pic: Modular Gearbox

[Ari423]

Originally Posted by Chris is me

While this is a cool design and all and I don't mean to discourage you, it certainly has been done before. One of AndyMark's early products was the StackerBox (now discontinued?), which did essentially the same thing as this. One could also argue that this is basically what a VersaPlanetary is, minus the planetary part of course.

I would switch the bearing holes to 1.125". You can get both 3/8 and 1/2" hex in that size, so you can pick what size shaft to use based on the stage of the gearbox.

I can't find the StackerBox on AndyMark's website, but from what I can tell they seem like similar ideas (great minds think alike!). Any idea why the Stackbox wasn't successful and was discontinued? My idea was to have something similar to a VersaPlanetary in its versatility except a spur gearbox not planetary, so it can be used in higher torque situations like a drivetrain or a heavy arm. Also spur gearboxes allow you to make smaller changes in gear reducation than planetaries, which can be useful to maximize the mechanism's efficiency.

The bearing holes, aka the two in the center where the axles go in the example, are 1.125" for exactly that reason (well 1.123" for a tight fit but yeah). All of the 3/4" holes are just lightening holes because the full aluminum tube was unnecessarily heavy IMO. I suppose you could put a bearing in them for some reason or other, but that's not their intended purpose. The only holes that are intended to have bearings in them are the two 1.125" holes in the center (or the four 1.125" holes in the double block).

[Chris is me]

Originally Posted by Ari423

I can't find the StackerBox on AndyMark's website, but from what I can tell they seem like similar ideas (great minds think alike!). Any idea why the Stackbox wasn't successful and was discontinued? My idea was to have something similar to a VersaPlanetary in its versatility except a spur gearbox not planetary, so it can be used in higher torque situations like a drivetrain or a heavy arm. Also spur gearboxes allow you to make smaller changes in gear reducation than planetaries, which can be useful to maximize the mechanism's efficiency.

The bearing holes, aka the two in the center where the axles go in the example, are 1.125" for exactly that reason (well 1.123" for a tight fit but yeah). All of the 3/4" holes are just lightening holes because the full aluminum tube was unnecessarily heavy IMO. I suppose you could put a bearing in them for some reason or other, but that's not their intended purpose. The only holes that are intended to have bearings in them are the two 1.125" holes in the center (or the four 1.125" holes in the double block).

I just misjudged the scale of the render and thought the bearing holes were 7/8, my bad.

I can only speculate on what happened with the StackerBox (I never bought one), but I suspect the main problem was that it had limited use cases. If you wanted 2 stages, you'd just get a ToughBox; if you wanted more than that, you could get an AM or GEM Planetary. There was also less variety in gear sizes then so it doesn't have as many possible ratios as the 84T Vex spacing allows.

Any way you could move to maybe 3/4" wide tube instead of 1" tube? At the moment its not very compact, and this would thin things out a bit.

[Ari423]

Originally Posted by Chris is me

Any way you could move to maybe 3/4" wide tube instead of 1" tube? At the moment its not very compact, and this would thin things out a bit.

The tubes there are actually 1.5" wide. Inside each tube needs to fit one gear (.5"), one shaft collar (.25"), and two bearing flanges (.0625" ea). 1" tube with a .100" wall gives .8" of inside width, which is .075" too small to fit everything. I was able to reduce the outside tube width to 1.125". I could further reduce it by removing the shaft collar and replacing it with two retaining rings, but that would require the shafts to be machines which I didn't want to do.

You can see the updated model at the same grabcad link.

[GeeTwo]

Originally Posted by Ari423

The tubes there are actually 1.5" wide. Inside each tube needs to fit one gear (.5"), one shaft collar (.25"), and two bearing flanges (.0625" ea). 1" tube with a .100" wall gives .8" of inside width, which is .075" too small to fit everything. I was able to reduce the outside tube width to 1.125". I could further reduce it by removing the shaft collar and replacing it with two retaining rings, but that would require the shafts to be machines which I didn't want to do.

I agree that you would not want retaining rings, even if machining were not an issue - they're a high-stress point.

Did you consider using thunder hex stock? Not because it's rounded, but because it has a bore. You could retain the shafts with self-tapping screws and washers (or tap the hole and use elevator bolts) and save a few tenths relative to shaft collars. There may be other pre-bored hex stock out there as well. Churros would work for demonstration purposes, but are not good for transferring the torque you'll need in a drive train, much less an arm.

[Ari423]

Originally Posted by GeeTwo

I agree that you would not want retaining rings, even if machining were not an issue - they're a high-stress point.

Did you consider using thunder hex stock? Not because it's rounded, but because it has a bore. You could retain the shafts with self-tapping screws and washers (or tap the hole and use elevator bolts) and save a few tenths relative to shaft collars. There may be other pre-bored hex stock out there as well. Churros would work for demonstration purposes, but are not good for transferring the torque you'll need in a drive train, much less an arm.

I thought of that, but I don't think it would work. Right now, the flanges are on the inside of the tube and the shaft collars are inside of them keeping the bearings from falling inwards. The screws on the end of the shaft would only be able to be outside of the bearings, meaning the bearings would have to be flipped so the flanges are on the outside of the tube. Because of that, the rest of the bearing is sticking through the tube wall .15" into the tube, taking up more space than the shaft collar.

*Let me know if I'm not explaining that clearly*

[Monochron]

Originally Posted by Ari423

I thought of that, but I don't think it would work. Right now, the flanges are on the inside of the tube and the shaft collars are inside of them keeping the bearings from falling inwards. The screws on the end of the shaft would only be able to be outside of the bearings, meaning the bearings would have to be flipped so the flanges are on the outside of the tube. Because of that, the rest of the bearing is sticking through the tube wall .15" into the tube, taking up more space than the shaft collar.

If you remove the collars from the inside you could replace them with a spacer so that the bearings (with their flanges on the inside) are flush against spacer or gear. Make the spacer long enough so that it presses against the gear and the gear presses against the other bearing. Doing that, you would only need a bolt/washer combo to retain the shaft axially if it isn't already held stationary by something else.

[Ari423]

Originally Posted by Monochron

If you remove the collars from the inside you could replace them with a spacer so that the bearings (with their flanges on the inside) are flush against spacer or gear. Make the spacer long enough so that it presses against the gear and the gear presses against the other bearing. Doing that, you would only need a bolt/washer combo to retain the shaft axially if it isn't already held stationary by something else.

That's a really good idea that I hadn't considered. With your suggestion, I was able to reduce single block to 1" wide, .125" wall and the double block to .75" wide, .100" wall. The 1 CIM gearbox is now 2 lbs and the 2 CIM gearbox is now 4 lbs (without motors). That makes the whole 2 CIM gearbox 2.75" wide.

Updated CAD models are at https://workbench.grabcad.com/workbe...ulyGgP-1FFgvMU

The other big problem I can't figure out how to avoid is that the CIM motor needs a trimmed shaft and a spacer (could be CIMcoder) to fit in the thin stage. I was really trying to not need any machining after the blocks are CNC'ed. Does anyone have any ideas how to get around this problem?

[Chris is me]

One more thing I want to remind you (and sorry to dominate the posts of this thread!) is that you want to make sure all of the tubing sizes you're using are readily commercially available. For example, .100 wall tubing is very uncommon outside of what Vex sells; most of it is 1/8" wall and 1/16" wall. Also, some of the odder sizes are only available in 6063 tubing, which, albeit weaker and crummier to machine, it is adequate for this application in 1/8" wall if there isn't a 6061 alternative.

[Ari423]

Originally Posted by Chris is me

One more thing I want to remind you (and sorry to dominate the posts of this thread!) is that you want to make sure all of the tubing sizes you're using are readily commercially available. For example, .100 wall tubing is very uncommon outside of what Vex sells; most of it is 1/8" wall and 1/16" wall. Also, some of the odder sizes are only available in 6063 tubing, which, albeit weaker and shittier to machine, it is adequate for this application in 1/8" wall if there isn't a 6061 alternative.

I don't mind your help at all! I thought 100 wall tubing was a common size but it seems I am mistaken. Also, I couldn't find any 3/4x6" tubing of any wall thickness so I changed the double block to 1x6x125". I can find that on McMaster, if not somewhere else cheaper. It makes the 2 CIM gearbox a little bit heavier and wider, but only slightly (.25" wider, .18 lbs heavier).

[nuclearnerd]

My idea was to have something similar to a VersaPlanetary in its versatility except a spur gearbox not planetary, so it can be used in higher torque situations like a drivetrain or a heavy arm

Having spent a season fixing exploding 3 stage versaplanetaries as they struggled to lift a heavy arm, I think the idea is great, especially if you can stack ratios to at least 200:1

[Ari423]

Originally Posted by nuclearnerd

Having spent a season fixing exploding 3 stage versaplanetaries as they struggled to lift a heavy arm, I think the idea is great, especially if you can stack ratios to at least 200:1

The big upside of planetary gearboxes is their huge reductions in small packages. Compared to the maximum 10:1 stage on a VP, the biggest reduction you can get from a spur gear stage with this system is 4.5:1 (9:1 over two stages using a combination of 3/8" and 1/2" gears). That would mean using only spur gear reductions to get up to 200:1 you would need 45 stages, or about a 4 ft wide gearbox.

Using a final 12:60 chain reduction (max for #35 chain w/ Vex sprockets) after the gearbox, it would take 9 stages. You could then further reduce that by putting the CIMs through a 10:1 VP reduction before it goes into this gearbox (which should be a lot more manageable than a 200:1 VP reduction). You would only need five stages direct driven with a 10:1 VP on the input side. With both the 10:1 VP reduction and the 12:60 chain reduction, you would only need two stages. If you want a 200:1 reduction without a final chain reduction, you can put a CIM through a 50:1 two-stage VP (the max it's rated for) and then into a two-stage spur gearbox. That will result in a max reduction of 672:1.

So in summary, yes you should be able to get a 200:1 reduction, it just takes more space than a VP. But, since it's (probably) rated for those high loads, it shouldn't explode every time you try to use it.

P.S. - All these huge reductions are theoretically possible, but are practically limited by the max torque capability of the shaft. I'm too lazy to look up the max torque rating but I'm sure it's online somewhere or other.

P.P.S. - VPs with a CIM input are rated for a max of 50:1, and only with a 1/2" hex output shaft. I'm not surprised a CIM through a 200:1 reduction exploded.

[asid61]

Originally Posted by Ari423

The big upside of planetary gearboxes is their huge reductions in small packages. Compared to the maximum 10:1 stage on a VP, the biggest reduction you can get from a spur gear stage with this system is 4.5:1 (9:1 over two stages using a combination of 3/8" and 1/2" gears). That would mean using only spur gear reductions to get up to 200:1 you would need 45 stages, or about a 4 ft wide gearbox.

You're not doing that calculation correctly. Gear ratios are multiplied, not added, so if you stacked 4 stages you would get a 4.5 * 4.5 * 4.5 * 4.5:1 ratio, which is about 400:1. Using a chain reduction for the last stage would give you even more.

[nuclearnerd]

Originally Posted by Ari423

P.P.S. - VPs with a CIM input are rated for a max of 50:1, and only with a 1/2" hex output shaft. I'm not surprised a CIM through a 200:1 reduction exploded.

We used Bag motors, which vex rates up to 300:1 if you stack the stages in the right order. That said, we found the VP assembly pretty unreliable at the limit of the torque "rating", so your point stands.

[Jon Stratis]

Originally Posted by GeeTwo

I agree that you would not want retaining rings, even if machining were not an issue - they're a high-stress point.

Sure, it's a stress point... but is it really that big of a deal? We've used some AndyMark Toughbox Nano's in several applications (including drive train) without any issues at all, and they use retaining rings. In fact, we have one pair of Nano's that's made it through two seasons on two separate robots, with a season in between where it was used for practice. That experience would seem to indicate that using retaining rings may not actually be all that bad.

[Chris is me]

Originally Posted by Jon Stratis

Sure, it's a stress point... but is it really that big of a deal? We've used some AndyMark Toughbox Nano's in several applications (including drive train) without any issues at all, and they use retaining rings. In fact, we have one pair of Nano's that's made it through two seasons on two separate robots, with a season in between where it was used for practice. That experience would seem to indicate that using retaining rings may not actually be all that bad.

It is a big deal if the snap rings are placed between loads on a shaft. They are a major stress riser then. They are fine on the ends of a shaft, but I would never (again) put them between torques, especially on higher reductions.

In this application - can you just have bearing / gear / bearing on the inside of the tube, and shaft retention handled elsewhere? That would require an inner width of only .75, which lets you use 1" wide .125 wall tubing.