Shifting Gears

[George1902]

What if you had treads driven by a 3" pully instead of 6" wheels? I'm pretty sure that would satisfy the 10 fps and 150 lbs requirements without burning up.

[Matt Adams]

Originally Posted by JVN

150/2 = 75 lbs per gearbox. Using a 6" wheel we can see that we need 50.84 N*m of output torque from each box to achieve 150lbs of linear pushing force.

Ah ha!

John is using the ole.... "I'll state a forumla and then use different units on the output" trick.

You need 75 lbs on each gearbox. You have a 6 inch wheel, with a 3" moment arm, which means you need 225 in-lbs, or 25.42 N-m. Maybe John used the diameter instead.

This halves the rest of the current requirements down to around 42.5 amps on the drill and 35 amps on the atwood... with a little tweaking they could be both modified to get under the 40 amp limit.

Admittedly, the numbers show that this will be pushing 160 amps altogether... which is a little too high for my blood.

and so continues the thread that never ends...

Matt

[Chris Hibner]

Originally Posted by Matt Adams

To meet the two "competitive" objectives:
A. Top speed of 10-12 ft/s and
B. A pushing force of 150 lbs.

I also want to re-emphasize something that Joe mentioned: the definition of "competitive advantage" really depends upon the game.

I agree that in your typical game, the numbers you present above (especially the 150 lb pushing force number) would be about your pushing limit. However, in the 2002 game, we determined that our robot would need to push with a force of about 720 - 960 lbs before our treads would start to slip. For this game, having 150 lb. of pushing force would not be an advantage at all.

The point is: assuming a peak pushing force number before the game is announced can get you in trouble... unless you have a shifter. Then you can just pop in a new ultra-low ratio. However, your strategy may be such that a shifter is irrelevant.

[IrisLab]

Originally Posted by Chris Hibner

The point is: assuming a peak pushing force number before the game is announced can get you in trouble... unless you have a shifter. Then you can just pop in a new ultra-low ratio. However, your strategy may be such that a shifter is irrelevant.

The what-if discussion has been very important for me (as a rookie mentor). I appreciate the hashing out of the various "theories". It may seem irrelevant to discuss strategies without knowing the game, but the experience that each of you have with regard to prior years and the debate relative to those years is very enlighting.

With regard to shifters, I'm assuming that teams do not design to shift on the fly. This thread has mainly discussed the design arguments and complexity of multi-motor/shifters.

My question would be what are the operational "gotchas" and "goodies" about the various theories. How well do the student operators handle shifting during competitions? How much awareness (training) does shifiting require? How does shifting affect autonomous mode?

[Matt Adams]

Originally Posted by IrisLab

With regard to shifters, I'm assuming that teams do not design to shift on the fly.

Actually, many top tier teams do design to shift on the fly. One of the primary reasons teams use transmissions that do not shift on the fly is is due to the additional complexity in machining that's typically required for these sorts of designs.

However, if I had more time, I'd press the arguement that though most definitely a nice feature, shifting on the fly isn't a game breaking feature.

However let it be known that there's a difference between:

1. Shifting on the fly.
2. Shifting after being stopped momentarily.
3. Stoping, shifting, and slowly moving forward for components to engage.

I would strongly suggest that transmissions of type 1 or 2 be used...

If I didn't have finals this week.. there'd be much more to this post. I'm sure others will follow through with their thoughts as they have done so well before.

Let me also note that most of the time shifting is actually done through software, so "learning how to shift" for drivers isn't exactly the same difficulty of learning to drive stick on a car. When to implement it in a match, however, is a different matter.

Matt

[Stu Bloom]

Originally Posted by IrisLab

The what-if discussion has been very important for me (as a rookie mentor). I appreciate the hashing out of the various "theories". It may seem irrelevant to discuss strategies without knowing the game, but the experience that each of you have with regard to prior years and the debate relative to those years is very enlighting...

I would like to STRONGLY echo what IrisLab says ... Although this is my third year as a mentor, it is the first year I am heavily involved in drivetrain design. While my career experience has been limited to static structures, these discussions, and especially this thread, have greatly increased my awareness of the important details to consider.

THANKS GUYS !!

[sanddrag]

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

[Gregory Frechou]

^could you elaborate

[Bill Gold]

Gregory Frechou:

http://www.chiefdelphi.com/forums/sh...ht=Technok at

[DanL]

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.

[Mark Hamilton]

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

[Specialagentjim]

Originally Posted by Mark Hamilton

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

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.

[Rickertsen2]

<rant>
There is something that i feel every member of a FIRST team and anyone involved in engineering in general should read and take to heart:
http://www.seattlerobotics.org/encod.../classics.html

That said i agree with previous posts, which state that the benefits of shifting should be carefully weighed against its disadvantages. As hard as it may be, decisions should be made not on emotion, but rather objective analysis and common sense. One thing is see all to often is something which is done merely because it is cool.

The first step in determining whether a shifting transmission (or any feature) is right for your team is to ask why do we need one? Every good design should start by clearly stating what problem(s) it is meant to solve (not the inverse!).

Are you trying to achieve higher speeds? higher torque? faster accereration? A better balance between the three that does not involve compromising at least one of the others?

After assessing the need for a feature, one must then ask "do we have the resources to design, fabricate, test, and pay for this feature all within the set time constraints?"

If the above criterion are met then i would say go for it!... But remember a good bot, or anything else for that matter, is built just as much if not more by careful planning at the systems and functional level just as much as by engineering prowess. This is a lesson that our team learned the hard way last year, which could have easily been avoided.
</rant>


this thread seems to have forked off into a discussion of two discrete but related topics and thus i will address them induvidually.

>1 motor per side
If weight permits and current is not an issue, i see no reason not to use multiple motors on each side. There are 2 advantages i see in using multiple motors. First and most obvious, you get more power to work with. Depending on how your drivetrain is designed this can translate into more speed or greater acceleration and pushing power. Depending on the game this year speed and or torque may be of great advantage. Secondly, someone earlier mentioned redundancy. If One motor fails, your robot will still be able to move (to some extent) rather than being completely disabled. Given the high mortality rate of the drill motors redundancy is definately a good thing( the reasons for this are another topic). In the IT world, redundancy is pretty standard on "mission critical" servers etc. The disadvantages of multiple motors seem minimal. Yes there is added weight, but not much. <rant>Most of the gearboxes i have seen in FIRST are ridiculously overengineered and much heavier than need be. using 1/2" aluminum plates and gears capable of transmitting in excess of like 50hp under shock loads is just not necesary. Look at the gears in the drill gearboxes and the Technokats transmission. Have you ever seen one of these break?</rant> The weight added by multiple motors and gearboxes to couple them really isn't that much.

Shifting Transmissions
I think that most people would agree with me that a good shifting transmission has many benefits, for example your bot can have a normal gear for general driving and then a low gear for pushing matches and more delicate menuvers. There can also be big disadvantages associated with shifting. Shifting gearboxes are (generally) heavy, difficult to engineer and fabricate, and possibly unreliable. That said, if there is good reason to have a shifting and your team has the resources to implement such a design, i see no reason not to have one.

[Matt Adams]

Originally Posted by Rickertsen2

That said i agree with previous posts, which state that the benefits of shifting should be carefully weighed against its disadvantages. As hard as it may be, decisions should be made not on emotion, but rather objective analysis and common sense. One thing is see all to often is something which is done merely because it is cool.

Absolutely correct. However, I think the past two years have really demanded a hearty drive train. See below.

Originally Posted by Rickertsen2

>1 motor per side
If weight permits and current is not an issue, i see no reason not to use multiple motors on each side. There are 2 advantages i see in using multiple motors. First and most obvious, you get more power to work with. Depending on how your drivetrain is designed this can translate into more speed or greater acceleration and pushing power.

To clarify, if you are referring to the part of my previous post where I brought up using multiple motors (which you very well may not be), I'd just like to clarify that I believe that multiple motors on a shifting transmission is unnecessary. However, I whole heartedly believe that it's very practical and essentially required to have multiple pairs of motors if you do not have a way to change your gear ratios.

I would fight to the tooth (with numbers of course!) that you can not have a robot that is both competitively fast and competitively strong using only one pair of drill motors or the chips without a mulit-ratio transmission.

Perhaps that'd be a post for later.

[Rob Colatutto]

Originally Posted by Matt Adams

I would fight to the tooth (with numbers of course!) that you can not have a robot that is both competitively fast and competitively strong using only one pair of drill motors or the chips without a mulit-ratio transmission.

That statement depends very highly on the game that you are playing. For example, last year you did not need to have pushing power if you weren't solely a ramp dominator. Take team 25, one speed, 4 motors, 12.5ft/s. All you need to move someone off of something, top of the ramp for example, is momentum (p=mv). No where in the formula for liner momentum is torque taken into account. Then you could also argue that team 25 was a strong robot for defense because of thier brake system. Or like any other robot with a high coef. of static friction, you can just turn sideways for defending a stack, or in previous years a goal. A 4 motor drive geared correctly for one speed can always be competitive.
Take last year for example, FIRST made it a little easier on everyone by making the ramp top out of hdpe so that way everyone would have a fair fighting chance for it at the end of the match. In most cases a fast robot could charge the ramp at the last few seconds and push down a stronger robot with thier momentum,which lead other teams to develope a method to keep thier robot attached to the hdpe by means other than just a drive system. Innovative new ideas can beat widely spread old ideas.
When picking your drive and wether or not your team wants to shift gears, analize the game first and then decide wether or not you actually need to shift to accomplish all the things you want to do in the game. If you can think of more reasons not to shift gears and concentrate somewhere else on an arm or such, then you should probably put a 2 speed tranny off to the side untill needed. If need the most pushing power or the most speed to accomplish your goal, then concentrate in there if you do not have the resources to make a shifting tranny. If you make a list of things you want your robot to do and high mobility is on the list, then I would suggest investing time in a swerve/crab drive rather than a 2 speed 1 dirrection drive.
One thing to keep in mind for everyone designing your multispeed tranny's: the shifting mechanisms can wear down to the point where they become unshiftable. You may want to keep that in mind with some methods so that if you are forced to lock it into one speed, you'll still be able to be competitive in that speed. Meaning don't set a goal of your robots high speed to go as fast as possible without tripping breakers, but then also try to be fast enough where you can be competitive if need be to lock into low gear. For that reason you also may want to make your multispeed tranny have 4 motors, since in many cases adding in another motor will only require 1 or 2 extra gears and so the power added is well worth it. You can also design it so if need be, you can add a motor or easily take out a motor if you decide to use it for a mechanism or if you just can't make weight. Just some of my thoughts on the subject...