01-09-2003 15:02

George1902

Someone's gonna have to explain this one to me. I just recently got my mind around your last creation, Tytus... Are you trying to drive me crazy?

Joking aside, care to give a more detailed explaination than in the caption?

01-09-2003 15:36

sanddrag

The top of the picture is a bevel gear differential that is coupled to a different size gear on either side. These gears are rigidly mated with a shaft, one on each side. The shaft's are separate from left and right sides.

In the lower portion of the picture are two different size gears (opposite of the ones on the differential shaft) that ride on bearings and mate with the other gears. The shifting dog is rigidly attatched to that shaft. And engages either gear to drive the differential.

But I'm still oblivious as to why there is even a differential there, this is just sort of the technokats design + one bevel gear differential.

I'm confused too.

01-09-2003 22:16

Tytus Gerrish

the diff is to compincate for any discrepinces in the Rpm's of the two motors over the entire power arc

01-09-2003 22:33

sanddrag

Oh I see now, the motors input the differential, not the shifter. We've gotten into this whole "difference compensation" discussion before.

01-09-2003 22:38

Tytus Gerrish

ill draw the motors in and post that later


Another thing about this box , is its semetrical , so ot can be used on either side of a robot without being mirrored

i also need to know the free rpm for the bosch without its gearbox
i know the CIM is around 5000

01-09-2003 23:08

Madison

Approximately 19,700 RPM and 5,500 RPM, respectively.

02-09-2003 00:20

Andy Baker

Tytus,

Nice helical gears. I absolutely love the fact that you are trying something new here with the reversing differential.

Since you posted this, you obviously want input on how to make it better. Here are some suggestions:

1. Give us a cross-sectional picture in order for us to see what is going on with the differential.

2. I am also confused about the qualities for the differential. From what I see, there are 2 input shafts. There are 3 bevel gears mated so that the rotation direction is reversed from one side to the other.

However, I do not see the rpm compensation you are referring to. For example: If the motor on the right input shaft inputs a speed of 4,000 rpm, then the left shaft will turn at 4,000 rpm. If the left shaft inputs a speed of 16,000 rpm, then the right shaft will be spinning at a rate much higher than the free speed limit of 5,500, causing the motor to be very, very inefficient.

3. Your attention to detail with regard to texturing and how the gears look is admirable. This is good stuff for graphical design and animation, but not needed for detailed mechanical design. It is entirely acceptable to model a gear (helical or spur) by simply creating a cylinder with the o.d. as the pitch diameter. This way, you are saving time on the "eye-candy" and spending more time on the nuts and bolts design.

I like your differential idea. But, in order to work, you need to look at changing your ratios on the bevel gears. You don't need to exactly match the free speed ratios of 19,700 : 5,500. If you hit the ratio from anywhere between 3.2:1 and 4:1, then you are doing OK.

Andy B.

02-09-2003 10:49

Tytus Gerrish

there are actual 4 bevels in the diff,1 for each motor and 2 connected to the diff, i think it will be more balanced, and instead of having 1 bevel connected to the diff casing i have 2, since there is twice the surface area i think there will be half the friction, I THINK!, correct me if I'm wrong

a wider tire has more traction than a thin tire right?

02-09-2003 11:05

sanddrag

I'm really sorry to break this to you but that is kind of what it looked like in the picture. I was just waiting for someone else to say it. Your diff is now, well, not a diff anymore. It will be just as good as a solid axle with all four bevel gears in there. Think about it or try it with legos. Trust me, you need to take one of the bevel gears out that is attatched to the diff housing.

02-09-2003 17:52

Tytus Gerrish

legos dont have a peg for a fourth

And it will work, 4 is fine

02-09-2003 20:32

sanddrag

And I wonder why they don't... Trust me if it worked, you wouldn't be the first to have made one. I've tried. You cannot have a 4 bevel gear differential. Sorry.

02-09-2003 21:34

Tytus Gerrish

i gont think you have grasped this concept


if you made a cube with 6 bevals in all, it of course would not work
but there are only 4 in here 4 will work

02-09-2003 21:45

sanddrag

Fine have it your way. You know, some people here are trying to help you but if you don't want it that's your own decision. I'm sorry to everyone for these back and forth posts but I thought it was worthwhile for everyone to read.

02-09-2003 21:53

Tytus Gerrish

Me too, sorry
I argue about my gears too much anyways,
ill just draw it better

02-09-2003 23:20

rbayer

There's absolutely nothing wrong with a 4-bevel gear differential. Check out http://auto.howstuffworks.com/differential2.htm for an animation of what it would look like.

As for reducing friction, I don't think it will do that. What it will do, however, is distribute the force more evenly over the two gears, thus reducing the chance of tearing teeth off, etc. Then again, I don't really know all that much about differentials so I could very well be wrong on that last point.

--Rob

03-09-2003 00:08

sanddrag

Oh my gosh. I feel so embarrased. Thanks for correcting me though Rob because I would not want a potentially good design to not be tried because soemthing I said. Tytus, it looks like I'm the one that needs some help. I am so sorry for not believing your idea would work and not listening when you said it would. I never really took as much time to sit down and think about it as you did. I guess when I get so caught up in all these advanced mechanical things I forget the basic priciples and need to revisit the Legos but never do.

Anyway, I'm really sorry for everything I said and the 4 gear bevel gear differential does work like you said it would. I think your drawing is really great and you should keep up the good work.

And with that, I stand corrected.

And for all of you who follow my postings often, these past few are just due to some pre-school (pre as in before, not preschool) anxiety. You do not have to be concerned with the truth to all of my past and future postings about various mechanical things.

04-09-2003 04:36

sevisehda

My only question is has a reverse differtial ever been tested? On paper or monitor it looks good. But in the real world how will the motors react? Will the load be divided between the 2 evenly? How will the motors react as they throttle up and down. I'm not saying it won't work I'm just saying this is something you've got to build and test to check feasability.

04-09-2003 06:51

Anthony Towne

I see where you are going and I have to admit that it looks like it will work. It took me a sec to realize that the 2 back helical gears aren't attached to the input shafts, they are just attached to each other through the cage structure and rotate on bearings. This isn't a new idea, however. All it is is a differential working backwards. Teams in years past have tried this. Many teams don't do this because it requires 2 motors for each set of wheels, and most teams use 2 sets of wheels, so the robot gets heavy fast and run out of room in the drive area to fit 4 motors with everything else.

04-09-2003 10:03

Tytus Gerrish

Who's tried it before?, Show me i could learn from them

04-09-2003 11:15

Anthony Towne

there was a team at the UCF regional that had a differential in their drive system. I don't know their team #, I just know it wasn't a UTC team I know that doesn't really narrow it down). When I was involved back in High School (6-7 years or so), I saw them used at a few regionals, but all the team numbers have changed since then and I can't even remember what their original team numbers were. Sorry I can't help you with that further.

04-09-2003 15:01

P.J. Baker

This looks like a pretty cool idea. Thanks for sharing. Here are my thoughts:

Without doing too much analysis, I think that this works as follows: The two motors adjust their speed until both are exerting the same amount of force on the two gears on the output section of the differential. The driven speed of the output will be the average of the two input shaft speeds. Essentially, it forces the motors to output the same torque, rather than the same speed – which is what happens with the more traditional dual motor designs. I don’t see any reason why it wouldn’t work in a robot.

However, it will suffer from the same problem as a car differential if one motor loses power – almost no torque will be transmitted through the drive train even if though there is still one good motor. The dead input shaft (unless it is somehow locked up) will be backdriven by the good input shaft, and the output of the differential into the treads or wheels of the robot will barely move.

I wouldn’t consider this to be a fatal flaw in the design, just one to be aware of. If one motor burns up or some how becomes mechanically or electrically disconnected from the transmission, the other will be just as useless.

Also, rather than matching free speeds between the two motors, I might suggest matching the stall torques. The differential design appears to have problems when motor A is supplying a lot more torque than motor B. Therefore, I would suggest gearing the motors to have similar torque performance rather than similar speed performance.

Good Luck,

P.J.

04-09-2003 16:32

Bill Beatty

Differential Transmission

Sorry, the design does have a fatal flaw. The higher torque motor will overpower the lower torque motor.

I am not sure why all the concern of matching motor speeds. Just gear the two motors to output the approximate same speed and both will contribute very nicely.