This has been quite a project but it is finally done. It is a modified VEX ball shifter with a bevel drive.
Calculations done with a 4in wheel
High: 20fps (Predicted)
Low: 9fps (Predicted)
We know that 20fps is a little fast but it was the best we could figure out without rethinking the whole gearbox. Our idea was to just put a software limit if it is too fast but we will have to see how fast it actually goes. It is going into a custom drive base that we are still working on. We will be manufacturing one as soon as our grant money is put in our account and our mill is up and running!
I think it might be a smart idea to 3D print the side plates, that way you don’t have to use all your grant money machining something that might not work
Remember, folks, 3D printing can be plenty strong, but if you 3D print plastic sideplates for gearboxes without redesigning from a design made for metal, you’re probably going to shoot yourself in the foot. Gearboxes with plastic sideplates are few and far between as it is due to durability concerns. Now you’re adding in that the sideplates would have a built-in weakness–the layer boundaries–and that they’re not going to be all that filled in… I’d make sure to run the numbers prior to attempting a 3D print.
I highly recommend against this.
3D printing does not hold the required tolerances for work such as bevel gears, particularly plastic 3D extruders. The printers normally found in FRC are NOT suitable for work such as this, plus they have a “step” of 0.003" vertically; you will end up with a lot of slop in your gear train. You have to double or triple stack the tolerances due to the existence of side plates.
Not to mention press fits with bearings is barely possible, as a decent press fit in plastic may weaken or destroy the frame during pressing due to the layering (depending on the printer).
Plus, you still have to do heavily loaded tests to make sure that it can hold up due to the plastic and layering.
Now, some actual questions:
What is the final weight of this gearbox without motors?
What is your reasoning for going with a bevel gear setup?
What advantages does this design hold over 192’s gearbox design from 2014? I still haven’t seen a shifting gearbox design that beats theirs in terms of weight or size.
20fps is too fast. You can limit the top speed in software for the driver, but your acceleration will be very poor with only 4 cims. Seeing as you are running bevel gears, would it be possible to add a 3rd cim sticking stright up?
Your final gear reduction looks like it can be reduced to a much smaller/better ratio. Top speed for a 4 cim drive shouldn’t hit above 17-18fps if you want to optimize distance/time. OC it depends on the game, but it would be a very rare game that requires 20fps on four cims.
I like the idea of 3-D printing the plates. I don’t know why everybody is so against it, last year we (1885) used 3D printed parts to join an axle for our intake, it worked very well and only “broke” when we didn’t check it and a shaft collar was loose. Then the axle would fall off and we would have tighten the collar up again.
Just use a high quality plastic and use a dense pattern while printing AND MAKE EXTRA’S!!! Better to get into 3D printing now while it is still being developed heavily, we all know it will only become more and more common, so this is a great foundation to build off of in the future.
If they were using metal 3D printing, I think most people wouldn’t bat an eye. The problem is that those machines are generally 6-7 figures, not counting material, and most schools won’t go for that.
Now, note this, I’m not against 3D printing, or even necessarily against 3D printing gearbox plates. I’m against 3D printing gearbox plates WITHOUT accounting for all the stresses and loads in the design, and “just because it’s the next thing”. If you think the plates should be 3D printed, please, by all means, print a set of plates for a given gearbox–just make sure to document the process and the failure.
A coupler sees very different stresses from a gearbox sideplate. I’ve seen 'em used in very high-load applications when 3D printed, but as I recall that particular coupler broke about a month after it entered service, due to a combination of factors that were not necessarily directly related to the loads it was seeing, and was replaced by a machined part.
Now, in case you need help reading a block of text, here’s a list of reasons not to print the plates:
3D printing can’t hold as tight of a tolerance as machining, and tight tolerances are REQUIRED in gearboxes.
The printer prints in layers. If you orient them right, you can get away with the built-in failure modes this creates. Otherwise, ouch.
Press-fits can damage the plates. What exactly is press-fit? Only about every bearing in the gearbox!
If extra spacers will cause an issue, you don’t want to print, because the screws to hold the gearbox together will need washers, which will act as spacers.
tl;dr: I wouldn’t 3D print a gearbox plate unless I had specifically designed the gearbox for 3D printed plates, and specifically designed the plates to be 3D printed. This gearbox is, as far as I know, NOT designed for 3D printing. Therefore, it is a very BAD idea to 3D print this gearbox.
What you have described to join an axle isn’t the same as gearbox plates.
Gears (especially bevel gears) require very accurate positioning. If the gears are to close together, or the shafts aren’t parallel, the gears wear much, much faster. Team 254 made some delrin side plates a long while ago and one of their mentors said that the flexibility of the plastic didn’t keep the gears well aligned. The thrust loads from the bevel gears will only worsen the problem of the deforming side plates. Also, you can’t really get a press fit for the bearing with a 3d printer. First of all, the printer isn’t accurate enough to get the right size hole, and secondly, the plastic has a high chance of cracking if you try to press in a bearing.
The side plates can be made on a mill and a bandsaw. The outer profile doesn’t need to be accurate, so you could just print and cut out out a correctly scaled drawing, and trace it with sharpie on a piece of 0.25" thick aluminum to cut on the bandsaw. The mounting holes/bearing bores should be fairly easy to do on a mill.
Personally, I wouldn’t go with the ratios you have chosen for your shifter. I’d recommend using smaller wheels, or adding an extra stage of chain reduction first. 9 feet per second with 4 CIMs in low gear is on the fast side, and you’ll experience breaker tripping in low gear. Also, 20 feet per second with 4 CIMs will have very poor acceleration, and will really wear down your batteries quickly.
Using 3D printed plates instead of metal is somewhat like using zip tie instead of steel wire. Zip tie is nice, you can quickly fasten most things and reasonably strong, but if some applications will have so much dynamic forces they will just get ripped.
Using 3D parts for lighter loads as in to feed a ball is different than gear box or any structural parts in a drive system.
May be just for prototyping and quick test it may be ok, but for competition robot… its not just worth the risk.
I like the gearbox-- it looks like some nice work. I’ll add my voice to the people that would recommend against printing this for actual use on a robot-- I think the others have done a pretty good job of explaining why that’s a bad idea. I can say firmly from experience that it isn’t nearly as good an idea as it first seems.
That being said, printing certainly has its place in the design of an all-metal gearbox if you want it to. Last year I had a decent amount of success doing a high shell count and drilling/reaming essential holes, and it’s always nice to test the geometry in practice as well. With a little bit of adjustment and a small amount of work, you could probably do a manually powered version to print and test before you made a metal version. I’m not entirely sure how beneficial that would be, but it would certainly be cool/interesting to show off.
Alternatively, you mention that you’re getting an inhouse mill in your first post. Do pre-season testing on a variant without all the fancy lightening and contouring, and then use your external resources and funds to do the full, lightened version for the season.
This is an excellent point to consider. In my opinion, a custom gearbox isn’t worth the trouble unless you’re sure it’ll work, you’re sure you can make it, and it’ll be superior to an off the shelf gearbox in some way.
This gearbox likely won’t be cheaper than an off the shelf wcp ds or ball shifter gearbox, but it doesn’t hang low to the ground, and I’d bet you could direct drive a 3.25" wheel with the gearbox, something that can’t be said about most off-the-shelf gearboxes. Its unique shape also frees up the space where motors would normally go in a typical configuration
If you think that this is the best way to improve your team’s robot (this totally depends on your team’s unique situation), then it’s a great idea to try to build one in the offseason first.
+1000. We are getting sidetracked. Any discussions about 3D printing might be kept to another thread. OP should do the research and decide that way if the teams wants to print.
Anyway, my post above asked a few questions about this gearbox.
EDIT:
The questions from ym last post:
What is the final weight of this gearbox without motors?
What is your reasoning for going with a bevel gear setup?
What advantages does this design hold over 192’s gearbox design from 2014? I still haven’t seen a shifting gearbox design that beats theirs in terms of weight or size.
20fps is too fast. You can limit the top speed in software for the driver, but your acceleration will be very poor with only 4 cims. Seeing as you are running bevel gears, would it be possible to add a 3rd cim sticking stright up?
Your final gear reduction looks like it can be reduced to a much smaller/better ratio. Top speed for a 4 cim drive shouldn’t hit above 17-18fps if you want to optimize distance/time. OC it depends on the game, but it would be a very rare game that requires 20fps on four cims.
And,
How you’re dealing with the thrust loads from the bevel gear and pinion? If they gear isn’t supported with thrust bearings you’ll probably destroy the radial bearings you’re using.
The cim shafts can move a bit in and out, do you have a way of making sure the pinion is in the correct place?
I wouldn’t say it was anything specific more than the combination of many smaller issues compounded. We had a couple of Replicator 2’s that we were just starting to get familiar with, and there wasn’t a ton of experience designing for printing on the team then. We printed in the wrong orientation, broke a couple of plates trying to fit in gears, and once we had it all together it made sounds I hadn’t heard come out of a gearbox before or since. I think it probably could have worked with some more effort (and experience), but when there’s so many nice COTS options out there it just didn’t seem to be worth the effort. In retrospect, I wouldn’t necessarily say it was a bad idea, just that it wasn’t nearly as good of an idea as we thought it was at the time.
We learned a ton about designing for printing from that gearbox though.
I’m not trying to denigrate what 192 did last year, because they made a gorgeous, highly functional gearbox, but it should also be asked “what advantage does 192’s gearbox design hold over a COTS or modified COTS solution?”
Moving the motors out of the way isn’t a good enough reason for most teams, IMO. What is all that extra work and potential compromise of reliability really buying you? An extra 8" in the interior of your robot that you probably don’t really need anyways?
192 had the benefit of doing something similar (with worm gears) to OP’s design in 2012. They didn’t do it again after that. They have at least 4 revs of their 2014 gearbox (as I recall they made 2 prototypes in the 2012 offseason, plus the 2013 gearbox, then the 2014 gearbox).
There are so many better obstacles for most teams to tackle than making custom gearboxes. If 254 were starting a new team right now, I highly doubt we would make custom gearboxes. Maybe custom sideplates to get the right ratio, but that’s about it. The stuff that’s out there now is so high quality that if you have any question about your ability to solve every other aspect of the game challenge, you really shouldn’t be going custom.
I could vary well be biased, but tend to think that building something unique and ambitious in the off season is almost always a good idea.
Some reasons:
Building something unique tends to get students excited about off season work.
Building something ambitious will force to team to expand their resources.
The experience wI’ll improve students’ CAD and machining skills.
There is very little risk of failure.
