Building Drivetrains by Week 1

Technical Technical Discussion
#1

After looking into how to strategize, design, and build quicker during build season, I came across a bunch of 1678 workshop videos from last year and this year that got straight to what I was looking for, here is the link to the ones I’m referencing:

(2016) https://www.youtube.com/watch?v=MW6KqtDR9_U
(2017) https://www.youtube.com/watch?v=SVacrE4sKig

One part of the video really intrigued me, but got me slightly confused. In the 2017 video, at 34:30, he states that by Day 3 the drivetrain design is done. So I understand that 3 days is more than enough time to design a WCD, but what I dont understand is how you adapt your other mechanisms to fit ontop your drivetrain if it is already out to build? Looking at their GrabCad renders:

https://grabcad.com/library/pomelo-pummeler-frc-team-1678-2017-1

It seems 1678 had their intake gap already added, so is that a decision you have to make pretty much day 2? Also, do you just add a bunch of holes and mount everything to those preset holes? If anyone (possibly even someone from Citrus Circuits) could shine some light on this that would be awesome, I would really like to start manufacturing as early as possible next season ::ouch::

#2

My guess is they simply went with as wide of an intake as allowed by their frame dimensions and bumper rules. Then worked around that. Notice how their ground claw for gears (the design of which is driven much more by game piece dimensions than a wide roller intake for fuel) deploys over their rear frame, rather than thru it.

That being said, being able to arrive at general dimensions required by day 2 is incredibly impressive. Doubly so with the two different dimension possibilities this year (short vs tall). I don’t think most teams come anywhere close to this level of speed in their design decisions.

#3

I think that’s a good guess. Certainly that is what my team did; however, it took us a bit longer than 3 days to arrive at our drive train / chassis layout. We had prototyped an 8WD two-speed skid steer chassis in the off season and used that to begin driver training, but even so it took us a couple of weeks to nail down wheelbase and track dimensions, and select the gearbox. Although we had the intake dimensions ‘frozen’, we didn’t actually use that space for anything until MSC. It ended up becoming a floor pick-up for gears rather than a ball intake.

1678 has a very good process. It is worth study and emulation.

#4

Often (but not always), the design difficulties of drawing mechanisms around your frame are a small price to pay for the advantage of having a drivetrain / frame set in stone so early. Sure, it won’t necessarily be perfect and you might have to work around something you’d prefer not to, but the advantage is often too big to pass up.

#5

Sometimes yes, sometimes no. Being able to quickly (and correctly) identify critical dimensions is a bit of a talent and art form.

To give an example of when 1712 got it wrong, our 2016 may have had a successful scaler if our frame was just 2" longer than it was. We set our frame dimensions around our drivetrain choice for defenses, but didn’t fully consider the implications it would have on our scaling mechanism design. If we had opted for a slightly narrower and longer frame, it would have opened the door to being able to fold our scaling design much more neatly into our low-bar capable robot’s frame (while still being tall enough to reach the bar).

#6

You can usually make a lot of your drive train and chassis decisions early by having a basic idea of how you’re going to interact with the field/game pieces, even if you don’t have those mechanisms designed. If you go in saying “We’re going to have an over the bumper intake”, then you know to design your front frame all the way across to support the bumper. The specifics of positioning, angle, motion, and how it connects to everything else can be figured out later. On the reverse, if you want to pick up between the bumpers, you’ll likely want to maximize the gap. When you have a fixed box to fit in (like this year), that’s very easy to design to. When you have a total frame perimeter to deal with (like in Stronghold), it becomes more involved (like the example Sean gave).

So figure out your basic design concepts, then use those to figure out what dimensions you need on the frame. Cross your fingers and hope you’re right.

It also helps if you build a practice robot. For us, the practice bot is the first one we build. The chassis and drive train start as soon as possible and we build onto it from there. We then can start on the competition robot a week or two later once we have a better idea of what we need. As a result, the practice robot can have stuff that’s less than perfect (but serviceable), with noticeable improvements on the competition robot. The trick is making sure you finish on time :slight_smile:

#7

1/2" hole patterns are magical :slight_smile: We drill 1/2" hole patterns into all of our tube, which lets us mount our mechanisms wherever is convenient. We have a jig on our router for this, I believe there’s some info on it in the recent thread on routers.

More like day 3, if I remember correctly. It’s in the video, but basically our process is:

Day 1: Decide on most of the “what” questions (what tasks to do, what autos to do, match strategy, etc)
Day 2: Finish “what” questions and decide on most of the “how” questions. Start talking about specific mechanisms and make a list of things that we want to prototype. At this point we usually have enough information to start on CAD of the drivetrain.
Day 3: Finish “how” questions enough to have a complete list of things that we will prototype. Start prototyping and finish drivetrain CAD.

It is a really tight timeline, so you need to make sure that your strategy discussions are really focused. 2 days of 9-5 brainstorming is a lot if you use it right :slight_smile:

Exactly. This year we were pretty lucky in that the things that required most prototyping (the shooter and indexer) were pretty much independent of the design of the frame.

The different dimension choices was actually one of the least contentious things during our strategy discussion - I don’t recall anyone advocating for the tall design this year.

#8

Actually usually day 4 or 5 the CAD is completely finished. We take extra care to lay out the electronics properly and then CAD the bellypan, while the smaller parts are being fabricated (such as shafts, spacers). Then the more complex parts like the side rails, gearbox plates, and bellypan parts are sent to fab.

#9

This is exactly what 2579 did on Kickoff this year and as a result by day 3 our CAD (dimensions finalized) looked like this: http://i.imgur.com/3qB5DWol.jpg

#10

Wesley and Josh did a great job addressing the pertinent points of the discussion.

I’ll only add one comment.

Trade offs and estimations are a critical component of the engineering process. Drivetrain design is one of the many areas where we take a calculated risk in committing early to keep our desired schedule with the resources at our disposal. I am very proud of how fully the students on our team have embraced and strengthened our design philosophy over the years.

-Mike

#11

Very impressive! This year was the fastest we settled on our drive train design and dimensions, and we made a tweak (added 2" of width) at the beginning of week 2. As we knew we wanted to be light to maximize single-speed performance for gear delivery, we built for what we called “tiny” dimensions, which meant that we built the drive train small enough laterally to support “tall” and pushed the decision for “tall” or “broad” back a few weeks; **definitely **happy with this decision. The thinking in week 1 was that we might have some “tall” elements to feed fuel to other robots on our alliance. Once we saw how many teams were focusing on gears rather than fuel, and how difficult it could be to hang a gear on a field full of fuel, and learned some of the peculiarities of climbing a rope, we opted instead for horizontal extensions that worked in the “short” configuration around week 3 or 4. We were rather happy with our progress, having a properly dimensioned drive train with a passive hanger during week 3 to support driver training, and the climber and active hanger working for weeks 5 and 6. Our major issue this year was programming, and we’re going a summer programming camp (lead by Gixxy, our founding programmer) in an attempt to do better!