View Full Version : Drivetrain Help

11-26-2001, 08:13 PM
Team 166 is looking for some assistance and ideas regarding drivetrains.

For teams that have used variable speeds, what methods did you use?

Did any teams use more than one motor type for a drive system?

Teams that used most of your motors, what did you have left over? Did you have to compensate in other areas because of weight? What motors did you use for what purpose?

Any other assistance you could give would be greatly appreciated.

--Jordan McNerney
Captain, Team 166

Clark Gilbert
11-26-2001, 08:19 PM
May i direct you to the WhitePaper sections?

There are 2 drive assembly word documents that include some explinations and full cad prints. One is called " Dual Motor Gear Switching Assy" and the other one is called "Gear Switching and Motor Mount Design"

Also notice that the first one uses a fisher price motor that is "piggy backed" to a drill motor (not an easy task to accomplish, but it works great) and the other one only uses one drill motor, both include gear switching.....The "Gear Switching and Motor Mount Design" has been out for acouple years and has been tested and used by numerous teams....

Hope this helped...

(EDIT: You can also get all of the above at http://www.technokats.org/robot.html )

Dan 550
11-26-2001, 08:36 PM
Talk to the TechnoKats and Joe Johnson about motors. Check out CD's White Papers (http://www.chiefdelphi.com/forums/papers.php) for a bunch of motor stuff. You can download the TechnoKats' motor manual in the White Papers (http://www.chiefdelphi.com/forums/papers.php) or click here (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=21) for direct download of the PDF. For a slightly condensed version, check out the Word (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=9) file. For regular drill motors, consult Joe Johnson's Drill Adivise, a Word file here (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=16) , and his torque-handling drill shaft (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=8) . He's the real King of Motors, and he has a really nice spreadsheet (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=15) and presentation (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=14) . He's so great, he had to have a second presentation (http://www.chiefdelphi.com/forums/papers.php?s=&action=downloadpaper&paperid=7) ! Well, you see the hyperlinks. Go ahead, use them well!

Andy Baker
11-26-2001, 10:04 PM
I'll try to help. You probably already know most of this stuff, but I'll post it anyway.

Variable speeds:
Many teams use variable speeds. There are a few ways to do this.

..Using the Bosch Gearbox gear switcher:
We've been switching gears by using the drill motor gearbox gear switch since '99, but team 73 used it before we did. Many teams use the "servo" provided in the kit to swtich gears, but team 71 used a small air cylinder to do it last year... that method is probably quicker than the servo. This is a easy way to switch gears that uses almost no robot weight, but your robot can only switch gears while standing still.

..Have a separate set of wheels turning at different rpms engage the carpet when you want to change speeds:
Teams 33 and 65 did this very successfully in 2000. This is VERY effective, since you can switch gears on the fly, but it takes alot of hardware and uses up valuable weight.

..Build your own transmission gearbox:
We built a CVT (continuously variable transmission) for our 2000 robot, but it didn't work too well and had too many losses in it. I hear that there are some other teams (Paul??) who may have some neat designs up their sleeves in this area.

The holy grail for changing speeds is to make something that lets you change speeds at full power and full speed, and it has to be simple, easy to maintain, and lightweight.

Many teams have learned that the best motors to use for the drive system is the drill motors. Also, the fisher-price motors are pretty good. Some teams use all 4 of these motors in their drive system. If any FIRST team wants to be the fastest or most powerful, then they better use 2 Bosch drill motors and 2 Fisher-Price motors.... that is, of course, IF FIRST provides the same motors in next years' kits.

--Tank style systems are used by the majority of the FIRST teams. This is a simple systems that can be used for power and speed. All teams who switch gears use some sort of tank style system. Most of these systems have 4 driven wheels or treads, but some systems use only 2 driven wheels accompanied by 1-2 casters.
--Crab style (or swerve) drive systems provide better maneuverability than the tank drive system, but are more complicated to engineer and develop. Some very impressive drive systems have been developed over the years, led by teams 47 and 11 (and others). While highly maneuverable, there has not been a crab system which can mechanically switch gears yet. So, while agility is high with the crab style system, it is difficult to get both high power and high speed.

--The van door motor is a good motor to use for the toughest job other than drive system. Many teams' arms are actuated by 1 or 2 van door motors. Be careful that these motors have plastic gears in them and can be stripped.
--The globe motors and window lift motors are also good for actuations. The rpms are low and manageable. Be careful to not put the mounting screws in too far on the globe motor... somehow these screws can suddenly grow and seize the gearbox of the motor.
--Seat motors are mainly used these days to collect balls (by driving a spinning roller) or to actuate something by driving a leadscrew. These motors have high rpms and have nice hardware which go with them in the kit.
--Torque motors were only used as triggers last year... if they were used at all.

Mounting a motor:

If you want your motor to last, take care of the shaft. Don't assume that the motor can take any side loads. Use ball or oillite bearings to support the exposed shaft of the motor so that the shaft is not cantelevered out into space.

Take care and good luck.

Andy B.

Ken Leung
11-26-2001, 11:31 PM
A lot of the presentation in the white paper talks about the science behind motors. While it is great to understand those science, sometime it is more important to experiment with the actual motors itself.

Experiences tell us that the drill motors are great for drive train, and the science tells us that drill motors are the most powerful motors among the collection of motors in the kit. You should at least have a pair of drill motors in the drive train to achieve a good combination of speed and pushing force, and it is up to you to gear the drive train to the speed you see best fit for the purpose of competition.

Take last year's competition for example. If your robot isn't pushing the goals up the bridge, chances are, you want speed more than pushing force to get across the playing field quickly (although you still need certain pushing force to get up the bridge or across the center barrier).

A lot of teams shift gears or use alternate wheels because they don't have enough power from the drive train motors for both high speed and high pushing force. Which is a pretty good solution, since you probably won't need both speed and force at the same time. But there are chances when the gear shifting fail... Well, let's just say it won't be 100% successful rate. And, you will need some percise machining to make a good shifter.

As Andy said, the drill gearbox gear switcher is pretty good at shifting with servos controlling it. You just have to mount the servo the right place to keep gear switcher in place.

But I've seen HOT bot's shifter (67) last year, which switch outside of the gearbox with pneumatics, and it works really well. Basically, imagine 4 gears on two shaft, A B C D with A and B on shaft 1, and C and D on shaft 2. Shaft 1 is the one powered by drill motors, and shaft 2 connect to the wheel. So, by sliding C D on 2, you can only engage C onto A or D onto B at a time (because AB and CD have different distance apart). And AC have one ratio, BD have another.

1-------A-----------B-------- 1-------A-----------B--------
2-------C------D------------ 2------------C------D--------

But my understanding is that the HOT bot didn't need to shift gear that often.

And, it is pretty tough building a great shifter... even tougher for the "holy grail" for changing speed.

MY feeling is that more motors on the drive train with more power is easier to build. Although it will use up some of the weight and space, you won't worry about stoping the robot half way to shift gear, and being pushed while shifting... With 4 motors, you can build up quite some momentum AND push other robots with full force if needed ;) ;) ;)

I thought a really good 2 powered wheel with two idle wheel system worth mentioning is 131's 2000 system. Their drive train was a 4 motors system, with two drill and two fisher price (second most powerful motor), and another pair of idle wheel.

--------Drill powered--O----fisher price O-----Drill--O----idle--O

Their 4 back wheels are mounted on two bar on a pivot, so they can either rotate the two fisher price powered wheel on the ground for more power, or two idle for greater manuveriability.

Dan 550
11-27-2001, 05:45 PM
Hey Andy, what did you guys make the CVT out of? Was it a hydraulic type?

Andy Baker
11-27-2001, 10:04 PM
It was a toroidal disk-type. Two steel disks transferred rotary motion between the concave toroidal disks. We used steel disks, but there was alot of friction loss and efficiency loss, so it wasn't worth it.

But, as a result of this effort, one of our student alumni recently got hired on at GM Powertrain as a CVT co-op engineering student (yeah, Phil!).

Andy B.

Matt Reiland
11-28-2001, 11:43 AM
This is a little off topic but is anyone out there up to giving out hints on manageable gear ratio's for drive systems. I have seen some really fast set-ups? are they too fast to drive effectively? Last year we used the Bosch and Gearbox with a chain to 3:1 to 8" wheels and we had more power than we needed but not enough speed. It would be interesting to hear some of the ratio's people use if it isn't a total secret

Joe Johnson
11-28-2001, 01:06 PM
To a rough order of magnitude, the following is a good rule of thumb that I tell folks often:

A good starting point for gear ratio for the drill motors is shown below:

(Diameter of your wheels measured in inches) * 10 = Ratio from the armature of the motor to the wheels

For example: Given an 8 inch wheel, you need an 80:1 ratio from the armature to the wheel. Given a 2 inch wheel you need to have a 20:1 ratio.

Note that the drill transmissions have a ratio of ~20 in "HI" and ~64 in "LO"

Again, this is only a rule of thumb but it is a pretty good one as a starting point. It seems to balance pushing torque and top speed pretty well.

You need to really do the math if you want to use some motor other than the drill motor as a drive, but that is beyond the scope of this note.

Hope this helps. Others please comment.

Joe J.

P.S. Generally speaking it is better to run the transmission in "HI" and provide an additional ratio via chain or by some other means. Essentially, the main benefit is reduce side loading on the drill mounting points & drill shaft. Again, this is not a hard and fast rule, but it is a good place to start -- moving away only when there is good thought and valid reasons behind the move. JJ

Matt Reiland
11-28-2001, 01:58 PM
Thanks Dr. Joe

Paul Copioli
11-28-2001, 05:09 PM
In reply to Andy's post. The CVT you used will actually work quite nice, but you have to add a lot of grease/oil (too much, in my opinion) in order to get the efficiency up ... it is a very good idea. The variable speed transmission 217 will use this year (if the additional parts list allows it) is base on a dual motor planetary gear system. Picture this ... the output from the drill motor transmission is the sun gear for another planetary stage. We use the fisher price motor to rotate the ring gear (as opposed to being fixed), which yields many gear ratios at a constant drill motor speed. The trick is to match the torque curves between the fisher price and the drill motor, because the fisher price hase to be able to react with the torque being produced by the drill motor. The reaction torque relationship is HIGHLY dependant upon the speed relationship between the ring gear and the planet gears and I have spent most of my time figuring out ways to predict the relationship. I have the drive 75% completed and should have results before X-mas holiday. I know this is a confusing post, but I am trying not to give too much away and at the same time trying to explain how the transmission works.


Dan 550
11-28-2001, 07:12 PM
Hey Paul, I hope you either get that curve of the FP to match the drill really close, or you have a lot of FPs...

PS to any TechnoKat: Do you guys have a pic or drawing of some sort for your CVT? I'm a very visual guy, and kinda missed how your CVT was put together.

Clark Gilbert
11-29-2001, 04:25 PM
Guess what Dan!!!

Your in luck.....i brought the CVT home today and took some good pics of it.....i'll post them on my website so i can link them here...

:) :D


EDIT: (so theres a little rust on it..hey its been sitting around for a year) If you get a darn red X, it seems that freewebspace.com is having problems again....so i'm only uploading 1 picture. Dan or anyone else if u want to see a back picture or some from the sides just IM me with AIM at twinattack

Joe Johnson
11-29-2001, 11:11 PM

I don't see how the planetary gear set up you describe can give you what you are claiming it gives you.

Here is my dilemna. One of the chief benefits from my point of view of a CVT is that for a given speed & torque of the output shaft, I can pick a ratio and voltage that will essentially have my motor running at some desired point on its speed torque curve (for that voltage). For example, I can choose to run my motor at its peak efficiency point or at it peak power point. The key idea is that I can always choose a ratio and a voltage that will keep my motor running at this desired point, regardless of what the speed and load conditions are on my output shart (within the bounds of my CVT ratio of course).

I don't see how this is done for the case you describe.

I am willing to be impressed when I see it work.

Joe J.

Ken Leung
11-30-2001, 03:41 AM
I actually dreamed about how to make this kind of CVT work last right, and it kept haunting me every time after I open my eyes in terror from the major confusions/problems I had to face every similar ¡§nightmare¡¨... Ever since then I wasn't able to get any decent amount of sleep... I wanted to figure this out so badly¡K

So I spent a lot of time thinking, and here is what I learned...

The way drill transmission switch gear is to slide one of the ring gears in one of the planetary gear stage to two positions: one where that ring is locked into place, and another where the ring spin along with the planetary gears...

So the speed of the ring gear determine how fast the over all transmission go, and thus the torque it can generate. If we are able to control the speed of that ring gear, then we can change the ratio of transmission on the fly...

There is a point when the ratio is just right such that the load on motor makes it most powerful/efficient, but it would be hard to reach that ratio just by running the ring at different speed. So, why do that? Why not gear the motor such that it let the motor to be more powerful/efficient when the ring gear is locked in place by the motor? And then spin the ring different direction to get more speed or torque?

For the drive train, given different situation there motor will require a different ratio to be most powerful/efficient...

Say free speeding, as the robot go faster and faster until motor get closer to free speed, and there's less and less load on the motor until the friction in the system is only load left. At that point the gear ratio can decrease so the wheels can go even faster... And that will get more power out of the motor (with bigger load and faster speed). And the high speed is the most important thing we can get out of this.

OR, if we are pushing against other robots. The drive train is often stalled because the opponents¡¦ robots have a similar ratio/traction. So, we increase the gear ratio until the motor stop stalling or opponent start moving. We won't be able to tell when we are getting the most power out of the motor, but at least we are getting something instead of just stalling motor. And the most important we get out of this is the torque...

I think a CVT will be beneficial for the robot, if we do it right. Say we pick a good ratio when we lock the ring gear in place such that the component is running a optimally (like a default ratio), then a CVT will be able to help you adjust the ratio at different time for different situation.

Well, about matching the speed torque curve of F-p with Drill's, I didn't really understand why at the beginning.

All I knew was that the ratio is changing as long as the ring gear is moving. It¡¦s like two racks with a pinion in between. When one of the rack is fixed, the gear move along with the other rack... When both rack move toward the same direction, the pinion travel along with them without spinning... Or when the two racks go opposite direction, the gear will be spinning but won't be moving in any direction... So, the rotation of the ring gear determines precisely how fast the output would rotate.

But that's an output shaft spinning with no load. When the output shaft is under load, the ring gear and the sun gear would react differently. So, a matching speed toque would help to get the reactions from ring gear and sun gear match up. It would require setting up a right gear ratio between F-p and ring gear, and the planetary gear ratio just right as well... And a lot more stuff you have to do to get it just right (which I can't think of)... But the main issue is to keep the two gears¡¦ reaction the same, so that the over all ratio stay the same. Unless the ratio increase with more load, which will be a good design because that setup automatically great more torque with more and more load onto it.

Hmm... This only raise more questions I have about this... Starting to involve physics and other science¡K But I will continue to study this until I find the right answer, or when someone posts it in the future¡K

Paul Copioli
11-30-2001, 11:46 AM
Dr. Joe,

Everything you said is true, but if you are willing to give up the constraint of keeping the motor running exactly at a specific desired point, then the planetary setup is viable. I admit that there is a lot of testing and compensating involved, but I have a lot of data so far and am really close to getting it to work as I have predicted. My biggest obstacle is pushing my way through P-basic (I am, afterall, a mechanical guy) to get the control system to do what I want.

Matt Attallah
12-02-2001, 11:14 AM
Team 5 used the speed changers that came with the drill motors. It worked ok, but our main problem was getting that main hit in-between speed and power. On "Low-Gear" we could pull both goals filled up and balance with no problem, but our tank traks didn't have enough friction to get up the ramp. We had no problem turning with them in low gear. When we went in to "High-Gear" our robot couldn't turn it's self, just go straight very, very fast. (Well, we could turn, but mabye at 15 degrees at a time, after that, we poped breakers!) So, this up comming year, if we can get some $$money$$, we will try to design a nice gear ratio for both sides of the equasion...