During the tuning process of our flywheel shooter, we burned a few 3d printed pulleys as their hex holes are grinded into round ones. Here’s a picture showing one of the damaged pulley:
Our 2021 shooter uses similar 3d-painted pulley wheels and such damage has never occurred.
At the same time, we also found our shooter’s performance constantly degrading. The trajectory of the Cargo got generally flatter and shorter as our pulley worn out. For reference, it takes 3690 RPM for our shooter with a possibly damaged pulley to shoot at 7m into the high hub, while a new one only takes 3150 RPM, with much more stability.
We first thought that it would be a shooter or feeder PID issue, but we had confirmed that both subsystems are stabilizing their speeds decently. Yet, when our DriverStation sometimes loses connection to the robot, the flywheel may indeed get instant accelerations/decelerations. The motor temperature was also a concern as our shooter runs hot compared to our 2021 robot, yet after periodic cool-downs using falcon’s cooling port, shooter performance remains mostly the same.
Has anyone encountered damaged pulley wheels before? Is this kind of phenomenon normal in Rapid React (as Cargo is much more massive and inelastic compared to Powercells), or it’s a result of bad DS connection or shooter settings? How might the pulley issue be related to shooter performance? Should we switch to steel/aluminum/markforged/highly infilled pulleys?
We are indeed actively considering using that. Thanks!
But I still think this kind of damage to the pulley itself is abnormal (or the 2020/21 Powercell shooting task is indeed “light duty” ?)
What material are you printing from? We use Onyx (Nylon with chopped carbon fiber) for pulleys. We also make them solid and increase the infill. We make extras to swap when we notice they are wearing. Aluminum for the hole is a good idea for critical hard to change pulleys.
Wether or not the pulley is being used in a “light duty” fashion is not just about the mechanism that it is being used in, but also about where in the gearing it is used. For example, it’s probably fine to use a 3D-printed hex bore pulley on the high speed/low torque side of a lifting mechanism, but probably problematic to use it on the low speed/high torque side. For example, if you have an arm that has a 100:1 total gear ratio in it, achieved by multiple gears and pulleys, using 3D printed pulleys “close” to the motor will generally work fine, but “close” to the arm will not.
The reason for this is that it is the torque which wants to rip the points of the hex through the printed hex hole - not the speed.
This makes a big difference. More force is required to accelerate a heavier ball. Also, the load is increased when compressing pressurized Cargo compared to a foam Power Cell.
When your shooter drive shaft starts to slip on the rounded pulley bore, the speed of the shooter wheel slows down relative to the shaft. A 3150 rpm shaft speed no longer corresponds to a 3150 rpm wheel speed.
A free-running shooter that only occasionally shoots a ball should probably not be making the motor(s) hot in the timeframe of a match or even during a reasonable amount of practice. If there is a lot of binding or drag in the shooter machinery, that will put extra load on the pulleys or gears or whatever is in the drive train. How hard is it to turn the shooter shaft by hand? It should spin easily.
On a topic not mentioned, is there any chance that the 2021 robot is using actual hex shaft through the pulley and the 2022 robot is using rounded hex shaft (“thunderhex”)? Every bit of engagement matters and rounded hex shaft is closer to round and will increase the chance of slipping on the pulley bore.
If you can get aluminum pulleys, they are going to be more robust in a shooter application. Inserts already mentioned would make a big difference in a printed pulley. A lower cost but less robust insert option is to use 0.75" aluminum hex tube as an insert. The larger contact area between the aluminum and plastic helps reduce slippage. Nylon pulleys and especially nylon+CF pulleys are going to be much stronger than PLA, for example.
The shooter spins rather smoothly all the time before the burn down (or the increase in friction is hard to notice as we never did the comparison).
I’m pretty sure that we are using thunderhex for both 2020/21 and 2022, hex-pulley assembly should run well all the time.
BTW, thank you for such a detailed answer! It really helps!
I want to ask the exact same thing @paulonis answers. Foam Powercells might require less torque from the shooter, thus is more “light duty” - even with gearing considered.
The acceleration and deceleration, to be clear, is referrring to sudden torque changes as the shooter is command to spin up and shut down rapidly.
Consider switching to PLA+ (tough PLA), Nylon or ABS. PLA+ will be the easiest to print. Nylon absorbs moisture so we use a food dehydrator before loading material. Definitely increase the infill settings. There should also be a setting for the amount of thickness before infill is used. Increasing that makes a huge difference. Now is the time of year to experiment and test!
Our team has had issues with pulleys breaking printed in PLA or even PETG. We used COTS aluminum pulleys this past season which have worked really well for us. Now, we have the ability to print PCCF, so we’ve swapped back to printing pulleys again with this new filament. If your team is still set on 3D printing pulleys, I think the thrifybot hex bore inserts or a small gear pressed into the gear combined with a higher infill is the way to go. Otherwise, you might find that COTS pulleys are more reliable.
