View Full Version : Team 871 - Homebrew Segway
We call it the "Craigway" named after the student in charge of getting this project rolling.
After the Nationals a few of Team 871's students were getting restless. After seeing a couple homebrew segways at the Nationals, they decided they wanted to build one, but they wanted to have it actually run on two wheels like the real machine. Honestly, I thought it was never going to materialize, boy was I wrong.
The prototype does not yet support the rider. The rider must try to remain balanced, but when the rider leans in one direction, the "Craigway" moves in the desired direction.
Hopefully we will soon have a web site up following the project. Our goal is to have a self balancing (which it does) machine to move a human rider around like the real segway (i.e. only 2 wheels and no training wheels)
Enjoy the videos at the site below. Be kind and post a mirror if you can, this is being hosted from a cablemodem.
.30 MB Divx and 60 MB MPEG 1 (http://wihs.zapto.org/~mrb/craigway/) clips are available.
E-mail me if you are interested in the 120 MB MPEG 2 video.
We would like to here everyone's comments. I've seen the other treads, but didn't want this getting LOST in the tread :-)
I should also point out that thet built the prototype in about 2 hours (power wheels wheels, 3/8" plywood, fisherprice motors and gear box, 1 gyro)
They spent about 2 weeks programming until they got the autobalance to work. Once balanced, a slight lean in that direction moves the "Craigway" in that direction.
We plan on testing the final design (see video (http://wihs.no-ip.com:8181/~mrb/craigway/) or pics in the above link) with the CIM or Van door motors..
Final project ETA is in Fall 2003...
http://wihs.no-ip.com:8181/~mrb/craigway/pics/P1010005.jpg
Matt Krass
20-06-2003, 03:00
My Team is also working on a homebrew Segway. Maybe we could get together and work on it with each other? I have some Inventor drawings of the base and controls I could share, I'd like to see the programming involved in that. You can reach me here or on AIM (ace32456 or CodeMonkey263).
I'm hoping to have a web page up within a few days, as soon as it is up we will post all specs, designs, future ideas and current code.
Then we can collaberate a bit :-)
Gadget470
20-06-2003, 17:23
I don't know if Segway, LLC would appreciate giving out info on a homebrew ripoff.. but meh
well, we aren't trying to mass produce a competiting project.
Here is a temperary site with some pics and info...
http://frohawk.no-ip.com:8181/craigway.htm
Matt Krass
20-06-2003, 18:13
I'd like to talk to your programmer about the autobalance code, any contact info available?
the URL in my last post is for his web site.. his e-mail is on there.
GregTheGreat
20-06-2003, 20:19
I guess I am not the only other person that has been working on a Segway summer project. I to would like to get in contact with your programmer. I couldent find his e-mail on the website, maybe I am just not looking in the right place.
Matt Krass
21-06-2003, 00:30
I looked over your website, but I don't believe you mentioned exactly what you were using as a control computer. We're looking into using the edubot RC as our controller. Also, in some pics I noticed a radio modem, why not just connect the OI directly via tether, since with the radio you can't legally use it at competitions. Plus, with a remote control base it limits your range to that of the radio. All info appreciated. TTYL
Hey all this is the programmer/web site guy. If you have q's about the programming i'd love to help you develop your own code.
wvxc400@optonline.net
you can see some code hidden in the pics of my notebook on my site. However since the bulk of this project is programming I will hesitate on giving all that info away so quickly. It takes the fun out of it. :D
Willum
sanddrag
21-06-2003, 01:20
I know the Robot controller holds all the programming to run the thing but I'm not quite sure how the radio and possibly OI come into play. It is controlled by the rider right? Could you explain this a bit. Also, how do you turn? I know how it would work mechanically but how does the user input his preference?
Matt Krass
21-06-2003, 01:32
Well generally the OI and RC function as a couple, I don't believe the RC will let anything happen without an OI giving it the ok. Correct me if I'm wrong
Jeff Waegelin
21-06-2003, 07:29
Originally posted by Matt Krass
Well generally the OI and RC function as a couple, I don't believe the RC will let anything happen without an OI giving it the ok. Correct me if I'm wrong
That is correct. If the radio (or tether) link is not established, the RC won't do anything. Even autonomous.
D.Fahringer
21-06-2003, 09:03
Quote: Gadget470 ...
"I don't know if Segway, LLC would appreciate giving out info on a homebrew ripoff.. "
This is true I believe, I have been trying to get ANY response from Segway, pro or con, on the design sharing matter since the
Championship! I have been asked to share the design of our TWO wheeled "segway" by a lot of teams and I have been trying to do it fair and with the blessing of Segway.
No answer at all... What does that tell you?
I would not be suprised to hear a FIRST ruling for next year... "No homebuilt "Segways" at the events!" That would be very typical.
Good job MrB! Looks like your summer team is on the right track to get a self-balancing rideable "Craigway"!
Thanks, but I didn't do much :-)
As for the RC and OI, you guys are right, I would guess, in the final drivable version (this was just a crude prototype) we would tether the RC and OI, and run them off the battery. It is either that or go to a completely different micro controller, and that doesn't seem like it is worth the effort until this machine is almost 100% operational.
I'm still worried that the Innovation FIRST equipment will not be able to react quickly enough to support a rider. It was never designed to work in real time. But then again, i've seen teams build robots as sophisticated as what were are trying to do...
To respond to the turning question, it doesn't not yet :-)
The prototype only gies forwards and backwards and is controlled by the input from 1 gyro. Will's plan is to add a potentiometer for CW/CCW rotation. That will probably be put into effect as soon as we build a new platform, in Sept.
Bingo. The code has what i call a correction factor that adjusts how fast the motors throttle to react to the sway of the rider. By intergrating potentiometers we will be able to manually adjust the correction factor to add more or less acceleration and decceleration.
Hence turning by adding more acceleration to one side while removing from the other side.
However you can't spin in place like this. From what i see in the segway videos the 3 angular rate sensors help detect lean into the turns so one doesn't fall off the side of the seqway.
Will
Rob Colatutto
21-06-2003, 16:23
Originally posted by D.Fahringer
I would not be suprised to hear a FIRST ruling for next year... "No homebuilt "Segways" at the events!" That would be very typical.
Actually, I recall hearing an announcment in the pits at Houston asking teams not to ride around on 'personal transport devices' in the pits. I'm guessing that was geared towards the teams who had thier own segways (or gokarts).
Matt Krass
21-06-2003, 16:50
How'd you guys interface with the gyro, did you do just connect it directly to the analogs and do some programming miracles or did you use some kind of special circuit? I was looking in to using a resistor/capacitor circuit to emulate the behavior of a "real" gyro, any info appreciated.
just attached the gyro to the analog inputs..
nothing fancy on the device yet :-)
JLambert
21-06-2003, 18:21
Originally posted by Rob Colatutto
...(or gokarts).
Umm... Gokarts?
I bet that was the same team looking for a flux capacitor... what devious plot could they be cooking up.
As for the segway thing, a kid from my team is convinced that we could build one using drill motors that would go at least 10 fps. As for wether or not we could, that's a very questionable matter.
Matt Krass
21-06-2003, 18:58
We're using the drills but I dont think we can get 10 fps, heh thats too fast anyway, I'd hurt myself. Now to justify this post, are you guys bringing that to SBPLI? We're hoping to have a working prototype with us there and I'd love to get together and exchange thoughts and ideas.
well, our main motivation was to build one just to show off the technical prowess of the robotics team and have some fun, but if we build one it would most definiately accompany the team to all competitions and demonstrations..
GregTheGreat
21-06-2003, 23:51
I have a question for those of you that have actually completed a FIRST built segway.
I have been working on some blueprints of the one I am trying to build, and I am having a hard time finding out how to connect the drive train to the "handel bars". My thinking is to forget about programming all the gyro stuff (I'll have to work hard enough to build it) and to just create a variable speed dial for controling the forward and reverse speed.
If anyone has any pictures or blueprints on how to connect the drive train to the "handel bars", I would love to see them.
Thank You.
GregTheGreat@attbi.com
Matt Krass
22-06-2003, 00:30
Also, do you guys know how to build one of those little twisty throttle things? like the ones it uses to turn?
D.Fahringer
22-06-2003, 10:27
On our KnightKrawler, the window motors and both wheels are attached to a two 3/4 thich aluminum motormounts which are attached to the standing platform which has the battery box pop rivited to it and the handlebar mounting tube is attached to the battery box.
The question about the PVC throttle mechanism ...the throttle and turning twist grip control is done by slipping a second tube over the handle bar tube and containing it with a press fitted third section of tube. a bit of the handlebar tube is left sticking out to allow for precise control. when your hand is part on and partly off the twist tube you can more accurately control the input while you are trying to keep your balance. Of course if you have a dynamicaly balanced system that works well this will be less of an issue.
The potentiometer has a wheel attached with an "O"ring to touch
the twist tube....it doesn't touch directly though, it needs additional traction and we use bicycle grip tape to give a spongey
positive reaction to the pot. If you slip the potentiometer for any reason you will lose balance. You need to be sure that the pot is free to slip in cases of overturning or you will break the pot.
When it is all working right and the rider is experienced it is hard to tell that there is no gyro!
Need pictures?
sure I'd really like to see a picture, its sounds like a great throttle setup..
Matt Krass
22-06-2003, 16:58
Originally posted by MrB
sure I'd really like to see a picture, its sounds like a great throttle setup..
Ditto, also does it use any sort of spring to revert back to center/null position?
sanddrag
22-06-2003, 18:54
Certain electric scooters have a throttle control that works sort of like that. You might be able to just buy the part if you don't want to tweak around with making one.
Matt Krass
22-06-2003, 18:57
Maybe, but I'd like to make my own, more for bragging rights than anything else really, but if it doesn't work out, got any links or suggestions as to where I can purchase one?
GregTheGreat
22-06-2003, 22:15
For those of you that have built you're own...
What type of motors did you use?
Thanks.
D.Fahringer
22-06-2003, 22:49
The grey tube that twists is bored to fit nice on the handlebar tube so it turns easily but no sloppy fit. (for the sake of the potentiometer relationship)
There was plenty of talk about a return system but if you think about the potentiometer and how it can slip a little now and then the return position may be off a little. If you have a more positive grasp on the potentiometer you could do the self return thing.
Personally I am controlling the rate of throttle advance and shutdown so completely as I balance I would prefer not to feel any resistance. I could get used to it I'll bet! And it would have to be a good thing for a new rider. It can get confusing when the machine is set to moving forward or backward in a turn and you have to figure out which way to turn the grips to get it to stop.
I marked the pot wheels to indicate the center position and that
is where you set it before turning the machine on.
You could use a zip tie as a resetable visual check temporarily I guess.
Just checking, but that's not gyro driven correct?
D.Fahringer
22-06-2003, 23:37
Yep no gyro installed yet, see?
On the subject of drill motors, there is considerable slop in the drill transmission that will be hard to deal with on a two wheeled "segway" This is a problem because the reaction time
to correct imbalance before things get out of hand is small.
It is a good powerful motor though worthy of a custom transmission but you have enough work ahead of you.
The most popular choice seems to be the Fisher Price.
I thought it had a lot of slop too but that one I based my opinion on had been modified...they are acually very good when new!
Moe will have one! (Moe Go)
GregTheGreat
22-06-2003, 23:50
That helps a lot, very good design.
Where can I purchase the Fisher Price motors?
Thanks.
Matt Krass
23-06-2003, 00:38
You guys are scaring me, my design uses the 2003 bosch and tranny as drive. Are you sure it isn't good enough? If it isn't it means a big redesign of the base....Just how much slop are we talking?
Victim of Fate
23-06-2003, 01:14
Ive been seeing quite a few homebrewed segways(indeed, I plan on building my own). And was thinking of maybe "segway races" next year at nationals, would be quite a sight.(hehe)
So. Getting back on track.
http://frohawk.no-ip.org/
check here now and then for updates, new pictures, etc.
Has anybody given much thought to the gyro programming process or am i alone on this?
sanddrag
23-06-2003, 02:21
Do not go to the above link unless you like clicking your mouse or pressing enter about 100 times.
The messages were kind of funny though.
<edit>Go to this link instead http://frohawk.no-ip.com:8181/craigway.htm </edit>
D.Fahringer
23-06-2003, 07:15
Quote: "You guys are scaring me, my design uses the 2003 bosch and tranny as drive. Are you sure it isn't good enough? If it isn't it means a big redesign of the base....Just how much slop are we talking?"
It's a pretty big deal because there is very little play in a good window motor but I had a gear set that was sloppy (about an inch of free movement as measured at the floor) and it was a lot harder to keep balance and look smooth. That is because by the time you make a correction it may have to take all the slack out of the drive train first and that means more time before the correction is actually stopping the fallover. Also... the faster your machine is the more important it is not to have slop because tall gearing amplifies the problem, the wheels will travel further freely
with that slop or backlash in the transmission.
If I remember right, the 2003 drill had about 20 or 30 degrees of rotational free play? Check it out maybe yours are better.
All this means less if you are always on the go, and most noticeable when stopped and balancing.
Dave
oh yeah .... sorry about that, i'll get that javascript out of there asap.
edit... all better.
I went into school today and took a lot of pics of the underbelly of the scooter. You can see how simple the layout really is.
http://frohawk.no-ip.com:8181/upclose.htm
Check it out ^^^
If you haven't done so already take a peek at the videos on the first page of this thread.
D.Fahringer
23-06-2003, 14:44
Power Wheels maker Fisher Price is now owned by Mattel and they have a service number ......800-348-0751
Couldn't get a look at the motors themselves in their pdf manuals
they should be in there but I couldn't find them!
Matt Krass
23-06-2003, 16:50
Anybody have an ideas on how to lock down the drill motors and eliminate the slop? or make it less severe to a point where it can function? I don't really want to change the design to use the FP motors, I have doubts they can handle a large load swiftly.
D.Fahringer
23-06-2003, 17:37
How bad is the backlash slop on your drill motors?
Roughly in degrees. Does it have the same amount in high or low?
I am no authority on FIRST motors by any stretch but.....
Team "Segway" Motor choice Pros and cons as I see them.
Fisher Price: pros: Comes in left and right hand so motor speed is matched, good or reasonable power, cons: a little loud, motor can sticks out in the way,
Drill motor: pros: excellent power, durable, has cooling fan and vents, cons: motor assembly is long and will get in the way of either ground clearance or rider platform minimum height. lots of backlash in drive train,
Window motor (1998-2001) pros: slim and allows for the lowest center of gravity mounting of all motors, quiet operation, low power consumption, left and right version matched motor speed, cons: not powerful enough! a little delicate for abusive learners.
Vandoor motor pros: slim and mounts pretty low, should have good power, easy to mount, cons: only one hand available may veer at top speed, has plastic worm gear inside possibly delicate,
Chiaphua: pros: Big motor, motor itself is quiet, easy to mount,
cons: must have custom transmission, probably will get in the way unless a worm drive is used.
Seat motor: Forget about it!
Matt Krass
23-06-2003, 17:52
Pretty interesting, question, can you get a good amount of speed out of the FP motors? I'm a little concerned about that, they don't seem to pack much punch.
i think with a 12 - 16" wheel the FP motors have good speed,
BUT they do not seem to have the torque required to hold a rider upright and compensate their weight to keep them balanced
we will be trying the CIM and possibly the van door motors..
sanddrag
23-06-2003, 18:07
What's interesting about the Craigway is that the rider stands far ABOVE the axis of the wheels passing the task of balancing largely to the motors and software. I just thought I'd point that out.
nice observation :-)
that was the goal of teh project, to see if we could actually build a self balancing device. As stated above, in its current state (early engineering prototype) is needs a skilled rider to try and balance, but that is only b/c the motors can't handle the load. But the autobalance/propulsion code works perfectly... I'll try to get some vids of the Craigway wthout a rider so you can see it balance itself.
Our hope is that in the next egineering model, the motors will e able to support the rider, and look and feel like a segway HT.
Our programmers is very pround of his programming skillz :-)
Matt Krass
23-06-2003, 18:53
Hmm, thats interesting I jsut realized that as well, wouldn't it be considerably easier to balance with the weight below the wheels? Above them the motors are sortof playing catchup...
Matt Krass
23-06-2003, 19:00
Sorry for the double post, but I hate editing them, it leaves that nasty little tag at the bottom. I was also wondering for the team's that've done this, how good are you with battery life?
D.Fahringer
23-06-2003, 19:14
Quote: "how good are you with battery life?"
Ours will go a few hours per charge!
Just how many Team Segways are there out there? We have two, a team behind our pit at the Nats had two (fourwheeler) and another team had one (fourwheeler) then the Craigway prototype (I like it by the way I just saw the video!)
I have a vandoor design that is a vast improvement over our current KnightKrawler! 16" Skyways, narrower track, battery is inline and upright, more ground clearance and more CG offset
to the platform. I'll show it soon......
Matt Krass
23-06-2003, 19:31
So on the Knight Krawler, you used FP motors and their respective gearboxes correct? Did you mount the wheels straight to the gearboxes? Sprocket and chain? Also, did you do any more tweaks to the ratios? More speed? More torque?
D.Fahringer
23-06-2003, 19:57
No, the KnightKrawler uses window motors from the 2000 kit. I was just relating that all other teams have been using the Fisher Price motor/transmission on theirs and I have to think it is for a reason....
Personally I would like to see what a Chiaphua could do.
But I would want a near zero backlash transmission and would want the motors pointing inline so we are talking a worm gear (friction) or a 90 degree bevel gear setup.
Look at the pictures on page three and you should pick out the windowmotors and how it is set up.
The motors are mounted to the upright motor mounts as are the axles and the gears on the motor mate with the gears on the wheel.
We have five different gear ratios to choose from.
Matt Krass
23-06-2003, 20:01
Ok, sorry, misinformed. So how is it on speed? Is it snappy? Sluggish? Also, hows it balance, are you wobbling around a lot on it?
D.Fahringer
23-06-2003, 20:27
Here is the Vandoor powered two wheeler....
very nice design...
did you use inventor for that?
D.Fahringer
23-06-2003, 20:45
Quote: "hows it balance, are you wobbling around a lot on it?"
Ha Ha! No way! I don't wobble around on anything! (Ego kicks in) :)
I would like to get out some clips of this running around but don't have a way to post them because they are too big about 1.5 megs. Does someone have a way for me to get them linked for everyone to see? I can send them to your E-Mail address.
I have some animated gifs that were taken with my digital camera
in my living room. Getting on, figure 8's, circles, takeoff, etc.
Can I put files that big here in a different area?
e-mail them to me at b3@buonomo.org
I will post them on my server and then post the URL where people can access them...
Matt Krass
23-06-2003, 21:04
Well, you can send them to me for my personal webserver but unfortunately it doesnt have that good a track record for staying online :-). In a few weeks I hope to have a linux webserver up and running, if it works out I'll let ya know.
Oh btw, you know there is a quote button right?
GregTheGreat
23-06-2003, 21:05
Fot those of you, that have attempted using drill motors....
How did you attach the motor to the wheel. I have been working on the drawings of the one I am trying to build, and I can not find out how to attach the motor to the wheel.
Thanks.
D.Fahringer
24-06-2003, 16:40
The Vandoor version was done in Anvil 1000, an old 80's 2D drafting CAD program! This is not a school design, they dropped the summer program for Team 122 students also. If it gives direction or sparks thought and imagination that's great!
MrB helped set up a link for my animated gifs of the KnightKrawler I will add more soon.
Find them here...Thanks MrB and Team 871 !!!
http://wihs.zapto.org/~knightkrawler/
Matt Krass
24-06-2003, 16:45
Nice, it's a little hard to to tell but you do like quite steady, hope our design comes out that well, I'm in the process of redesigning to use the FP motors :mad: So we'll see how they handle it. Thanks for all the help and thanks for giving me a good excuse to get my post count to 150! :D Check out the custom title all!
Rickertsen2
24-06-2003, 16:55
What kind of control algorithm is being used to balance your "Craigway"? PI, PID? Something else? How well is it working?
I'm not sure what pi or pid are. But the algorithm contains a few complex steps or reverse engineering.
Right now it balances decent, but i think it's balancing for a weight somewhere between what the average rider is and what the unloaded craigway is.
Hence it overreacts to unloaded sway, underreacts to a load. Our test driver, Craig, was able to do a very good job controlling it. Staying upright, swivel turning, and moving across the room. Check out the videos on the first page of this thread to see for yourself.
I'm currently working out a far more complex algorithm to determine speed, with a configurable user weight input. I only wish i payed attention in ap physics. Stupid torque problems!!!
Matt Krass
24-06-2003, 18:37
So is there any way to turn it without pushing off a wall or something? How does one control that?
<EDIT>I have no clue what pi or pid is either so don't feel bad ;) </EDIT>
Rickertsen2
24-06-2003, 20:17
PI and PID are simple algorithms for stabalizing a system. In your case this could be applied to preventing your sooter from toppling over. For a simple explanation of PI/PID as well as some stuff about diferential steering click here (http://www.seattlerobotics.org/encoder/200108/using_a_pid.html). Scroll down to the section entitled "Differential Steering and PID".
</edit> I just found this (http://www.geocities.com/chemforum/undpidcontrol.htm). Its another pretty good explanation of PID.
Originally posted by D.Fahringer
MrB helped set up a link for my animated gifs of the KnightKrawler I will add more soon.
Find them here...Thanks MrB and Team 871 !!!
http://wihs.zapto.org/~knightkrawler/
No problem, my pleasure. I'd be more than happy to host all Craigway/knightkrawler/Homebrew Segway pics and vids....
D.Fahringer
24-06-2003, 21:13
OK ! What did I do wrong in setting up the link in my post?
Yours works ...mine gives an error message.
The tilde looks different in yours ...is that it, MrB?
WernerNYK
24-06-2003, 22:08
The "text" of your link is fine. But for some reason the actual CODE of the link is just telling it to go to "http:///" which obviously isn't going to go anywhere.
I'm not really sure what you did wrong in order for that to happen, but for some reason it did. :p
sanddrag
24-06-2003, 22:10
Just edit your post and take out your url vB code and just leave the url and make sure "Automatically parse URLs:" is checked. :)
Matt Krass
24-06-2003, 22:45
sliding a little OT here, so on the Craigway and Knight Krawler, how do you turn? on the Craigway can it turn without losing its ability to balance gyroscopically?
Andrew S
24-06-2003, 22:58
(To the makers of the Craigway)
Just a thought . . .
I looked at your video clip and I noticed something. You still were having a hard time balancing on the platform. (This is just a suggestion but . . .) I would try to lower the center of gravity on the machine a bit more. Most importantly put the platform under the wheel axle.
I don't know if this has been mentioned before, or what problems others may find with this idea . . . its just an idea.
Good luck :)
WernerNYK
24-06-2003, 23:04
Having seen both these devices:
Knightcrawler: Turns by adjusting the speed of the motors (ie, tank drive on 2 wheels) through the handlebar throttles.
Craigway: It doesn't, yet. The turning you see in the video is simply caused by the rider shifting the balance of his weight. And no, the turning in the final product should have no affect on the gyro at all. That's more a programming issue than anything else, but if you think about it, the gyro is only 1-axis. When the device turns, it is simply rotating that single axis that the gyro can read, so it should not be affected at all. :)
D.Fahringer
24-06-2003, 23:20
Thanks Sanddrag! That advice was right on the money!
The link works now!
The KnightKrawler does turn like a tank but the right grip throttles forward and backward and the left twist grip controls left and right steering.
The cool thing is that when you are spinning around in a circle, you sense your lack of balance, and make corrections with the right grip like normal! I suspect this is identical in a gyro segway and on our KnightKrawler this feels good! Just maintaining your balance manually feels good, I wonder how fast and how powerful can a manual machine get assuming good control charactaristics and low CG platform design. Can we overtake a standard Segway's 12.5 MPH limit? Manually?
sanddrag
25-06-2003, 00:02
Originally posted by D.Fahringer
Thanks Sanddrag! That advice was right on the money!
The link works now! Cool. Glad I could help.
For the speed, I'm sure we could push it past 12.5 mph with a couple chias under there. Manually, I don't know. I've never ridden a manual one.
everything regarding the craigwat is correct.
1) it does not turn, yet :-)
2) the rider still has to balance it manually, the code works, but the FP drivetrain is not strong enough to support a human riders applied weight.
3) a lower CF would help balancing, but while testing we needed ground clearance to allow for great changes in angle to test the gtro code. it increases with change in angle and we needed to test it as such..
4) WERNER, where is your "871 ALUMNI" TAG!!!!
(you spend years helping create a genius and then they pretend not to be associated with you :-)
hmmm, as for the 12.5 ft/second i'm doubtfut...
the segway is powered bt two 2kw specialty motors, i doubt we will ever come close to its output :-)
i'm more interested in skillful balancinf and asthetics.. we'll worry about speed in bersion 2.0 :-)
did i mention we are only at engineering prototype bersion 0.001 :-)
D.Fahringer
25-06-2003, 09:12
That is 12.5 mph... so all we have to do is 13-15 mph. Ahh the power of Chiaphuas! or the 1/2" drill with a custom transmission would have power enough to reach that speed but It would not leave any extra power for recovery of rider or terrain bobbles.
It would be a combined effort of the gyro control system and the riders effort similar to the Craigway prototype...if the rider doesn't allow the machine to get in a fight for balance it could be coaxed to get up to a higher speed.
The real Segway won't get up to it's speed potential if the rider doggedly leans forward trying to force it to go...
GregTheGreat
25-06-2003, 12:26
Originally posted by D.Fahringer
That is 12.5 mph... so all we have to do is 13-15 mph. Ahh the power of Chiaphuas! or the 1/2" drill with a custom transmission would have power enough to reach that speed but It would not leave any extra power for recovery of rider or terrain bobbles.
It would be a combined effort of the gyro control system and the riders effort similar to the Craigway prototype...if the rider doesn't allow the machine to get in a fight for balance it could be coaxed to get up to a higher speed.
The real Segway won't get up to it's speed potential if the rider doggedly leans forward trying to force it to go...
You might be able to pull of the 13-15 without the gyro balancing. The programming and motors can only compensate so fast. If you were to wish to make one that fast, I would recommend drill motors, but remember you would need casters in the front and back (that or an exceptional programmer). If you're programmer can make it operate at over 12.5 with the drills and be able to maintain balance, I will be hiring him to teach me his unbelievable skills. The drills are sloppy, and if you are attempting a speed such as that, use a Chiaphua, I think you might be able to use that with a high gear ratio (that and a very stable rider) and you might be able to pull it off. If you find a way, you can send me some video, and I will even host it for you. I don't think its possible, but if it is, I am sure you guys will figure out how. I wish you the best of luck on creating this unique project.
sanddrag
25-06-2003, 14:29
Originally posted by MrB
the segway is powered bt two 2kw specialty motors, i doubt we will ever come close to its output :-) 2kw for each motor? My electric scooter has one 450w motor and it goes 16 mph.
Matt Krass
25-06-2003, 18:07
Hey guys, can any of you electrical geniuses out there provide some info on Resistor-Capacitor circuits and how to use them with the edubot? I remember being told it could be done but not how. I want to use it to stabilize gyro output.
I worked out the physics behind the Craigway and essentially any gyro driven balancing scooter. I'll post it on the website asap. But it's 1am now, and i have a graduation rehearsal at 8am, and need to leave for prom at 2pm, so it may be a day or so before you guys see this stuff. It's neat and something to ponder.
I may have figured out how the Segway allows for 3 speed maximums, with three keys.
Will
Matt Krass
26-06-2003, 13:11
Originally posted by MrB
go willium!!!
Heh, same here! I need all the info I can get!
<EDIT> Can you guys tell me the sensitivity of the gyro? I want to make sure it has enough clearance to react, like how far do I have to tilt it and how fast before its measurable? </EDIT>
WernerNYK
26-06-2003, 19:17
The gyro provided in the kit is _alright_. It provides pretty accurate readings at slow speeds of rotation. However, it can only sense so quickly... I dont know the numbers at all, but ie only 15 degrees of rotation/second, than it would max out and have to "reset" and that takes some time.
I would imagine for these self-balancing HTs that you all are trying to build the stock gyro will not present a large problem, but just FYI there are much better ones available from Digikey.
Matt Krass
26-06-2003, 19:56
Great, that helps a lot, I guess I better put a notice on there to nto snap it in any direction. Having the stabilization gyro pause to reset might result in a GW Bush! :D I still need to learn though, how far and fast does it have to turn before it will register? and what do you recommend for freedom of movement?
D.Fahringer
27-06-2003, 07:51
I was curious for some time about the exact cause of this drill motor slop in the transmission. I found it to be part of the variable load clutch mechanism and it can be taken out completely but you will have a little job on your hands. The slop results from the internal gear sleeve is prevented from rotating under load by balls that are spring loaded. These balls pass through the holes in the housing shown in the picture and the bumps on the back of the gear sleeve are so far apart that it will spin until the ball comes into contact. The fix? weld on more bumps! then it won't rotate at all. Minimal play.
The unit shown is a 2001 model but I suspect the same feature in the 2003.
The only problem then is the mounting of a long motor/transmission on the Segway gets in the way of where you would want to stand if you are going for the ultimate in low CG
of the machine.
Adrian Wong
27-06-2003, 09:05
Hello. Excuse me for intruding, but something has been bothering me since I started reading this thread.
Isn't the Segway and these other gyroscopically stabilized platforms an example of the common "inverted pendulum" control theory experiment?
If so, you wouldn't want to have a low center of gravity for the system. You would want the center of gravity to be higher (e.g. farther from the wheels).
Just a thought. Best of luck to everyone on these projects; they're very interesting to follow.
D.Fahringer
27-06-2003, 10:57
The lower the CG the easier it is to remain upright it seems to me. The fact that I can stand on the KnightKrawler without turning it on displays some inherent stability. The students and I have a little trick we do on it where we set the speed and a mild turn and then let go of the controls....(the "No Hander Meander").
Take a base where you stand above the axles of the wheels
and one where you stand below and try to balance....no contest!
If the wheels are free to turn the first one will get you in trouble right away!
I am keeping in mind that a yard stick is easier to balance on my hand than a 6 inch ruler and the yard stick's CG is further from my hand but I feel that is just a reaction time matter.
Anyone else have thoughts about this? It is interesting!
GregTheGreat
27-06-2003, 12:12
I was just wondering what type of wheels the KnightKrawler uses? I have been looking at many different types of wheels on skyway and other sites, but really haven't found anything that can support the weight and is thin enough.
Thanks
Originally posted by Adrian Wong
Hello. Excuse me for intruding, but something has been bothering me since I started reading this thread.
Isn't the Segway and these other gyroscopically stabilized platforms an example of the common "inverted pendulum" control theory experiment?
You are 100% correct! That is exactly what it is ....
D.Fahringer
27-06-2003, 12:49
The wheels are one of the best things because they are so good and so cheap! I got them from Northern Equipment, a hardware chain and they were only 7 bucks each...with two 1/2" I.D. bearings included.
The wheels are 12" in diameter, and get fantastic traction.
Skyway makes this type of wheel and tire in many sizes, they sponsor FIRST, and will give you a deal if you need a different (larger) size.
How much weight?
GregTheGreat
27-06-2003, 13:41
Originally posted by D.Fahringer
The wheels are one of the best things because they are so good and so cheap! I got them from Northern Equipment, a hardware chain and they were only 7 bucks each...with two 1/2" I.D. bearings included.
The wheels are 12" in diameter, and get fantastic traction.
Skyway makes this type of wheel and tire in many sizes, they sponsor FIRST, and will give you a deal if you need a different (larger) size.
How much weight?
Well I was looking at skyways tires, and I couldn't find something in a 14". I think that 14" would be about the right size. I was also considering wheel barrow wheels, because they are very inexpensive. I was actually considering having our mentor contact them to see if they could get us some 14" wheels. I think that they might be the best choice. Thanks much Fahringer.
Matt Krass
27-06-2003, 15:28
How much would 14" wheels cost? We were looking in to 9" skyways cause we had them, but 14" would benefit us much more, so if the price isn't too insane maybe we could swing them.
Thanks Guys!
I liked the size of teh wheels on the knightkrawler, but was also thinking of going to 14 or even 16" wheels.
I may get one set of each just to experiment with.
...oh yeah, and I agree, Northern us a great company to deal with...
Matt Krass
27-06-2003, 16:04
How are the prices on those? Are we talking 20 bucks area? 'Cause I know I can swing that even if I pay for it myself
D.Fahringer
27-06-2003, 16:06
I vote for 16" wheels from Skyway! When you only have two window motors for power it is best to stay small and light. (12" wheels)
If I were to make another, I would go for that big motor now.
Matt Krass
27-06-2003, 22:25
Anybody know of inexpensive 14" wheels? Also, how big is "too big"? I'm thinking around 3 feet the wheels might be stretching it :D
Matt Leese
28-06-2003, 11:01
Originally posted by D.Fahringer
The lower the CG the easier it is to remain upright it seems to me. The fact that I can stand on the KnightKrawler without turning it on displays some inherent stability. The students and I have a little trick we do on it where we set the speed and a mild turn and then let go of the controls....(the "No Hander Meander").
Take a base where you stand above the axles of the wheels
and one where you stand below and try to balance....no contest!
If the wheels are free to turn the first one will get you in trouble right away!
I am keeping in mind that a yard stick is easier to balance on my hand than a 6 inch ruler and the yard stick's CG is further from my hand but I feel that is just a reaction time matter.
Anyone else have thoughts about this? It is interesting!
It's going to be easier to balance something with a higher CG (within reason). They physics behind it are fairly simple. If you have an inverted pendulum (which this is the same thing as), you can consider all of the weight of the pendulum to be at the CG. Now, in order to balance the pendulum, you have to keep the CG in roughly the same place. Because of this, you move the bottom of the pendulum back and forth. Now, if you have a higher CG, you can move the bottom of the pendulum a greater distance while keeping the CG in the same place. This means that you can have less precision in your movements. It gives time for the control loop to settle down to a more stable state. So that's why a yard stick is easier to balance than a ruler.
Matt
D.Fahringer
28-06-2003, 12:50
It's going to be easier to balance something with a higher CG
So it would be easier to balance a machine with the rider standing over the wheel axles than under the axles? Hmmm...
A pendulum platform stabilizes at least some of the weight it is supporting, of course the overall CG is very much above the axles. It does this naturally and should require less power from the motors to keep a rider upright. In an inverted pendulum base all weight is in an inverted pendulum status and all of it requires stabilization.
The following example shows two yard sticks one supported at its end the other also supported at its end (superglued) but it is nested in a balsa pendulum.
Any difference between the two?
It's going to be easier to balance something with a higher CG
The opposite is true I thought. I always thought that tightrope walkers carried that big stick is that it lowers there cg and makes it really hard to fall.
Matt Krass
29-06-2003, 00:29
My brother explained to me like this:
With the platform above the wheels, the motors are always playing catchup when trying to balance. With the platform below the wheels it allows the motors do what they need and have the platform follow instead of the other way around. I hope that helps, it definitely cleared it up for me.
hmmm...
well, it wouldn't be hard to adjust the deck height to change CG on the design we are going to build, so we'll probably try both.
But the Segway looks as if the deck is above the wheels, so that is our goal :-)
sanddrag
29-06-2003, 14:41
Originally posted by MrB
But the Segway looks as if the deck is above the wheels, so that is our goal :-) Uuuuummmmmm? On the Segway, the entire deck is below the axis of the wheels' rotation and the surface on which the operator stands is just a tad below the axis of the wheels' rotation. www.segway.com
DanLevin247
29-06-2003, 14:57
With the platform below the wheels....when you start moving, the front of your'e platform will dip down upon acceleration...correct? If you correct this initial lean by compensating in the opposite direction wouldn't, a pendlious action occur...rocking back and forth as the rider tries to maintain ther own balance? This act of acceleration would cause you to lean even further foward before you had a chance to compensate...and if you had true segway like function...go even faster?
If you had your'e platform on top of the wheels, I think it would be harder on the gyro, but easier to maintain balance upon acceleration.
D.Fahringer
29-06-2003, 17:18
The deck dips down in front in advance of going forward not because of it. A gyro or manual Segway has to tilt accordingly to the amount of acceleration required. This would happen by the way even if your Segway has it's platform over the axles.
Yes, the Segway has its platform well under the axle line and although it is harder to make them like this it is for good reason.
The heaviest things on it are at the lowest possible position...that is the batteries, then the motors and the aluminum casing etc.
If they were trying for a high CG they would have put the batteries in the steering stalk like the batteries in a flashlight.
I finally made a dv .mpg clip of the KnightKrawler at it's VCU debut. (I apologise for the poor quality animated gifs) I wish it were lighter but you can adjust that I hope. I'll make a link to it soon.
The pendulous action yuo describe certainly can happen on our machine as a result of over correcting and in the case of the Segway as a result of a new rider not trusting the stablility of the machine. I will say that in both cases it is experience and anticipation that will fix that problem for the most part.
Looking at the Video from March 2003 of me on that knightKrawler I can say I am smoother than that now!
And a Segway with the base over the axles will probably be fine provided there is power to make it work for a more novice rider.
sanddrag
29-06-2003, 19:37
Originally posted by D.Fahringer
The pendulous action yuo describe certainly can happen on our machine as a result of over correcting and in the case of the Segway as a result of a new rider not trusting the stablility of the machine. Oh, yeah. I remember that from the first time I got on the Segway. As long as you relax you can pick up the feel for it in only a couple minutes. Then the machine moves precisely as how you would want it. Ah, what memories...
yep, i was wrong.. for some reason i thought the deck was higher...
D.Fahringer
30-06-2003, 09:08
Here is the only DV Mpeg of our KnightKrawler debut at the 2003 VCU event, the video software doesn't have a working provision for lightening the picture....it came with the capture card.
If anyone has a video of our KnightKrawler it might be worth a Team 122 Omniwheel Pendant! If you have the video of Dean Kamen riding our KnightKrawler at VCU, it is worth THREE pendants!!!
Check out the video....KnightKrawler Debut at VCU
http:/wihs.zapto.org/~knightkrawler/ (wihs.zapto.org/~knightkrawler/)
that is a nice clip... and out of curiosity, what are these pendants that you speak of? They sound like a prestigious prize :-)
D.Fahringer
30-06-2003, 11:47
Thanks! And yes the Omni Badges are cool...any color bead you want too! The omni rollers actually spin also!
I put up a little insight to what i'm thinking about the center of gravity on the website. Pardon the connection problems you may have the server isn't 24/7 (it's my personal pc).
I'm trying to post as fast as possible, but i've spent the last 4 days running between prom and graduation parties. It's been nuts!
Will
WOW, that pendant is pretty nice...
Makes me wish I had the video clip your looking for :-)
ok, Willum back to CG
D.Fahringer
02-07-2003, 15:05
Here are team 471's segways, two different prototypes, one is controlled with a central joystick and the other uses two drill housings for speed control. They move around pretty quick
with those FP motors, and you can hear them comming because the motor gears make some noise. They have casters on the back now, but I think they were originally trying a tilt switch balance system where the motors would be compensating for the tilt of the base due to a smooth teflon slipper switch mounted under the platform that would ride on the floor.
Matt Krass
02-07-2003, 15:17
Oh hey this thread is still here good. I had an idea for my solution to the sloppy drill problem. I'm gonna try this whether or not you guys like it ;) but I'd like an opinion. Also, much thanks to Andy Baker for this idea (in other words, if it blows up and hurts someone blame him! :D ) Anyway, since the drills have the power I need to move this thing and the FP are tight enough for the balance I was thinking about joining them. Basically both motors with a sprocket on an output shaft leading to a loose shaft for lack of better words. Chains link them and I think I can play around with the ratios to get the RPM close enough. If it works then I will have a sprocket take the output and chain it to the wheels. The idea is basically the FP will keep you steady with the Bosch throwing in horsepower so you wont be crawling. Any comments? and keep in mind even if this is a doomed failure I'm still trying, I'm stubbornly optimistic plus I need an excuse to get myself to learn ratios and the like.
sanddrag
02-07-2003, 15:40
You would definitely have a quite powerful machine. However, chaining the output to the wheels would give you a lot of unwanted play (or little play and excessive chain tension). The best way I've found so far is to just make a custom drivetrain for the Chias. Or just settle for less power an use the window motors (but not the new ones, they get to have 45 degrees of play after a bit of use).
D.Fahringer
02-07-2003, 17:30
Not a bad idea...and you could mount a dual chain tensioner.
This is why these forums work so well...ideas can just flow out
and help everyone who participates. Get going Matt, build it, it'll be fun and you won't regret it. And you ask for comments...one I could give is that when you settle on a matching gear ratio between the two motors, consider that the top speed matchup is not that important unless you are making a four wheeled "segway" because you seldom will reach top speed. This is because you need to hold some speed in reserve to make tilt corrections as you ride.
Can't fix the slop in the drill motor? I take it that the drill motor you are using is the new one then? Is the clutch different in it?
I'm begining to think that i'm the only graduated high school senior crazy enough to try and build a mostly functional self balancing robotic scooter.
Is there any knowledge of any other teams trying to do this, or is everyone just building segway look-alike's???
WernerNYK
02-07-2003, 18:24
The J-Way as seen (and built only the night before) at Nationals. Sorry, it's the best picture I have. J-Way v.2.0 is expected to have balancing and speed controlling capabilities :p
Now, WHEN J-Way v.2.0 actually is developed, well that's another story...not very easy to find time when FIRST is a full-year program. :cool:
D.Fahringer
02-07-2003, 18:43
Was The "J-Way" a 190 WPI project or Team 871's?
It went everywhere at the Nationals I remember!
Matt Krass
02-07-2003, 18:58
Originally posted by Willum
I'm begining to think that i'm the only graduated high school senior crazy enough to try and build a mostly functional self balancing robotic scooter.
Is there any knowledge of any other teams trying to do this, or is everyone just building segway look-alike's???
I'm also trying to make mine two-wheeled and self balancing. But I just passed ninth so I'm crazier than you.
And thanks for the encouragement guys, these forums rock!
The "J-way" is a team 190 project....
Team 871 has no affilation with Team 190....
ppppppttttt on the J-way :-P
(WernerNYK is a former Team 871 student team leader)
WPI yada yada yada :-)
Meet the ev warrior!! (http://www.surpluscenter.com/item.asp?UID=2003070220561702&item=10-1761&catname=)
http://www.enigmaindustries.com/Bosch_EV_Warrior.htm
These powerhouses are designed to move a human on a bycycle. An alternative to the motors that you see in the kit. It may not be as powerful at 12volts but you can definately overvolt these to 24 volts. You really can not do that with the drill motors.Is there any knowledge of any other teams trying to do this, or is everyone just building segway look-alike's???
Im not sure where my post went but I have decided two wheels are too easy. One wheel. That is where the challenge lies.
Here's a technical paper on a "mobile inverted pendulum" that could be of some use.
JOE (http://leiwww.epfl.ch/publications/grasser_darrigo_colombi_rufer_mic_01.pdf)
And some videos of it in action.
http://leiwww.epfl.ch/joe/
seanwitte
03-07-2003, 10:39
I remember seeing a two-wheeled balancing robot made with Mindstorms that used IR proximity sensors to balance instead of a gyro. We used some of the Sharp IR sensors this year, I don't have the part number handy but they're the ones with 6" - 6' range. Using one on the front and back, pointed at the floor, you can figure out your angle of tilt. They're pretty fast, you might be able to make a balancing platform without a computer, just a closed-loop with the sensors as feedback.
I've considered a single point balancing device. You couldn't use only one wheel, you'd need a sphere to balance on. You'd need to come up with a fancy turning mechanism or program because leaning like on the segway will only provide motion in that direction, not spin the rider around.
otherwise it's a good idea. just a little out of the range of the parts we're supplied with.
Will
I've considered a single point balancing device. You couldn't use only one wheel, you'd need a sphere to balance on. You'd need to come up with a fancy turning mechanism or program because leaning like on the segway will only provide motion in that direction, not spin the rider around.
Actually there are quite a few designs to building a monowheel robot. One design involves contructing a giant wheel and then balancing the person inside the wheel. To move the wheel all you would need to do is shift the devices weight like a mouse in its little wheel.
Originally posted by seanwitte
I remember seeing a two-wheeled balancing robot made with Mindstorms that used IR proximity sensors to balance instead of a gyro. We used some of the Sharp IR sensors this year, I don't have the part number handy but they're the ones with 6" - 6' range. Using one on the front and back, pointed at the floor, you can figure out your angle of tilt. They're pretty fast, you might be able to make a balancing platform without a computer, just a closed-loop with the sensors as feedback.
My concern is that the IR sensors would be prone to errors depending on what type of surface you are running on.
I would guess and ultrasonic signal would be better, like what they use in wall "stud finders".
I like the gyro idea the best though, The Craigway's goal is to mimic the Segway, with the standard KOP.
This way other teams can duplicate the results (after we are finished that is :-)
Rickertsen2
03-07-2003, 23:17
I have used SHARP GP2D12 IR proximity sensors and surprisingly, they are almost entirely impervious to to differences in color, texture and angle. Their man drawback is that they are not linear and this must be corrected for. The problem is especially bad at large distances. Other than that, though these things are great. Sonar could be another interesting option.
Ok i finally revealed some stuff for the holiday weekend. This includes some of the physics and two key lines of code.
Enjoy! (it's on the incomplete, but available MASSIVE PHYSICS page)
Will :D
http://frohawk.no-ip.org/physics.htm
Rickertsen2
04-07-2003, 12:03
Ahh a fellow case modder as well!
Matt Krass
04-07-2003, 23:45
Alright, working on the design guys. Does anybody know of a drill motor model for Inventor besides the FIRST one? The FIRST one has a tendency to bring my system to its knees in crashes. Anyhow, using a 3:1 reduction on the drill to sync it with FP at 100 rpm, expecting about 65 rpm under normal load. I'm looking in to using 14" bicycle tires and wheels. Anybody have an opinion on how well the FP can handle the tweaks back and forth to balance? Can it at all? The drills wont kick in until you start moving, that darn slop.....Thanks
we used the FPs on the Craigway, they respond well, just not powerful enough for dynamic balancing. And I would take the stop pins out of the drill motors, otherwise you are going to wind up throwing someone off the device when it comes to a screeching stop.
D.Fahringer
05-07-2003, 13:50
That is a good point MrB, you have to be able to coast and use "engine braking", and that reminded that the picture I put in of the 2001 Drill motor and it's source of slop does not mention that we build our own shafts and the original shaft and it's related parts ALSO may contribute to play...I think it is the "Hammer" feature.
By the way I got a big kick out of the Joe series of balancing robot videos...high tech and what a riot! And it would do no harm to take a 30 LB weight and add it to the stalk to see what happens!
I think we talked about this before school let out B.
You need to keep feeding motor power out and not let the scooter freewheel. Gaining undesired momentum will add forces that are unexpected to the code. Say for instance going up or down a hill. The program, based upon angle of tilt, must tell you how fast to go.
DON'T USE FREEWHEEL!!!
btw, the FP's seemed powerful enough to dynamically balance, but just weren't programmed to come back hard enough under a limited tilt. I'm working out the bugs, but without the prototype to play with all i got is theory and roboemu code.
Will
Matt Krass
05-07-2003, 19:48
Ok, quick question, with 150lbs total load and 100rpm free speed with drill and FP together, what do you think it will run at under load? I'm hoping between 60 to 65rpm but I'm no expert...thanks.
Ok see here's where we start diving into unknown territory.
You never will have 150lbs of load with a 150lb rider. Following what I did. 666.4 newtons is the full force of the rider, multiply that by the cosine of the angle of tilt. your answer to that is your horizontal angle of force. Try it, follow what i did, just change the angle and you'll see how the force increases as the angle decreases.
Figuring out how torque, rpm, and theoretical force all come together... that's where i'm having a problem.
Matt Krass
06-07-2003, 15:25
Originally posted by Willum
Ok see here's where we start diving into unknown territory.
You never will have 150lbs of load with a 150lb rider. Following what I did. 666.4 newtons is the full force of the rider, multiply that by the cosine of the angle of tilt. your answer to that is your horizontal angle of force. Try it, follow what i did, just change the angle and you'll see how the force increases as the angle decreases.
Figuring out how torque, rpm, and theoretical force all come together... that's where i'm having a problem.
:confused:
Uhhh.. I can add....? OK, wanna give us that in laymans? please?
hmmm... in physics there's a thing called normal force. That is, the force that pushes back at you. When you stand on the ground, your body pushes into the ground, but the ground pushes back up.
When riding the Segway or the Craigway the normal force is much different. Your body pushes straight down, always, but the normal force pushes back up from the axle at the angle you are tilted.
What I tried to say is that when you break up that angular force into it's two direction, horizontal and vertical, the horizontal force is never a full 150 pounds. If you tilt 10 degrees forward your horizontal force is about 26lbs.
I'm getting lost on how i get that force out of the motors... I've calculated out momentum, rpm, torque yadda yadda yadda...
So i guess those engineers get big $$$ for a reason. I hope that helped a little.
sanddrag
06-07-2003, 16:21
You probably know this already but I just didn't want anyone to forget the fact that the normal force is always perpendicular to the supporting surface.
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.