I have gotten lots of questions about this so I thought I’d just make a post.
Feel free to also use this to ask me about our strategies for defense.
The drivetrain and control system is 10 wheel (all 6" diameter) 2 sliced up stealth wheels on the big cutout side and the rest are pneumatic wheels. The drop is an 1/8" W drop with the center wheel being direct driven and the rest being chain driven. We run this all in a west coast style and we use this gearbox (
WCP 3 Motor Flipped Gearbox (Dog Shifter)
) with 6 full cims. gearing puts us at 7ft/s low and 17ft/s high. One motor on each cluster is run with a talon and the other two with victors that are followers of the talon. I use an XBOX ONE controller to drive with a split arcade format. Left stick Up/down is forward/backward. Right stick left/right is left/right. The sticks themselves are linearly mapped for up to .25 percent motor output. The remaining .75 is locked behind the right trigger, also linearly mapped. This means I can throttle control much easterly to put max power down without slipping the wheels. Without that functionality we cannot push anything. We just spin the wheels.
How do you control brownouts?
Also how does driving at only 25% output still have pushing power while not slipping the wheels?
For the most part, I don’t push the robot to the point where it could brownout. The robot is manually controlled so as far as any backup or brownout control, that’s up to feel. As for driving on low power, its slow but that’s really assuming you actually drive with no throttle. Most of the time you’ll be giving it at least a little extra power.
Edit: Of you were looking for some advanced PIDF control scheme or some massively innovative design, we just didn’t have the time. I wrote this software myself and I kept it simple since I had to be practicing with it right away.
Cool controls idea for defense and brownout prevention.
Torque is only part of the story in pushing. Traction is the bigger battle. You need to not slip your wheels and hope your opponent slips theirs. 7 ft/s is around 20:1 on a CIM with a 6" wheel. This means they get 2.41 Nm * 6 (CIMs) * 20 (ratio) * 0.25 (% voltage) * 0.80 (guess of % efficiency) * (0.7375621 ft lb / Nm) / 0.25 ft (wheel radius) * 1.0 (CoF) = 170lb of pushing power still. which means you can barely slip your wheels at full throttle. I might have messed up the math but you get the idea. You’ll only draw minimal current at this voltage and that will prevent brown out.
I’m curious how often you pull the trigger while playing defense / pushing.
It’s really almost constantly. To get a bot moving I found that pulsing the throttle to about half and then holding 33% until i got some speed and then moving to full power once i was moving and then reeling back when i felt slip was the best method.
For reference, .25% power is a crawl. Barely moving. and in low gear full power is 7ft/s so a moderate walk to slow jog.
If you had the ability to automate drive train controls to reduce slipping and brownouts would you do it?
Also for anyone out there reading this, I have been told that 3 cims very commonly brown out. Is there a way to stop this without sacrificing power output?
We actually had very few issues with brownouts. If you don’t just mash the throttle it had no problem making it through a full match.
It seems to me that the biggest drop in power happened when trying to start full force from a stop. Ramping the throttle manually to allow for some speed up seemed to make a massive difference.
While, yes, improving and automating would likely result in better performance, I prefer driving manually simply because it feels more involved.
More from a strategy/scouting point of view, but what characteristics did your team look for to identify robots you wanted to defend against? Pit scouting for drivetrains/robot weight? Match scouting to find specific weak points? Watch previous events beforehand?
All of the above. The first thing our scouts were tasked with finding out was the drivetrain and wheel type/position of any bots we’d be facing. Next is over all scoring potential. Then defense delta. Lastly strategic randomness.
This data was found through match scouting, pit scouting, team communication and conversation, match review videos, and the like.
Appreciate the response. One more for you, did you have to train your scouts on what to look for?
We tried pit scouting for the first time this year and struggled to get quality info from some of our students. We historically have relied more on the match data, but I would like to gather a complete picture of a team’s capabilities in the future.
Yes and no. For the majority of the year our scouts measured the same things everyone else did: cargo/hatches that are scored/where. Pieces held vs pieces scored. But at worlds we had them entirely switch gears to a full sandstorm period analysis as well as extensive defense notes both for the teams defending and those who were being defended. Really helped us find out what worked and on who.
Yes, we trained our scouters what to look for while pit scouting. Things like drive train type and collector types, from that data we can got a general idea of robots that would have a higher defense delta. But where the majority of the data we got for who to play defense on was watching matches. Specifically watching for patterns robots did, stuff like did they go to the same rocket every match, and how the bot and drive team performed under defense.
The easiest way to prevent brownouts with 6cim is to make sure the center drop, in the case of tank drive, is big enough so that not all the wheels touch the ground at once. In addition recommend making sure to ramp your controls to your drive. Going 0 to 100% instantly is the quickest way to browning out.
Can you provide any pictures of your chassis/frame? Particular focus on any frame cut-outs, the bellypan, and any cross bracing/supports used.
Our cad can be found on GrabCAD. I’d give you the link but i just don’t know where it is.
ill have to pull pictures off the grabcad Nathaniel the project is broken at the moment and still private
Makes sense why I can’t find the link then xD