Drive Train Gearboxes

[thinker&planner]

Background: We have had CAD of a 6/8 wheel Drive Train for 6 months that revolves around the AM Sonic Shifter, which just went out of stock

My first question: What other shifting gearboxes are there out there (MUST be shifting)? I only know of the two Vex shifters, the WCP DS and the Vex Ball Shifter.
My second question: What are the advantages/disadvantages of each of the Vex shifters? What experiences have you had with them?

My third question: What is is the problem with making custom gearboxes? I know how to calculate center distance and all of the other design aspects, and I heard from one of the top engineers at NHMFL (Our local "National High Magnetic Field Laboratory") that the limiting factor is the machinist. Is this true?
What are the most crucial things to consider when designing and manufacturing a custom gearbox? What should we keep in mind when deciding to make custom gearboxes or order COTS ones? (I am planning to design the custom one with all the same gears/parts as a COTS one, just with a different plate layout.

Thanks so much in advance.

[TheFrozenSlink]

Originally Posted by thinker&planner

Background: We have had CAD of a 6/8 wheel Drive Train for 6 months that revolves around the AM Sonic Shifter, which just went out of stock

My first question: What other shifting gearboxes are there out there (MUST be shifting)? I only know of the two Vex shifters, the WCP DS and the Vex Ball Shifter.
My second question: What are the advantages/disadvantages of each of the Vex shifters? What experiences have you had with them?

My third question: What is is the problem with making custom gearboxes? I know how to calculate center distance and all of the other design aspects, and I heard from one of the top engineers at NHMFL (Our local "National High Magnetic Field Laboratory") that the limiting factor is the machinist. Is this true?
What are the most crucial things to consider when designing and manufacturing a custom gearbox? What should we keep in mind when deciding to make custom gearboxes or order COTS ones? (I am planning to design the custom one with all the same gears/parts as a COTS one, just with a different plate layout.

Thanks so much in advance.

AndyMark also has a couple of other shifters, namely the supershifters. They work on the same principle, the sonic shifters are just the newest, most updated.

Our team bought ball shifters a couple weeks before kickoff, and still hasnt gotten them yet ( ), so I cant really comment on how well they work yet, thought from what CD tells me they do their job great.

[Adamz_]

My team use the VEX 2 Cim ballshifters last year, and we had 0 problems, they worked great, i cant speak for much else tho

[cad321]

In terms of the placement of the gears/bearing holes, if you decide to go with the stock ratio or any other one supported by the shifter, you can just take the layout from the CAD model provided by the manufacturer. This way you know that everything is layed out correctly. The ball shifters work great however if you make a custom enclosure (we are doing this and did it last year as well), make sure you have all of the proper dimensions. Last year we didn't have one dimension (distance from the pancake cylinder to the 44 tooth gear) and it screwed up the entire shifter. Finally for the machining aspect, we CNC'd the plates last year in our shop and they turned out fantastic! I wouldn't even want to imagine having to do it on a manual mill though. I would definitely suggest having access to a CNC, water jet or laser cutter for this.

[magnets]

Originally Posted by cad321

I heard when we were at Waterloo last year that 1114 also ran into this issue last year when making a their custom PTO (correct me if I'm wrong). Finally for the machining aspect, we CNC'd the plates last year in our shop and they turned out fantastic! I wouldn't even want to imagine having to do it on a manual mill though. I would definitely suggest having access to a CNC, water jet or laser cutter for this.

IIRC, 1114 didn't have a PTO in 2013.

As for the transmissions, they aren't actually that difficult to do on a manual mill if you don't include fancy lightening holes, or you use a thin plate and add bends for rigidity.

To make them on a manual mill, you need holes for the motor/motor mounts (2 holes, 4 mounting holes), then holes for your bearings, then mounting holes for the gearbox. If you want press fit bearings, drill the holes undersize and ream them.

I can get a much more accurate part on a 60's bridgeport that on a 20k CNC machine I've used on a robotics team. On my bridgeport, I can always approach my hole from the same way, meaning backlash in the ball screws doesn't affect my accuracy.

[Answer42]

At this point in the season, it would most likely be easiest to just buy a COTS gearbox. The options on the market now are pretty fantastic, and IMHO it would be really hard to match their effectiveness and reliability if you've never made one before. Maybe you should save custom making them for the off-season. The vex pro 3 cim ball shifters are a great option for your teams current needs.

[Seth Mallory]

Definitely buy gear boxes at this time. Team 192 has built different gear boxes in 12 seasons. Each was a great project and in final form work well . We always built prototypes in the fall and then built improved ones after testing and finding faults each time. The games ones are built during build. If we could not build and test in the fall we would buy ones for reliability.

[Ether]

Originally Posted by Seth Mallory

Defiantly buy gear boxes

You will likely get friendlier customer service if you buy them courteously

[thinker&planner]

Thanks for all your responses.
When we talked to machinists that make 45T hybrid magnets, they said that we would need precision down to at least .0005". Is this truly necessary for FRC gearboxes? I think that they are used to metric gears with much smaller diametral pitches where that kind of precision really maters.

We were told repeatedly that you can't waterjet gearboxes, because you just can't get the precision that you need for press fit bearings.

Basically, all our machinists said that they would not make a gearbox for us unless they would be mass-producing them. I understand that calibration and setup takes a long time, but if there's no fancy lightening pattern, couldn't you just put two plates together and make all the holes relative to each other? Later, you could cut it out to the right dimensions (roughly), as long as all the holes were correctly spaced.
Thoughts? Thanks!

[apalrd]

A waterjet has adequate tolerances for an FRC gearbox.

We can tolerate some amount of slop by adding 0.003" or so between centers.

We usually waterjet the holes a tiny bit undersized and then manually file them to be on size for the bearings. We oversize the non-bearing holes slightly (bolts and such), or undersize them quite a bit and drill them out.

IMHO, we will probably never do that again for drive gearboxes now that we can buy a plastic case VexPro shifter that's lighter, has better ratio spreads, handles 3 CIMS already, and is cheap. We used to make custom plates for Andy's Gen1 gearboxes because the AM plates were steel, then when he started making aluminum plates we continued to make our own so we could integrate them as part of the chassis panels.

[JamesCH95]

For FRC applications I would be very surprised in 1/2-thousandths tolerance was required... frankly I would stunned if any OTS gearboxes were near that tolerance.

I've heard of teams water-jetting gearbox plates and carefully using a deburring tool to get the bearing holes to the right size. You could also water jet the plates and ream the bearing holes to a final size. There are many other fabrication methods proven to work in FIRST, include good old-fashion milling.

Remind your machinist(s) that the gearbox needs to last a matter of hours/days, not decades.

[magnets]

Originally Posted by thinker&planner

Thanks for all your responses.
When we talked to machinists that make 45T hybrid magnets, they said that we would need precision down to at least .0005". Is this truly necessary for FRC gearboxes? I think that they are used to metric gears with much smaller diametral pitches where that kind of precision really maters.

We were told repeatedly that you can't waterjet gearboxes, because you just can't get the precision that you need for press fit bearings.

Basically, all our machinists said that they would not make a gearbox for us unless they would be mass-producing them. I understand that calibration and setup takes a long time, but if there's no fancy lightening pattern, couldn't you just put two plates together and make all the holes relative to each other? Later, you could cut it out to the right dimensions (roughly), as long as all the holes were correctly spaced.
Thoughts? Thanks!

If you're talking about cutting plates, then your machinist is very wrong. You can easily get by with 0.003" (6 times what he said). The issue he may be worrying about is if the gears get too close to each other. They rub, have a lot of friction, and heat up. It's one of the reasons teams add the magic 0.003" between center to center distances, so that if it's off, it still works. The extra slop also makes the gearbox run better for applications with little wear in time and low usage. remember, the thing only has to work for less than 100 hours, which is why companies like vex pro can get away with selling aluminum gears, which are more than good enough for FRC, but not for a transmission in your car, or something like that.

The only thing you really would need crazy tolerances for are press fit bearings, but you can ream them yourself.

You're right in saying that the plate is nothing more than a few lined up holes. The rest of the plate (mounting holes, lightening holes) don't really need to be anything special. I'm not a machining expert, but I've been able to make made our practice robot's gearbox plates in under 15 minutes each.

[AdamHeard]

Originally Posted by magnets

If you're talking about cutting plates, then your machinist is very wrong.

He's not wrong, he's absolutely correct... just for a different industry and application.

FRC robots operate for short time periods with near constant maintenance (and in the presence of their techs/designers near 100% of the time). Most industry gearboxes run for several orders of magnitude longer and need to operate w/o any maintenance or failure and the techs/designers aren't't on site for any fix.

As for opening holes, purchasing reamers for the bearing sizes you is a great investment if waterjetting is available to you.

So many good COTS options now though, making your own just isn't as appealing as it used to be.

[Mr. Van]

I agree that you should probably simply buy a pair of gearboxes. The primary FIRST vendors (AndyMark & VEXpro) have fantastic customer service, and they ship fast.

I've actually lost count as to the number of robots that our current AM Supershifters have actually been in. It's at least 3. Quite a return on our initial investment!

- Mr. Van
Coach, Robodox