Team 254 Presents: Deadlift Technical Binder 2015

[Cherry254]

Originally Posted by thatprogrammer

1.How are the carriages driven up and down? They appear to be attached to timing belts through some type of clip?

2. What type of bearings did you use for your elevator?

3. What types of belt did you use in your elevator?

4. How did you install your chain in the tubing?

5. Are you using the talon srx or victor SP?

1. Via HTD timing belts (endless) joined with a custom timing belt clamp that was integrated into the carriage bearing block assemblies.

2. Standard R3 and R4 equiv bearings.

3. 5mm Pitch HTD x 9MM wide

4. We held the chassis vertically and lowered the loop of chain down the end of the tube with the sprockets already on it and slid the shafts through the bearing blocks.

5. Victor SP

From what I can see, you guys have almost all of the robot in CAD before you start making parts for systems other than drive. At what point in the season do you finish drivebase CAD and the rest of the CAD? How many people do you have on your CAD team, and how long do they work to get it done?

Finally, does your team use any type of version control for SolidWorks, either EPDM or workgroup? If so, what type of server/computer does it run on?

It looks like you had no way to adjust chain tension this year. Would you recommend running a chain in tube drive system with no tensioners in a more aggressive game, or would you probably switch back to an adjustable and accessible design?

I've also seen your team run 3mm pitch pulleys off the output shaft of the RS-775 motors many times. Are these pulleys you make yourselves with a press fit for the motor shaft, or are they something else?

What material was used for the friction brakes?

Drive CAD is done early by a group of a few students and a few mentors. It ideally is complete in the first 4-6 days. The rest of the CAD can linger for many more weeks as prototyping continues. Deadlift wasn’t even completed in CAD until the last minute with some parts were designed days before stop build.

We use PDM workgroup to manage SolidWorks files.

Undecided on the chain in tube. An adjustable and accessible design has worked well for us in the past. We didn't like that the chain rattled inside the tube, but otherwise we didn't have any issues with it. Assembly was not difficult and once it was installed we never had an issue.

3mm pulleys are available from WCP. http://www.wcproducts.net/gt2-timing-pulleys

Friction material was from McMaster: High Friction Molded Brake & Clutch Lining Sheet.

[GuyM142]

Thank you for sharing! I find this really helpful

[Michael Hill]

I'm curious, what's the point of dropping the center 0.100 when you have omni wheels in the corners? Just a failsafe if you decide to go all traction later on?

[Jack S.]

Originally Posted by Michael Hill

I'm curious, what's the point of dropping the center 0.100 when you have omni wheels in the corners? Just a failsafe if you decide to go all traction later on?

IIRC, their original release photos had all colson wheels...

http://team254.com/first/2015frc/

[Nathan Streeter]

You guys had an absolutely fantastic robot; thank you very much for sharing through this resource!

The 254-style of very elegant, clean, box-oriented mechanical design makes your robots look deceptively simple... but let no one confuse elegant either for 'easy,' for not being full of design time/thought, or for not having complicated components.

I am intrigued as to what motivated you to use your particular elevator drive setup... I can see why a setup with 2 RS775's geared very fast with a pneumatic brake for holding is more 'mechanically ideal', but it seems like similar performance could be attained with a CIM (or two) at higher reliability and with less complexity. Was the elevator speed needed with the weight/CG of the RS775 solution? It doesn't seem like the power draw of a non-braking solution would've been too bad. Am I just standing on the other side of the 'elite powerhouse' line where the difference between an elevator at 5.85fps and one at 3fps is highly significant?

Also, am I correctly understanding the design that the elevator carriages were driven from only the right or left side (i.e., only from the right for the RC-carriage and only from the left side for the tote-carriage)? I see 4 long, vertical belts on each side which would make me think otherwise, but I don't see how the power for each carriage is transmitted to both right and left on the elevator gearbox.

Fabulous design and execution! Tough ending at CMP, but everyone obviously still has the utmost regard for Deadlift!

[Anupam Goli]

Great read! Every year these are released, I learn a little more each time. I can't wait for the software release, because the whole simulator concept sounds really neat!

[moneyisgreen]

Originally Posted by Jack S.

IIRC, their original release photos had all colson wheels...

http://team254.com/first/2015frc/

We originally designed the drive without the intention of using omni-wheels. However we decided to replace the outer colsons with omni-wheels after we found out that the robot turned much better with them.

[Michael Hill]

Originally Posted by Jack S.

IIRC, their original release photos had all colson wheels...

http://team254.com/first/2015frc/

Ah, that makes sense now. Thanks.

[Jared Russell]

Originally Posted by Nathan Streeter

I am intrigued as to what motivated you to use your particular elevator drive setup...

FWIW, we seriously contemplated adding a third RS775 to the bottom carriage after SVR to improve loaded speed when carrying several totes.

[Cherry254]

I am intrigued as to what motivated you to use your particular elevator drive setup... I can see why a setup with 2 RS775's geared very fast with a pneumatic brake for holding is more 'mechanically ideal', but it seems like similar performance could be attained with a CIM (or two) at higher reliability and with less complexity. Was the elevator speed needed with the weight/CG of the RS775 solution? It doesn't seem like the power draw of a non-braking solution would've been too bad. Am I just standing on the other side of the 'elite powerhouse' line where the difference between an elevator at 5.85fps and one at 3fps is highly significant?

Also, am I correctly understanding the design that the elevator carriages were driven from only the right or left side (i.e., only from the right for the RC-carriage and only from the left side for the tote-carriage)? I see 4 long, vertical belts on each side which would make me think otherwise, but I don't see how the power for each carriage is transmitted to both right and left on the elevator gearbox.

The goal was to have an elevator fast enough and powerful enough such that we were never waiting on the elevator while stacking. Two RS775s have a greater power to weight ratio even with the additional half pound from the pneumatic brake assembly than one CIM.

The gearbox transmits power to all four corners of each carriage. Its hard to tell from the render but each pair of motors drive two sets of pulleys which transmit power to the left and right side of the elevator.

[Tom Bottiglieri]

Originally Posted by Anupam Goli

Great read! Every year these are released, I learn a little more each time. I can't wait for the software release, because the whole simulator concept sounds really neat!

The whole robot package of robot software (Robot code + robot sim code + sim lib) should be out soon, but for now feel free to look through the WPI shim we wrote to get our robot code to run on x86.

https://github.com/tombot/FakeWPILib

On the other side of this, we had a simulated robot which could monitor PWM motor output values and set sensor values based on physics simulations. We used the debugger/monitoring/logging mechanisms built in to our main robot code to get everything working before it hit the real hardware.

[marshall]

Originally Posted by Tom Bottiglieri

The whole robot package of robot software (Robot code + robot sim code + sim lib) should be out soon, but for now feel free to look through the WPI shim we wrote to get our robot code to run on x86.

https://github.com/tombot/FakeWPILib

On the other side of this, we had a simulated robot which could monitor PWM motor output values and set sensor values based on physics simulations. We used the debugger/monitoring/logging mechanisms built in to our main robot code to get everything working before it hit the real hardware.

Seriously impressive stuff. Thank you for sharing this.

[Thad House]

Getting good simulation is the next big thing I think for FRC robots. Its hard to build a general sim for everyone, but even just having a simple backend that's extendable is a great start, especially since most FRC robots can be simulated with fairly basic physics. An elevator can be simulated without any calculus as long as you have a brake, or you lift counterbalance well. And even the calculus to simulate gravity is not very difficult. I really like how the Python port did this, and I wish someday that could be ported to the official languages as well. But without dynamic typing, it becomes a lot harder to do the way they did it.

[mman1506]

Did you do any current monitoring of the intakes to prevent the totes from jamming?

[Nathan Streeter]

Originally Posted by Jared Russell

FWIW, we seriously contemplated adding a third RS775 to the bottom carriage after SVR to improve loaded speed when carrying several totes.

Fair enough. :-)

Originally Posted by Cherry254

The goal was to have an elevator fast enough and powerful enough such that we were never waiting on the elevator while stacking. Two RS775s have a greater power to weight ratio even with the additional half pound from the pneumatic brake assembly than one CIM.

Sounds like a worthy and intuitive goal... necessary for the ~8s 6-tote stack from RAMPage (which is just absolutely fabulous!). Having thought about the pneumatic brake aspect a little more, it really isn't that much complexity/risk... not any worse than (and very similar to) a pneumatic shifter. Had you guys used one before, or was it essentially all-new?

Originally Posted by Cherry254

The gearbox transmits power to all four corners of each carriage. Its hard to tell from the render but each pair of motors drive two sets of pulleys which transmit power to the left and right side of the elevator.

Makes sense... I was having a hard time imagining how it could work without something like what you mention. I think I can see a bit of the 'hidden' pulleys now that you note them.