![]() |
Slickest drive transmission yet...
For everyone that has attempted, or wants to attempt, to design their own gearbox/transmission design for the FIRST competitions, sit up and take notice - there is a new sheriff in town!
I had a chance to see the drive system designed and built by Team 33 this past weekend at the Richmond regional competition. It is, without a doubt, one of the most elegant designs I have seen in the nine years I have been working with the FIRST program. They have built a dual-motor, four-speed, shift-on-the-fly, AUTOMATIC transmission that is amazing. It is the ultimate in simplicity - you can build it with a lathe and a drill press with an X-Y table. The gearbox itself is very small and compact - under two pounds. The gearbox and combination of the motors are separate subsystems, so the gearbox can be used with a single-motor (or triple-motor) design without any redesign. The shifting gears are all cut from 12, 14 and 20-tooth, 16-pitch pinion wire - the large tooth profile, small DP-to-diameter ratio, plus 15 mils of designed-in backlash makes on-the-fly shifting trivial. But the best part is the automatic shifting function. The system detects wheel speed to determine the output of the gearbox. They have calculated the power curves for transmission in each of the four gears, and determined the optimal crossover points for the curves. The control system has been programmed to sense when the system reaches one of the crossover points, and automatically fires one or both of the shifting pistons to bump up or down one gear. All the robot driver has to do is push the throttle, and the system just goes. It auto-shifts all the way up to fourth gear for high-speed runs. Bump into something and need to start pushing harder? No problem - it just senses the drop in speed and auto-shifts back into first gear. The Killer Bees have set up a great demo board with a prototype of the design all laid out, along with their design specs and performance analysis. They are very open with the design, and indicated that they will be writing up a white paper on it later in the year. If you are going to be at the Detroit or Canadian regionals, make sure you find a way to see this drive transmission in action (and if you end up against them in competition, you will definitely see how effective it is). My only regret is that I didn't get a chance to get any photos of their gearbox - if anyone has any, please post them! -dave |
Re: Slickest drive transmission yet...
I spent a good amount of time just staring at that thing, it was absolutely fantastic. If anyone has a chance to, I highly reccommended checking it out. I'm new to FIRST, so I don't know how common this kind of thing is, but I was blown away.
My team took some video of it, along with their coach explaining it and it will be up on our website soon. I wonder how they could be topped... (A CVT perhaps? ) :D |
Re: Slickest drive transmission yet...
Quote:
-dave |
Re: Slickest drive transmission yet...
Speaking of Drive systems.
I would like to thank Dave Lavery and the crew at Team 116 for the awesome shift on the fly gearbox they shared with the FIRST community. We used that this year, and so far have been unstoppable to get our bot up on that platforms! (Both of them) We modified it a tiny bit, and included a little extra spiffy design work in the end plates and the gears. Thank you Team 116, and be sure to stop by our pits in Atlanta to see what kind of full drive system we came up with including your gear box! |
Re: Slickest drive transmission yet...
Thanks for mentioning this. I can't wait to see a whitepaper. I've started to look for a new drive system for next year and something like this sounds great. We don't have the best machining ability so something that can't be done in a high school shop would be great for every team. Since next year will be our third year it would be nice to make a transition from a single speed chain drive system to a multi-speed transmission. I will have to look into this.
|
Re: Slickest drive transmission yet...
actually our team(701) built a CVT transmission much like thunderchickens without that much machining capabilities UC davis did housing plates BUT it wouldn't have been hard to do it w/ out mill but when the mill came we didn't know how to use it ...lol (it came like 1 week into the competition)
|
Re: Slickest drive transmission yet...
If I remember correctly, your system only gave you a constant torque at all speeds, rather than the Thunderchickens, which gives them variable torque at different speeds. (It's still a very cool pseudo-CVT, and does work well for it's intended purpose, congrats for pulling it off!)
To get back on topic, 33's drive system sure sounds amazing, and I can't wait to see pictures of it. Cory |
Re: Slickest drive transmission yet...
Quote:
-dave |
Re: Slickest drive transmission yet...
1 Attachment(s)
I agree it is the ultimate in simplicity and function. Great job.
Enclosed is a picture of their demo board - I did not want to intrude for one of their actual gearbox. |
Re: Slickest drive transmission yet...
Thanks, I'm glad everyone likes our design.
We have made two speed transmissions for the past several years. Each year we try to make them simpler and smaller. I found that with the kind of high/low shift range that I like to have (3:1 or more) that the robot would tend to lurch severely if you downshifted at high speeds. I really want to make a 3 speed with an intermediate gear to help this. It turned out it was easier to build a 4 speed than a 3 speed ( its really two small two-speeds in series). We focused the design around reliablity and ease of construction, the demo unit on display was made entirely by high school students in a few evenings using no special tooling. Once I had a design with 4 gears I was pretty much forced to automate the shifting in order to keep from overloading the driver. I talked with a few Automatic Transmission experts at DaimlerChrysler on how to do a shift scheduler and viola....a fully automatic electronic robot transmission. It really is pretty neat, providing a nearly flat power output across a very wide RPM band with a 5:1 total shift span. I plan on posting the design and engineering here as soon as I get a good set of drawings made up (I'm not a CAD guy, I made the ones in the robot from a napkin sketch). |
Re: Slickest drive transmission yet...
This may sound wierd.......but we still haven't tested our low speed yet........haha....we will hook up the current sensors etc...AT COMPETITION!!! (funny i know) ....but they actually said their transmission doesn't give a different torque......torque remains constant....as for ours we still think it might give dif torque outputs but they say it won't..........we probably have to test it....... lol...
btw AWESOME transmission........just saw the pic |
Re: Slickest drive transmission yet...
I do belive that a lot of the teams that saw the 4 speed where very impressed. By the end of the day i was tired of explaning the thing. :)
|
Re: Slickest drive transmission yet...
Is there a picture of the real thing? That would serve us both better than me asking a bunch of questions. Thanks.
|
Re: Slickest drive transmission yet...
Hey, if u have any questions about our trannys or how they were built ask me because I built them!
Have fun @ all ur regionals. -Eric Y P.S. Killa Bees in da club!!! |
Re: Slickest drive transmission yet...
I saw the robot in competition, and it was great. I was really impressed with how you were able to speed accross the field at high speed, then push the mobile goal with little effort. I am guessing that this is a shift on-the-fly transmission. Could someone tell me how some teams have transmissions that cannot be shifted while driving, and others can? How are they diffrent? Sorry if these sound like stupid questions to you but, I do not know that much about drive systems.
|
Re: Slickest drive transmission yet...
They are all different........most of the shifting ones i've seen have 2 sets of gears and a pneumatic slides an input gear into one gear and out of the other, one that we built a while ago based on Technokats had dogs which catch on to the gears and engage them...the gears are on bearings so they don't do anything without a dog...
its kinda hard 2 explain...check out some pics and you will understand |
Re: Slickest drive transmission yet...
Sam, I know that some teams shift the drill transmission with servos, and to the best of my knowledge, these cannot be shifted on the fly.
|
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Very Nice, Very simple. The best part is definaly making it automatic. It is something new and refreshing so to speak. Very cool that you guys actually took time to calculate the power curves and shift at the right points, props to you guys!
Quote:
|
Re: Slickest drive transmission yet...
I really must say that this is an awesome system. Congrats to team 33.
I've thought of automatic transmissions for a couple seasons now, and I never thought of monitoring output RPM. In my theoretical designs for an automatic transmission, I would have monitored current to the motor, and shift based on that information. At different gear ratios, the stall current of the motor is constant, and if you set upper and lower thresholds, you can shift up or down, depending completely on whether you need less or more pushing power, respectively. I'm sure it's more complicated than I'm making it sound, but as I said, I'm stuck in the realm of theory, here. Just my take on the theory of an automatic transmission. With all of the talk of constructive vs. destructive criticism that went on recently, I want to say that I'm very impressed by this system. Again, many congratulations to 33. |
Re: Slickest drive transmission yet...
Quote:
http://www.pbase.com/image/26818929 |
Re: Slickest drive transmission yet...
Just Remember to Use Hard Metal!!!!
|
Re: Slickest drive transmission yet...
Hardened Steel is what he means. We ground down our high gear on one gearbox in Detroit.
|
Re: Slickest drive transmission yet...
well our team, 1024, design a CVT VST (continious velocity, torque, variable speed transmission). it uses bevel gears and three planetary gear systems. then we ran into the problem: it was gonna weigh around 30 lbs!!! so i was told to design a casing to replace the alumnium box it was going in (around half the weighht). well we ran out of time, but i did get a nifty looking case made from a FDM prototyping machine. Hopefully (crosses fingers) we will have it done at IRI. we had two allision transmission engineers helping us design, so hopefully we can have something to show for my 10+ straight weeks of doing CAD for this. and yes i did start in the offseason..and yes it is still not done :(
|
Re: Slickest drive transmission yet...
I was wondering if I could get some peoples opinion over weather a variable speed/torque transmission is a big advantage, or a lot of work for nothing. Our team has a CIM/Drill motor combo for our drivetrain, with no special shifting gearbox, and we seem to have pleanty of speed and torque. :)
|
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
rforystek:
That might be a debate for a different thread. To answer your question, though this thread from last summer contained some good debate on that topic (4 motor vs. shifters). The fourth page of replies in particular is well worth a look. http://chiefdelphi.com/forums/showth...g+motor+debate |
Re: Slickest drive transmission yet...
team 151's drive this year was awesome as well... they designed the system with simplicity in mind, as well as sleek and compactness as well..
a far shot, but their tranny are smaller than the aluminum tubing they made a new gearbox for the drills using the previous planetary gears, and made a new housing...it was gorgeous...they also implemented the pneumatic shifting, and ran the pneumatic shifter inside the drive shafts to save space...tthis thing was gorgeous... :ahh: |
Re: Slickest drive transmission yet...
Quote:
If any one wants the name of this person who is against Gracious Professionalism and against thanking other teams, PM me. |
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Team 33 was right across the way from our pit in Atlanta and I was over there any spare second I got to look at that demo board. When I first found out that they had constructed an automatic transmission, I freaked out. Seeing the transmission in person was something else though. Unfortunately I never got a chance to look inside their robot and see that articulated chassis that allowed them to climb the platform. I've been wanting to design my own drivetrain for the last year or so but with the lack of foundation knowledge and facilities it hasn't quite happened yet. With the new robotics club I helped start at URI hopefully it won't be long now.
I'll be awaiting for that whitepaper with great anticipation. You were without a doubt my favorite robot this season. Good luck at any off-season comps and keep up the good work! -Tom |
Re: Slickest drive transmission yet...
Hey guys, any news on that whitepaper yet?
|
Re: Slickest drive transmission yet...
I am from team 481 and i am tryuing to build a 3 speed
shift on the fly atomated pinion gear transmission. My teacher saw your design and he was blown away and he wants me to build one for next year i am just looking for suggestions and a little help(pictures would be nice) on how to build one of these transmissions. |
Re: Slickest drive transmission yet...
The lead engineer, Jim Zondag, is in the process of moving into a new house and is swamped at work from what I hear. I'll talk to him at the Sweet Repeat (October 9th) and see when he's planning on posting the papers.
|
Re: Slickest drive transmission yet...
At Nationals my team was right next to the Killerbees, so I saw their gearbox demo thingy a lot. I think it is the coolest one I saw. Also, I just don't say that because of how it works, I say that partially because they explained to some of us how it worked etc. I wonder who will build off their design next year? :rolleyes: Or who will go bigger and better? :yikes:
-Kyle I would also like to thank 33 for being great "pit mates". :] |
Re: Slickest drive transmission yet...
1 Attachment(s)
Well, I finally learned Inventor well enough to draw this transmisison up,
I will be posting this with full details in the white papers soon. here is a sneak peak. |
Re: Slickest drive transmission yet...
I have posted the design for the Killer Bees' 4 speed transmission on the White Papers board. Because the Inventor Files are so big, I could not post them all here, they are available for Download from the Killer Bees Website at http://www.ndprep.org/robotics/archive.htm.
|
Re: Slickest drive transmission yet...
Jim,
GREAT paper! Thanks for sharing. I'm sure this will help out quite a few teams in the 2005 season. I love the way you illustrate the power curve comparison for a 2 speed vs. 4 speed design. This is something I will show our students to help them understand these principles. Thanks again, John |
Re: Slickest drive transmission yet...
If I were drooling any more, I would short out my laptop. Great work!
|
Re: Slickest drive transmission yet...
I used to not like gears slamming into other gears, but I've seen it done so much that now I'm starting to really like this thing. It is totally not the way I would design a transmission, but the way I would design it would require twice as much time to build. This thing can be busted out in a few good nights at the machine shop. I really like the simplicity.
Do you have any (could you make some) pseudo-code on how it is programmed to shift? |
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Jim,
Thank you for posting the white paper on the Killer Bees 2004 4 speed automatic transmission! The paper is excellent! I have read, re-read, and studied the paper and all of the CAD drawings until I think I just about have all of the dimensions memorized! I am amazed at the elegant simplicity of the transmission. It is truly a work of art! Our team has never built any type of multispeed transmission before but I think we might give it a try this year. We have been gathering the various parts and materials in the BOM in your white paper and are working on building a prototype to play with and see if we think it will be something we can tackle for the robot this year. Hopes are high! Needless to say, we are eagerly awaiting the companion software paper on the automatic shifting code. In the mean time I have 4 specific questions that I hope you or someone here on CD can answer. 1) The BOM in the paper calls for qty 4 3/8" wave washers. I have studied all of the drawings and can't seem to figure out where they go. What am I missing? 2) The BOM also calls for qty 16 5/16" aluminum washers which are used on the shift shafts. Is there a reason these should be aluminum (other than to save a tiny amount of weight) or could we just use standard steel or stainless steel washers? I ask because I assume there must be a reason you specify aluminum but I just don't understand at this time and that would be one less special item we have to order. Maybe it will become clear when we get the prototype/test gearbox built. 3) I am sure that this will be covered in the companion software paper but I am wondering about the Grayhill 61K64 64 pulse per revolution encoders. The 61K64 encoders have output pins designed for a PC board, which seems rather inconvenient to use on a robot unless the interface to the controller requires additional circuitry not covered in this paper. Digikey also sells a 61K128-050 encoder which appears dimensionally equivalent in all of the critical dimensions but is 128 pulses per revolution and has a cable output which seems like it might be easier to wire up than PC board pins. Is there any reason a 128 PPR encoder wouldn't work just as good or better than a 64 PPR encoder? Since these are about $50 each I haven't ordered any yet until I am sure the 61K128-050 will work. 4) In the paper you mention that you had some of the gear/shaft assemblies hardened after welding. Could you provide more specifics on the hardening process you used? Our machine shop says they can heat treat "up to 2000 degrees" but need to know what "process" you used. This is not my area I am just trying to pass along the question from the machine shop guys. Is this a special process that must be sent out to a special heat treater or is it something we can do ourselves? If it must be sent out where do we send it and about how much does it cost? Again, thank you very much for posting the white paper on the transmission. I think there are many teams drooling over it! Thanks, Chuck |
Re: Slickest drive transmission yet...
Deepwater, Thanks for the compliments: to answer your questions:
1. The wave washers are used as needed to take up lateral slop in the assembly. If you make everything perfectly, you don't need them. If you have student welders and lathe operators like I do, you may find you need some. 2. You can use steel washers if you like. We typically use aluminum washers for everything on FIRST robots. It is amazing how much easy weight you can save if you have hundreds of washers on you machine. 3. The Grayhill sensor is just a suggestion...you can use any similar device here. The number of counts per revolution you choose will depend on much resolution you need/want. I like this one and we made a interconnect so that we can disconnect it right at the transmission. 4. Hardening, I had our heat-treat lab do this for us....I told them to make them hard and they did. Being a programmer I am somewhat ignorant on the details of metallurgy...another great opportunity to learn through FIRST! I will get details and reply later. |
Re: Slickest drive transmission yet...
Jim,
Thanks for the additional information. That's what I was assuming on 1, 2, & 3 but just wanted to make sure. I don't know anything about hardening/heat treating either but if you can relay the info from your folks I'll pass it on to our folks. Thanks, Chuck |
Re: Slickest drive transmission yet...
Quote:
Funny you mention this. Heat treating came up the other day in a 229 design discussion. If you (or anyone else) knows more about this process. Please share! I have the vaguest understanding. I'm especially interested in "home" heat treating. I think it is possible to temper a shaft using only a torch and a coffee can of oil. Is this true? John |
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Quote:
I had seen that before, but obviously forgot COMPLETELY about it. Thank you for bringing it to my attention. Remember folks, don't be like JVN -- search before you post. John |
Re: Slickest drive transmission yet...
Quote:
If you do a water (instead of oil) quench for the hardening/tempering steps, you can improve the results and consistency of the hardening by using an appropriate surfactant in the water. As you quench the part, bubbles form in the liquid at the interface between the hot part and the quenching fluid. You want to move the bubbles away from the hot metal as quickly as possible, and and keep a solid surface of quenching fluid against the metal. The surfactant will break the surface tension around the bubbles, and allow them to quickly seperate from the metal and float free in the fluid. This will allow the fluid to restore contact with the metal, and lets it do its work cooling off the part. The "magic mix" that I use is 1/4 cup of Cascade dishwasher powder per gallon of water. DO NOT use a liquid dish soap or laundry detergent! You want something with a low "suds content" that will not foam up when you stir it around (remember, we are trying to avoid bubbles here!). If you are using a low-carbon or mild steel, you will want an alternative to the quench-and-temper process (which is really only effective on high-carbon steels like tool steel). For low-carbon steel you can do small-job case hardening (which will basically add carbon to the surface of the metal and increase its ability to be temperature hardened). Heat up the part until it is cherry red, bury it in a surface hardening compound, reheat it and cool by quenching. The compound (I use Kasenit, available from MSC, and there are others) will provide a surface case hardening a few tens of mils deep, which should be good for most applications that we will run across. The case hardening will give you a harder surface, but the effect will not penetrate as deeply into the metal as the quenching/tempering process. You will have to figure out where you want to be in the hardness/depth trade space based on your application. (note: there has been a lot of discussion recently in the hobby machinist community on this topic recently - if you can grab a few of the recent issues of the Home Shop Machinist, there are some good articles that will provide additional information) -dave |
Re: Slickest drive transmission yet...
i am sorta new to all of this. would it be possible for someone to e-mail me an inventor file for a 2 speed transmission. that would help us out alot!! also if its possible please send the specs of the transmision (i.e. gear ratios)
thank you!!! |
Re: Slickest drive transmission yet...
Quote:
|
Re: Slickest drive transmission yet...
Is there any timeframe on when the companion automation software paper will be available? Our team would really like to see how you automated the shifting. Will it be available soon so that we may decide if it is something we can tackle this year or not? Even just some pieces of code would be helpful at this point.
Thanks! Chuck D. |
Re: Slickest drive transmission yet...
I too would love to see the code for the automatic transmission. Would any team be willing to post the automatic transmission portion of their code on this site? Even pseudocode would be helpful with our designs since we have never attempted this before. Thanks! :D
|
Re: Slickest drive transmission yet...
Having a shifting transmission is deffinatly a helpfull and worthwile modification; especially when you don't even have to worry about shifting. The only problem that i can see that might go wrong is that if you hit something at full speed and the wheel speed sensor gets knocked off or is rendered unable to sense wheel speed, you would be stuck in the gear that the robot was in before the sensor was disabled. Especially if you were in 4th gear. That battery might not last much longer! :eek:
|
Re: Slickest drive transmission yet...
Quote:
which sensor is that? Isn't it built into the motor? off topic: has anyone built that spacer for the single sprocket if so where did u find a diagram and what matters to use |
Re: Slickest drive transmission yet...
In response to the above posts:
artic_raven: Cut one of these in half and you have a VERY easy to machine 2 speed transmission, not to mention tiny and light. DeepWater and cabbagekid2: I have no idea if Jim intends on posting a software whitepaper, and if so, when. 1028_Machinist: There is no wheel speed sensor. If you look at the whitepaper, the output speed is measured by a shaft encoder. That is what the automation uses. If you knock that loose, I don't think you'd have to worry about driving around in any gear....I don't think you'd be driving around at all. You can put a manual override if you want, though. Nitroxextreme: Wheel speed sensor would be a banner sensor (light sensor) aimed at a tone wheel. A tone wheel is basically just white and black stripes, and the banner sensor reads how fast the wheel is going by "seeing" the change between white and black as the wheel rotates. In our design, there is no wheel speed sensor, just a shaft encoder which reads how fast the output shaft of the transmission is spinning. There is no sensor built into the motor. |
Re: Slickest drive transmission yet...
This is a nice easy to manufacture transmission design but the problem is that there is very little->no reduction. In high gear it is 1:1 which means you have to use some external reduction like a Dewalt drill gearbox and on the input end then maybe some sprockets and chains on the output end. Don't forget about all this stuff when looking at how light and small it is. What you see in the picture is not the only part.
|
Re: Slickest drive transmission yet...
I'll repeat my question from the other thread here since I haven't seen the software paper yet. Now that the season has kicked-off, I expect the last thing on your mind is documenting last year's work.
Jim, could you answer some general questions about how your software works? 1. Is shifting based on RPM alone? What other factors are used to determine when to shift? Current? Voltage? The commanded PWM value? To put it another way, are the shift points fixed, or are they different when you are accelerating at 50% throttle? 2. I assume there is some kind of hysteresis. Is it a simple pair of shift-up/shift-down values? Or do you add something like a "dwell time"? 3. Do the left and right transmissions shift at the same time? I'd imagine that keeping both sides in the same gear would make it easier to match speeds and keep the 'bot going in a straight line. Of course I don't expect you to give away all your secrets, Gracious Professionalism or no. But some general guidance about the operation of the shifting software would be helpful. Once again, great job 33! It's a sweet design. |
| All times are GMT -5. The time now is 11:24. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi