I’m too tired for this. Are you being funny and referential to the semi-absurd discussion about the fairlane wheels and safety wire discussion from 2020 or are you being serious with this question?
It’s a chunk of steel. The hex bearings are going to give out long before the steel or the wheel does.
I’m assuming the more critical concern is that, due to the mounting solution, it’s probably pretty hard to balance these flywheels. Take the question to mean: Assuming a worst case mounting job, how fast would someone be able to spin these without excessive vibration?
I’m being genuine about this. A bearing failure does not contain the energy of the flywheel. Nothing about the other discussions of flywheel safety is “absurd.” We’re talking objects with Kinetic Energy comparable to a bullet. Just because we haven’t experienced a safety disaster yet doesn’t mean we should be asleep at the wheel here, especially with products now being explicitly marketed as high speed flywheels (rather than drive or manipulation wheels that are being repurposed). Having a rating for a component in its intended behavior is standard industry practice, and is even a relatively common feature of FRC vendors. Having assurances from the manufacturer about practices being put in place to verify flywheel balance and/or material defect issues would go a long way to inspiring consumer confidence.
Seriously? You think AndyMark can slap an RPM rating on there that accounts for the misalignment and assembly of a system that they didn’t design, fabricate, or construct?
I’m really not sure what people want from the overworked and awesome engineers at our favorite FRC vendors. It seems like there is just absolutely no limit to the amount of silly specifications that they are asked to adhere to and people just keep coming up with more, not because they actually want anything meaningful to come from the discussion but because they just want to get one over on them.
I hope they don’t give an RPM rating for this or if they do it, it’s for the speed at which air resistance will cause the steel to melt. This just seems like a fools errand.
My point about the bearings isn’t that they will contain any failure but rather, they are the failure point for these systems and no one is asking about them or the hex shafts or whatever else. The normal smattering of FRC bearings are NOT rated for what we do to them - particularly with shooters.
Then why are you asking about a chunk of steel when you know what the failure point is going to be?
Because a failed bearing doesn’t contain the energy of a flywheel, and there are multiple failure modes in play. Moreover, this isn’t just a “chunk of steel.” It’s a chunk of steel that it is explicitly designed and advertised to attach to an injection molded polycarbonate wheel with $1.80 worth of provided 10-24 thread forming screws.
Then give me your address and I’ll buy you a couple to test. We can use this thread to come up with the testing you will perform. You can write a paper, document the process, and we will be better served as a community.
I am not the group attempting to make a profit from this, nor do I have an enclosure/barrier I would trust to perform such testing.
Ok. Well, I’ve got friends in low places. What testing do you want done? What level of shaft concentricity is applicable for the “normal” application with an FRC robot? Should we account for sudden high torque loading like when a ball is passed through?
If you’re serious about performing the testing, the most informative testing would be to use it in its advertised configuration, with the flywheel attached to a 4"AndyMark Stealth wheel with the provided 10-24 thread forming screws. The stated radial dimensional tolerance of both products is +/-0.005".
Would the bearing failing not result in an unbalancing of the flywheel and thus result in a very similar failure as what you’re concerned about? So don’t we really need to understand the lowest rated piece in the whole system and design around that constraint?
(I got no skin in this game, I’m gonna go do even more comically dangerous things with my remaining 9 fingers)
It may, but it is also not guaranteed to be the first failure point (and the fact that we’re operating in a vacuum of knowledge of the failure point of the flywheel assembly is exactly the point we’re getting at). Round bore FR8ZZ bearings are rated for a 1100lb dynamic radial load and 16500 RPM. I don’t know if I can trust this assembly up to those rating numbers. Similarly, I don’t know how this assembly will behave in the event of a bearing failure.
First should really get better at not forcing low resource teams to build these dangerous mechanisms. No more yoga balls or climbs over 5 inches. Vex is a much better program for this reason
I thought this was interesting so I took a look. Of course a bullet has been designed to maximize the conversion of energy into human harm, whereas an FRC shooter has been designed to shoot a ball into a goal.
Our am-4019 has mass
m = 0.8 lbf = 0.363 kg
And dimensions
r_1 = 11.938 mm
r_2 = 41.91 mm
For a thick-walled hollow cylinder,
I = 1/2 * m * (r_1^2 + r_2^2) = 3.447E-4 kg*m^2
Kinetic energy stored in such a cylinder can be expressed as:
E = 1/2 * I * omega^2
A 22LR bullet apparently has a muzzle energy of 135 ft*lbf or 183 J
For our am-4019 to have the energy of a 22LR bullet, the speed needs to be
omega = sqrt(2 * E / I) = sqrt[2 * (183 kg * m^2 * s^-2) / (3.447E-4 kg * m^2)] = 1030 rad/s = 9840 rpm
Which definitely is a reasonable speed for an FRC shooter.
Note, a 135 lbf mass raised to a height of 1 ft above the ground also has the stored energy of a bullet!
I know you’re being facetious but there is a discussion to be had that “how do teams safely test solutions” should really be a criteria of game design.
I think this is very off topic for the thread though.
As a side note to the safety of a flywheel, I would like to know from our European FRC members if there are additional safety regulations they need to adhere to due to CE regulations. Having had to go through CE certifying equipment that was sent to the UK and France, the hoops that had to be jumped through were extensive and extremely thorough. It would not surprise me in the slightest if there were testing that would need to be done to allow a flywheel to be CE certified for sale in the EU, unless there is some loophole that bypasses the CE certification process.
This was a lot more fun when we did this type of stuff from scratch and the only “safety” oversight was common sense in your local shop. Asking vendors to assume liability via blanket RPM ratings is ridiculous. Robots are dangerous. People can get hurt doing this. Keep yourself and your kids safe.
I said it two years ago and I’ll say it again: the faux safety culture in FRC is out of control.
I will say the only other “flywheel” vendor I know does list the speed that it has been tested up to.
Its a steel wheel…so as long as common sense is exercised in the design, I don’t expect it to ever go flying at any reasonable speeds. I think safety is very much in the hands of the teams (and robot inspectors if things get that far)
That said, It would be interesting to know how balanced these wheels, and how fast they can go before vibration would become excessive. I wouldn’t call this a safety rating, but more of a design consideration. Assuming standard hex shaft and no “balancing of the wheel” from a team, can this setup spin at 6k rpm without excessive vibration? 7k? 8k?
It’s a difficult thing to rate, but I think Andymark saying its been tested at X rpm, it would help teams know what to expect. Certainly not required, but would be kinda nice.
Our team already bought 2. We will be testing up to some TBD speed, and deciding if its adequate. If not, we are only out $40 and learned something.
@Lil_Lavery please avoid saying ‘the vendor didn’t provide a rating, so I don’t know if this will explode or not’ and crying wolf in the name of safety.
We can do the math instead.
There are tons of flywheel stress example problems out there. It is a common calculation. Here is one that I found with clear variable/value descriptions.
AndyMark lists the dimensions and material on their drawing. The material is A36 steel, conveniently listing the yield strength in the name: 36KSI. Documentation on A36 properties is easily found from many sources.
I plugged these numbers into the stress equation and guessed 20krpm was a stupidly high speed to run. Then I dumped the whole mess into wolfram alpha (I don’t even need to understand algebra or unit conversion).
So, at 20krpm (or 2100rad/s) we have a factor of safety of about 5.
We will start to see yielding around 44,000rpm.
Even then the material is extremely ductile, with an UTS around double the yield strength even 44krpm isn’t likely to break, just stay expanded a bit.
So, children, please try to keep your AM flywheels under 44krpm. Even better, keep them under 20krpm for a lovely 5x factory of safety. (5x FoS is what is recommended for lifting rating of rope and chain, failure is hysterically unlikely at this threshold).
