We have a catapult based build, but even after hours of driver practice, the accuracy is very low.
For e.g. even after 4 hours of practice with the hero bot, we are finding it difficult to accurately aim from a distance.
How are teams doing this?
You may wish to provide pictures of your robot and/or code but I will assume you’re using Swish with no modifications.
Have you checked that the rubber bands are the same?
I’m not sure what you’re saying, are you saying that the robot is inconsistent or that the driving is inconsistent.
I would suggest you set the robot in one spot and try to fire at the goal.
If it’s with driving, then you may wish to make this modification to your robot.
There are Ben Lipper videos about it but you could attach a motor to the rubber bands and then use that to make the rubber bands more or less tense. Then you can drive right up to the goal, and score in the low and high
Can you say from how far you are trying to throw? If you look at the floor pattern, and the max allowed robot dimensions (23” wide x 73” long x 15” high) and your supply of sensors, can you think of a way to add sensors to help your robot line itself up?
1 Like
Yes. It is Swish with no modifications. The rubber bands are the same (we did try with both types of rubber bands).
By inconsistency I mean this:
- We take a few mins to figure out the boundary in which teh robot should stay to score.
- Once we have figured out the boundary, we get the goals a few times (50% chance)
- But over a period of time, the accuracy slowly goes down. and we need to adjust the position of the robot again…going back to step 1
My theory is that this is due to rubber bands losing their elasticity over time, which affects the projectile path.
Wondering how this will play out in competition scenario where the rubber band has to stay consistent over the entire day.
1 Like
As in a 50% accuracy rate?
Are you sure?
I wasn’t in pitching in maybe someone who was could confirm this, but it would be pretty shocking if the rubber bands were to wear out after just minutes of match play.
Maybe send the code?
You could be not pulling the catapult down enough.
If you’re going the catapult route then I think your team needs to work on eliminating variables and thus reducing the likelihood of human error and the negative implications of “ghosts” in the code for autonomous. Trying to rely on perfect driving is always going to be difficult and that’s only magnified in teamwork challenge when you have another bot on the field that you have no control of or experience with.
Nothing stops you from replacing rubber bands between matches.
1 Like
Answering multiple questions here:
-
we are not using any code here. we are only trying with the hero bot with zero modifications. Not talking about auton mode at all…jsut basic driver skills.
-
So far, we have had 2 rubber bands on the catapult snapped and another is half way through getting cut.
-
After talking to another team who are testing with the hero bot, we find that they are aiming from a couple of inches different from where we are aiming from…Both of us have arrived at this distance after testing the bot’s performance. I dont think it is a build issue. It has to be the rubber bands, isnt it? which leads me to think, if we change rubber bands in the middle of the game, its not a simple switch and make it work like before…The drivers need to test it properly multiple times to figure out the new aiming area.
Can you provide more guidance here? we do have distance and optical sensors. Can you give me an example of how we could be aligning our robot during non-autonomous driving?
right now, we are aiming from 5 squares (~60 inches) away from goal zone to get the balls inside.
Yes. even if the robot hasnt moved from the previous try and we just feed the ball and shoot, its a hit or a miss.
Can I see a picture of this launcher/catapult? This would help let me see what is going wrong.
uploaded a few pics. we have been experimenting with rubber band positions to get more distance, but regardless of band positions, the issues I have mentioned in this thread remain
how do we eliminate the human error factor. We do have sensors, but we havent figured out how to use them to reduce or eliminate minor inaccuracies.
Ok now that I see it, personally I would add a bar to wrap the rubber bands around to keep the bands from fraying and making sure it shoots more consistent.
This is what the rubber bands should look like if you set it up with the bar.
THanks. could you provide a few more pics on how it looks from under? I didnt understand how the rubber band passes around the bar
Although this is for pitching in, all you would need to do from these images is modify it to make it stronger (add more bands and or move the bands).
I understand your frustration with rubber bands and catapults we have all been there once. Anyways here are some key takeaways from the pictures you sent:
-
You should add more support towards the back of the catapult. The reason is the two separated beams will be slightly different so could be leading to inaccuracy.
-
Your rubber bands are way too far and you’re adding tension for no reason. This is if you wrap a rubber band against a tight corner and there is enough tension the rubber band will only be using 1/2 its power instead of all. So I would recommend using more rubberbands with overall less tension mounted to those pins at the bottom not the front which will help the longevity of the bands.
-
Use one long axle instead of two and support it from the other side as well not just one side. This is because if you load it the weight will be distributed to the sides bending it slightly again reducing efficiency. Also one axle will make it stronger and straighter.
Hope this helps