Design Accelerator Data Collection

Hey there Chief, don’t post much on here but here we go!

I’m the developer behind the KLib Design Accelerator and I’m currently working on the newest version of it. Such features include an inbuilt shot-arc calculator in the same sheet as the shooter calculator, pneumatic forces chart (credit to FRC team 4201 for the presentation they posted a while ago), and I’m working on creating a pneumatics center distance calculator with Kremer, among several other things.

Some of the things I’d like to solve here today are:

How much do different kinda of wheels slip with different compression and different RPM’s on a game piece while trying to possess it?

  • This would allow intake calculations to be more precise, something that always irked me. I also do strategy calculations and intaking calculations can be way off. In a game like Rapid React where you are intaking 20 balls a match, if your intaking # is even 1/4 of a second off, that’s 5 seconds, and it adds up.

Whats the minimum voltage a robot requires before it can drive?

  • I would like to, in the future, create a very accurate drivetrain acceleration calculator using the co-efficient of friction of several different types of FRC wheels. Creating a library of this would help several teams with their drivetrain calculations.

So, how would you provide this data to me here?

  • To discover the amount of wheel slip is taking place, I’d need a video (slow-mo preferably, but I can work with standard speed) of the intake intaking a game piece. It can be from any year. I would also require the RPM that the wheels are spinning. I can source the wheels myself to figure out durometer & eyeball compression if the shot is good enough, but if you can provide that information too that would be sick

  • The drivetrain must have all wheels contacting the ground to have the same traction or this # won’t be accurate. Swerve drives is a perfect testing playground for this. Just slowly ramp up the voltage to the motors until you get movement, and then post that number here. I’d most interested in blue and black nitrile treads, but other tread types would also be welcome!

Thanks for reading! :))

3 Likes

I’ve had a blast bouncing around ideas with Kevin on how to improve and add features to the KLib Design Accelerator, it really is becoming a one-stop-shop for all my FRC robot design needs!

@DaMoose3512 do you think we can set aside some time next week to do this using the 2022 bot:

1 Like

Sure, can do. :+1:

2 Likes

Thanks! I’d use the bots at 1902 but having more than just a few variables from a single team can really up the amount reliability of the collected data.

2 Likes

Don’t you also need the weight of the robot?

I think collecting the outputs from SysID would give you this and more, and not require teams to write code. Specifically the kS value is the voltage for movement to start. But I think kV and KA would also be helpful for your model. You can see one attempt to get this data here: SysId Data Sharing Thread

1 Like

This is probably hard/impossible to do “very accurately” in the general case. Friction is nightmarishly complicated.

1 Like

For personal enlightenment:

Is this because there are other friction sources besides the wheel-ground? (Such as the axles.) Or is the wheel-ground too complicated as is?

Thanks

Yes. The ground-wheel interaction is complicated because there is physical interference between the “hairs” of the tread and the fibers of the carpet. A lot depends on the freshness of the tread and carpet, carpet orientation, wheel material/size/width, robot weight, etc. Then there’s additional friction from the drive gearbox, chain/belt, wheel axles, motor internals, etc all of which respond differently to changes in speed, torque, temperature, and who knows what else. Modeling the system like sysid is probably your best bet here.

As far as projectile and pneumatics calculations, you might want to take a look at my AMB calculator which has those features.

1 Like

Feels a bit in bad taste to plug your own design calculator on a post about a different design calculator.

Both.

1 Like

It’s not like we’re selling competing products here. The goal is to help design robots and teach engineering. If someone came to this thread looking for one calculator to learn from and found another, now they have two calculators to learn from.

I tagged my calculator here because I spent the time to document the derivations for everything to make it easier to understand the principles behind the equations. That should definitely be helpful for someone looking to add similar features to their own calculator, as now some of the math is done for them and they can work on improving it where they see fit.

1 Like

Friction is the least reliable engineering parameter, at least in MY professional career…

That said, I did a series of tests pulling on pipes with grippers (coiled tubing for those in the oilpatch) out to the point of squashing the pipe skinnier. I got good linearity between total interface force and slip load. Avoiding slip crushing and reaching full tensile yield required MANY diameters of grip length.

Things as simple as bead blasted vs polished surface finish could change the loads by a factor of 2!

However, similar tests with BOP slips give you radically different results. These bit into the pipe on the order of hundredths of an inch. With that engagement you can part pipe on a 1 diameter engagement!

These cases illustrate the difference between simple friction and mechanical engagement; shear.

The smooth rubber wheels on new or minimally damaged carpet will be pretty linear vs vertical force. Bumpy/knobby tires actually reach into the carpet, like the BOP slips biting into the pipe. With that the interaction becomes RADICALLY more complicated with wear and slipping speeds factoring in.

That said, non-slipping or incipient slipping data on wheel to carpet interactions with controlled loads would be hugely valuable. A test rig would be the right way to get it. Robots have very complicated load transfer between the wheels during acceleration and motion due to the CG being well above the wheels.

2 Likes

Ah, gotcha. I was under the impression that getting the coefficient of friction would be advantageous enough to warrant collecting data from other teams. Although I am still interested in building a database for slippage on intake wheels of various types and compressions

It would be super awesome!!! It’s just extremy tricky to isolate…