As has been noted before (NEO 550 and Pressed Pulleys), the NEO 550 is sometimes prone to having the output shaft snap off.
What I would like to see is the details of how the NEO 550 was used. Specifically the type of reduction (belt, planetary, gear), gear ratio, use of the motor, material and material thickness of what the motor was mounted to, if it was outside the frame perimeter, and if something happened to the robot (impact, fall, etc) that might have caused this. Pictures and CAD screenshots would be helpful. Hopefully, by looking at a bunch of different cases, we can narrow down the failures to a specific use case, and we can all know what to avoid in the future.
OP from the original post here–like mentioned, we used a 5:1 reduction, also with GT2 timing belts. Our team is pretty new to belts, and we had problems throughout the season because our belts always seemed too tight.
Our NEO 550 shaft ended up breaking early on in our testing. Here are some pictures of the break.
We used 4 NEO550s.
Two for ball handling, one for shooter hood, and one for shooter azimuth.
All four were mounted to versa-planetary gearboxes.
I believe that the mounting path was through the gearbox.
No shaft issues with this setup.
Sorry, no other details available.
Hmm… looks like the snap ring was well-named in this case.
It’s only three data points, but the biggest difference right now seems to be transverse loading. A planetary input will have essentially none, belts definitely do.
We had the shaft of a N550 shear off this year. The motor was mounted with a 1/8" alu plate riveted to 1x1 tube. The shaft it was connected to was running on TTB bushings. The intake had several impacts but as far as I know, the shaft did not break directly because of these impacts, although it is possible that an impact bent the shaft which lead to a fatigue failure in regular operation. We ran it on a 1:3 reduction with a GT2 3mm belt. I’m working on a design for a printed pulley with a bearing on the other side, which would hopefully prevent this failure mode.
We had a failure similar to 1425’s - identical intake architecture, similar reduction. Unfortunately no photos of the fracture pattern.
Designed to exact center-center belt distance.
Our intake is floppy as all heck. I’m sure that belt tension is fluctuating wildly as the shaft that the output sprocket is on moves around.
I was planning to suggest that our students move/stiffen the design so that the first reduction doesn’t get moved around, and reduce belt tension by modifying the spacing to subtract maybe ~0.020" from the exact center-center.
We used a 550 in versa for our hood, 550 in versa for our turret, 550 in versa for a ‘side’ conveyors in our ball hopper. Gear ratios with something like 340:1, 200:1, and 5:1. No issues.
Has anyone ever had a shaft failure with a NEO 550 that was NOT using it in a single stage belt reduction? Specifically 5mm pitch HTD? These failures all look the same.
I was thinking the same thing… My takeaway (thus far) is to not use the shaft with any sort of side load on it. If a 1:1 (or other ratio) VP keeps the shaft from breaking, it’s well worth it!
Looking closely at the pictures of the failed shafts, it’s pretty clear that these are fatigue failures. (example) Fatigue failure comes from periodic stresses on a beam, e.g. side-load on a rotating shaft. It’ll work fine for a while, but after a certain number of rotations it will suddenly break. A high-speed motor with significant side load (like a well tensioned belt reduction) is the perfect storm for a short life before fatigue failure. There’s a good bit of math and engineering you can do to predict exactly how long it will take before the shaft breaks (accurate to about an order of magnitude), but for something like this it’s much easier to put a blanket “don’t side-load NEO 550 shafts” and call it a day.
Joining the club to say that we had a shaft snap in a single stage belt reduction on our funnel, mounted on 1/8" polycarb and 1/8" aluminum plate sandwich (the aluminum was added as a stiffener to keep the C-C constant). The holes were poorly drilled, so we chalked it up to poor belt tension, but it’s a near-identical failure mode as everything else here.
It seems like an ultraplanetary is a must when using these motors with belts. It’s disappointing to lose out on some of the really tight spaces you can fit an N550 into (like our intake) but it’s not worth random failures of critical mechanisms. I suspect a gear pinion would probably work but that simply isn’t possible in a lot of designs, especially 2" VIW intakes.
Alternatively here’s a quick concept for supporting the shaft. A printed pulley that fits over a 6t 20DP pinion turned down to 3/8" round.
To be contrary for a minute, we ran a very similar set up as a lot of people here only difference is with another belt stage and never had the motor shaft break, even after a considerable amount of abuse.
Come to think of it, we did also have a two stage belt reduction on our turret conveyor, and despite being located on the shooter, we never had issues with it
Another alternative would be to mount the pulley to the rear housing. Hmm… a small timing pulley with one oversize integral flange and six holes on a Rev 16mm motion circle?
The number of periodic cycles that a beam can undergo before it fails is determined by an S-N curve. Here’s an example curve for a type of steel, though every alloy will have a different curve. On the x-axis is the number of cycles before failure (N), and on the y-axis is the max stress for that number of cycles (S). It’s clear that N is highly-dependent on S: for every extra 100 MPa the shaft will break ~10x faster. And if the stress is below a certain level then the shaft can work basically forever. With the free speed of the NEO 550 being 11k rpm, if the robot has a 2 hour life the shaft will spin 10^6 cycles. So if one belt is more tensioned than another that can be the difference between failing and not.
And I want to restate here: just because your shaft hasn’t failed yet doesn’t mean that it’s safe. Fatigue failure comes on suddenly; you won’t know it’s going to fail until after it’s already failed. If the shaft stress is above the infinite value on the S-N curve then your shaft will fail eventually. Since it’s very difficult to actually measure the belt tension here, it’s hard to say when. I’d be cautious of using them in a situation similar to ones that have shown to end in failure, even if others haven’t seen failure with a similar setup (yet).
We ran our intake off a 3mm GT2 belt using a pulley on the NEO550 pinion shaft. The reduction was fairly small with subsequent belt reductions protecting the motor from any shock loading from jammed balls or such. We never experienced a shaft failure.
We used four “baby NEO’s” in our robot this year. Two competitions, 30 matches plus practice - no shaft failures. All of ours go directly to a versaplanetary. We did have some issues with the set screw coming loose on the shaft and therefore the gearbox did not turn.
