I love BB-8 videos, they're my favourite videos of the week! It's so great to see it driving and stabilising. It's going to look awesome when it's finished.
I love your BB8 videos. I dislike people who think they should comment with technical hints. You are an amazing robot builder who does not need technical "help" from people here. I enjoy and learn so much from your videos. My deepest respect for your skills and the willingness to share them with us.
Most technical and building projects throughout history have been collaborative efforts. James asks for comments at the end of every episode, so I believe he's comfortable with people leaving their thoughts, advice and criticism of his efforts. He's also at liberty to respond to commenters and take or leave any "advice" if he wishes to do so. The videos can also be enjoyed by NOT reading the associated comments if they cause you anxiety.
Watching the bonus features for episode 7, not one of the BB-8s they made for the movie was self contained. They had one with a a puppeteer pishinh it with poles attaches to the head and axis and an RC wheeled rig attached to the outside with an arm attached to the head. You sir have managed do something that all of the resources of both Disney and Lucasfilm couldn't. Well done sir.
Gf thinks I'm being cute staying up until almost 3 am talking to her on skype when really I'm staying up to see this new video update on the bb-8 build
It's rather fun watching how much better you are getting at your craft! I don't understand any of the programming portions, but I enjoy seeing the process. Keep up the great work!
I must admit when the droid fell off at 5:34 I got a worried that it broke something inside. Fortunately you built it really sturdy. A lot of good progress in this episode and a lot of small problem solving, which I like to watch. When things don't go exactly according the plan then it gets the most interesting as it requires a lot of assessing the situation and coming up with simple and creative ways how to bypass the issue. Thanks for the video. Looking forward to next one.
I hope we see many BB-8 versions. I just want to see an army of BB-8s rolling down a street ;D. Seriously, amazing work as always. Excited to see what the solution is, a build video with a cliff hanger!
brilliant to watch you develop this. I would get the main ball trued up those ridges are adding a lot more wobble as the head rides over them maybe even 'turn' the ball with some sort of lathe before you try skinning it
James, just had a thought!!! in order to gain more space and lower the fly wheel; rather than have a large main drive wheel and idlers on the banana.... create a "tank tread" concept. You will have tons of surface contact, alleviate the banana "push/pull" issues during acceleration. #JustAThought" Awesome work keep it up!
Hey James! I enjoy seeing your process throughout these bb8 videos. I'm curious to know how you would approach building a Dalek from the Doctor Who series!
If you bring the flywheel down you might have clearance issues. Adding weight on the banana's bottom looks like the better option plus it should help to not tip over when tilting in turns :)
Another solid video James. I think adding more weight lower down would be the better option, not sure how much difference lowering the flywheel only 10mm will make, plus you risk it fouling on something.
To help it be less top-heavy, I would suggest just attaching some extra weights along the bottom of the 'banana'. That would lower the centre of gravity for the bot and all that weight in the flywheel can still make it lean / rotate, but the overall centre of mass could be lowered. Granted, this might come at the cost of some responsiveness and increase your turning radius.
Hi James, for the issue of lowering the CofG have you considered replacing the lead shot in the flywheel. Easiest option would be to use cast lead rather than shot in the flywheel. You might get as much as 15% more mass depending on the density of the resin you currently use to stop the shot from moving. And decent local fisherman would be able to show you how to cast lead, it's pretty simple job for the home kitchen.
Also, if roundness of the ball is an issue have you considered making it from wood? A competent wood worker with a router and an idea of how to make a jog could make you a sectional ball without the ridges you currently have, any hardwood would be ok. It would have extra mass, but the CofG of the ball would still be centred, so the only effect would be increased inertia.
+James Bruton (XRobots) I understand it will be skinned. I was offering a suggestion that would also give you a more spherical surface for gluing as well as removing the bumps. One of the things I find when making things is we all get comfortable with the techniques we know and don't often explore other methods. For example I'm much more comfortable in wood and plastics but don't 3D print, whereas I cringe whenever you get a tenon saw out. I cannot deny your abilities and the huge successes you have with your projects. Very entertaining and educational. 👍🏼
The wobble is likely the result of the servo controller attempting to only balance for current rotation, resulting in jerky movement of the servos, rather than taking into consideration how momentum will change over time as a result of the balancing.
Looking really good. I can tell this version is far more refined. SO! The Force Awakens blu-ray drops in a few days and I hear it has a vid on the creation of BB8...I'm guessing you will be buying it. (Before Ver. 4 gets started lol )
May have to consider a redesign to a simpler and more elegant system. Maybe consider a floating gyroscopic inner core that drives the ball body/shell in variable directions with a variable direction friction ball of ninjaflex. That would negate the inherent instability in this current design and wouldn't require as much weight since it would effectively do away with any need to spin a flywheel to turn it in place. The flywheel can be done away with completely. The floating magnetic head should still work fine as long as the floating core clearance is minimal. The biggest challenge then would be down to figuring how to design the variable direction driver that doesn't cause drag in more than one direction at a time. You're smart enough to figure out the electronics on this current design so figuring out that driver shouldn't be too big of a chore.
I would say the mass in the bottom is more ideal since the flywheel already exerts to much force into the droid. As you said the flywheel is overcompensating which is only gonna get worse if you increase its mass even more. The mass in the bottom could compensate for some of the force the flywheel exerts.
Hey James, maybe to increase the weight on the bottom you could drill some holes in the banana along it's bottom middle. Then fill those holes in with a lead puck of the same diameter as the holes.
Hi, And a huge thank you for your upload, and the great way you explain what is going on, inside an increasingly complex ball and what relation it has to the actual roll and turn of it. You have to also be super thankful for the many different types of plastic and what you can do with it to make things even more tightly controlled, and again thanks for your info on what’s happening as you click a button or joy-stick, because there is so many pieces moving at times in opposite directions, not including the ‘trousers.’ Of course one thing you could do in not make any other adjustments, and instead take your ‘furniture’ desc etc with you to shows and let your ball bounce off the barricade? ;=) I am going on what I am able to recall only here having not filed the previous second ball you made so I can’t find them easily, but this version 3 seems a lot more ‘controlled’ with the help of the different ‘suspension’ plastics and have an ever variable software that you can as it seems adjust easier?
Nice video as always ! I think you could really benefit from using composite such as glass fibre or carbon if you can afford it. You could make your parts stronger and lighter (like the head structure). And it's really easy to do with a 3D printed mold :)
Great video as always. I have to say it amazes me how much content you pack in each week. Often times other channels just add filler to create longer videos. You're longer videos are just as engaging as the shorter ones. Keep up the excellent work, we all appreciate it! Also, i think adding weight would be simpler. Its to bad we can't print lead molds. It would be simple to CAD up a mold that would fit all extra space if we have a material that could withstand the heat of molten lead. I wonder if you could print, old school ink on paper, templates to cut out steel strapping or lead bars to fit the dead space? Not sure what type of heavy materials are available to you.
because your ball is not perfectly round, you should use spring loaded tension on the wheels like on a car or truck or tank track wheels. that will make the friction even at all times no matter the position. not only that, you can control the friction with a piece of all-thread through a spring using a jam-hut set up for infinitely adjustable tension on the wheels and drive wheel.
+James Bruton (XRobots) A bipedal would be amazing, should go for a terminator! But back to the non walking robots for a sec, what ever happened to the Johnny 5 robot? I remember seeing the 3d printed head but don't know if you finished it or if Johnny 5 isn't alive...
While watching previous videos I thought to myself "How does he have all these servo horns? Do the servos come with so many different ones?" Only now has the realization struck.
I think you should add more ballist to the bottom, I think that would solve the issue and be less work and less redesigning than moving the entire enternal structure down.
if you put a rod through the middle, then suspend the ball between two objects... maybe it would be possible to remove the ridges with bondo and sanding? The ball would move (like a lathe) and you could make it very smooth with minimal effort
i've thought from the beginning that the roundness of the ball might be an issue.. that makes it a lot more unstable then if it were a perfect sphere without ridges.. one other thought: maybe not use blocks to push those wheels up against the rim? how about strong springs? i am totally amazed by this build.. it reminds me of the techies in cyberpunk novels.. or the tinkerer from bicentenial man. James is building a real robot here.. out of commonly available parts.. it won't be long until someone builds a robot to take over his mundane day to day job.. looking forward to the law-suit when he tries to get his pay even though his robot does the work.. i know.. i am strange ^^
+Ben “Paracidic” Anderson perhaps two axial flywheels? Four Flywheels should get rid off all turning moment if they are put like this: Clockwise Counter-Clockwise Counter-Clockwise Clockwise Then you could change the speeds between them to turn?
James you do amazing work. I think your head movement is too limited in this version. With the original BB-8 you get the illusion that the head is driving the ball by shifting it weight around on the ball. I think with this version you lose that illusion because the head stays for on top of the ball. For version 4 you should combine the aspects of version 2 good head movement, with version 3 leaning flywheel(maybe even do away with the flywheel and just have a leaning counter weight). Make it inherently unstable with the head. so if you shift the head forward the unit wants to tip on to the head but the drive wheel counter acts that by driving the wheel. then you can drive it around just by moving the head. The lean can be initiated by the head and controlled with leaning counter weight. Anyway I really enjoy your videos. I don't know how you find the time, but please keep making stuff :)
+James Bruton (XRobots) have you considered putting it on a lathe and removing the excess material with a chisel? that way you can also work on the ridges inside.
You have an alien suit so why not make a predator suit with most of the features of the helmet and the three laser aiming system with the gun that tracks with the head? Just an idea for a future project that I would love to see :)
What if you sanded down a ring around the outside of the cheese. It would be flatter on the rim and therefore wobble a bit less (instead of a sphere balancing it is closer to a wheel balancing).
you still need to add the fly-wheel motor. Consider that before rebuilding the whole main-frame. The previous version had speakers fitted in the ball. The main CPU loop is also not really efficient. I am not sure if it is possible, but you have to check for new data on the blue-tooth and not read all channels during each loop. I think you also need to add constrains for the critical positions of the ball. Currently you are limiting the motors, and not the ball itself. In order to manage the ball you need to describe it in the software and then transform it to the mechanical components. The will allow you to manage the complex states that occur with the head positioning and the trousers. The bast way to do that is a state machine- define the main states of the whole robot: 1. move forward - excludes states 2 and 5 2..move backward - excludes states 1 and 5 3. move right - excludes states 4 and 5 4. move left - excludes states 3 and 5 5. stand still - exclude states 1-4 6. emergency stop 1 - on lost comm with remote 7. emergency stop 2 - on fail condition(ball tilted above permitted angle for example) then add the head freedom grades as sub states for each of the states listed above. 1. head locked - head position will be controlled from the drive logic 2. head unlocked - head position will be controlled from the user define the whole robot as a simple set of dinamic and static parameters(for example). dynamic: 1. Side-side PID parameters 2. Front-Back PID parameters 3. speed limits 4. channel amplifiers(sensitivity for the user input) 5. loop speeds(will allow you to prioritize particular operations like reading user inputs, managing motor speeds, switching between states) static(fail conditions and parameters to describe those): 1. max tilt angle for the ball 2. tilt prevention trigger angle 3. max trousers offset 4. max head offset angle 5 Use the state machine to adjust the parameters, and then feed those to the drives. The approach will allow you to simplify the code and keep it easy to manage. The last time you had a separate accelerometer in the head which allowed for more direct approach, this time you have to do some software push-ups ;)
Have you considered putting the entire ball on a leithe to true up the roundness of the ball.. the problems I keep seeing is the hop in the uneven ball plates may confuse your stabilization program. I think everything would work much better if the ball were more true to roundness.
+James Bruton (XRobots) Thanks, I forgot that detail from your last build. lo...l You are doing an awesome job on this channel. You have become an inspiration to home droid builders everywhere. Thanks and keep up the great work!
Just wondering are you gonna ad the bb-8 thumbs up ?? i am not sure if its possible to make that in the droid, but i would be epic keep up whit your epic videos i been watching your channel for a long time =)
Once you get a smooth surface on the outside and on a level surface wont you have to do all the coding again? Is it easy enough to make adjustments? Seems like you thought of everything to get it running smoothly. You going to be at SW celebration Europe in July? Also with this nearing completion and Hulk Buster basically done is there another new project on the horizon?
how far is the range before the bluetooth disconnects from the droid to the controller? if it is quite powerful with a strong range it would be very clever to add a small lens onto the droids head and connect it onto a phone display you could add to the controller so you can control the droid without following it everywhere?
Why are you not using the massive flywheel for stabilization?? Just keep it rotating at constant speed? It would probably require completely different bearings though
Hi James, you spend a lot of time talking about PID controllers but I'm wondering what your strategy is for tuning the controller. Are you essentially just fiddling with the values or do you actually use something like the zeigler-nichols method or even a more complicated method based on control theory. I bring this up because I think if you tried to tune the controller using something like matlab (which can automatically tune and create arduino compatible code for you) you should be able to get far better results. Thanks, Andrew (Aerospace engineer)
Porque no pones un tope en el joystick entonces no podrás moverlo mas de lo que debe? Así no se cae. Saludos desde Argentina! Why you do not put a cap on the joystick then you cannot move more than you should? So it doesn't fall. Saludos from Argentina!
the front wheels... could have just retained a nut in their truck and adjusted them in and out with a bolt, easily turned with everything in place... would have been simpler than trying to make the truck pivot and make a new block to push against
+James Bruton (XRobots) mm no, it doesn't seem to be a led. It is on the blue circuit board, looking at the video, in the right side of the robot, the side near you. I can even hear a "tic" noise similar to the noise of a sparkle... You can see a little white flash
James, I finally got to see the Building BB-8 special on the Bonus Disc from the film, and you will be very shocked pissed off and laugh how they actually pulled off the droid's movements in the film. It is NOTHING like what you are trying to accomplish or even theorized how they did it!!! Not even your Version 1or 2 is anything remotely what they did!!! The picture of the metal framed sphere is legit and they used it on all of their different ones they built and that we all saw is an actual screen shot from the Building BB-8 video on the Bonus Features. they finally reveal how theyactually built the working filmed BB-8's yes plural they made many different ones. they used an external Tricycle motorized system that is attached around the outside attached to the left and right side axis with gears on the shaft inside behind the sphere that also keeps it so stabilized and they used the same thick full metal frame inside its moving spinning sphere and a 3rd arm attached around to the back of the head and the head is not attached at all to the sphere its just above it from behind, that is why there is the bezel that doesnt spin and the head spins only!!! also a puppeteer man pushing from behind or the side out of camer with arms attached to the tricyle they used different angles that doesnt show the tricycle in shots!!! when they did the shot in the desert, the puppeteer pulled it backwards with the head spun around and the tricycle was pulling from the front hiding it!!! They digitally removed these external tricycle attachment systems they used!! Now they do not show the fully working one they used on stage at the Celebration that you are basing your theories off of on the axis tilting. Judging from all the ones they showed they used on set for filming, This confirms now they had to have made a giant SPHERO design for the Celebration BB-8 showing!!!! They did not use the Sphero design for the Filming like they confirmed said, probably didn't have the Sphero's final design theory and final engineering plans completed back while they first conceived BB-8 idea before filming. The BB-8 at Celebration does not have the 2 part independent Head feature like in the film droid scenes that you are implementing on version 3.
+Phar2Rekliss c3.staticflickr.com/6/5766/20864841850_41292d8864_m.jpg On a serious note, yeah they did say before it came out (as James has stated ^^^^) that the celebration was a droid built after the movie and that Abrams just wanted one "real" droid that would combine all the droids they made for the film. As you stated, there's quite a few BB-8s they used, each with a specific use. Very eager to watch the special features :).
+James Bruton (XRobots) Thanks for the clarifications on the Sphero internal function theories I jumped on. I just got all upset confused after seeing the prop droids not working anything like how you are making it work so it made me think the Sphero internal functions must be the only way it works. So I take it being friends with Matt and how they did it, did he mildly hinted you in the right direction on the physics theory and you are hitting it correctly now.? I think you got it! the axel front to back like you are doing functions like how a motor cycle turns by leaning left and right to turn.
Would you ever consider building Mettaton from Undertale? I've been asking this for a while, but I never got a response. The regular version, EX and NEO are all good, it's up to you to decide which you pick if you do pick one.
When this is done the difference between it and BB-8 will be it not having true AI (and maybe the lighter and grappling cable hooks but we don't need those)
You should fix the whole bot sideways onto a lazy susan and center it on it, then you can use a dremel (or similar) while turning the bot on the turntable to make it nice and round. Great videos, keep up the good work ;
![GUILTY GEAR -STRIVE- Season Pass 4 Playable Character #1 [Queen Dizzy] Trailer](http://i.ytimg.com/vi/nXDsAgYXXv4/mqdefault.jpg)
I love BB-8 videos, they're my favourite videos of the week! It's so great to see it driving and stabilising. It's going to look awesome when it's finished.
I love your BB8 videos. I dislike people who think they should comment with technical hints. You are an amazing robot builder who does not need technical "help" from people here. I enjoy and learn so much from your videos. My deepest respect for your skills and the willingness to share them with us.
Most technical and building projects throughout history have been collaborative efforts. James asks for comments at the end of every episode, so I believe he's comfortable with people leaving their thoughts, advice and criticism of his efforts. He's also at liberty to respond to commenters and take or leave any "advice" if he wishes to do so. The videos can also be enjoyed by NOT reading the associated comments if they cause you anxiety.
Watching the bonus features for episode 7, not one of the BB-8s they made for the movie was self contained. They had one with a a puppeteer pishinh it with poles attaches to the head and axis and an RC wheeled rig attached to the outside with an arm attached to the head. You sir have managed do something that all of the resources of both Disney and Lucasfilm couldn't. Well done sir.
+James Bruton (XRobots) wow I didn't expect you to make that kind of rebuttal. (Get rekt Pib Smablert)
Gf thinks I'm being cute staying up until almost 3 am talking to her on skype when really I'm staying up to see this new video update on the bb-8 build
Hope she doesn't see this...time to delete browser history and all that jazz..anyone of the female variety asks I wasn't here.
hahahahahaha
+mynamemaycontainAtoZ This isn't the user you are looking for... *waves hand*
It's rather fun watching how much better you are getting at your craft! I don't understand any of the programming portions, but I enjoy seeing the process. Keep up the great work!
Always a pleasure to watch James, we do appreciate all your hard work :) looks like your getting there and I have no doubt you will work it out :)
I must admit when the droid fell off at 5:34 I got a worried that it broke something inside. Fortunately you built it really sturdy.
A lot of good progress in this episode and a lot of small problem solving, which I like to watch. When things don't go exactly according the plan then it gets the most interesting as it requires a lot of assessing the situation and coming up with simple and creative ways how to bypass the issue.
Thanks for the video. Looking forward to next one.
Worth it all just to hear you say "just for Bling".
When you were test driving the droid, I could help thinking that ol' version 2 in the background was saying "Why are you replacing me?" :D
Movement looks quite amazing now! A job well done!
I hope we see many BB-8 versions. I just want to see an army of BB-8s rolling down a street ;D.
Seriously, amazing work as always. Excited to see what the solution is, a build video with a cliff hanger!
I don't understand anything that he says in any of these but I still enjoy watching - great videos!
"Cunning Pivot" would make a good band name. Or a nasty insult.
brilliant to watch you develop this. I would get the main ball trued up those ridges are adding a lot more wobble as the head rides over them maybe even 'turn' the ball with some sort of lathe before you try skinning it
Ninjaflex should sponsor you! Some superb examples of why it's handy to have
Yay! This episode was really cool..
I really like how its moving. I'm sad that the parts doesn't fit on my printers so I have do redesign parts to build one.
James, just had a thought!!! in order to gain more space and lower the fly wheel; rather than have a large main drive wheel and idlers on the banana.... create a "tank tread" concept. You will have tons of surface contact, alleviate the banana "push/pull" issues during acceleration. #JustAThought" Awesome work keep it up!
Hey James! I enjoy seeing your process throughout these bb8 videos. I'm curious to know how you would approach building a Dalek from the Doctor Who series!
James I'm finding myself very anxious for your Ultron robot . I realize you have several projects but man o man I can't wait to see more ULTRON :)
If you bring the flywheel down you might have clearance issues. Adding weight on the banana's bottom looks like the better option plus it should help to not tip over when tilting in turns :)
Another solid video James. I think adding more weight lower down would be the better option, not sure how much difference lowering the flywheel only 10mm will make, plus you risk it fouling on something.
by the end you will have a whole BB8 family. Btw I really like the controller.
Who needs CGI when you have a crazy Brit? =D
The things you do are amazing
Adding the windscreen wiper motor for the flywheel should help keep the mass at the bottom
Awesome stuff man ;) I hope skinning will fix the wobble issues.
To help it be less top-heavy, I would suggest just attaching some extra weights along the bottom of the 'banana'. That would lower the centre of gravity for the bot and all that weight in the flywheel can still make it lean / rotate, but the overall centre of mass could be lowered. Granted, this might come at the cost of some responsiveness and increase your turning radius.
This is great! Also needs the option of tilting the head to make it move like V2 :)
Funny how many wheels are required to get rid of the "wheel"
Great work Mr. Bruton! ;) It seems, that you don't need the flywheel at all. Keep on with your wonderful videos! :)
Greetings from Germany!
regarding ballast, magnetic fluid & electromagnetic "breaks" could make for some interesting balancing/maneuvering performance
Hi James, for the issue of lowering the CofG have you considered replacing the lead shot in the flywheel. Easiest option would be to use cast lead rather than shot in the flywheel. You might get as much as 15% more mass depending on the density of the resin you currently use to stop the shot from moving. And decent local fisherman would be able to show you how to cast lead, it's pretty simple job for the home kitchen.
Also, if roundness of the ball is an issue have you considered making it from wood? A competent wood worker with a router and an idea of how to make a jog could make you a sectional ball without the ridges you currently have, any hardwood would be ok. It would have extra mass, but the CofG of the ball would still be centred, so the only effect would be increased inertia.
+James Bruton (XRobots) I understand it will be skinned. I was offering a suggestion that would also give you a more spherical surface for gluing as well as removing the bumps. One of the things I find when making things is we all get comfortable with the techniques we know and don't often explore other methods. For example I'm much more comfortable in wood and plastics but don't 3D print, whereas I cringe whenever you get a tenon saw out. I cannot deny your abilities and the huge successes you have with your projects. Very entertaining and educational. 👍🏼
The wobble is likely the result of the servo controller attempting to only balance for current rotation, resulting in jerky movement of the servos, rather than taking into consideration how momentum will change over time as a result of the balancing.
I wasn't able to line up my garage door sensors today. James made BB-8.
Looking really good. I can tell this version is far more refined.
SO! The Force Awakens blu-ray drops in a few days and I hear it has a vid on the creation of BB8...I'm guessing you will be buying it. (Before Ver. 4 gets started lol )
May have to consider a redesign to a simpler and more elegant system. Maybe consider a floating gyroscopic inner core that drives the ball body/shell in variable directions with a variable direction friction ball of ninjaflex. That would negate the inherent instability in this current design and wouldn't require as much weight since it would effectively do away with any need to spin a flywheel to turn it in place. The flywheel can be done away with completely. The floating magnetic head should still work fine as long as the floating core clearance is minimal. The biggest challenge then would be down to figuring how to design the variable direction driver that doesn't cause drag in more than one direction at a time. You're smart enough to figure out the electronics on this current design so figuring out that driver shouldn't be too big of a chore.
I would say the mass in the bottom is more ideal since the flywheel already exerts to much force into the droid. As you said the flywheel is overcompensating which is only gonna get worse if you increase its mass even more. The mass in the bottom could compensate for some of the force the flywheel exerts.
Hey James, maybe to increase the weight on the bottom you could drill some holes in the banana along it's bottom middle. Then fill those holes in with a lead puck of the same diameter as the holes.
Hi,
And a huge thank you for your upload, and the great way you explain what is going on, inside an increasingly complex ball and what relation it has to the actual roll and turn of it.
You have to also be super thankful for the many different types of plastic and what you can do with it to make things even more tightly controlled, and again thanks for your info on what’s happening as you click a button or joy-stick, because there is so many pieces moving at times in opposite directions, not including the ‘trousers.’
Of course one thing you could do in not make any other adjustments, and instead take your ‘furniture’ desc etc with you to shows and let your ball bounce off the barricade? ;=)
I am going on what I am able to recall only here having not filed the previous second ball you made so I can’t find them easily, but this version 3 seems a lot more ‘controlled’ with the help of the different ‘suspension’ plastics and have an ever variable software that you can as it seems adjust easier?
Nice video as always !
I think you could really benefit from using composite such as glass fibre or carbon if you can afford it.
You could make your parts stronger and lighter (like the head structure).
And it's really easy to do with a 3D printed mold :)
Great video as always. I have to say it amazes me how much content you pack in each week. Often times other channels just add filler to create longer videos. You're longer videos are just as engaging as the shorter ones. Keep up the excellent work, we all appreciate it!
Also, i think adding weight would be simpler. Its to bad we can't print lead molds. It would be simple to CAD up a mold that would fit all extra space if we have a material that could withstand the heat of molten lead. I wonder if you could print, old school ink on paper, templates to cut out steel strapping or lead bars to fit the dead space? Not sure what type of heavy materials are available to you.
Print the part you want then use it to make a sand casting
I wonder if you can dial some flywheel in during a moving turn to aleve some of the need to tilt sideways.
Amazing bro nice work
because your ball is not perfectly round, you should use spring loaded tension on the wheels like on a car or truck or tank track wheels. that will make the friction even at all times no matter the position. not only that, you can control the friction with a piece of all-thread through a spring using a jam-hut set up for infinitely adjustable tension on the wheels and drive wheel.
Always so excited to watch episodes on bb8. Going for a 4th version after this? I have to see a version with a lighter arm regardless.
+James Bruton (XRobots) A bipedal would be amazing, should go for a terminator!
But back to the non walking robots for a sec, what ever happened to the Johnny 5 robot? I remember seeing the 3d printed head but don't know if you finished it or if Johnny 5 isn't alive...
+James Bruton (XRobots) so understandable :)
Dang, those magnets are strong.
While watching previous videos I thought to myself "How does he have all these servo horns? Do the servos come with so many different ones?" Only now has the realization struck.
Oh man I want to see the spherical operation of this so bad.
Seems to be doing much better than version 2 nicely done! Will you have to adjust the setup again with all the other little bits on it?
I think you should add more ballist to the bottom, I think that would solve the issue and be less work and less redesigning than moving the entire enternal structure down.
if you put a rod through the middle, then suspend the ball between two objects... maybe it would be possible to remove the ridges with bondo and sanding? The ball would move (like a lathe) and you could make it very smooth with minimal effort
You are amazing james :)
i've thought from the beginning that the roundness of the ball might be an issue.. that makes it a lot more unstable then if it were a perfect sphere without ridges..
one other thought: maybe not use blocks to push those wheels up against the rim? how about strong springs?
i am totally amazed by this build.. it reminds me of the techies in cyberpunk novels.. or the tinkerer from bicentenial man.
James is building a real robot here.. out of commonly available parts..
it won't be long until someone builds a robot to take over his mundane day to day job.. looking forward to the law-suit when he tries to get his pay even though his robot does the work..
i know.. i am strange ^^
Did you think of making the head pole extensible to fix the problem of the head pulling the whole mechanism?
I feel like spinning up the flywheel should help a lot with the stabilization.
Have you considered keeping it spinning at a constant speed? I think it would be much more useful being used as a gyroscopic stabilizer.
Damn.
+Ben “Paracidic” Anderson perhaps two axial flywheels? Four Flywheels should get rid off all turning moment if they are put like this:
Clockwise
Counter-Clockwise
Counter-Clockwise
Clockwise
Then you could change the speeds between them to turn?
spring loaded drive wheel maybe? so when the head pulls up?
James you do amazing work. I think your head movement is too limited in this version. With the original BB-8 you get the illusion that the head is driving the ball by shifting it weight around on the ball. I think with this version you lose that illusion because the head stays for on top of the ball.
For version 4 you should combine the aspects of version 2 good head movement, with version 3 leaning flywheel(maybe even do away with the flywheel and just have a leaning counter weight). Make it inherently unstable with the head. so if you shift the head forward the unit wants to tip on to the head but the drive wheel counter acts that by driving the wheel. then you can drive it around just by moving the head. The lean can be initiated by the head and controlled with leaning counter weight.
Anyway I really enjoy your videos. I don't know how you find the time, but please keep making stuff :)
if your ultron project goes well you could try to make C3PO which is basically the same concept
Perhaps a c3po costume for ultron to wear?
maybe you could put in a brake on one of the guide wheels, I imagine on the center wheels, cause brake force on the sides would pull it
When does the blu Ray come out in the UK? Can't come soon enough... Looking good..Is there a plan for the ridges...?
+James Bruton (XRobots) have you considered putting it on a lathe and removing the excess material with a chisel? that way you can also work on the ridges inside.
You have an alien suit so why not make a predator suit with most of the features of the helmet and the three laser aiming system with the gun that tracks with the head? Just an idea for a future project that I would love to see :)
You should try TPU filament
What if you sanded down a ring around the outside of the cheese. It would be flatter on the rim and therefore wobble a bit less (instead of a sphere balancing it is closer to a wheel balancing).
you still need to add the fly-wheel motor. Consider that before rebuilding the whole main-frame. The previous version had speakers fitted in the ball.
The main CPU loop is also not really efficient. I am not sure if it is possible, but you have to check for new data on the blue-tooth and not read all channels during each loop.
I think you also need to add constrains for the critical positions of the ball. Currently you are limiting the motors, and not the ball itself. In order to manage the ball you need to describe it in the software and then transform it to the mechanical components. The will allow you to manage the complex states that occur with the head positioning and the trousers.
The bast way to do that is a state machine- define the main states of the whole robot:
1. move forward - excludes states 2 and 5
2..move backward - excludes states 1 and 5
3. move right - excludes states 4 and 5
4. move left - excludes states 3 and 5
5. stand still - exclude states 1-4
6. emergency stop 1 - on lost comm with remote
7. emergency stop 2 - on fail condition(ball tilted above permitted angle for example)
then add the head freedom grades as sub states for each of the states listed above.
1. head locked - head position will be controlled from the drive logic
2. head unlocked - head position will be controlled from the user
define the whole robot as a simple set of dinamic and static parameters(for example).
dynamic:
1. Side-side PID parameters
2. Front-Back PID parameters
3. speed limits
4. channel amplifiers(sensitivity for the user input)
5. loop speeds(will allow you to prioritize particular operations like reading user inputs, managing motor speeds, switching between states)
static(fail conditions and parameters to describe those):
1. max tilt angle for the ball
2. tilt prevention trigger angle
3. max trousers offset
4. max head offset angle
5
Use the state machine to adjust the parameters, and then feed those to the drives.
The approach will allow you to simplify the code and keep it easy to manage.
The last time you had a separate accelerometer in the head which allowed for more direct approach, this time you have to do some software push-ups ;)
+Tim Stoev not sure if I understood anything but sounds cool suggestions
+Sami Mustonen I will explain as good as I can. What was difficult to understand?
+Tim Stoev It probably won't be for James. :)
Have you considered putting the entire ball on a leithe to true up the roundness of the ball.. the problems I keep seeing is the hop in the uneven ball plates may confuse your stabilization program. I think everything would work much better if the ball were more true to roundness.
+James Bruton (XRobots) Thanks, I forgot that detail from your last build. lo...l You are doing an awesome job on this channel. You have become an inspiration to home droid builders everywhere. Thanks and keep up the great work!
When you're done, can you put insides in the globe without stops on the sides?
the panels should be skinned in upholstery vinyl. Will add grip and durability.
Have you ever considered driving a constant flywheel for stabilization?
You have stability from the head and ballast but will you be spinning up the flywheel too?
Just wondering are you gonna ad the bb-8 thumbs up ??
i am not sure if its possible to make that in the droid, but i would be epic
keep up whit your epic videos i been watching your channel for a long time =)
When hulkbuster gets done do you think you can do a iron spider suit from the comics or from the civil war movie?
more ballast first...easier to remove than lower the whole flywheel and undo if not necessary
you should put a PIXHAWK FC in it so it can be a true autonomous robot
Once you get a smooth surface on the outside and on a level surface wont you have to do all the coding again? Is it easy enough to make adjustments? Seems like you thought of everything to get it running smoothly. You going to be at SW celebration Europe in July? Also with this nearing completion and Hulk Buster basically done is there another new project on the horizon?
Sounds great. I'll come find you on the Friday! :)
how far is the range before the bluetooth disconnects from the droid to the controller? if it is quite powerful with a strong range it would be very clever to add a small lens onto the droids head and connect it onto a phone display you could add to the controller so you can control the droid without following it everywhere?
there is still space for an arm that can pop out with a lighter. that would be awesome
Noticed when you have any input into sideways movement the self balancing stops working, seems to only work when its stationary
Hi James you can make your Banane lager and make a gape to put ballas inside will this not help so you dont need to change your flywheel?
Are you planning on facing the head forward with 90 degreed of turn to either side? Similar to a human head
James, for the power distribution, can I get a way with 12 slots on the large one and 5 slots on the small one? (I notice yours are 15 and 8)
Why are you not using the massive flywheel for stabilization?? Just keep it rotating at constant speed? It would probably require completely different bearings though
Oh right, of course. Needs two flywheels spinning in opposite directions, but then they cancel each other out completely so that won't help either
Hi James, you spend a lot of time talking about PID controllers but I'm wondering what your strategy is for tuning the controller. Are you essentially just fiddling with the values or do you actually use something like the zeigler-nichols method or even a more complicated method based on control theory.
I bring this up because I think if you tried to tune the controller using something like matlab (which can automatically tune and create arduino compatible code for you) you should be able to get far better results.
Thanks,
Andrew (Aerospace engineer)
Porque no pones un tope en el joystick entonces no podrás moverlo mas de lo que debe? Así no se cae. Saludos desde Argentina!
Why you do not put a cap on the joystick then you cannot move more than you should? So it doesn't fall. Saludos from Argentina!
the front wheels... could have just retained a nut in their truck and adjusted them in and out with a bolt, easily turned with everything in place... would have been simpler than trying to make the truck pivot and make a new block to push against
I've noticed a spark at minute 4.38. Is that ok? :)
+James Bruton (XRobots) mm no, it doesn't seem to be a led. It is on the blue circuit board, looking at the video, in the right side of the robot, the side near you. I can even hear a "tic" noise similar to the noise of a sparkle... You can see a little white flash
+Davide Alessi It was just a reflection on the circuit board.
James, I finally got to see the Building BB-8 special on the Bonus Disc from the film, and you will be very shocked pissed off and laugh how they actually pulled off the droid's movements in the film. It is NOTHING like what you are trying to accomplish or even theorized how they did it!!! Not even your Version 1or 2 is anything remotely what they did!!! The picture of the metal framed sphere is legit and they used it on all of their different ones they built and that we all saw is an actual screen shot from the Building BB-8 video on the Bonus Features. they finally reveal how theyactually built the working filmed BB-8's yes plural they made many different ones. they used an external Tricycle motorized system that is attached around the outside attached to the left and right side axis with gears on the shaft inside behind the sphere that also keeps it so stabilized and they used the same thick full metal frame inside its moving spinning sphere and a 3rd arm attached around to the back of the head and the head is not attached at all to the sphere its just above it from behind, that is why there is the bezel that doesnt spin and the head spins only!!! also a puppeteer man pushing from behind or the side out of camer with arms attached to the tricyle they used different angles that doesnt show the tricycle in shots!!! when they did the shot in the desert, the puppeteer pulled it backwards with the head spun around and the tricycle was pulling from the front hiding it!!! They digitally removed these external tricycle attachment systems they used!! Now they do not show the fully working one they used on stage at the Celebration that you are basing your theories off of on the axis tilting. Judging from all the ones they showed they used on set for filming, This confirms now they had to have made a giant SPHERO design for the Celebration BB-8 showing!!!! They did not use the Sphero design for the Filming like they confirmed said, probably didn't have the Sphero's final design theory and final engineering plans completed back while they first conceived BB-8 idea before filming. The BB-8 at Celebration does not have the 2 part independent Head feature like in the film droid scenes that you are implementing on version 3.
+Phar2Rekliss c3.staticflickr.com/6/5766/20864841850_41292d8864_m.jpg
On a serious note, yeah they did say before it came out (as James has stated ^^^^) that the celebration was a droid built after the movie and that Abrams just wanted one "real" droid that would combine all the droids they made for the film. As you stated, there's quite a few BB-8s they used, each with a specific use. Very eager to watch the special features :).
+James Bruton (XRobots) Thanks for the clarifications on the Sphero internal function theories I jumped on. I just got all upset confused after seeing the prop droids not working anything like how you are making it work so it made me think the Sphero internal functions must be the only way it works. So I take it being friends with Matt and how they did it, did he mildly hinted you in the right direction on the physics theory and you are hitting it correctly now.? I think you got it! the axel front to back like you are doing functions like how a motor cycle turns by leaning left and right to turn.
Would you ever consider building Mettaton from Undertale? I've been asking this for a while, but I never got a response. The regular version, EX and NEO are all good, it's up to you to decide which you pick if you do pick one.
two of these makes a Tron motorcycle =)
When this is done the difference between it and BB-8 will be it not having true AI (and maybe the lighter and grappling cable hooks but we don't need those)
Definitely need to find a solution to smooth out the head movement - looks a bit too jerky
+Nazgull123 I'm guessing it's mostly caused by the ball it's sitting on not being a smooth surface yet
+James Bruton (XRobots) bondo it up then mount it on a car wheel while it is jacked up :-P or a meaty motor
You should fix the whole bot sideways onto a lazy susan and center it on it, then you can use a dremel (or similar) while turning the bot on the turntable to make it nice and round. Great videos, keep up the good work ;
Slamdoxicalz Yeah I noticed that was some of the cause, looking at the video.
+Niocio01 good one..
It April fool guys also James where do you store the hulkbuster and ultron
James is it posibul for you to creat a bionic arm in BB-8 that CONTAINS A LAZER!!!!!!!!!!!!!!!!!!!!!!??????????????????????? ;)
I guess it would be dangerous! Thx Anyway!!! :)
Where do you get all of this money James?!?
Why i can't watch this video in SnoopaVision before 2280 ?
I not sure what your day job is but it should be this, nasa give this a job doing mars rovers or similar, awesome. is as awesome does
+Brendan Haslip (Doone, daknaki) he is a toy designer for BladezToys
I kept waiting for the april fools joke...
I'm sorry James, but you have failed your driving test. @4:17 haha
Did anyone else notice that the writing in the thumbnail says part 8
thumbnail says episode 8 :p