Custom gearboxes

Technical Technical Discussion
1

What exactly goes into a custom gearbox? I am specifically interested in ones for manipulators, as I plan on using AM headboard in the drivetrain.

2

I mean, I don’t want to sound like a smart-alick or anything but essentially identical components go into custom gearboxes as anything you’d buy off of andymark.com.

The thing that makes the majority of custom gearboxes “custom” is the fact they redesign the mounting plates to get motors in a more desirable location, offer easier mounting to their manipulator/frames, and get the gear ratios to something suitable for their needs.

3

This.

http://www.chiefdelphi.com/media/papers/2059

And this.

http://www.team1323.com/resources/manafacturing/team1323/all_in_one_gear_calculator.html

The bottom one. (Hint: 20 DP)

Enjoy.

4

another reason to do a custom gearbox is weight. and a lot of custom gearboxes are derived from cots components. this year we made a custom gearbox for our lift. we made a new housing for a dewalt drill gearbox.

5

If you’re interested in making your own custom gearboxes, it would help to have access to two machines:

-Lathe (For shafts)
-Mill/CNC (For side plates/ facing-off gears)

A gearbox is simple: Side plates, stand offs, bearings, shafts, spacers, gears, and keyways/keys.

A good place to buy quality gears would be Martin Sprocket. You can use either hardened steel gears from that manufacturer or 7075 aluminum gears (from a different source). It is safe to go with 24 DP, but I believe some people have had success with 28 DP.

Stand offs you can get from McMaster-Carr as well as keystock for keys. You can make aluminum side plates from 3/16 6061 alum. or steel 1/16 sheet metal (we don’t use sheet metal).

Autodesk Inventor makes it fairly easy to design gearboxes, if you are serious about making them CAD is the place to start. Inventor has gear generators, key/keyway generators, and bolted connection. If you input the right numbers into it it’ll calculate the ratio and it will also tell you if the gearbox will fail or not.

Gearboxes are harder to make the smaller they are but it’s not a matter of design it’s more of where to get the parts; tiny bearings are easy to find but working with such small shafts and gears may be difficult. It is not impossible to create a custom planetary gearbox but it is a challenge.

When making your gearbox always pick your gear ratio based on the motor curve- look at the point where torque and speed are balanced and decide where to go from there. Perhaps for and arm you need torque but a gripper may need speed. That’s probably the first step to determining how you design.

Hope this helps.

-Eric

6

From past experience, try to use as many parts from a common gearbox. If you are looking for a shifting gearbox look at http://www.andymark.com/product-p/am-0114.htm. The super shifter is on the slightly heavy side. This season, my team built custom gearboxes, we repackaged the majority of the Super Shifter parts into a lighter/more compact gearbox.

Hope this helps

7

For varied reduction single speed gearboxes, we’ve done something that I have poorly named in CAD as “AM based lasered gearbox plates”.

It’s a neat concept, we decide that regardless of gearing needs, we will ONLY use andymark toughbox gears. That means spurs with a .4 overall thickness, and pinions with a .6 overall thickness. Both already have machined nubs so they don’t rub on the bearings.

Then, you can machine gearbox plates by whatever means (we like waterjet/laser as it’s obviously fast and allows more flexible design).

Design the plates to be 1.125" apart (we usually bolt onto a 1" face with a .125" spacer.

For assembly, press bearings in, and you can use exclusively AM shafts if you desire. We machine our own to perfectly match our needs and eliminate spacers, but this is not necessary.

Doing this allows you to create custom gearboxes FAST and with minimal machining. We’ve shown you can get just about any reduction, from 1:1, to several hundred to one.

For my first custom geaarbox, I would go in this direction; utilize COTS gears and shafts. Machine only plates, and possibly standoffs. Spacers should not be necessary if you design it right (standoffs as well possibly)!

If you want the CAD for these to make the concept more clear, let me know.

Oh, last little tip, we also try to exclusively use 50T spurs. this allows us to use the aluminum gears AM sells. We cut weight off with our lathe and mill and get them down to .07 lbs each (versus .5 for the stock steel!).

8

Adam,

Thanks for mentioning this. You got me interested in designing gearboxes, and I’m pleased at how accessible it is to create a gearbox design that only calls for custom side plates and COTS gears and shafts. I’ve spent a bunch of time playing with various gearbox designs this summer, and I also showed some students how to design gearboxes. It’s really fun stuff. I’m also pleased at how easy it is to come up with a gearbox that’s lighter than the standard Toughbox or Super Shifter. It’s really cool that AndyMark sells their gearbox components individually to make this possible.

I think I also read a post that Art Dutra (Team 228) wrote about the same general concept, so thanks go out to him as well.

I noticed that the AndyMark gearboxes are designed with an extra 0.003" added to the center distances between gears. I’m curious whether teams typically add that extra clearance when designing custom gearboxes.

9

Team 254 typically does this.

10

Adam, please send the CAD on your gearbox designs to [email protected]

Rookie team #3481 Brandeis Bronc Botz says THANK YOU for your valuable help!

11

Adam,

Thanks for mentioning this. You got me interested in designing gearboxes, and I’m pleased at how accessible it is to create a gearbox design that only calls for custom side plates and COTS gears and shafts. I’ve spent a bunch of time playing with various gearbox designs this summer, and I also showed some students how to design gearboxes. It’s really fun stuff. I’m also pleased at how easy it is to come up with a gearbox that’s lighter than the standard Toughbox or Super Shifter. It’s really cool that AndyMark sells their gearbox components individually to make this possible.

I think I also read a post that Art Dutra (Team 228) wrote about the same general concept, so thanks go out to him as well.

I noticed that the AndyMark gearboxes are designed with an extra 0.003" added to the center distances between gears. I’m curious whether teams typically add that extra clearance when designing custom gearboxes.

I’m glad it worked out for you, it really takes out all the trivial work in making generic gearboxes. We still go all out for specialty gearboxes like shifters however. It certainly makes it easier for students to design gearboxes. I’d reccomend looking into easy ways for you to lighten gears, or just use the Aluminum gears w/o lightening. The 50T aluminum gears pocketed are about .07 lbs for us, versus .5 for stock steel!

The .003 is kind of a FIRST specific trick. You are actually supposed to not add to the c-c for gears, but since our run times are low and we usually don’t have an adequate wear in period it works. The .003 makes up for rougher surface finish on gears, etc… The tradeoff is you’re introducing backlash and slightly decreasing strength (very slightly, it’s not like .020"). Neither of these are usually a huge concern.

We tried exact C-C + a wear in period on or 2010 drive gearboxes for the sake of “doing it right” and we’ll never do that again, it was a waste of time for FRC purposes.

Since that post we’ve posted our entire 2009-2011 robots. The kicker on the 2010 and the shoulder on the 2011 both demonstrate such gearboxes. We’ve made multiple other for misc projects and prototypes.

12

Adding .003 of backlash is a plus for us mainly because of our low run time with gearboxes. These see maybe 10 hours of run time, so adding that bit of space makes things run smoother since gearboxes don’t really get to break in much. It also gives you a touch of wiggle room if you can’t hold thous very well.

13

I feel the 0.003" adder is most beneficial in terms of artificially shifting your tolerance band. For a gearbox thats used in FRC, its better to be a little bit too far away in terms of engagement, as opposed to too close and bottoming the teeth out.

We use it on all of our gearboxes.

-Brando

14

Adam, Thanks for your clarion gear guidance. I was wondering … Do the bearings fit with only water-jet cutting or do you undercut and ream for a perfect fit? We need to build our first custom gearbox this year and we’re planning to use your workflow.

Also, if you’d send your gearbox CAD file to [email protected], it would be supremely appreciated!

All the best,
Michael Lucido
Mentor, Team 2838 Eaglebots

15

973 and many other teams vary their ream based on the fit they want. If they want a slip fit they will go with the exact bearing OD ream, if they want a press fit they will ream .001 under the OD. Reamers are a very worthwhile investment. The main bearing sizes used for FRC are R6 and R8, so you would want to get .874", .875", 1.124", and 1.125" reamers(we got from maritool, and maritool doesn’t sell 1.124" reamers. Does anyone have a good source for reamers this size?)

16

You could invest in all those reamers, or just one boring head. Its a lot more versatile and a lot less spend on tooling.

17

I see that most teams are adding the 0.003" to the center-to-center gear distance. What’s the best way to calculate the starting center-to-center from which to subtract 0.003"

  • Inventor Gear Calculator? … Which seems fairly straightforward
  • Something better?

Thanks,
Michael

18

(gear1 / 40) + (gear2 / 40) + 0.003

(50t / 40) + (30t / 40) + 0.003

1.25 + 0.75 + .003 = 2.003

19

(number of teeth on gear A + number of teeth on gear B) / (2 * dp)

For AM gears DP is either 20 or 32, usually 20.

FYI, for a given c-c distance, any combination of gears with the same combined number of teeth will fit. E.g. 25-40 will fit in the same c-c distance calcualted for 20-45.

20

Thanks Andy and IndySam. I was just working through the pinion relationships to understand IndySam’s magic number of 40 and my calculation was off by 20 (on which Andy shines a bright light!).

Thanks!