Gearbox

If you want to have a four motor drive system with the cims and the drill motors do they have to be running at exactly the same rpms or just about the same?

*Originally posted by Lord Nerdlinger *
**If you want to have a four motor drive system with the cims and the drill motors do they have to be running at exactly the same rpms or just about the same? **

You should try to match the free speeds as close as possible. I think ours are within 5% error.

Since the characteristics of the motors vary so much (in my experience) it is not critical to be overly accurate.

Anyone wanna back me up on this?
How much error does everyone else have in their speed matching?

John

Right, the speeds don’t have to be exactly the same, but closer is better.

Just out of curiosity, what is your objective of using two motors?

  • Patrick

see my reply here

I have wondered the same thing.

This is the conclusion I came up with.

If you want the two motors to run in unison; thus acting like a single super motor then yes they have to be geared so that their free wheeling speeds match closely.

However If you want to have one motor be the loaded motor in the low RPM output range and the other motor be the loaded motor in the high RPM output range then no. The only concerns in this setup are. First that the motor that is gear for the high range doesn’t stall out in the low RPM range I.E. don’t power it until it comes with in its own range. The other concern is the motor that is geared for the low range is uncoupled at the high RPM range so that it doesn’t explode from over revving.

The next question one might ask is how is this accomplished?

First you need two motors, Motor A will become the motor for the low range and Motor B will become the motor for the High RPM range.

Next you gear the motor for they desired ranges and couple them to the output shaft.

Next you’ll need to attach an optical encoder to the output shaft so that you know what the output shafts RPM is. This reading will govern the power for the high range motor (Motor B) turning it off at low RPM and for the low range motor (Motor A) disengaging it form the system at high RPM.

Here’s the hard part. You need a device that disengages the low range motor (Motor A) from the system when the output RPM goes too high and the motor is in danger of exceeding it’s RPM max. Possible devices are a clutch either frictional over positive engagement or a device that I call a “captured ratchet clutch” meaning that the only time that the motor is coupled with the system is when the motor is powered.

The captured ratchet clutch or CRC has an Input plate and an output plate. The input plate has ratchet teeth on the face of it these teeth engage similar depressions in a receiving plate that is held to the input plate by a spring. So that when torque is applied the ratchet teeth on the input plate try to ratchet over to the next hole. While in the process of doing this it forces the receiving plate to push out however this action of pushing out the receiving plate causes the receiving plate to engage with the output plate using dog teeth and receiving holes on the output plate. However the distance between the receiving plate and the output plate is less than what is necessary for the ratchet teeth to disengage and move over to the next hole but that same distance is just enough so that the receiving plate doesn’t engage the output plate when no torque is being applied.

The previous explanation is best understood through a diagram but I don’t currently have time to draw one. I will try to post a follow up with a diagram included.

An example of a combined low and high drive system is this. There is this grinder that we use at school and because of the motors gearing it doesn’t have enough stall torque to start the grinding wheel spinning so we have to use either our hand or the object the we need to grind to give the grinder a kick start.

This is a fairly difficult concept that needs some exploration. I’m not sure how advantageous this system is compared to a same speed two motor system. But I think that it deserves some exploration.

Just food for thought, anyway you’ll be better off just gearing two motors the so that the speeds match.

Before you go ahead and design a multiple-motor drive system, think it through. Multiple motor systems do not achieve what most suspect.

The physics:

  1. The amount of force that can be transferred to the ground is limited by the effective coefficient of friction of your wheels (or treads). Once this force is exceeded, your wheels are now slipping and provide a smaller, constant force with the ground.
  2. The amount of force that a motor can transfer to the ground is a function of gear ratio and wheel size (and applied voltage). This force is limited by the voltage your battery can provide (~13V)

Combining points (1) and (2) yield my main point (3):
3) There is a physical upper bound on how much “pushing power” (by that I mean force your robot can apply at stall) your robot has! As long as you choose an appropriate gear ratio for your robot, and you have wheels or treads that maximize your traction, you CAN NOT improve your pushing power.

Other lessons to be learned:
4) For given motors, you can maximize your pushing power AND speed simultaneously by measuring your coefficient of friction. Once that number is obtained, you can gear your robot such that at stall, it provides exactly the amount of torque required to make the wheels just begin to slip. This ensures that you can push as hard as your robot is capable of, and at the same time, maximizes your robot velocity.
5) The only thing to be gained by adding multiple motors in a drive system is velocity. For example, if I put two drill motors in series instead of one, I can now cut my gear ratio in half, doubling the acceleration and maximum velocity, without sacrificing any pushing power. If geared properly, you CAN NOT improve pushing power by adding motors.

I have seen many teams over my the years add multiple motors only to see their wheels spinning non stop in a pushing war, while their opponent is moving them with ease. Multiple motors in series can be used to give you speed. But if pushing power is your objective, look elsewhere (i.e. get better traction with the ground). Once you do that, then you can recalculate your gear ratio, and if your robot moves too slow, only then consider multiple motors.

  • Patrick

P.S. There are three assumptions I make when choosing the appropriate gear ratio for maximizing your pushing power. You can actually get more pushing power than my point #4 solution above (without changing the coefficient of friction). Can anybody figure out what my assumptions are, and how you might actually get better pushing power than my solution?

Gee, the above post came kind of shocking to me. I seem to recall that word of team 696 having incredible speed and torque was flying around the pits of the Phoenix regional (and LA too for that matter) Strangely enough, they were running a four motor drive train (drills and chias). I wonder how they did it? Searching my memory, they actually pushed both opponents at once in one match and pushed a bin under the bar in another match. Not to mention a total of something like 11 opponent flippings in the one regional. Of course, all of this performance couldn’t have been because of the 4 motor drive, I mean c’mon.

*Originally posted by sanddrag *
**Gee, the above post came kind of shocking to me. I seem to recall that word of team 696 having incredible speed and torque was flying around the pits of the Phoenix regional (and LA too for that matter) Strangely enough, they were running a four motor drive train (drills and chias). I wonder how they did it? Searching my memory, they actually pushed both opponents at once in one match and pushed a bin under the bar in another match. Not to mention a total of something like 11 opponent flippings in the one regional. Of course, all of this performance couldn’t have been because of the 4 motor drive, I mean c’mon. **

Everything Patrick said is valid.
Your analysis of what he said is either flawed. Or you do not understand the physical principles involved. Perhaps you should read it again. Ask for help if you do not understand.

Sarcasm is not necessary. Neither is being rude.

John

*Originally posted by patrickrd *
**P.S. There are three assumptions I make when choosing the appropriate gear ratio for maximizing your pushing power. You can actually get more pushing power than my point #4 solution above (without changing the coefficient of friction). Can anybody figure out what my assumptions are, and how you might actually get better pushing power than my solution? **

Increase the normal force acting on the wheels, either by lifting something, adding more weight to the robot, or pulling down on something (like suctioning to the HDPE). This will result in better traction (F = Mu N) and the ability to push more.

*Originally posted by sanddrag *
** Of course, all of this performance couldn’t have been because of the 4 motor drive, I mean c’mon. **

It has a little to do with the 4 motor drive train of course ;). Just not as much as you think it is.

Consider the wheels for the moment. If you have frictionless wheels for the sake of the argument. Then you will have NO speed or pushing force at all, no matter how many motors you put on the drive system.

Take a look at the formula of friction for a moment:

F(s) = Mu(s)N where F(s) is how much friction force you can get out of your material, Mu(s) is the coefficient of static friction of your material, and N is how much weight you put on the material.

If you have lots of Mu(s), and no N, that means you have really good wheels with high friction, but no weight on the robot, which will yeild a low friction force. If you have lots of N, but no u(s), then that means you have a really heavy robot, but no friction on the wheels at all, which again yeild a low friction force.

Should be fairly easy to understand, right?

So, now that you know what wheel you use and how heavy your robot is limit how much force you can put on the wheels, let’s look at the motors side. Every motor can output a certain amount of mechanical power, and mechanical power = force X velocity. You may set up the gear ratio to setup different combination of force and velocity, but at the end, you can only get so much power out of the motors. With more motors, you have more mechanical power. But that’s just power.

Consider this simple example:

If you have 10 unit of mechanical power with one motor, you could setup a gear ratio such that you have 2 unit of velocity and 5 units of force, or 5 units of speed and 2 unit of force. If you have 2 motors, you have 20 unit of mechanical power, and you can setup a ratio and get 4 units of speed and 5 unit of force, or 10 units of speed and 2 unit of force.

In both cases, you can see how 1 motor and 2 motor setup could yeild the same amount of force. Only the one with more motors will go faster.

So, it’s really comes down to how you setup your robot. If you have lots and lots of pushing force from the motors, but very crapy wheels, you won’t be able to push anything around. If you have good traction wheels, with lots of pushing force, but only 1 motor, you can push everything around all right, just at 1 ft per minute.

In your case, when you have a strong AND fast robot, you can out push people because you can build up your momentum under a short distance. Otherwise, if you are in a deadlock with another robot with just as much pushing force but with only 2 motors instead of 4, and equal amount of traction, you won’t be able to outpush it.

If you really want to learn all these, check out my notes in whitepaper. The title should be “WRRF motor selection lecture notes”. Then you will understand in Engineering we don’t do too much guess work ;).

http://www.chiefdelphi.com/forums/papers.php?s=&action=single&paperid=23

We used four motors in our drive system this year, 2 drills, 2 CIMs.

We used the first stage of the drill motor gear box, which matched NL speeds pretty closely between the two.

The drill motor is the more powerful motor. So, we chain drove a set of front wheels off the drill and direct drove the middle set of wheels.

The CIMs were connected to the rear set of wheels.

We had a total of 10 skyway wheels in our drive system (2 sets of duallys per side and one single on the passive axle).

We picked a gear ratio that gave us about 4.1 fps top speed with six in diameter wheels.

We had a ridiculous amount of pushing power, although we could not quite push the entire field boundary when we ran into it in autonomous.:wink:

We used the six axle approach, rather than the four axle approach to avoid high-centering on the ramp.

We are in the process of doing some sensitivity tests right now to tune the vehicle speed/pushing power versus motor commands.

As an aside, the drill motors have a distinct assymetry in forward/reverse. More so than the 2002 drills.

Some things which have not been raised in this thread (or others) about speed selection.

  1. gearboxes have an efficiency. Torque out does not equal torque in * gear ratio. The more stages you put in your gearbox, the lower the efficiency. (not to mention machining time, weight and cost of gears) In an extreme case of a poorly made gearbox with ultra high gear ratio, the thing can stall your motor without providing any output torque.

In a FIRST system, you have a limited energy rate source. Ie. you can only draw so many amps per second. The less power you throw away due to inefficiencies, the better you are. This is especially true in the drive system, which is probably responsible for 60% of your resources.

Even though you may achieve the same GR with a worm and gear or a bevel gear or a spherical gear or a planetary gear, you should probably only use spur gears arranged in a conventional train. You should choose the smallest diameter wheel to achieve your overall goals.

  1. Motors have an optimum efficiency point and a max power point. As you approach the max power point, you draw more current. In the case of the drills, this is an insane amount of current.

This will drain your main battery quicker, heat up your circuit breakers, heat up your motors. As your motors get hotter, they become less efficient. Ie, you start to lose power over time. You also start to degrade your motors. This last effect is seen in the NL current.

Bottom line, you don’t want your drive system to just slip at your maximum pushing power. You want to have excess capacity.

Then, if you can “acquire weight” through some kind of redirection (picking up a goal, bins, whatever) you can make use of the additional drive system power.

Because of this excess capacity, you will be running closer to your “max efficiency” point during the majority of operation. Hence, when you finally do get into a pushing contest, your components will be less stressed than the guy who was faster, but who heated everything up just running around.

For those of you interested, I uploaded a newer version of the lecture note in the white paper section about motors. It’s at: http://www.chiefdelphi.com/forums/papers.php?s=&action=single&paperid=197

Although it’s lacking specific advices about drive system, it should be enough for the basic of motors. Let me know if you have any questions.

This is just my opinion but, I believe that using two different types of motors is bad news waiting to happen.

*Originally posted by dddriveman *
**This is just my opinion but, I believe that using two different types of motors is bad news waiting to happen. **
It was good news for us. Funny thing - our four motor is the only robot of ours running right now.

In a real engineering project you would just use more of the same motor or a more powerful motor but in FIRST, you learn to work with what you have.

True that.

This is just my opinion but, I believe that using two different types of motors is bad news waiting to happen.

I predict that four motors in the drive system is going to become a “must” as time goes on.

I’m still reserving my opinion on shifting.

I believe that Andrew is correct. I don’t want that to happen but, as the games get tougher and as FIRST wants to see what we can make out of what we are given it could happen in the near future.

I used to be a multi motor skeptic.

This year, I became a believer (at least in some instances).

With this year’s rules allowing pneumatic tires, it was easy to get enough traction to make multiple motors worth the bother.

We initially had a 2 motor system, after 2 regionals and a ton of hours on our practice robot, it became clear that more power to the wheels was a big factor (and this with a shifting transmission).

The MAIN improvement (imho) came from having multiple current sources.

When you have a lot of traction and a reasonable top speed (10 ft per second, for example), turning requires that the motors run too close to their stall torque for too long, resulting in tripping breakers.

With multiple motors you share the torque somewhat – this helps – but more importantly you have multiple paths to draw current from. Rather than being limited to 40 amps (nominal) you may have twice that. If you are geared to be on the fast side and you have a lot of traction, the 40 amp breakers become your power limiting factor rather than overheating the motors or breaking the tires free.

From our experience this year, it can be worth the bother of designing multi-motor drives.

As to matching free speeds or matching stall torques or matching some other speed-torque point, again, I restate that there is no real magic here: The two motors act like a single motor with different characteristics.

You do not have to match free speeds – Really.

The story is made a bit more complex by the fact that you can give different voltages to the different motors but it is still a fairly simple arrangement.

I feel a white paper coming on… … stay tuned.

Joe J.

The MAIN improvement (imho) came from having multiple current sources.

Hear! Hear! This is the thing which drove us to multi-motors.

Another advantage (especially if you direct drive separate wheels with your multi-motors) is that you have additional failure modes. Even if you trip your breakers on one set of motors, you can still limp around. It takes a double failure on one side of your robot to put you out of commission.

*Originally posted by JVN *
**You should try to match the free speeds as close as possible. I think ours are within 5% error.

Since the characteristics of the motors vary so much (in my experience) it is not critical to be overly accurate.

Anyone wanna back me up on this?
How much error does everyone else have in their speed matching?

John **

I’ll back you up. We have about 1.2% error at max power output and about 3.5% on free spinning. This drive train is planned for next year though, so I can’t say too much. Last year was about 4% error on average, using the drills and cims.

It is easier just to match the max power speeds of the motors, since that is the time when you really need the error percentage to be less.