pic: 2791's Chassis

[Justin Montois]

Looks great. I love the belt running through the tube. Very slick.

[Leav]

Originally Posted by 340x4xLife

Looks great. I love the belt running through the tube. Very slick.

I would actually say that is my least favourite feature about this chassis.
it's cool no doubt, but looks like a pain to maintain if you throw a chain...

A lesson I learned from my mentor a while back was to design for maintainability, as well as performance and looks.

[Chris is me]

Originally Posted by Leav

I would actually say that is my least favourite feature about this chassis.
it's cool no doubt, but looks like a pain to maintain if you throw a chain...

A lesson I learned from my mentor a while back was to design for maintainability, as well as performance and looks.

It took about 7 minutes to change a belt on our prototype chassis.

This one will have a removable easy access window over the middle wheel. That and practice / belt changing drills should bring it down to 5 minutes - twice as fast as our previous dead axle drivetrain.

We also stress tested the drive. Stalled the wheels multiple times, ran it forward and back a bit with no signs of slipping or failing.

[rahilm]

Looking at the CAD with electronics, I just wanted to point out that the only legal bridge this year is the DAP-1522 included in the kit.
Looks like a pretty cool drive design with the 6 motors

[Matt H.]

Why did you choose to use six motors in your drive train?

4 CIMs at a 7.4:1 ratio gives a maximum of 634 in-lbs of torque at the wheels. With 4 inch wheels, that becomes ~315 lbs of force to accelerate the robot. However, with a generous coefficient of friction of 1.5 your robot will spin its wheels at 180 lbs of accelerating force.

[Chris is me]

Originally Posted by Matt H.

Why did you choose to use six motors in your drive train?

4 CIMs at a 7.4:1 ratio gives a maximum of 634 in-lbs of torque at the wheels. With 4 inch wheels, that becomes ~315 lbs of force to accelerate the robot. However, with a generous coefficient of friction of 1.5 your robot will spin its wheels at 180 lbs of accelerating force.

We're assuming about 1.3 for traction.

Essentially, pushing power. 4 CIMs geared at 12 FPS draws a LOT of current. 4 CIMs and 2 BB isn't quite traction limited (60some amps last time I checked) but at least we won't stall out pushing mecanum bots and kitbots around a bit, and we'll have great acceleration.

[JesseK]

Originally Posted by Chris is me

We're assuming about 1.3 for traction.

Essentially, pushing power. 4 CIMs geared at 12 FPS draws a LOT of current. 4 CIMs and 2 BB isn't quite traction limited (60some amps last time I checked) but at least we won't stall out pushing mecanum bots and kitbots around a bit, and we'll have great acceleration.

Looks like a very solid setup.

If you use the line following sensors, where would you put them?

I only ask because the first words out of the student programmer's mouth was "I want my line followers here, here, and here", which correlated to the front corners of the robot and the center of the robot. In this setup, it looks like those sensors would interfere with the live axles (they're about an inch tall), though other modifications could be made.

[sdcantrell56]

Id be real careful gearing at 12fps. We did that last year with 4cims and 2fp's and often times our battery was totally dead with around 30 seconds left in the match. We later dropped down to 9.5fps with the 6 motors and it was much better. Just be ready to swap out gearing after testing. Otherwise this looks super awesome!

[Teched3]

I've always been an advocate of using smaller diameter wheels for a little better grip on the carpet. You also obtain a little lower CG, which helps on Sat afternoon. Was there a reason that your standoffs appear to be reversed from one side to the other, or just an oversight. are you using belt or chains to drive your front/back wheels? Your rendering is awesome, and is motivating me to learn a 3D modeling program. any suggestions for an old-timer?

[Chris is me]

Originally Posted by JesseK

Looks like a very solid setup.

If you use the line following sensors, where would you put them?

If we opt to use line followers (I'm thinking an ultrasonic sensor and a gyro for dead-straight tracking is as good as we'll need), we'll probably cut a hole in our electronics board and mount them a little above the bottom of the front of the chassis.

Originally Posted by sdcantrell56

Id be real careful gearing at 12fps. We did that last year with 4cims and 2fp's and often times our battery was totally dead with around 30 seconds left in the match. We later dropped down to 9.5fps with the 6 motors and it was much better. Just be ready to swap out gearing after testing. Otherwise this looks super awesome!

One of the really cool things about this transmission is the belt reduction - we can adjust with a new belt and output pulley between 9 and 14 FPS without taking up as much space as a chain run would. It also lets us mount our motors in a very compact configuration - the two-motor model is literally 3 inches from the ground.

That being said this is a valid concern - does anyone know how I could predict battery discharge rates using math? I have no idea how to do it.

Originally Posted by Teched3

I've always been an advocate of using smaller diameter wheels for a little better grip on the carpet. You also obtain a little lower CG, which helps on Sat afternoon. Was there a reason that your standoffs appear to be reversed from one side to the other, or just an oversight. are you using belt or chains to drive your front/back wheels? Your rendering is awesome, and is motivating me to learn a 3D modeling program. any suggestions for an old-timer?

Our drivetrain's CG last year was about 5 inches from the ground.

This year, it's just under an inch and a half.

With the 9 foot tall arm we're working on, that could be pretty important.

[Matt H.]

Originally Posted by Chris is me

That being said this is a valid concern - does anyone know how I could predict battery discharge rates using math? I have no idea how to do it.

Looking at the spec. sheet for the batteries:
http://www.alliedelec.com/Images/Pro...A/610-0002.PDF

The last graph gives discharge characteristics. The C-rate or CA on the graph can be roughly calculated as Current Load/Rated Amp Hours so at your quoted 60 amps
CA~3.4
Visually extending the curve; you're right on the edge.

If you current usage rises to 80 amps with a greedy manipulator you'll almost certainly run of battery life.


Disclaimer: I'm not an expert on batteries and am only ~70% sure my interpretation of the C-Rate is correct.

Another method you could try is using Peukert's law.

Where the time to full discharge is approximated by:

t=H(C/(IH))^k
Here t is time to full discharge, C is the rated capacity, I is the current, H is the rated discharge time and k ranges between 1.1-1.5 for lead acid batteries (it is generally empirically determined, but we can use the discharge curves to estimate it)

Using the spec sheet again:
C=17.4 H=20 hours I=60 amps k~1.34

t~4.1 minutes

For 80 amps

t~2.8 minutes

So it all comes down to how much you want to trust the math. To me it looks like a very border line case especially if you have a high current (always on rollers, heavy lifting etc.) manipulator.

For more information:
http://en.wikipedia.org/wiki/Peukert's_law

[Chris is me]

That's pretty compelling, really. Though I doubt the drive will spend more than 30 seconds a match drawing maximum current, it is still something we should watch out for.

Our current plan as of today's meeting is to oder a set of pulleys to gear us for the more conservative 9.5 feet per second and to do extensive battery life testing. We've got a simple, probably low load manipulator planned - but we will still be careful. If during practice or competition we figure out we're going through batteries too fast - we'll swap those right out. Should take under 10 minutes.