Motors DIY Segway

Hello All,
Thinking about making my own Segway, and want to see what you all think.

The brains-
http://store.makezine.com/ProductDetails.asp?ProductCode=MKARDIE

The muscle
http://banebots.com/pc/MOTOR-BRUSH/M4-R0062-12

The brains for the muscle
http://www.robotmarketplace.com/products/IFI-V883.html
OR
http://www.robotmarketplace.com/products/0-IBC1.html

The balancing act
http://www.sparkfun.com/commerce/product_info.php?products_id=741

For the transmission I was going to use bike gears, probably just a one set although I could use a mountain bike setup if I wanted to get tricky.

26" mountain bike wheels.

Now here's what you come in, is it possible to do a segway with 26" wheels? The center of gravity would be above the wheel axes, but it would also be on an autobalancing platform (the wheels would have a mount that goes down where the rider stands).

And a few other things I still have to work out before posting.



Thanks in advance.

 

I have no idea, but i'll be impressed if you can pull it off. It would be a good way for you to get started here.
And by the way, welcome to the forums!

 

I say no to 26" wheels. If your center of gravity is way above the wheel axis, the autobalancing platform would need to be in full effect all the time. I believe that there would be too much risk in such a system since the tendency of the platform with your body on top will be to lean to one side over the other.

I highly doubt that a $35 microcontroller will be capable of constantly autobalancing the system.

First of all, if you're going to go with affordable microcontrollers, consider using several in a network so that each one is only responsible for a particular task. That way, your autobalancing system is not monopolized by the controls input or another system. Then, drop the center of mass as low as you can. If the center of mass will be near or below the wheel axis, the platform's natural tendency would be to align itself to a perfectly horizontal plane. This way, both the workload on the autobalancing system will be reduced, AND the necessary correction times are reduced. If the platform were above the center axis, it would require constant correction and counter-correction. With the plaform and additional body weight below the axis, smaller, less frequent corrections may be needed.

What's your educational background by the way?

Also bear in mind that the Segway cost a lot of money to develop, and even now is still quite an expensive machine.

Another thing you should realize is that 26" tires are a bit too narrow and may cause stability problems. If you place them too close together, you'll also encounter turning/steering difficulties.

Good luck!

 

you're overthinking this.

 

Aww, i need to have a look at maddox's site again. It's awesome.

Regarding the segway, , cool - have a try at doing it. Just don't expect it to work the first time around.

 

Woo! Good luck, this sounds awesome! I assume you've seen Trevor Blackwell's scooter? Post pics!

 

These systems ALWAYS need constant correction. The segway is an inherently unstable design. The actual problem of self balancing is not that hard and does not require much computing power. The basic theory is that you detect which way you are falling then use the motors to torque you back into not falling. It requires CONSTANT input from the micro-controller but in essence is not a particularly hard problem.

I would say that the micro is 50/50 to work but the mountain bike wheels are probably a bad idea because its going to make it hard to get on the thing!

 

It's not hard to keep the centre of gravity low with 26" wheels, having a decent width for stability and getting on or off wouldn't be hard either.
The biggest problem with wheel size, is that the larger the wheel, the more accurate the motor control needs to be, as the motor's movement is amplified by the radius of the wheel just as a lever does. Any play in any gears or gearbox assembly also becomes more critical the greater the radius/wheel diameter.
So some smaller wheels from a kid's bike simplifies things

 

I'm digging the diversity here!

My friend is going to help me program this, but I really don't think that it's an intense program, all the controller needs to do is say hey, we're tipping bump them motors up! and then interpret the next signal from the gyro to decide if it is helping or not. I don't really see a way to use more than one controller, it seems like it all kind of needs to go in the same controller, and I am definitely not smart enough to incorporate more than one.

Little bike wheels are a good idea, but with the 26" wheels I was hoping to beat the standard 12.5 mph that segways and DIY segways alike get, as well as increase stability and ability to handly bumps in the road.

I am a senior in high school.

There's a good thrift store near my house, I always figure if it doesn't work it's only $10 to change it.


here's my original design (i know i left a lot of things out, don't worry).

(pardon the terrible drawing)


And I know this won't work (i keep this mindset to avoid being discouraged when it doesnt, and when it does I'll be the happiest man on earth).

 

Cool design! How will you keep the passenger safe from cogs/gears/whatnots and wheels?

 

You should check this one out: http://tlb.org/scooter.html

It may give you some tips.

 

Passengers will be protected by some lexan, the case will be bed frame angle iron (very cheap, but also kind of heavy) (nice to use because it's already in the angles I want, no machining, maybe Aluminum after I finish my CNC machine). The inner casing will be some more lexan (going to use this next year in college so I need to weld most of it together to help prevent someone from accidently walking away with it, not entirely sure how this is going to work. The problem I see with Segways and DIY segways is the turning mechanism, if you're going to have something where you lean forward to go, there should be something to lean to stop. I am thinking of two gas pedals from a go kart which when both are 100% pushed, it will go straight, then as you gradually lift a foot it will turn in the direction with more "push". Still need to refine that, I think it would be cooler if I could incorporate that into the lexan sheets, have to sheets and then just step on them to make it go around, so it is a little more stealthy than a lame gas pedal. *patent pending*

The nice thing I foresee with my setup of parts is that I'll be able to use a motorcycle 12v battery. I don't really like how that guy uses RC car batteries, not enough cushion for the pushin'. Also the motors and everything else run off 12v (as well as a few other common computer accessories cough CCFL cough) so I don't need to regulate the voltages outside of a distribution block with some breakers. I still need to find a way to round up some money, I don't have a job and only have 100 on my person, although I do graduate in a few months might be able to take out a temporary loan from my dad. Current expected cost is in the ballpark $600, but I was part of the FIRST Robotics team at school so I might be able to borrow some parts until summer when I can buy some more. I should be able to get the materials for cheap, a mentor at the robotics team has the hook up.

All for now, time to get CNC'ing.

 

That design idea has some problems. The centre of gravity is nice & low, and the width is good, but the big wheels & drivetrain is bad.

Did you read what i wrote?

 

Well, since you're scavenging bike parts for the wheels, why not use bike gears and chains too?

They come in all sizes, so if large wheels are a problem, put a 6" diameter gear on the wheel, and a 2" diameter gear on the motor. If you want different gearing (low gears for uphill, higher gears for cruisin', etc) then that would be really easy to implement. The parts would be cheap... just pick up a pair of used bikes at a flea market. You'll get all the gears you need, and your shifters, and your wheels, and your brakes. Your entire drivetrain for what, $20 for a pair of older 10-speed bikes? The drivetrain is pretty rugged, and you'll be able to transfer the power from the motors down low to the wheels way high up without much loss at all.

 

mvagusta- I did read it, but it's an easy switch if the 26" wheels don't work (i'll probably try some small guys first anyway, just throwing some radical ideas out there) and I already have some little kid bike wheels so I'll be able to use those if/when the bigger ones don't work. If it's just a gear on the motor to a gear on the wheel, where would the play come in? If i can find a cheap gearbox I would indeed use that, but they're just so expensive why not try a bike transmission?

Altron- My thoughts exactly.

Here's some more resources if anyone wants to do this endeavor on their own, or just to give me some fabulous ideas.
http://web.mit.edu/first/segway/

http://cobweb.ecn.purdue.edu/~477grp12/media.html

 

Every pair of gears has play, fine/small gears have little play, large/coarse gears have more play. The relatively large chain also has plenty of play with the sprockets - all this play means the motor has to move more than what is needed in one direction, because first all the play is being taken up, then there is drive. The same thing happens in both directions of course! All this play might only take a fraction of a second to take up, but that's a big deal with a segway, as the control box is trying to make heaps of fine adjustments per second.

A similar problem is had with the big wheels, where the fine adjustments, are being amplified into large adjustments. So when the slack is finally taken up, a huge & late movement occurs!! The computer will respond quickly with a movement in the opposite direction, which takes longer than required because the slack has to be taken up again from the opposite direction, and again, a larger than required movement is had.

Notice the small diameter of wheels used on segways, and the geartrain used - they are as direct as possible, using the least amount of gears as possible, and the gearbox looks to be a planetary gearbox which have multiple gears engaged simultaneously, which equals very small amounts of play.

 

Hmm okay. Well this project might not get off the ground. After much thought, I think this project would be extremely awesome, but I am looking for something a little quicker (i'm going to use this around on the college campus, and to get home ~3 miles). The ants crawl of ~12 would be a little annoying everyday. Also, this will cost ~600-700. The nerd points would be phenomenal, but the functionality wouldn't be there. So, onto my next idea. A DC scooter (don't want gas), They make some really cool stuff, but it's all so expensive, so again the DIY route. DC motors are extremely expensive for 1 hp, but AC is not at all. So I'm thinking, 12v motorcycle battery -> 1500w power inverter -> router speed controller -> AC motor. Then once I realize that it gets terrible battery life, small gas chainsaw motor -> alternator -> battery. And so goes the cycle. (i'm hoping it gets okay battery life, but i have no idea if it will). Now there's some problems with this design as well, as far as i know they don't make twist throttle AC motor controllers. So I would have to move my hand off to adjust the speed.

 

I know that it seems really exciting to build and design something yourself.

First of all, I think all that electric crap is a bad idea. It's too many inefficient things. You're talking a small gas motor, a small alternator, a battery, an inverter, and an AC motor. That's easily 50 pounds. Scooters to me seem like a bad idea, not very stable at high speeds. I think a bike would be better. Grab a bike, and behind the seat install a motor. Swap the gears around, and just bolt the pedals in place so you have a place to rest your feet. It will be more stable at faster speeds, and will have more braking capabilities, especially if you outfit a set of disc brakes on it.

I think that you'll run into a lot of problems trying to design and build something... have you thought about just buying a kit and installing it? Something like this? http://www.bicycle-engines.com/

 

a cheap micro is ideal you can get an SX that can do 80Mips for under £10 I doubt even if you will need that (pic running at a few Mhz should be more than enough).

 

Oh, well I was planning on mounting this to a mountain bike or a moped (not sure why i said scooter, i actually meant moped anyway) The gas motor and alternator would most likely not happen on this setup, but if I find that it is something that is necessary I may turn this project into a gokart or something with more real estate. The motor and battery will be mounted as low as possible (under the bottom bar of the frame, which would keep it as balanced as possible, but still not really. That kit is pretty awesome, but I'm expecting this AC kit to be $100-150, pretty cheap. Already bought the motor, a 1 hp motor with three speeds (still going to try the router controller). I think I'm going to get some peltiers and some disc brakes, install the peltiers on the brakes to get some of that power back (i know it's not an incredibly efficient process, but it would help some), and possibly some solar panels (on the back rack of the bike). These are just some ideas i'm kicking around, nothing is set in stone yet. Also, with the proper setup (soldering wires together, and ensuring there's no possibility of water) would the 115 acv be a terrible idea (it's always going to be a bad idea, but terrible?)
I have a mountain bike with a 3 speed internal hub gear, i think this would be ideal for the setup (the bike is really old, hoping to just rip the gears off and throw the bike).

Thanks for the replies.

EDIT:
the motor is pretty much the same as this one
http://americanhvacparts.com/Merchant2/graphics/00000001/Blower Motor.jpg

And, it actually is a bit more than one HP (i think?) It has a 115vac voltage, and 14 amps. That is 1610watts / 746watts = 2.15 hp.