Shifting Gears

Hello all. I’ll keep this short and sweet.
My team and I have decided that we are going to build a multi gear gearbox so that we can shift gears on the fly. We have 2 solid ideas on how too do it and were currently doing all the math. I would just like to here from others any difficulties you had with changing gears. Any words of advice or ideas would be greatly appreciated.

Keep it simple :smiley: so when it do brake down it would be easier to fix.

Make it robust so it won’t break down. (then it can be as complex as you want:p)

Shifting on the fly is a challenge.
Bevel the teeth (grind them off at a 45 degree angle) of the mating gears. It makes the shifting smoother. Not all the way, mind you. Just on the edges.
We have tried programming a “synchro-mesh” system that would match the motor speeds. But up to now the computer power wasn’t there. Maybe this year.
And you don’t have to really shift on the fly. Think about what you are doing at the time you want to shift from high gear to low gear and low gear to high gear. You might be surprised what’s really going on. I was.

You don’t have to mesh gears. You could have a shifter dog mechanism.

^could you elaborate:D

Gregory Frechou:

I second the whole bevel-gears thing… we had lots of problems with the gears meshing until we did this.

Another thing I suggest goes to the programmers. When shifting, even our team’s better drivers had problems getting the speed right. In the heat of the battle, it’s hard NOT to be going in full throttle. As a result, when we were shifting, all of our drivers were putting too much speed in between gears, and the gears were making the clang-clang-clang noise of dhoom from sliding past each other and not meshing. What we did was we used two digital inputs and four limit switches (wired as shown in the 1-min Paint diagram below) to tell us when we were engaged and when we weren’t engaged. At any point in time, the High-Gear OR the Low-Gear input must be 1. If both are 0, that means we’re not engaged. When we’re not engaged, the program limits the speed to something like 1/5th or 1/6th of the maximum speed - slow enough for the gears to finally mesh. This system is nice because its fully automatic - the driver can push as hard as he wants on the joysticks because the robot automatically limits the max speed.

Depending on how easily your system shifts, you might also consider having the program automatically go forward and backwards. We found that if you jiggle between forwards and backwards, it shifts faster. It’s something you need to test out on your system and see how it works for you.

Hope all this made sense… it’s 1am and I’m falling asleep, so my grammar probably isn’t up to par. If you don’t understand something, post and I’ll get back to you in the morning.



Perform all the math, figure out exactly how strong it needs to be, then build it 3x stronger.

I’d like to mention that this design is courtesy of the technokats.

Here are some ideas that we have. Please offer any and all words of advice it’s greatly appreciated. We are rookies at building gear boxes. We’re not sure which equations to use or even were to start. Please help.

*Originally posted by MBosompra *
I would hate to have to make that. Although I would feel very proud after I did.

Why shift gears when you can shift wheels? :slight_smile:

Yeah kickers. 190 did this very sucessfuly in 2001. But i think it takes up amuch more weight than a transmission.

*Originally posted by Rpifirst *
**Yeah kickers. 190 did this very sucessfuly in 2001. But i think it takes up amuch more weight than a transmission. **

It is heavier, usually, but can offer more advantages as well. Not only does it allow you to change speed and output torque, but other things like wheel base, tread, width and type.

I’ve helped do it very successfully in 2002 and, with luck, I’m going to do it even better in 2004.

I just wanted to point out, that on the shaft we had bend, when we redesigned it to avoid the bending, we made it 8x stronger. :wink:

Alrighty, it appears that you’re in sort of a bind (bad pun intended) with where to start, so I’ll sort of give you some (or a lot) of background on transmissions. Please don’t take my words as absolute truths, but instead use them balanced with some other veterans to come to your own conclussions. Some people more talented than myself may disagree with some of what I have to say.

Thoughts on Transmissions in FIRST.

When choosing a transmission design, you have to weigh a few different aspects before making your choice: The most important aspect is your justified purpose of having the transmission in the first place. Robots can maneuver reasonably well without the potential headache, cost, and design time needed to create a transmission.

The foremost purpose of any transmission is to effectively provide an adequate balance of speed and torque for a given situation. Typically in FIRST, a transmission implies more than one torque and speed setting which is adjustable during a round.

In FIRST robotics, your drive train must be able to do two things effectively:

  1. Move from one point to another as quickly as possible, typically in a specific orientation.
  2. Maneuver in “high torque situations” against field playing objects and opposing robots.

Without going into the details of the analyses that I’ve performed, the amount of torque required in “high torque situations” is approximately 2.5 to 3 times as much torque as required for standard and high speed maneuvering situations. Essentially, pushing a robot requires significantly more torque than is required for simple turning, which requires more torque than just moving forward and back.

Some quick facts on transmissions are written below. When I spit out numbers, these are calculated. Please ask questions about any and all of it if you’d like more information.

  1. The proper gear ratio for using the drill motor in low speed, with 6 inch wheels, a 130 lbs robot with a coefficient of friction of 1.2, while not pulling more that 37 Amps is about 2.5 to 1. Using any less than this ratio with those assumptions will lead your robot to trip the 40 Amp circuit breaker while in pushing matches, no questions asked.

  2. The limiting factor in almost all drive trains using a single pair of motors is tripping the circuit breakers of the individual motors. The circuit breakers for the chips and drill motors are 40 amps, while the stall current, (the amount of current that a motor is using while it is exerting is maximum (stall) torque), is over twice that of the breaker. Thus, even if the wheels are spinning due to slip, you can still be pulling more current than is permitted by the breakers.

  3. If you choose to couple a drill motor and an Atwood motor in single side of a drive train design, (using 4 motors total) your danger lies in pulling more than 120 Amps which would trip the master circuit breaker.

  4. Transmissions are (relatively) engineering and manufacturing intensive, as gears have much tighter tolerances than most other features of a robot, such as some sort of macro level manipulator. They are also much more detrimental when the fail, since when your drive train is not working, neither is your robot. Hence, special care and adequate resources must be applied such that a transmission performs effectively. Failure of your drive train and transmission can completely ruin an entire competition or potentially an entire season.

  5. Transmissions, obviously, have a weight associated with them. Typical transmissions have weights between 3 and 8 lbs. per side. Often, this is primarily determined by the quantity of steel required in the gearing and meshing components. Using more than a single pair of motors can further increase weight by 8 to 10 lbs., depending on mounting and gearing required.

  6. Transmissions have relatively high financial costs. Since each gear average ~$25, and you’ll be requiring between 6 and 10 gears per side, the total cost adds up quickly. Often times, for fiscially conservative teams, this means that having replacement gears may not be an option.

With those facts put out there in the open, I’ll talk about my personal thoughts on transmissions. Again, please comment as you agree or disagree.

It should be noted that there is a finite (limited) amount of force that a robot can push on the field. In a four-wheel drive system, this is equal to the weight of the robot times the coefficient of friction (stickiness of the wheels to the ground). Even if you have a max output torque by your motors equal to that of a V8 automotive engine, you won’t be able to push more than the weight of the robot times that coefficient of friction. (which is typically in ideal situations, about 1.1 to 1.5)

With all of this being said, I have calculated that you can move at a relatively high speed, ~10-12 feet per second, as well as push other robots with the use of only a single pair of motors, and a pair of properly chosen gear ratios.

To further extrapolate on this, I believe that it is (though admittedly very debatable) wasteful to use a second pair of motors. Below is my justification.

As I stated, a drive train with a single motor pair using proper gear ratios can reach the maximum amount of torque required, which means that a second pair of motors will only provide a different (higher) maximum speed. With that said, I would like to state the primary disadvantages of using a second motor pair.

  1. The additional actual weight of a motor pair, the gearing needed to mesh with the first set of motors, as well as the required mounting of these motors could add between 8 and 15 lbs to your robot.

  2. You run the risk of tripping your main (120 A) circuit breaker, which means that the entire robot will shut down for the rest of the round. Two motors could pull a maximum current no higher than 80 Amps, while 4 motors could pull 160 amps, dangerously above what will trip the main breaker.

  3. At higher speeds, it is more difficult for the driver to accurately control the robot.

  4. Twice the motor count means at least twice the complexity. Calculation requirements double, nearly twice the machining for motor mounts double, and at least twice as many pieces will require the high tolerances needed for an efficient drive train.

Hence, it is in my opinion that with the current rules that are in place with FIRST robotics, a second motor pair is a waste of resources, since the primary purposes of a transmission can be achieved with only one pair.

However, YOU need to decide: is it worthwhile to spend twice (or more) the machining, twice (or more) the cost, twice (or more) the design time, and twice (or more) the weight simply so that your robot can move FASTER. You need to decide what is fast enough.

Currently, Team 461 is working on a design that could potentially completely level the playing field for many financially strapped and machining limited teams by designing a transmission that:

  1. Requires no machining beyond a drill press and bandsaw.
  2. Has a total cost for a PAIR of transmissions less than $250
  3. Has weight total less than 4.5 lbs per side.

More details are coming soon regarding that design.

Feel free to ask questions, make comments, or corrections.


I agree with 80-90% of what Matt has to say, but the above quoted comment is one that I feel I need to say something about.

In the 2002 season, many of the matches ended up being 120 second pushing matches. In addition, many teams lifted the goal(s) in order to get a higher normal force with the floor (resulting in more pushing force for the same coef. of friction). Finally, FIRST provided 80 Amp household breakers (read: “not rated for mechanical impacts”) for the main power.

This combination resulted a number of teams with N motors per side (N>1) having issues with the breaker tripping, especially if they got bumped during or shortly after a pushing match.

Last year, FIRST switched to 120 Amp automotive aftermarket breakers. Between the higher current rating and the designer’s expectation that cars see more shock and vibration than typical homes (West Coast teams excluded ;-), I don’t know of any teams that had problems with the main breaker tripping last year, whether they had N motors per side or not.

This is not to say that there are not multiple reasons NOT to use N motors per side – there are many, but tripping the 120 Amp breaker is not one of them (imho).

By the way, I have it from usually reliable sources that given the current breaker and the individual 40 or 30 or 20 amp circuit breakers, it is a very close call as to whether it is even possible to trip the breaker short of a metal bar shorting your main power feeds.

Depending on the state of the charge of your battery and the variation limits of your particular battery’s internal resistance and the lengths of your wiring harnesses, it may be almost impossible to trip the main breaker.

Having said I agree with Matt, I will point out one reason for multiple motor drives that he ALMOST points out but does not quite complete the thought.

As Matt points out, in many cases, the 40 Amp breaker is the limited factor on the output of your drive system.

The consequence of this is that if you have 2 motors per side, you have 2X the 40 Amp current limit.

This was the case with our robot last year. In normal driving (i.e. practicing), our robot worked just fine. It had a good balance between top speed and turning/pushing torque. But after a while of “competition driving” the breakers would poop out on us at the whimpiest of pushing matches.

So… …we retrofitted an extra set of motors per side for the sole purpose of pushing back the limits due to the single 40 Amp current path.

For what it is worth…

Joe J.

I’ll take an 80-90% from Joe any day. :slight_smile:

To comment on what Joe is saying, I understand his point. In addition, I have a few hunches as to why there were teams that used multiple motors and didn’t trip the 120 Amp breakers:

Most of the time, teams who are using multiple pairs of motors have excellent designers in the first place, or copied the designs of excellent designers. In pushing matches, the teams were in their properly designed high torque gear set. This should have caused them little fear of tripping much of anything.

However, I would bet the farm that a robot in their high speed gear set pushing up against a wall would trip the 120 Amp breaker before any individual motors would give out.

So I’ll definitely agree that Joe has a viewpoint that is very valid. My point was just a heads up: It theoretically possible to push that 120 Amp limit if your gear ratios are not designed properly. You must design your ratios with a limiting factor of the 120 Amp breaker if you use more than two motors.

If you assume that since you’re using twice the number of motors, you can make your speed ratio twice as high, you’ll be in trouble, since you don’t have twice the power due to the 120 Amp circuit breaker. If the game changes this year to requiring some sort of manipulator that could be pulling another 10 or 20 amps… be alert!

To summarize, I’m giving a theory viewpoint for the new designer, while Joe is calling it like he’s seen it (which is essentially that teams have designed multiple motor drive trains well! Woo hoo! :)).

Thanks for the comments, Joe!

Also don’t totally forget about another reason for multiple motors -


We used 2 per side last year and had the drills fail twice in the season, the previous year the fisher price failed 3 times. However the robot kept on moving since more than one motor was linked to the drivetrain (Although one side had less power we were not a sitting duck that drives in circles)

For our robot last year with Tank Treads, we were able to use a version of the 45 design that gave us the power we needed to push hard and spin the tracks against an immovable object and in high cover the field at a very quick pace. Both of these would have been somewhat less if we didn’t have both the motors teamed together. We still would have had a high and low but they would have been slower. Also anyone considering building a transmission should be lightening out all of the gears to a geat extent to reduce the weight so that the total added is much less than 8-9 pounds per motor. We have shown that 20 pitch gears can be lightened nearly 80% and still hold up to the duty required of them in a 130 lb Drivetrain for an entire season.

So I guess I disagree that more than one motor per side is a total waste in a FIRST robot, it all depends on what you want your robot to do. Too fast for some people is not too fast for others. I have seen many people in complete control of RC Nitro cars at over 50 MPH and I have seen others that can’t drive straight a 5MPH.

One thing about redundancy: It can hurt you some times too.

On several occassions we have had one of our multiple motors fail or partially fail and not discover it for several matches.

The machine works, but it works at a lower level of performance.

This year, our team is planning on putting in a “diagnostic mode” in which every motor is tested while monitoring sensors (current, speed, etc.) for out of range values. We are also planning on having sensors similarly checked automatically.

I don’t suppose that it will find 100% of the problems, but it could help us avoid a full speed crash into the wall during autonomous simply because of an unpluged cable (this really happened last year).

Joe J.

P.S. Breaker ratings have safety factors in them. Essentially, they can carry the rated current for infinite time. At some percentage above that rating, they trip after some time depending on the percentage of over current.

Bottom line, I estimate that a robot could probably dead short the battery for 5 or 10 seconds before it would actually trip the 120 Amp battery (don’t try this at home as batteries do not like such dead shorts). Given our current battery, with the motors we have and the required secondary breakers, I think it would be hard to trip that main breaker.