pic: Dual FP gearbox



I need to get off my computer and go outside or something…

Anyway, this is a gearbox designed to take two fisher price motors and sort of goes along with my arm joint idea. I’m trying to minimize the amount of machining on gears that would have to be done (seeing as I’ve never actually designed or built a gearbox), so this gearbox uses AM toughbox gears and the second stage reduction from the fisher price gearbox that has come in the kit in the past. The parts that need to be machined are the 3/8" hex axles, the standoffs, and the two 1/4" aluminum plates (also the spacers, but I didn’t put them in yet). Total dimensions are 3.5" x 8.5" x 2.125".

Since I’ve never made one of these, I was hoping some more experienced people could give some criticism. From other gearboxes I’ve seen, there are always a few thousandths added to the nominal center to center distances between meshing gears. I added 3 thousandths to all of the meshes in this gearbox (just like on the Toughbox), is this too much? Too little? Not make a difference?

Also, has anyone had success in the past running two FP motors through one of the plastic gears provided in the kit?

Thanks in advance for any tips.

The gearbox looks great overall!

Just a few questions:

  • What is the overall reduction?
    *]Have you thought of using thinner plate than 1/4?
    My main point of feedback for this is that you should really consider putting some holes in the plate around the motor mount location for the FP to vent.

Also, you could consider using delrin instead of aluminum. For the same amount of material, it is almost half the weight. Since you wanted to do as little machining as possible, using a lighter material than aluminum could help alot.

Those distances added to the center to centers all depend on your plate material, stiffness, machining capabilities, and so on. The team I’ve worked with has always been fortunate enough to work with a machine shop that could hold true to .0001, so we never really saw a need to design around the machining equipment.

What will these be machined on? If you find what tolerances they hold, it’ll be easier to identify what spacing you need.

If not the plates, as DKong suggested to be delrin, at least make the standoffs out of delrin.

One thing you might want to consider is adding clearance holes for an Allen wrench to be able to access the FP motor screws. This would make it a lot easier to replace a motor if necessary.

Ryan

I always design with an additional 0.003" between axle centers. This has absolutely nothing to do with machine tolerances, but instead with gearbox efficiency.

They run a lot better if you run them a little loose…

I think designing them at “pitch circle + .003” is a pretty good bench mark. Should be fine.

-John

Listen to John instead of me. He’s a real Engineer, and knows a TON more than I do.

(out of curiosity, how is there more efficiency from more space? I’m visualizing the rounded tooth profile rolling across another tooth, and not seeing how more space can help this. Care to clarify?)

Forgot to mention; the overall reduction is about 190:1. This gives the output shaft a stall torque of about 126 foot-pounds. With a 12/60 chain reduction to the arm, the stall torque at the arm is 630 foot-pounds (to which we all said “wow!”) and the tension in the chain is about 2000 lbs (again, wow!). Ideally, though, the robot will never see these kinds of loads. The purpose of running two fisher price motors is to keep them well under their stall torques, so they hopefully never stall and release precious smoke.

We were considering the possibility of thinner plates, or possibly pocketing out the extra material or using delrin, but I didn’t get around to it in this rendering. The total weight of both aluminum sideplates is 1.29 lbs, so the weight really isn’t that bad (it could be reduced if necessary though). Total weight of the gearbox with motors and bolts is given by inventor to be about 5.3 lbs. This can be dropped to 4.6 by using aluminum AM cluster gears, and could probably get below 4 if we felt like pocketing sideplates and gears.

I also thought about breathing holes for the motors and clearance holes for their screws, but didn’t get around to putting them in this rendering. I doubt we would try to build a fisher price gearbox without them.

Thanks, this is the answer I was hoping for.

Our current plan is to try machining gearbox plates on a manual mill. We have access to a big one at our robot building facility, and there are a few high end ones in the AME department machine shop at my college. Hopefully we can keep all of the center to center distances within a thousandth, so it will probably work fine if we aim for 3 thousandths over nominal distance (I hope). If we get around to making something like this, we’ll see how it turns out.

Even if you can machine things to high tolerance, if you can design them to work under less strict tolerances you’d drastically cut down on the price of the component in the real world. I’ve seen instances where moving from 0.005" to 0.010" tolerances on all non-critical dimensions on a part cut the price nearly in half.

Does this mill have digital read outs?

If it does, then keeping tight tolerances isn’t generally too bad. But otherwise, you better be the master of feeling the backlash in the ball screws of the table to keep tight tolerances. But seeing how almost all of your holes are in the same line, even if you don’t have DROs, you can keep precise (though not necessarily accurate) tolerances as long as you never overshoot your target distance (read: you are always staying on one side of the ball screw backlash).

That thing could become MUCH more compact if you’re willing to have fancier side plates. There’s no reason that the gears need to be in a straight line. Also, you can make the whole gearbox thinner (less distance between sideplates) by flipping the small and large gears on some of the axles.

Does that small blue gear need to be there? Doesn’t look like it’s doing anything.

thanks, Vivek

p.s. I’m a fan of heavily lightened 1/4" al sideplates. Besides being light and effective, they’re absolutely beautiful. :smiley:

Someone more experienced than me can chime in if I am incorrect…

It was once explained to me that standard tooth profiles are NOT designed to run “happy” initially. They are designed to wear-in over time. Unfortunately for us, FRC Robots have a ridiculously short lifespan compared to more “industrial” applications. Adding a few extra mills to the axle spacing just serves as a replacement for this true wear-in.

I’m not sure if this is true, or if it is just some urban legend. I know that the person who told it to me is someone I trust and respect VERY much. I also know you can “feel” the difference when spinning a gearbox by hand.

The most important stage to reduce drag on, is the initial stage of the gearbox. This drag has the greatest effect on the output.

-John

No digital readouts, but it does have thousandths increments on the dial. I’m aware of the slop on most of the machines we get to use, so there’s a chance that the holes will end up where they need to be. :slight_smile: Earlier this morning I made one of the axles just for fun on my dad’s clunky old lathe (backlash galore), and I was able to get the ends turned down to within .001" of where I wanted them. We’ll see how it turns out on an endmill, though.

And you are right about them being in a straight line; I’d like to do as few adjustments as possible.

It definitely could be more compact, but I wasn’t aiming for anything fancy for a first attempt. Perhaps for future gearboxes.

The small blue gear is molded onto the big one (it’s one piece). I would shy away from cutting it off… it would make the hub much weaker. It may survive, though. We were planning on torquing the blue gear until failure with a torque wrench to see just how strong it is.

chiming in on what JVN said:

We have done pitch dia. + .003 in the past and it has worked out well, and you can most certainly feel the difference between one that was designed with and without.

I’ve also been told about the “urban legend” of gears being designed to wear in to a place where they are spaced properly.

In my head it makes sense to me, because I see the gear teeth not reaching as far into the groove of the other gear. If there is less frictional loss (meaning less area of contact between the tooth of the gear and its corresponding groove) it would make sense it has a better efficiency. Of course this only applies to a certain degree, because at some point the gears are not going to be meshing properly.

As to using the FP plastic gear with two motors…we did a test with the torque wrench, a plastic gear in a vise, and a 3/8" hex shaft, to see if the hex hub in the gear could withstand the extra torque.

Looks like a great way to move an arm. I’m wondering, though, if replacing a few of the stages with a couple of Banebots planetary gearboxes would cut down size and weight sufficiently to be worth the cost.

We were looking at how we could take a couple FPs, then use the BBs to drive a worm gearbox for moving an arm.

I know that kind of defeats the purpose of building your own gearbox from scratch… and I do have to admit that I enjoy seeing custom work more than just seeing yet another COTS gearbox… but when build time gets tight our team tends to head for the COTS solution.

Jason

yes but these are the kind of things that can keep you busy in the off-season. make these and other components, so that when build season comes, you really just need to assemble together and then program it!

(sounds good in theory, but doesn’t work out for our team) :frowning:

It’s also largely against the rules.

Exercises like this are great for following through on ideas and learning about complications that may arise and how to best solve them. That’s the stuff that useful during the build season – not having a plug and play system of robot components at the ready.

is it really? i didnt know that :frowning: i guess i need to read the rule manual again.

so you cant make like L brackets or anything before the season start?

You certainly cannot manufacture anything before the season starts and their are also rules governing the extent to which you may design components before kick off as well.