Prototypes printed in HIPS with a .8 nozzle to get things done in a reasonable time final version will be in Nylon 910
Neat stuff! At the risk of being “that guy,” I’d like to offer some advice regarding your posts. To preface this, I’ll say that most of the projects you’ve posted seem interesting and are a contributing factor into me getting into 3D printing as a hobby (as opposed printing at work).
I have found that your posts/threads often lack clear context which might help drive more/better discussion and/or might be better served if kept in the same thread (e.g: multiple threads regarding bevel gears in the last month). Some context I’m missing from this particular widget:
Ok the end goal is to have a gearbox/wheel combo that takes less space and is customizable. it will look something like this
Customizable ratios - the bevels are intended to come off the motor and input shaft easily so that
can allow to quickly change ratios from 6:1 to 11:1 in minutes.
Take less space “where it counts” like often the motors sticking out interfere with ball intakes.
Hopefully provide cooling for the motor as we will attach a fan to the gear that is mounted on the motor shaft. (That is currently under discussion)
Reduce Weight - without the motor (which could be CIM, Mini CIM, NEO) The whole wheel/gearbox combo including all mounting hardware and bearings (5-10 608) will be about 2lb.
Reduce cost. Including everything (if it works as intended) will be below $ 25.
The first version will go into a demo / community outreach bot. If it works well there it might make itself into the competition bots which we hope very much.
As for 3DP we consider it a tool to build robots rather than a hobby - unless you consider doing FIRST a hobby in which case - it would be.
Everything I post is at least a co-production. At this moment we spent going over gears/gearboxes and how to draw them. The kids figured out and designed/cadded the components. I fixed some things like today I changed the backlash (and told them) cause the thing was too loose and I print as currently all printers are at my house. We are remote only due to covid. I then take pictures/videos of the stuff they designed and I printed and post it on our Discord Channel that we use to communicate/hold meetings. I also go order/pick up parts and then drop them at the kids porches/front/back door and they do a task with it like disassemble, assemble, paint, Cut something, drill holes - depending on what tools they have at home and whats ok with their parents. Then I pick stuff up and either bring it someplace else or my place and I QC everything like I used to when we were “in person” before covid. If it passes/fine - if it doesn’t and its something minor I might fix it right there or bring it back or inform the cad people to change the design.
The prototype in this post is the 2nd iteration. The first one I posted some “problem pics” on discord with
Planet carrier does not fit over bevel gear mount there is no way to assemble it
Planet carrier is too flimsy.
There is no way to get a wrench in there either make a holder for the head of the bolt (like the did as is shown in the above picture as part of the thickening of the planet holder) or find some room on that side when the bevel is in place.
etc.
Then I got the above version and printed it and I put it together and there is a report that we need a bearing holder for the sun on the outside ring gear holder and the mount for the bevel on the outside (not shown) needs to be redesigned. So the whole thing has been going on for a couple of weeks now but now we actually have something to hold and look at (and take pics)
Now someone will redo the outside holder. I changed the backlash right now and refreshed the drawing cause I consider that minor I might have to change the backlash a couple of more times in different places until its neither too loose nor too tight. This is what happens with 3DP gear system at least the printers are quite consistent so once a part is “dialed in” you can print many and they come out close enough. There might be some flowrate and/or backlash adjustments necessary again when we go to Nylon. The reason we are doing them in HIPS right now is cause we pick up HIPS for <$10/kg including shipping. So its not that financially painful if you got to redo a part. We are even considering making the gear boxes for the demo bot out of HIPS and see how long that lasts running it over rough terrain.
So I guess its a both. As for who does it. Its usually a both in the “in person” world too. As you got to show them how to do it at least once.
Now the Bevels atm are all me as no one knows how to do them right including me and what is the best way. I have figured straight bevels out and they work great but am still struggling in properly designing a proper zero bevel or double helical one. Would be nice to use those as they are self aligning (like the double helical gears in the planetary above) and that would make mounting/changing/replacing them easier. I am currently preparing a “class” where I will teach them how to do a proper 3DP plastic straight bevel gear and then they will design one for that gearbox in place of the one I designed. So that is how that works.
Now the community outreach bot will look something like this
Thanks for the extra info! It’s clarified a lot of lingering questions I’ve had.
I hope my mentioning of 3DP as a hobby didn’t come off the wrong way. It’s an indispensable tool in my professional work at a consumer electronics R&D lab, but in that context it needs to be more or less bulletproof. I have assembled small-but-growing fleet of “prosumer” printers, primarily Markforged, for this purpose. I’ll admit that it does feel a bit boring at times but I’m greatful for this class of machines on most days.
To appease my Maker tendencies, I decided to hop into 3D printing as a hobby after seeing how hobby-grade/sub-$1000 hardware has come a fair way since I interned at MAKE: in 2015 both in terms of the hardware itself and the communities. I use the word “hobby” here to indicate that I’m just working on personal projects unattached to any company or activity like FRC.
The above was printed on an anycubic Chiron we have 2 we bought them for <400 each, The other one needs some TLC, This one already got it. printing 4400 prusa masks at 53 g each plus the robotics stuff took its toll My printer has by now emptied over 150kg worth of filament spools mostly abs and hips and you might have seen this for context
https://pirringers.com/3dp/?p=36
And FWIW it has helped us progress as a robotics team
You may want to revisit the number of planets, depending on desired ratio and how things are mounted (driving carrier vs driving sun) It may be a good Idea to up the number to 5 or 7. Why an odd number? Well as the ring gear is under load as integral to the hub you will experience deflection which will in turn could wreck havoc with your gear train (changing CTC distance resulting in excessive wear best case, worst case… bang). There are ways around this for sure, but just take it into account.
I have suspected that the (final iteration(s) of) Biscuit Technology in 2767’s 2019 bot could have adapted well into a drive train hub for low speed applications (due to the bushing). This used something like 10dp or 12dp onyx gearing with lubrication (shielded) (11t sun 12t planets (x3) if memory serves), a bearing on one side and a bushing on the other to keep things as true as possible as the whole sub assembly was cantilevered.
Outside of FRC:
Tantrum, of battlebots fame, has a good implementation of reduction hub in wheel (clean design IMO) if you are willing to pull out all the stops (which is kinda the point for BB). Note the massive roller bearing that was used to reduce deflection of the rotational axis to keep thing aligned. Maybe someone can chime in a bit more about this design. (@Aren_Hill)
I don’t see how? The ring is the hub in our design so it has to be the output If I make the planet carrier the input then I got to keep the sun fixed. The formula for a planetary is (R+S)Ty = RTr+STs
in your case STs would become 0, Ty 1 SO TR would be (R+S)/R so that means the output would be slightly faster than the input.We want to gear down not up.
If we add more planets we got to make the planets smaller that will cut the output ratio down And the reason we haver 4 planets is that more wont fit and make the hub 4 in max. If you 3DP the smallest module is about 1.2 mm with the setup we currently have we use 1.5 as the load will most of the time come from the ground up so we are basically riding on the ring and bottom planet and sun for the most part the remaining gears are holding the wheel in place when we accelerate or break or hit bumps We wanted to make the module as big as possible while still falling within the range of reduction we are looking for. In a way the planetary works like a bearing in this case besides being used to reduce speed/increase torque. That is why we are using double helical gears and tried to maximize the module to give it a chance to also work as a thrust bearing in case someone rams us from the side or we are taking a sharp turn. So the bigger the module the less likely the ring gets pushed off the planets of the planets of the sun or if it does a little then the double helical gears will in most cases realign the whole system. And if you ask why 3DP and not go out and buy all kinds of stuff. For one its the only manufacturing method we can afford and its cheap and with our budget constraints 3DP allows us to do things we otherwise never could and we can “afford” to take on extra projects like this gearbox and community bot for both of which and the other gearboxes including all the electronics (we will run it off a PI with own rolled PWM H drivers to drive the motors). And before you ask why that. First we can and 2nd this thing will have a minimum of 5 - most likely 7 motors. We can send a direction and PWM signal from the PI we have done that and a couple of pnp transistors and 4 MOSFETs in an Hbridge arrangement will drive the motor in both direction. And you are talking pennies for the transistors maybe $1 in resistors caps and diodes to take care of the back emf a buck or so for the pc board and about $5 for the mosfets and heatsinks. And if we want to splurge we can stick an Op amp on it and do some current sensing. All in all we can drive a motor for about $10 probably less. Now I know that is not possible on a FIRST robot but on a demo bot it makes a nice electronics project and helps to stay under budget
I’m not trying to tear apart the collaborative design process here, just pointing to similar stuff that exists in the wild. All I was trying to do was provide some other designs and ways groups have tackled similar mechanisms, for inspiration. I am not implying you should aspire to break the bank, those designs are documented and I know of them off the top of my head.
I will admit I typed the bit about deflection with out giving full consideration that (as a demo bot not hitting bumps at high speed) the loads will ultimately be taken through the ring gear and into the carrier fully supported. Deflection would probably be manageable in the nominal loading case for sure.
To put it more formally, things to consider:
I agree that is some concern too but each planet carries 2 bearings like this
Now we have one on each side the gears are 32 mm wide and also this was a 22 tooth gear with a 60 tooth ring that if reinforce was a tad too big so we dropped the teeth down to 20 with a 16 tooth sun and a 56 tooth ring. Anyway you can see that the teeth is just like a “tire” that transfers the load to the steel bearing which is a 608 which is a scateboard/rollerscate bearing and designed to survive that kind of abuse. So the plastic will have 2 force groups one will be the torque put through it from the sun to the ring which is at least somewhat shared with the other planets and the other one is shocks from the ground as its part of a tire which is compressive and nylon does very well (and HIPS to a lesser extend) under compression and has great recovery to the original shape Now we used HIPS gears in a rack and pinion setup at that pitch and size to lift 200 lb So I have no concern load wise. We also Had a similar setup before and under excessive impact it bent the 8mm or 5/16 steel shaft and in that case the nylon gear showed no damage even though we also broke the seal on the bearing due to impact. Now Nylon has the good/bad property of being flexible so at a too small of a pitch you can skip teeth as they bend out of the way under too heavy a load but snap back and the gear is fine whereas with HIPS you get less skipping but can break a tooth. So for that reason also we upped the modulus to 1.5 and maxed out on the addendum. If you look at it there is very little room for a bearing as for allignment of the planets for now the planet on he bottom people won over the turn it 45 deg people in which case it would ride on the ring which could slightly flex and deform and give it a little “spring” an make 2 planets carry equal loads. Another suggestion was made to not space the planets evenly which can be easily achieved and would just require the carrier to be redesigned and reprinted
When you design your parts do you consider thermal expansion of the plastic when determining the length of everything? If not what plastic are you using? I’ve tried 3D printing a few things on my school’s MakerBot 3D printer but the XY dimensions are always off.
On a planetary it does not seem to be a problem with either Nylon or HIPS. First we never really had any significant heat buildup except what was transmitted through the motor an then the bearings have more “slop” than any heat expansion differential we had. Cause if it mostly comes from ambient then everything expands a tiny bit but pretty much in proportion to each other. I don’t have any data handy but from observation metal expands way more than plastic. With Nylon you can have more of a problem with moisture expansion so the first run of the gearbox I do in Nylon I will add .1 to the backlash I am establishing for HIPS (average from experience) and then print the dehydrated Nylon to get a good print then I will wrap them in a wet towel and stick them in a plastic bag for at least 2 days and then test them and make adjustments if necessary accordingly
This is what I had in my head to take the radial (and thrust) load off of the planetary set, the carrier plates double down as some shielding for the gearset so rocks etc don’t mangle the gears, you could go further with the shielding concept for sure. I didn’t model any of the planetary set beyond the ring gear.
The flange on the wheel acts against the thrust loads, so you could ditch the herringbone pattern if you want (this should make for a better surface finish right off the printer).
If you went this route you would probably want to print the wheel/ring gear as two halves (which you may be doing already) and add some thru fasteners to make assembly significantly easier.
EDIT: There are a ton of small features left out on these screenshots that would dramatically increase the strength/ lifetime of the print (at least in Onyx w/ superlube). Screenshots for illustrative purposes only. Also, take everything w/ a grain of salt, just trying to provide a little outside feedback, not trying to shatter the collaborative design process of the team 
It will get an enclosure to keep dirt out. We will definitely not ditch the Heringbone setup. first any helical gear is superior to any spur gear as multiple teeth are engaged at the same time then they run smoother. On a car when you go in reverse those are spur gears because they are not used often - lots of noise - lots of wear but you dont drive 10s of thousands of miles in reverse at 55mph. Now fwd you have helical gears they are much better due to multiple teeth engaging an there is less noise an wear but they need thrust bearings. Now herring bone - double helical do not need thrust bearings, are self aligning and the best gears there is . Drawback is its extremely difficult and expensive to make so they are rarely used. its difficult to cut them or make them with subtractive manufacturing. Now with additive manufacturing like 3DP that is not a problem. So therefor I advise herring bone 3dP gears whenever possible.
Lots of work, not a ton of return, and not a high enough FOS for that application, we’ve upgraded to these weird 14$ off the shelf neoprene fairlane wheels that are awesome.
Managed to get a fair number of hours of drive practice and being able to quickly and cheaply throw on a new set of tires was clutch for that.
I don’t really buy putting planetaries in the wheels as very useful for FRC purposes unless you’re driving other wheels off one wheel with the gearbox in it, then maybe 
(even then, your time is best spent elsewhere if you’re looking to improve performance)
-Aren
Update with Tire. Still all in HIPS for fit purposes once everything fits we will print in Nylon/TPU etc
We will get the proper length bolts once its finalized It can use either 5/16 or M8 as they are close enough to nicely fit the 608 bearings
Spur gears are actually more efficient in power transmission, but are typically not used in roadgoing transmissions due to their noise and wear characteristics. The main exception to this would be in competition rally transmissions, which have a pretty recognizable gear whine. So there’s a fun fact for the day. I do agree though that the double helical gearing is ideal for a printed setup, and greatly increases reliability. They’re pretty cool too.
Many (most?) transmissions in road-going cars use a straight-cut reverse gear because they are so small. Many people will hear the gear wine in reverse.
Roadranger transmissions, found in tractor trailers, are straight-cut gears without synchros. Straight cut gears do not generally suffer from poor “wear characteristics” all else being equal.
Straight cut gear transmissions are found in racecars from F1 to Production class. Basically any application where you care more about efficiency, weight, and power-handling, than noise.
Was talking what I learned from literature in regards to plastic gears - particularly 3DP ones. But agree otherwise plus with a straight cut gear you don’t need thrust bearings which again with a double helical you dont need either so I guess like with always it depends on the application
Also we moved the planets as its going to be a wheel this pic is with a test planet holder to see if we got the spacing right
Riddle me this Batman,
It looks like the robot’s weight is passing through planets. Isn’t that breaking some gear law? I though you weren’t suppose to push the ring into the planets.
Don’t get me wrong, I’m all for breaking gear law which I view more as guidelines anyway, just wondering if you considered that.
BTW @Skiddy4795 @JamesCH95 … M22 Rock Crusher… Ya buddy
