To Jeff W.
 

Hey Jeff,

you wouldn't happen to be spinning the ring gear of a planetary gear set, would you?


-Paul

We were considering a dual motor CVT using planetary gears and had the basic design, but didn't have time or expertise to engineer all the details. Maybe next year.

Here is a link that describes the basic idea. http://www.solomontechnologies.com/articles/design1.htm

Andres

Paul,

A student on our team suggested spinning the ring gear of a planetary gear set. He argued that in low gear you would realize the benefit of the drive motor and the motor spinning the ring working together to give you high torque, while in high gear you would get increased speed with both motors working together.

I didnt believe this to be true, however. I agree that in one case you do get both working together to give you high torque, but its not a true transmission (since there is no gearing going on), so in the high gear both motors are not in fact working together.

Your thoughts?

Keith Sevcik
Mentor, #920

Planetary
 

Keith,

We are, in fact spinning the ring gear this year. I am not divulging how many motors/type (at least, not yet), but we are using the spinning ring gear transmission. The student on your team is right and wrong about the motors working together.

They actually don't work together. If you simply coupled the output of 2 motors together, you would get more torque, but at a fixed gear ratio. What the spinning ring gear gives you is variable gear ratio. Simply put, if you spin the ring gear and the sun gear in the same direction, you get a low gear ratio (fast output speed). If you spin the ring gear in the opposite direction, you get a high gear ratio (low speed output). The motors do not at all work together! The ONLY thing the ring gear motor does is spin the ring gear and react to the loads of the sun gear motor (input motor). This means of transmission is inefficient, because the ring gear motor has to react to forces normally taken out by locking the ring gear against the housing; but it is very effective in varying the gear ratio.

Bottom line is that we thought it was of value, but have spent a year and a half perfecting it and making it work.

-Paul

Thanks Paul, you've helped to end a long standing controversy. I was trying to attack the problem from a static approach but encountered quite a bit of difficulty and was reluctant to dive into the dynamic problem. But yeah, i figured it was something like that.

To Paul Copioli
 

Hey! I'm that student Keith is talking about, and I'm really glad that my idea was at least worth making, as I'm learning from you guys. Unfortunately we only have one engineer (Keith), and no money and time to develop such a sophisticated transmission, so I came up with a much simpler design for a dual speed transmission. I would really appreciate if after building is done if you could share you knowledge, so I could truly understand this design.

Eugene

Eugene
 

Hey Eugene,

Actually, we plan to release ALL the detail and assembly drawings right after the ship date. We did all the files in SolidWorks, and it has a free viewer that everyone can use. The viewer can be downloaded from www.solidworks.com.

I will put a zip file together with all the assembly models, part models, and detail drawings. If you download the viewer, you will be able to view and print all of the files.


-Paul

P.S. We will be at Chicago, Ypsilanti (MI), and the Championship.
If you are going to any one of these, just stop on by.

Fortunately there are the chiaphua's in the kit this year, that have a really close maximum power output to the drill motors... So matching speed torque curve won't be too hard to do. Gear them down to the same slope, and electronically lower one of them to match the y-intercept, if that's necessary at all. I guess the curves don't have to match exactly. I wonder how close they have to be.

Would've been hard to do with drills and fisher price last year. Guess you got lucky this year... Unless you knew something in advance before we do. ;) Looking forward to see this working. I hope you don't mind me introducing this design in the Technical newsletter.

To Paul Copioli:

Hopefully our team will get enough money to get to champs, so if that is the case, I would be very glad to stop by and talk with you guys.


To Ken Leung:

If you remember me talking about a mysterious way of changing speed without switching gears, then this is what I was talking about. Really sucks that I will not get chance to build it, but I'm still hoping that our current design will be sufficient to draw some attention to the robot.

Eugene

Re: Shifting Gears --
 

Team 27 is using an"air" shifter on the trans, just like my Camaro. We are also using 2 Taco motors and two hole shooter motors. We like big torque, Kinda like the torque the 471 in my Camaro makes. Better yet, Don't be suprised if you see our bot' show up with a fire breathin' Big Block Chevy stuffed between the frame rails!!!:D

Lots of drawing details are available in the following white paper.

http://www.chiefdelphi.com/forums/pa...per&paperid=21

This is also a stop-before-you-change design, but it's not too hard to automate the code to stop, change, and go again without the driver having to let up on the joysticks. Here's the code we used to do it last year:

'---------- Gear Change --------------------------------------------------------

if (p1_wheel > 57) then next_option
servo_gear = LowGear 'p1_wheel button setup for lo_gear

next_option:

if (p1_wheel <= 200) then GCnext7:
servo_gear = HighGear 'p1_wheel button setup for hi_gear
GCnext7:

if (servo_gear_old = servo_gear) then no_change:

for k = 1 to 50 'Stops main drive base
Serout USERCPU, OUTBAUD, [255, 255, 127, relayA, 127, relayB, 127, 127, 127, 127, 127, 127, servo_gear_old, servo_gear_old+GearOff, 127, 127, BSRollPWM, 127, BSArmPWM, 127]
next

for k = 1 to 50 'Changes gear with the servo drive
Serout USERCPU, OUTBAUD, [255, 255, 127, relayA, 127, relayB, 127, 127, 127, 127, 127, 127, servo_gear, servo_gear+GearOff, 127, 127, BSRollPWM, 127, BSArmPWM, 127]
next

for k = 0 to 128 'Ramp Speed up
Serout USERCPU, OUTBAUD, [255, 255, 127+(k/8), relayA, 127, relayB, 127+(k/8), 127, 127, 127, 127, 127, servo_gear, servo_gear+GearOff, 127, 127, BSRollPWM, 127, BSArmPWM, 127]
next

for k = 0 to 100 'Stop again
Serout USERCPU, OUTBAUD, [255, 255, 127, relayA, 127, relayB, 127, 127, 127, 127, 127, 127, servo_gear, servo_gear+GearOff, 127, 127, BSRollPWM, 127, BSArmPWM, 127]
next

no_change:

servo_gear_old = servo_gear

Am I alone in thinking that if you have to stop and shift that another robot could come up to you and push you with almost no resistance? Or is everyone content with their design and willing to take the chance?