would use the metal out of the spring in a clockwork toy ,or off an old mechanical clock spring chopped into four parts for the arms .hope this helps in some small way .
How about ditching the magnets and having two coils attract and repel each other. Maybe also add a DC DC boost converter onboard to get more current with more voltage
I didn't know there were pcb"s like that one
very expensive. they are called flex pcbs
You must be new here. Check out his previous FLEX AR videos it's a product he sells.
He sells them for like 2$ on his site each. Bargain imo :D
In fact, if you open your phone you may see this kind of pcb
@@Travis_Yu ribbon cables lol
You have to increase the grip of the foot on the table putting a thin sponge under the coils.
Have you considered using a different waveform to drive the coils? Sine and sawtooth curves arise more commonly in oscillation functions, and may yield better results.
You should try using some sort of Lipo watch battery! You obviously wouldn't get all that much run time, but it might have the perfect amount of mass for higher jumps!
maybe try to make more windings on the bottom coils would make it more powerful
Please make a v3, it's an awesome project and you shouldn't give up! Keep it up!
That is a nice pcb
Not sure you have them where you are, but there’s these flexible nasal strips for allergies, snoring & other uses.
Putting these on the flexible loops would give them more spring action, as it’s more this energy being used to “jump” than it is electrical. Unless you could program a reverse polarity “push” or repel sequence to help it hop.
It’d dramatically reduce runtime, but a (rechargeable) coin cell battery might be a happy medium between size, weight & ability to hop
You should add some little springs between the top and bottom part
Can you incorporate a hard stop while the top is moving up? You are absorbing all of the deceleration over time time with the flexible arms. If the mass comes to an immediate stop, the momentum will drag the coils into the air because it is lighter than the moving mass.
umm
i have an idea donno if its possible but hope it'll work well so the the coil attracts the coil and the once the coil is off it jumps because with the potential energy in the springy pcb okay so what if you use that coil to attract the magnets and then by reversing the current in the coil you can make the magnet repel which will make it jump higher
While this machine jumping My imagination is like: the kid pointing something, Mommy! mommy! I want that candy! And he is jumping too while pointing the candy.
Why using permanent magnets instead of another coil? This will likely reduce the weight and probably make it work better
I was wondering if this was explained in an earlier video while watching this.
Yeah v2 go for it!
Great work man love your work
I'm no expert in material and their flexibility, but you can try reducing the area of contact between the 4 circular coils with each other (that means all the 4 coils are a little distanced to each other and connected by strips of flex pcb) like a small square with coils on its corners.
I think it might work.
Would love to see a page of calculations for why you think the frog bot should be jumping 1cm (+-) into the air.
While I may be wrong, it seems to me that when the current is cut to the coils and the magnets "release" the energy is stored as elastic potential energy in the bend of the PCB. Thus, the bend unflexes (roughly equally) pushing up on the bot and down on the table. As a result this problem becomes one of the bot trying to jump by pushing off of itself which will not work.
If I am totally off base here, someone please correct me. I am just trying to understand the physical forces in effect here.
By that logic maybe adding weight to the top surface (in addition to increasing the stiffness of the 4 legs) to increase the momentum of the upward force would actually help, as it's the weight that 'carries' it upwards?
Very interesting and fun project.
Hey i think may be changing the pulse width to the coil while repelling would work
I love this. You DEFINITELY need to make a next version. I cant wait ti see it. Is there any way to get a smaller, lighter battery, i feel thats what is letting you down.
thanks!! it very hard to find reliable lipos with low mass and high discharge current rating
eat your heart out no so great at soldering Scott.. Lol . Nice tiny soldering work Carl.. Make V2 .. Make it now! :)
Really cool idea! Would definitely love to see the next version. How about making flippers using these actuators?
You might have some resistance from just the surface area of the pcb contacting the desk. Test it by adding a few grains of salt on the bottom of the pcb.