There are a couple of generalized statements in this thread that I’d like to post a few counterpoints so people reading this thread for educational purposes don’t take them as facts for all situations.

In many high inertial load situations (like drivetrains) aluminum gears are far superior to steel gears as it will reduce current draw during acceleration. There have been posts about that when the aluminum gears first came into proliferation. The teams I have mentored over the years have used aluminum gears in all of our drive trains with few exceptions (see #2 below).

When FRC practically standardized on the 20DP 14.5 pressure angle gears back when the 2005 kit bot came out, one mistake was made. We did not do any profile shift on the 12-20 tooth gears. This causes undercut for those gears which makes them significantly weaker than their profile shifted counterparts. Using steel gears for some of these sizes is necessary to not expose that weakness. I recommend using steel gears for the gears in those tooth ranges.

Do the math. There are resources everywhere to do the actual math for these cases. Warning, standard Lewis calls are very conservative for FRC robots because they have useful life calculations embedded in the formulaic sequence. Bending strength assuming 3-4x shock loads is appropriate for most cases.

Lubricate your gears. No exceptions. Lack of lubrication is the #1 reason gears fail in gear wear (not breaking teeth but wearing them to razor thin sizes so they eventually break).

The 10T and 11T motor gears are stronger than their 12T counterparts due to profile shift eliminating undercut. In most cases this doesn’t matter for strength t that load but they are a little more efficient.

Using chain or rope for a final climb stage is always a good idea since it takes out a lot of shock load in the gearbox. Putting those chain reduction final stage on a dead axle at around 1.25 - 2” shaft diameter will eliminate most all load problems at that joint for FRC robots.

EDIT: I wholeheartedly disagree with the generalized statement, “steel gears are better than aluminum gears” as that is simply not true. I’d be more than happy to show you a decades worth of robots that use aluminum gears almost exclusively (notable exceptions listed above) with 0 failures.

Dual Neos, not sure off the top of my head the reduction. All gears lubed. The teeth wore thin until they gave way. I’m not sure where we got it, but this gear is extraordinarily light weight. Identical replacement gears are significantly worn after a few matches. Steel gears on the way.

Reversing shock loads for sure, FRC is brutal in general in that regard and add in things like potentially loose chains, etc and it just gets uglier.

Oh, like rotational interia. Seems like something worth calculating given the non-obvious effect compared to all the energy needed to move the robot and wheels.

Yeah, it is because there is a non linear affect based on gear ratio. In most cases, the gear closest to the motor (not the actual motor pinion) should be the smallest diameter gear in your drive train. I will take a picture of the inertia math and post here.

My understanding is that positive addendum shift increases effective pressure angle, which might explain the increased wear seen with the modified pinions.

It also might be worth noting that some VEXPro pinions have more shift than others (e.g.: Both the 10T and 11T pinions are 12T Center Distance). In @AdamHeard’s terms, I assume this means the 10T pinions are “trickier tricksters” than the 11T pinions.

Yes, but in the particular case of the 11T the addendum modification is pretty small. It is only modified enough to eliminate undercut, which should not affect west performance.

In the case of the 10T, like I said, the details of how VEX did it matters and since I didn’t specifically design that gear (unlike the 11T) I can’t really comment on why it could show more wear as there are a few reasons thst could cause it,

In theory, the higher working pressure angle could explain it but we are starting with the lower 14.5 PA standard so we already have a reduced wear signature. The “normal” standard is 20 degree PA. 14.5 is the “older” standard.

Seems like a large first stage gear holds about 1.2 joules when made of steel, where the speed of the robot holds ~500 joules. I’m ignoring the energy stored in rotating shafts and wheels because those would only stack the deck even more against the first stage gear. So unless you’re trying to optimize out that last 0.24%, I don’t know if the lower durability of aluminum gears is worth it for that first stage reduction. Your other points still hold water but specifically the point about reduced current draw is not very likely due to flywheel energy alone.

You should give Juliacalc (based on JCN calc) a try. You can put your lever arm and mass in and it will calculate motor currents based on your gear ratio. For such a high torque I would strongly consider a very large #35 sprocket (or two, really) for the last stage reduction.

That’s not enough, because that gear ratio is barely larger than 1. You’ll shred 20DP gears that are earlier in the train. I was imagining something like 60:18.

Does this recommendation for steel gears on tooth counts 20 or lower apply broadly to all applications? Is there a difference when it’s a lighter stress application like an intake or flywheel drive versus a drive train or torquey arm joint being powered?

Ok so, I attached a spreadsheet with your initial input numbers and did the reflected inertia math. This is not about steady state energy. This is about reflected inertia to the motor and the load that it imparts during acceleration and deceleration. This is very important for drivetrains and any mechanism that has rapid accel and decel, including collisions.

Here is the public Google Sheet (View only): Reflected Inertia
Using this very low, single stage gear ratio it is only about 2% inertia reduction, but that affects every acceleration, every deceleration and can add up.

However, when you go to a higher gear ratio gearbox with more stages, then the robot affect at the motor goes down and the gear closest to the motor has more of an affect (not the motor pinion). Someone can do the math on a multi stage gearbox as I just ran out of time and can’t find my hand written calcs from almost 20 years ago.

PS - Here is a post where we talked about it almost 20 years ago: Reflected Inertia … man, how did I get so old?

That is not my total reduction. I was referring to your comment about needing large sprockets for the last stage chain reduction, and was wondering if a 30 toothe #35 chain sprocket qualifies as large.