View Full Version : Most Impressive/Advanced/Unique/Cool drivetrain you've ever seen
sanddrag
27-05-2005, 22:15
Post a pic/link of the most Impressive, Advanced, Unique, or Cool drive system you have ever seen. Also write a couple comments about why you think it is so good. Please nobody post their own, we all already know each of ours is the best.
Well, at Peachtree I saw a pretty cool robot from Austen High School (Alabama). I think it was four wheel drive and the back to wheels actually turned. I think it was rack and pinion steering in the back. Very small turning radius. Maybe it was just me that thought it was cool. I don't have a picture, sorry.
Kyle Love
27-05-2005, 22:34
After seeing what 116 looked like this year, I'm a fan! It just looks awesome to me...very...out of this world? :rolleyes: Anywho, here's (http://www.invisiblerobot.com/robotics/robot_c55h/index/p1080706.html) the link. Can't wait to see it in action at IRI!
Beatty 2002-awesome idea, completely dominant
SPAM 2002-crazy fast, and super powerful
25 2003-fast as heck. Fit their strategy perfectly
33 2004-Autoshifting 4 speed. Need I say more?
spears312
27-05-2005, 22:52
I have to say I was impressed by S.P.A.M. 180, team 65 and team 45's drive trains. I also noticed the unique wheels on team 207. And, of course, who could forget the chain stuffed omni drive of 118 Robonauts robot Chainzilla?
sanddrag
27-05-2005, 23:05
Pictures please, if you can.
Meyerman
28-05-2005, 00:47
i agree with cory 2003 team 25s drive train was insanely fast and very powerful pretty much unstopable that was the most dominant drive ive seen. i like the 22/60/254 drive they all had a great base that was very quick with power but as insane as 25's 2003 robot it is a nice drive tho.
Team 71 in 2002
The file cards were an ingenious way of turning The Beast into an unmovable object. They slowly pushed through any obstacle on their way to the endzone. In one match (Newton SF2), they took three 90 lb goals, and two 120 lb robots (67 & 68) and dragged them into the endzone.
The file cards proved to be so effective at gaining traction, that FIRST outlawed them the following year. (Also because they had a habit of damaging the carpet)
http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains007.jpg
Team 857 in 2002
At this point in history Omniwheels were a rarity, and holonomic drive was completely unheard of. The Kiwi drive turned heads across the country as news of it spread. A true innovation.
http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains040.jpg
Team 45 in 2003 -- Not their competition bot
The "Ball Drive". This drive system was similar to a traditional computer mouse. Probably the most unique drivetrain I've ever seen, up until Team 116 this past season.
http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains057.jpg
Team 111 in 2002
The most common complaint about "swerve" drive systems, is the lack of traction. Well, here's a simple solution. Why not add a drop down tank drive? This gave Wildstang the maneuverability from the swerve modules, and the ability to push with the tank drive. The fact that they crammed all this plus a ball pickup system into 120 lbs. of robot is beyond me. One of the best feats I've ever seen in FIRST.
http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains066.jpg
Team 157 in 1999
This unique 8 wheel drive system allowed 157 to glide onto the 1999 puck with ease. The drove over the puck as if it wasn't even there.
http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains102.jpg
All these pictures, and many more can be found at http://www.firstrobotics.uwaterloo.ca/drivetrains.php .
At this point in history Omniwheels were a rarity, and holonomic drive was completely unheard of. The Kiwi drive turned heads across the country as news of it spread. A true innovation.
Just one minor clarification, but it might be more correct to say "Omniwheels were a rarity in the FIRST community." This is not to take anything away from 857 or any of the other early adopters of omnidirectional drive systems. But omnidirectional drives, holonomic drives, and Mecanum have been around for quite a while. The patent for Mecanum wheels (also known as "Swedish wheels") goes back to the 1970s. They were developed by Bengt Ilon, an engineer for the Swedish company Mecanum AB (one paper describing some of the early Mecanum wheel history and designs can be found here (http://www.araa.asn.au/acra/acra2002/Papers/Diegel-Badve-Bright-Potgieter-Tlale.pdf)). The Navy used them on commercially-purchased cargo handling systems in the 1980s. Orthogonal holonomic drive systems were first developed in the 1980s. "Killough Platform Drives" were developed by Steve Killough in the early 1990s, based on the early omnidirectional holonomic drive work (http://www.ornl.gov/info/ornlreview/rev30-12/text/robotic.htm) he did at Oak Ridge National Labs in 1991-1992.
There have been some truly unique "first ever, anywhere" developments by FIRST teams (the Thunder Chicken's Chicken Drive comes to mind). But many of our developments have been based on our ability to see something developed for, or used in, other applications, and realize that they can be adapted to FIRST-style robots. The skill/ability to do this is extremely valuable, and forms the basis for a lot of creative developments. But we should not forget that these developments are based on the work of others, and we stand on the shoulders of true innovators.
-dave
http://www.invisiblerobot.com/robotics/robot_c48h/p1050963.jpg
Like Kyle45 said earlier, Team 116 had a very unique design this year. While working at Annapolis i had the pleasure to talk with some students from 116 and learn how it worked. I think they not only took a huge step this year by doing something no one else has ever done in FIRST before but made a memorable drive train that we will be seeing more versions of in the future. Oh and who else can say their drive train is Team Tested, Tony Approved ;) (http://www.invisiblerobot.com/robotics/robot_c48h/p1060086.jpg).
http://www.invisiblerobot.com/robotics/robot_c48h/p1050963.jpg
Like Kyle45 said earlier, Team 116 had a very unique design this year. While working at Annapolis i had the pleasure to talk with some students from 116 and learn how it worked. I think they not only took a huge step this year by doing something no one else has ever done in FIRST before but made a memorable drive train that we will be seeing more versions of in the future. Oh and who else can say their drive train is Team Tested, Tony Approved ;) (http://www.invisiblerobot.com/robotics/robot_c48h/p1060086.jpg).
Is that a picture of omni wheels placed sideways. Hmmm... I actually had that same idea but I didn't know if it would actually work. It's nice to know that it would work.
sanddrag
28-05-2005, 14:51
Can someone elaborate on exactly what the 116 drive is and how it works? This is the first I've ever seen of it.
richardp
29-05-2005, 00:35
Is that a picture of omni wheels placed sideways. Hmmm... I actually had that same idea but I didn't know if it would actually work. It's nice to know that it would work.
No those are on the right way. (but yes at a slight angle from being flat)
Dave could explain it better I am a control person. I thinks it's even patent pending. I have the sheet we had for the judges I might post it later.
Can someone elaborate on exactly what the 116 drive is and how it works? This is the first I've ever seen of it.
http://www.invisiblerobot.com/robotics/robot_c50h/p1060718.jpg (http://www.invisiblerobot.com/robotics/robot_c50h/index/p1060718.html)
Start with a standard holonomic drive system, with four omnidirectional wheels at the center of each edge of the robot. Tilt the wheels over at an 82-degree angle, and cram them as far into the corners of the legal footprint as you can. The wheels now effectively run on the "edge" rather than the "outer face" of the wheel. The wheel is driven by mounting a 144-tooth, 20DP gear (with spokes and hub removed) to the "upper" side of the gear with 5/8" standoffs. This ring gear is driven by a Chiaphua motor with an 11-tooth gear, which is mounted on the same strut that supports the wheel/gear combination.
When you do this, a couple of things happen. The points of contact between the robot and the floor are spread way out. The resulting conservative support polygon is about 56% larger than it would be with a conventional orthogonal holonomic drive. Thus, the entire robot is considerably more stable. Also, by pushing the contact points farther out from the CG of the robot, you create a longer lever arm for the application of force when trying to turn the robot. Finally, by tilting the wheels over that far, the entire CG of the drive system and robot is dropped by about 3 inches. This increases the stability of the robot even further.
We played around with a few similar design concepts before converging on this version. We also had to play around with the particular compound used in the smaller wheels of the omniwheels. While several other teams that were using the AndyMark omniwheels were after harder mini-wheels, we actually wanted the softer compounds to maximize the traction we would get. Otherwise, we would spin out and create these really neat cresent-shaped burn marks on the carpet. That is why we ended up with the unusual white, slightly lower durometer, mini-wheels on our omniwheels (the picture above is from the Annapolis Regional, before we swapped out the mini-wheels for the Atlanta competition).
-dave
About how much extra time did that take compared to building just a regular holonomic drivetrain? Also, how many feet/second did you get?
About how much extra time did that take compared to building just a regular holonomic drivetrain? Also, how many feet/second did you get?
The particular design we developed was very easy to build. There are only three machined parts in each drive module (the modified gear, the structural support strut, and a small bracket that holds the load bearing in place). The drive modules can be used in either the cambered holonomic drive configuration (our version) or the traditional orthogonal holonomic drive configuration. In fact, we used the modules in the traditional configuration on our test robot for software development. So the answer is that it didn't take any longer to make this version vs. a traditional version. With regard to linear speed of the robot, it was designed for 10 ft/second. But it was never actually timed to determine real performance, so we can't attest to the actual speed.
-dave
Lil' Lavery
30-05-2005, 21:16
http://www.invisiblerobot.com/robotics/robot_c26/index/p1010464.html
http://www.invisiblerobot.com/robotics/robot_c26/index/p1010469.html
Those two pictures show some of the design shots of the drive system, hopefully giving a bit better idea of what we did. In simple terms, its a normal holonomic drive tilted on it's side.
*edit*
http://www.firstrobotics.uwaterloo....vetrains022.jpg (http://www.firstrobotics.uwaterloo.ca/resources/galleries/drivetrains/drivetrains022.jpg)
Although it's only an external view, 33's 4 speed, shift on the fly, automatic transmission last year kicked butt.*edit*
CRAZYMADI
31-05-2005, 00:43
Not far away from team 116, Team 612 worked on their Aukerman Steering Drive System. (http://www.chantillyrobotics.com/gallery/2005/front_2)
Unfortunatly, only 2 days beofre the ship date, Team 612 end up breaking some of the major parts of the drive system during the DC scrimmage clash. Its kind of sad story. Team 612 worked all six weeks wokring on this drive system, and then they have to switch back to tank drive within 48 hours of time span.
Dave.Norton
31-05-2005, 12:05
Hi;
Couple of questions for you; why the 82 degrees (I have some ideas, by why speculate)? The second is how did you control it? We went with an orthogonal holonomic drive this year, but discovered the unfortunate fact that 3 points define a plane, but 4 are over-constrained. Which is to say every time we hit an uneven spot in the floor whoever had the traction won and the robot went in that direction. We solved it in two ways, first added a yaw sensor so at least it knew what a straight line was, and then added a suspension so that all wheels maintained traction. This design looks even more interesting because it is inscribing arcs; at least ours was in a vector.
Thanks.
Holtzman
31-05-2005, 12:26
I'm very surprised that no one has mentioned 47's swerve drive from the year 2000. It was, at the time, revolutionary. By todays standards, it may not seem all that cool, but 5 years ago, when it was one of the first strafing robots in FIRST, it was really something else. It gave 47 a clear advantage that year and served as the inspiration to many future swerve drives. It was well before my time coming to FIRST but I'm still inspired.
My thanks for FIRST Mechanism library for the pics
Lil' Lavery
31-05-2005, 15:51
Hi;
Couple of questions for you; why the 82 degrees (I have some ideas, by why speculate)? The second is how did you control it? We went with an orthogonal holonomic drive this year, but discovered the unfortunate fact that 3 points define a plane, but 4 are over-constrained. Which is to say every time we hit an uneven spot in the floor whoever had the traction won and the robot went in that direction. We solved it in two ways, first added a yaw sensor so at least it knew what a straight line was, and then added a suspension so that all wheels maintained traction. This design looks even more interesting because it is inscribing arcs; at least ours was in a vector.
Thanks.
As you can imagine, a completely flat omni-wheel wouldnt work properly, and because the surface isnt entirely flat, if you had your tilt too low, the same problems would occur. You don't want the wrong omni wheels to be toching the ground, or else you'd get unecisarry scrubbing of the wheels. So, we places the wheels at enough of an angle to eliminate that, but still greatly improve our stability.
We had some issues with wheels getting more traction than others, especially when we we're driving on and off the loading zones. But we solved it by "shimming" the wheels so all four we're completly co-planar with the floor, and by using the white neoprene wheels.
I like 71 in 2002, 111, 25, 16, and 65 in 2003, 111, I forgot what team number it was (im saddend now) but at detroit that had three front wheels that spun at one point, 65, and 2005 i liked 79, 71, 111, and once again 65. I love our drivetrain.
2003
111: http://www.wildstang.org/gallery/view_photo.php?full=1&set_albumName=2003_electrical&id=2003_02_01_Electrical_005
http://www.wildstang.org/gallery/view_photo.php?set_albumName=2003_electrical&id=2003_02_01_Electrical_011
25: http://firstrobotics.net/03gallery/images/0025-1_jpg.jpg
16: http://firstrobotics.net/03gallery/images/0016-1_jpg.jpg
65: if you seen this years drive trainf, picture that but faster and cool wheels
2004
111: http://www.wildstang.org/main/video.php that video should explain it
???: urg what was that team number
65: I have noticed that we have no pictures of our robots it past years any where to be found. I want to change that.
2005
65: finally http://www.chiefdelphi.com/forums/pictures.php?s=&action=single&picid=11238&direction=DESC&sort=date&perrow=4&trows=10&quiet=Verbose
79: http://krunch79.com/gallery/displayimage.php?album=7&pos=8
http://krunch79.com/gallery/displayimage.php?album=7&pos=7
71: i con't find any good pictures of them
111: http://www.wildstang.org/gallery/view_photo.php?full=1&set_albumName=2005_NBC_Demo&id=IMG_2493
ChrisCook
31-05-2005, 23:39
Not far away from team 116, Team 612 worked on their Aukerman Steering Drive System. (http://www.chantillyrobotics.com/gallery/2005/front_2)
Unfortunatly, only 2 days beofre the ship date, Team 612 end up breaking some of the major parts of the drive system during the DC scrimmage clash. Its kind of sad story. Team 612 worked all six weeks wokring on this drive system, and then they have to switch back to tank drive within 48 hours of time span.
Hey We did something similar but i need to find the pictures for it, Do you have any sensor feed back on your robot to control its direction? contact me and we'll compare notes.
This year, what we did is we set all of our 4 omniwheels in the corner of our bot and set them at 45 degres. This allowed us to be able to strafe and spin in place and spin while moving forward. I think we had 8 ft per second.
We were the team that spun in circles for 2 matches on the Newton Field this year. We also had a piston in the middle of the robot that pushed down which cause all of our drive modules to expand. It gave us 4 inches on each side. This helped us with out stability.
We used field forward, and the fact that there were so many factors that affected our compass, we had to turn to robot forward, and this caused us to be bad at controlling our robot. But overall, it was cool
NeedMoreEngines
01-06-2005, 19:00
Post a pic/link of the most Impressive, Advanced, Unique, or Cool drive system you have ever seen. Also write a couple comments about why you think it is so good. Please nobody post their own, we all already know each of ours is the best.
At nationals there was the coolest drivetrain i have ever seen. I think it was teem 738 or 278 or 378, I don't remember. they were in Newton and their colors were yellow and purple. They made their own wheel system and they called it Jester Drive. It was really called Mechanum Drive, some brilliant swedish design. It allowed the robot to go in every direction depending upon how the wheels turned. It was the coolest thing i had ever seen...way cooler than omni wheels or tracks or 6wd or anything i had ever seen.
You guys know who you are...it was insane.
Lil' Lavery
01-06-2005, 22:05
At nationals there was the coolest drivetrain i have ever seen. I think it was teem 738 or 278 or 378, I don't remember. they were in Newton and their colors were yellow and purple. They made their own wheel system and they called it Jester Drive. It was really called Mechanum Drive, some brilliant swedish design. It allowed the robot to go in every direction depending upon how the wheels turned. It was the coolest thing i had ever seen...way cooler than omni wheels or tracks or 6wd or anything i had ever seen.
You guys know who you are...it was insane.
That's team 357. While a mechanum drive is quite impressive, teams have done them before.
Holonomic drives, such as 1002's, and 116's. or kiwi drives like 857's also can drive in any direction instantly.
Can some one please explain Kiwi Drive?
sanddrag
02-06-2005, 00:08
Can some one please explain Kiwi Drive?
Couple Videos:
http://stuweb.ee.mtu.edu/~alkrajew/FIRST/kiwi.mpg
http://www.team229.org/site/video/1106259564kiwiparody.wmv
56k+ Online Videos = No Solution lol Thanks though
Can some one please explain Kiwi Drive?
its similar to what your team (1002) did, except instead of 4 omni-wheels 90 degrees apart, it had 3 omniwheels 120 degrees apart.
Okay thanks a lot. What are the advantages of that over 4 wheels? I guess you use lss motors, but isn't it more stable with 4 wheels?
sanddrag
02-06-2005, 02:43
Okay thanks a lot. What are the advantages of that over 4 wheels? I guess you use lss motors, but isn't it more stable with 4 wheels? I don't know if it is really an advantage or disadvantage but I will point out one difference between the two.
With 3 wheels at 120* intervals when going straight forward you have only 2 of the 3 motors driving. But each of the two motors is getting roughly 86% of it's speed being applied in the forward direction.
With 4 wheels at 90* intervals when going straight forward (or sideways or backward) all 4 motors are driving but you are getting roughly 70% of the speed of the motors being applied to the direction of travel.
So I suppose you could say 3 wheels is more efficient because there is less vector cancellation.
With 4 wheels it is naturally square while with 3 wheels it is naturally triangular
4 wheels/motors gives more traction
3 wheels/motors looks cooler?
When it comes down to it, I don't see any reasons for picking 3 over 4 other than less weight and less power consumption and possibly greater speed for the same gearing. If anyone has any other benefits of only 3 wheels I'd sure like to hear about them.
The main reason I'm building mine with 3 is so I can use one of these nifty little doo-dads http://www.robotlogic.com/product_omx3.html So I don't have to program a thing. :)
Meyerman
02-06-2005, 14:20
the drive 357 was cool but they were easy to push at philly they got pushed around, but it was an awsome drive to watch i still say 25 in 2003! lol sure it didnt move side to side but! it didnt need to
Jeffrafa
02-06-2005, 19:50
Okay thanks a lot. What are the advantages of that over 4 wheels? I guess you use lss motors, but isn't it more stable with 4 wheels?
A clear advantage of Kiwi drive over Holonomic that Sanddrag didn't mention is that the traction on each wheel isn't thrown off by slightly varying terrain.
It was mentioned with the holonomic that when one wheel would get up on the loading zone triangles it would throw off their traction on the wheels and make control a little shaky. With the Kiwi its like the difference between a four legged object and a tripod - the four legged chair/table or whatever will teeter if not on a flat surface, whereas the tripod doesn't care.
Daniel Brim
02-06-2005, 21:17
Kiwi drive has me intrigued a bit. I'm probably going to design one this summer and see how far I can get. My team is pretty hesitant to use omnidirectional because of lack of pushing power and ease to build. I'm still going to do what I can though. How would you calculate gearing for a kiwi robot?
Interesting ideas:
Drop down tank drive like 111's
Shifting kiwi (AM shifter, anyone?)
Thanks,
Daniel
My team is pretty hesitant to use omnidirectional because of lack of pushing power and ease to build.
When you say ease of construction, I take it you include programming?
From a mechanical standpoint, it shouldn't be all that hard, so long as you aren't making your own omni wheels.
heck, just slap 3/4 NBD's on there and you're good to go :)
I imagine the programming is the much more complicated aspect.
Billfred
02-06-2005, 21:51
When you say ease of construction, I take it you include programming?
From a mechanical standpoint, it shouldn't be all that hard, so long as you aren't making your own omni wheels.
heck, just slap 3/4 NBD's on there and you're good to go :)
I imagine the programming is the much more complicated aspect.
This has intrigued me about the NBDs. How hard is it to get the servos to shift everything at the same time? I've not dealt with a multi-speed drivetrain (yet), but I can imagine that having one side in first gear and the other in second would be less than desirable.
Alex.Norton
02-06-2005, 21:59
Okay thanks a lot. What are the advantages of that over 4 wheels? I guess you use lss motors, but isn't it more stable with 4 wheels?
actually when it came to making the drive train for our robot this year it gave us far fewer problems with basic drive than our previous drive trains had (we tried to add extra function that our gyro wasn't acurate enough for). The programing is relatively simple (just sine and cosine) and the mechanics was almost simpler to build than the drive train given in the kit, we had only one custom part per wheel which could be made in a half hour on a lathe which we used to adapt the wheel to the shaves of the kit transmit ions.
The other trouble that arises from three wheels is that to move forward (if there is a solid arm on the bot) you must drive one wheel across its face. This means that if it hits any bumps (especially if there is a large load on it) then it is far more likely to cause the robot to trip and fall. The other option is to have the wheels drive towards the center instead of perpendicular to it and while this type means that the robot is less likely to trip you can't actively turn the base and anything hitting the robot will turn is.
About four wheels yes they will never all touch the floor properly if they are simply shimmed it is just as easy to put a suspension on them than to try to solve the problems faced by a three wheeled design. This is why when we found that we couldn't control our chassis because one wheel would loose traction and it would steer off to the side instead of making a new chassis or modifying the one we had we put a suspension on the wheels. It is now extremely easy and fun to drive while it is still stable and it doesn't trip over its own wheels.
Alex
Daniel Brim
02-06-2005, 22:07
When you say ease of construction, I take it you include programming?
From a mechanical standpoint, it shouldn't be all that hard, so long as you aren't making your own omni wheels.
heck, just slap 3/4 NBD's on there and you're good to go :)
I imagine the programming is the much more complicated aspect.
Programming was included as well. NBD would work, as long as I can see how to do it. I am not too good with drivetrain construction, since we have just used the kit stuff in the past few years. I feel that we should start stepping outside of this now, but I can't do it alone.
Thanks a lot for the explainations.
This has intrigued me about the NBDs. How hard is it to get the servos to shift everything at the same time? I've not dealt with a multi-speed drivetrain (yet), but I can imagine that having one side in first gear and the other in second would be less than desirable.
You don't need to shift them, I just figured it would be the easiest, and lightest way to get the reduction without making your own gearbox (unless the kit gearbox is lighter than a transmission, which I doubt).
Billfred
03-06-2005, 09:31
This is true, nobody says you have to shift anywhere (just ask the folks that used the kitbot this year). However, I'm a sucker for taking things to The Next Level (tm) when I can. ;)
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.