2363 feeder station robot loading

[holygrail]

Cool!

Just a thought. Could you shave off a few seconds by just leaving the first tote on the ground and sliding the second one on top of it, then pick both up at once?

Good luck this year! Looks like a strong design.

[notmattlythgoe]

Originally Posted by holygrail

Cool!

Just a thought. Could you shave off a few seconds by just leaving the first tote on the ground and sliding the second one on top of it, then pick both up at once?

Good luck this year! Looks like a strong design.

I think a lot of that depends on what the throughput of a HP can be. Shaving the time off of the robot doesn't help if the robot is waiting on the human player. I think that will be the limiting factor for a lot of robots that feed from the human player station.

[Nathan Streeter]

Love the design! I'm always curious to see what Triple Helix will be building and how you'll do it... imo, you guys always do a great job picking a good strategy from game analysis.

Originally Posted by notmattlythgoe

I think a lot of that depends on what the throughput of a HP can be. Shaving the time off of the robot doesn't help if the robot is waiting on the human player. I think that will be the limiting factor for a lot of robots that feed from the human player station.

I disagree; I think *most* HP-loaded teams will be robot-limited rather than HP-limited. That said, the *good* HP-loaded teams will be HP-limited... and that with a well-practiced HP! I certainly think you guys will fall into the category of 'good' HP-loaded teams though! :-)

Seems to me although your HP is being smooth, he could still be 30-60% faster... perhaps by situating totes such that he doesn't need to walk around that would itself be a fair bit quicker. Then of course there's just being practiced and even quicker (think 2013 HPs)! Seems like he also is usually needing to wait to release the tote through the Chute Door, so that's a pause (although not currently his fault).

Seems like the two big slow-downs on your robot are retrieving the tote after it slides through (looks like 2-3s to me) and the requirement to raise and lower your stack repeatedly. There's certainly potential to speed up the first one quite a bit without dramatic re-design, but the second one's harder to speed up dramatically without changing the fundamental process.

[tr6scott]

Great work...

Your totes never seem to have the dreaded, stand up on end problem we have seen in other videos.

Wondering how the dimensions work out, and if adding a wheel in the middle of your arm would help if you ran it in reverse while dropping the crate to pull that leading edge out to prohibit the stop on end.

With the totes above, it looks like you could not stand it on end, but one that stalled with just leading point resting maybe a problem.

[Nick Lawrence]

Great work. This looks really familiar.

#teamdiscbrake

-Nick

[Pratik Kunapuli]

What is the passive system that you are using on your elevator arm?

[notmattlythgoe]

Originally Posted by Pratik341

What is the passive system that you are using on your elevator arm?

I'm assuming you are referring to the system that is gripping the totes. They are just lexan fingers that are angled in.

[Lil' Lavery]

Originally Posted by Pratik341

What is the passive system that you are using on your elevator arm?

Based on their previous thread, they appear to be bent polycarbonate/lexan tabs, similar to Ri3D Indiana.

[ToddF]

Originally Posted by holygrail

Could you shave off a few seconds by just leaving the first tote on the ground and sliding the second one on top of it, then pick both up at once?

Our early prototyping focused on comparing the two stacking methods of always stacking from the floor vs stacking on a "seed" tote. Based on our testing, we decided to focus on stacking from the floor.

[Michael Corsetto]

Originally Posted by ToddF

Our early prototyping focused on comparing the two stacking methods of always stacking from the floor vs stacking on a "seed" tote. Based on our testing, we decided to focus on stacking from the floor.

Could you explain your process for comparing these two stacking methods? I'm curious.

Excellent test videos, thanks for sharing!!

-Mike

[ToddF]

Originally Posted by tr6scott

Great work...

Your totes never seem to have the dreaded, stand up on end problem we have seen in other videos.

The tote starts with zero linear velocity and zero angular velocity. For the tote to tumble, the CG of tote must pass the end of the ramp. When the CG of the tote passes the end, there is a couple between the weight of the tote acting at the CG and the reaction from the end of the ramp. This couple imparts angular velocity to the tote.

Visualize the angled ramp from the feeder station as extending to an arbitrary height from the floor. Going to one extreme, imagine the dropoff from the end of the ramp being 1 inch. The sliding tote would contact the floor before the CG passed the end of the ramp, so no tumble. (neglecting the horizontal force of the point in contact with the floor) As the height of the dropoff increases, the more likely the tote is to tumble. The trick is to extend the ramp until the dropoff height is low enough that the tote doesn't tumble and end up on end.

[ToddF]

Originally Posted by Michael Corsetto

Could you explain your process for comparing these two stacking methods?

For the seed tote method:
1) Place a seed tote where we think the stack will be.
2) Feed many, many, totes onto the seed tote from the feeder station, doing everything we can to make it mess up.
3) Add structure to prevent mis-stacking.
4) If structure looks practical, and mis-stacking is reduced to an acceptable level, or if structure is too complicated, stop and draw conclusions. Else, return to step 2.

For the floor loading method:
1) Load many, many, totes onto the floor from the feeder station, doing everything we can to make it mess up.
2) Add structure to make totes end up in the proper spot for stacking.
3) If structure looks practical, and mishaps are reduced to an acceptable level, or if structure is too complicated, stop and draw conclusions. Else, return to step 1.

Final step: Compare the structure/mechanisms required to make the two methods work, and choose to implement the method which works best with how we want to play the game and how the rest of the robot is shaping up.

This pretty much summarizes how we do all our prototyping. We try to fail quickly, tweak, and repeat. The trick when prototyping is to know when something is failing because it's a bad idea vs poor execution of a prototype. The students are often far to quick to discard a good idea because the first try doesn't work. Often, just the right tweak turns a complete failure into a complete success.

We live for those eureka moments. Whether we win or lose competitions, it's those moments when a student's idea makes something just "click" that will stick with them and influence them to become engineers.

[waialua359]

That's a smooth looking feeder station robot.
Great job on the disc brake! We have one working at the moment as well.

[waialua359]

Originally Posted by Nick Lawrence

Great work. This looks really familiar.

#teamdiscbrake

-Nick

Haha!

[lcoreyl]

Originally Posted by ToddF

Now we just need to speed things up a little...

6 tote stacking at the feeder station

Very smooth. Ruined steak!! (Well done!!)

It looks like if you needed to you could stack 2 at a time from the floor--is that true or does only one fit under an elevated tote?