3D Printed Parts on an FRC Robot?

[Why_A_Username?]

Hey CD,

when I first started on my team, one of the many things that fascinated me was 3D Printing. The concept of additive manufacturing opens up a world of new possibilities for designing and creating intricate parts, but are they useful in FRC?

I would argue yes (though that may be somewhat biased, as I own a 3D Printer), but I've heard many argue against it. my reasons for why it's useful are as follows (In Relation to Fused-Deposition Modeling only):

1. In recent years, material choice has expanded hugely, ballooning from just PLA and ABS, to HTPLA, Carbon Fibre, Flexible materials, HIPS, Nylon, PETG, metal fill... etc.

2. People often assume that 3D printed parts are too weak or too brittle. largely this is a myth sprouting from basic PLAs which are extremely brittle. Nylon, ABS, and PETG are all viable alternatives which are much stronger and more shock-resistant.

3. 3D Printed parts offer shorter wait times than ordering parts from online retailers

4. printed parts are easily replaced when broken because they can be inexpensively replaced, and custom parts can be made easily, often in a shorter time period than if they were manufactured conventionally.

5. they can be used as quick and dirty spare parts in the event that an ordered part fails. (Team 2910 had several Rhino Track Pulleys printed out in case we had parts break during competition)

Obviously, I'm not advocating for printing things like motor mounts and chassis components, which are subject to incredibly high stresses during the game. But in the case of things such as custom sensor mounts, passive gears , and other relatively low stress positions, like turret ring-gears, printed parts can often be the way to go.

I'd love to hear your thoughts on printing parts, both for and against. My biggest issue on the subject is the 3D printing Phobia I see in many FIRST teams. I believe it has its uses and, when an option, should be exploited as a cheap and fast method of production.

[guniv]

Team 3824 in our region prints their entire chassis. Oak Ridge National Laboratories has been sort of a pioneer in large scale additive manufacturing, and 3824 gets to work out of their facility with a lot of mentor support. You can read about them in the book that came with kickoff kits last year (Behind the Design). Here's a video of them printing in 2014: https://www.youtube.com/watch?v=dvLxsRnA0Y0

[MailmanDelivers]

We [The LigerBots] love 3D Printed Parts! I've attached a document that describes most of our 3D Printed Parts from last year.

The file looks a little odd when opened it word as it is made from google docs.

If you have any questions I'd be happy to answer. :D

[BrendanB]

Spacers are a good part to use a printer for. Its nice for some of the more arbitrary sizes you need based on CAD files to just hit print and have a ton made consistently.

This year we printed a ton of 1/2" hex spacers in 1/8th increments. We had a box full and it made last minute additions and testing pain free.

We also used them to save some money in our drivebase. Instead of spending almost $100 on sprocket spacers we just printed them to cut costs.

[lorykzarr]

This year, we were having trouble getting hubs to fit our pneumatic tires. Our solution was to modify the AndyMark hub to fit our tires as a center piece then use 2 milled aluminum plates as beadlock rings to sandwich the tire and inner tube together. We printed several full sets as backups but at the competition we attended didn't need to change a single one. Even now, after doing numerous demonstrations they've held up very well. In addition to our wheels, we also printed some pieces to join tubing on our superstructure. These proved to be fairly strong but were reinforced with some gussets.

[Cothron Theiss]

I think more FRC teams are using additive manufacturing on their robots than you realize. Search around the forums here, and talk to all the teams when you go to competitions. I think you'll be pleasantly surprised.


Team 3824 in our region prints their entire chassis. Oak Ridge National Laboratories has been sort of a pioneer in large scale additive manufacturing, and 3824 gets to work out of their facility with a lot of mentor support. You can read about them in the book that came with kickoff kits last year (Behind the Design). Here's a video of them printing in 2014: https://www.youtube.com/watch?v=dvLxsRnA0Y0

Case in point. From 2012-2015, 3824 printed just about every component of their robots, especially the structural components. Even before ORNL developed their Big Area Additive Machine, 3824 used a Fortus printer to print their robots. Many of the surrounding Knoxville teams use those printers as well.

[ollien]

We [The LigerBots] love 3D Printed Parts! I've attached a document that describes most of our 3D Printed Parts from last year.

The file looks a little odd when opened it word as it is made from google docs.

If you have any questions I'd be happy to answer. :D

Would you mind posting the CAD for those Talon brackets? They sound awesome.

[tpvagnini]

3D printing definitely has it's place in frc and in the build season, but it shouldn't be used for everything and it's very easy to get carried away and 3d print too much.

My team uses professional Stratasys 3d printers, a uPrint, objet 30, and a new Fortus for this year. We print the majority of our parts in ABS with about 90% infill.

Like people have mentioned before, 3d printers are great for spacers, ball/object guide rails, and even lightweight mounting brackets.

During the build season it's fine to prototype anything with the 3d printer, but from experience, any criticals part should be replaced with stronger parts before competition.

For any 3d printed parts that make it to competition, make sure you 3d printed them properly, (correct orientation, infill, material) and made plenty of extras.

[cbale2000]

The only 3D Printed part we ever had an issue with was a PLA "spindle" we made for the elastic bands on the shooter of our 2014 robot (it prevented the bands from rubbing against the sharp 90 degree corners of the 1"x1" aluminum tube they attached to).

Needless to say our shooter was WAY over-tensioned for what it needed to be, so the part kept being crushed under the force of the elastic bands. After replacing it twice we finally just printed one at 100% infill and never had another problem with it (we had one of the free Cube printers, so our only options were basically: "no fill", "some fill", and "ALL THE FILL").


As long as the part itself is not too small for the application, when in doubt, add more infill. :rolleyes:

[The Ginger]

3840 in MN has used 3D printed parts since 2014 with great success. this year they made a incredible tank suspension system with all drive and running wheels printed. I would love to see more teams, including my own, utilizing this great technology.

[Bryce Clegg]

3196 uses 3D printed parts for prototyping, but when it comes to our robot competing we use ordered on milled parts. Since 3D printing with material used in FRC is becoming more available, I think we'll definitely see an increase in the future.

[Michael Hill]

I've been toying around with the idea of using 3D printed turret parts. I generated a large (either HTD or GT2) sprocket, put a hole in it and split it into 6 sections. Those six sections get 3D printed and then bolted onto aluminum.

Check out screenshots here (obviously I need a gearbox on the lower section to drive the belt, but that wasn't part of this exercise):
http://imgur.com/a/0ZGUS

I had one of the sections printed up by shapeways with their white, strong & flexible material (the section isn't flexible at all, they only call it that for thin parts). It's really just a nylon I believe. We had 3mm GT3 belt at our shop that seemed to mesh fine with the section, though I am a bit concerned that the printed part seems to be very slightly undersized, so if I stuck 6 of them together, I'm not a fan of how those errors are going to stack up and possibly ratchet my belt.

[Cothron Theiss]

I've been toying around with the idea of using 3D printed turret parts. I generated a large (either HTD or GT2) sprocket, put a hole in it and split it into 6 sections. Those six sections get 3D printed and then bolted onto aluminum.

That looks really interesting! Does it have to be 6 sections though? I feel like the fewer sections you can make it out of, the tighter your tolerances will be. Also, you might want to add in a different way of aligning and piloting your sprocket sections. It looks like you're just using the bolt holes, but those will not hold very tight tolerances, from the printer and to the actual fit. Maybe design in a boss or something similar to the Versakey pattern, or something where the edges of the sections meet so that they fit like puzzle pieces. Make it tight enough so it's a light press fit on an arbor press, and they'll align much more true to your design.

[Michael Hill]

That looks really interesting! Does it have to be 6 sections though? I feel like the fewer sections you can make it out of, the tighter your tolerances will be. Also, you might want to add in a different way of aligning and piloting your sprocket sections. It looks like you're just using the bolt holes, but those will not hold very tight tolerances, from the printer and to the actual fit. Maybe design in a boss or something similar to the Versakey pattern, or something where the edges of the sections meet so that they fit like puzzle pieces. Make it tight enough so it's a light press fit on an arbor press, and they'll align much more true to your design.

I used 6 sections to keep the physical size of the part small so it would be more economical to print. I'm not sure if making larger sections would make the tolerance situation any better (or worse). The overall size seemed to be around 98% of what the actual size should be, and with their printing process, dimensional stability isn't an exact science. The overall sizes can change due to temperature, humidity and other factors. But I would like to get 6 of them printed to see if it really is an issue though.

I'all have to check on adding a boss. The aluminum would get waterjet anyway, so adding holes in there to fit a boss wouldn't be too hard.

[Kevin Kolodziej]

FRC1675 has used 3d printed parts for 4 years now. Most of our parts have been done on a Stratasys Dimension 1200 but we have also had limited success with the the Ekocycle printers that were given out a couple of years ago.

Most of our parts are small, including spacers, sensor mounts, camera holders, etc. One innovative solution mated with a versablock to mount an encoder.

In 2014, we printed quite a few versablocks for quick prototyping. We did not use any printed versablocks on the competition robot, but I don't see any problem with it for certain applications.

We have also done a LOT of pulleys and sprockets for HTD belt and Vex EDR chain for intakes and sensors. We have used a 2-part pulley (ranging from 12 to 30 teeth) to eliminate the need for overhang printing on the flanges that fit together using steel dowel pins.

In 2013, we printed the plates for our custom drive gearboxes. They weren't very optimized, but it was our only way (at the time) to create high tolerance bearing holes. They were two stage reductions from Mini CIMs into 6" wheels with aluminum standoffs between the plates. Had zero issues with those over the course of the season.

[nuclearnerd]

This may not interest you, because it requires a specialized machine, but our local community college (Mohawk (http://www.mohawkcollege.ca/about-mohawk/initiatives/applied-research/additive-manufacturing-resource-centre-amrc)) printed our intake rollers from nylon powder using the "Selective Laser Sintering" process (SLS). We were worried these wouldn't perform, but after a season of smashing the intake rollers against defense dividers, I can report that the rollers were tougher than the hex shaft that ran through them. An FDM printed roller would have cracked or crushed in this application for sure. We were very pleased, and hope to work with Mohawk again next year to print more SLS parts.

[Chris is me]

3D printed parts are absolutely valid in FRC, it's just a matter of knowing when and where to use them, and how to design and manufacture them such that they do not see excessive loading.

Printing is particularly useful in FRC:

- When parts don't see significant load. 3D printed parts can be used with load as well, but that requires more active effort or reinforcement.

- When you need a specific geometry of part that is a pain to make through other methods. Things like electronics brackets, camera mounts, eccentric spacers, gearbox shields, even internal ball cradles can be printed to easily achieve a specific size and shape.

- When your other manufacturing resources are tied up, or needed for more valuable things. Spacers and standoffs are a common 3D printed part, not because the technology is required for it, but because the 3D printer has more downtime than the lathe and these parts are just boring to make by hand.

- When you are stuck in CAD at 2 AM trying to interface two internal assemblies and need to bridge some oddly specific gap between two parts. Personal experience on this one.

Honestly, if you get and have a printer, you'll find ways to use it. You can even make composites out of it, using the 3d printed geometry for shape and adding metal components for structure. It's just too handy.

[AdamHeard]

I've been toying around with the idea of using 3D printed turret parts. I generated a large (either HTD or GT2) sprocket, put a hole in it and split it into 6 sections. Those six sections get 3D printed and then bolted onto aluminum.

Check out screenshots here (obviously I need a gearbox on the lower section to drive the belt, but that wasn't part of this exercise):
http://imgur.com/a/0ZGUS

I had one of the sections printed up by shapeways with their white, strong & flexible material (the section isn't flexible at all, they only call it that for thin parts). It's really just a nylon I believe. We had 3mm GT3 belt at our shop that seemed to mesh fine with the section, though I am a bit concerned that the printed part seems to be very slightly undersized, so if I stuck 6 of them together, I'm not a fan of how those errors are going to stack up and possibly ratchet my belt.

What machine did you use to cut the two (presumably) aluminum rings?

If you lasered or waterjetted the plate, you can cut gearteeth directly into it and skip the printing portion.

[Brandon Holley]

This may not interest you, because it requires a specialized machine, but our local community college (Mohawk (http://www.mohawkcollege.ca/about-mohawk/initiatives/applied-research/additive-manufacturing-resource-centre-amrc)) printed our intake rollers from nylon powder using the "Selective Laser Sintering" process (SLS). We were worried these wouldn't perform, but after a season of smashing the intake rollers against defense dividers, I can report that the rollers were tougher than the hex shaft that ran through them. An FDM printed roller would have cracked or crushed in this application for sure. We were very pleased, and hope to work with Mohawk again next year to print more SLS parts.

SLS is definitely a really cool technology that makes great parts. Unfortunately, most teams will not have access to SLS parts as the technology is considerably more expensive than an FDM printer. FDM parts are plenty robust enough for that application you described.

We used FDM to print our intake roller, among other parts this year and had no issues. We actually played all of our ~160 matches with the same roller this year, so it took a beating, but held up considerably well.

http://i.imgur.com/ZtDuxAnl.jpg

As with everything, its about using the technology and material smartly. We printed the 'shell' of the part we wanted because of its complex shape. FDM allowed us to make this helical screw type of intake that would be unachievable otherwise. We liked this because it was a lot less complex and a lot lighter than a row of mecanums. It also would also be way more maintenance friendly than adhering a strip of something to a cylinder.

We took our shell and epoxied it to a very thin wall aluminum pipe, and printed some endcaps to plug into the ends, complete with thunder hex pattern.

The other parts you see in the below picture, particularly the drums are our hanging drum/winch for our robot. We used these drums to wind the winch up and scale the castle. Easiest winch drums we ever made!

http://i.imgur.com/sSm33A1l.jpg

-Brando

[Cothron Theiss]

An FDM printed roller would have cracked or crushed in this application for sure.

Are (https://www.sculpteo.com/blog/2014/05/14/material-considerations-choose-right-plastic-production-method-part-2/) you (https://www.stratasysdirect.com/blog/sls-vs-fdm-3d-printing/) sure (https://all3dp.com/fdm-vs-sla/)?

[Sam Skoglund]

In 2013, we printed the plates for our custom drive gearboxes. They weren't very optimized, but it was our only way (at the time) to create high tolerance bearing holes. They were two stage reductions from Mini CIMs into 6" wheels with aluminum standoffs between the plates. Had zero issues with those over the course of the season.

I have never seen 3D parts used for this before! A couple of questions: Were the 3D printed plates proving any structure for the robot frame? Or were they just mounted onto the frame? Also, how thick did you print them?

[Michael Hill]

What machine did you use to cut the two (presumably) aluminum rings?

If you lasered or waterjetted the plate, you can cut gearteeth directly into it and skip the printing portion.
First off, the aluminum rings haven't been made. I just did a 3D printing test as a proof of concept.

During the regular season, I don't know the exact model our sponsor uses, but I believe it's somewhat older machine (from what I was told, there's no head tilting for taper removal). If you can tell from a close-up picture, this is the machine (and sponsor) that does our waterjet work: http://www.rpgindustries.com/11301/11385.html

If there really is no taper reduction, I'm a little cautious about having gear teeth cut. He could do wire EDM (he's cut some gears for us in the past), but I really don't like to use up his machine time since he already loans us our build space. So we try to limit ourselves to just a single waterjet job per season, and it's usually a 1/4" plate job. Though I'm sure he'd do more for us if we asked.

The other thing 3D printing lets me do that waterjet wouldn't is adding the belt flange directly on the 3D printed part. It's not a huge deal though because it can always be added on after the fact. It's just another part I don't have to worry about making.

In the end, I've just made it sound like I need to send our sponsor an e-mail ;)

[rch328]

3D printed motor mounts on a horizontal flywheel shooter for stronghold held up fine. In my experience, problems with 3D printing models strength is usually a design flaw; though they are inherently weaker than non-3D printed parts, especially with ABS filament or stronger they are plenty strong, if well designed, for anything in FRC.

[AdamHeard]

First off, the aluminum rings haven't been made. I just did a 3D printing test as a proof of concept.

During the regular season, I don't know the exact model our sponsor uses, but I believe it's somewhat older machine (from what I was told, there's no head tilting for taper removal). If you can tell from a close-up picture, this is the machine (and sponsor) that does our waterjet work: http://www.rpgindustries.com/11301/11385.html

If there really is no taper reduction, I'm a little cautious about having gear teeth cut. He could do wire EDM (he's cut some gears for us in the past), but I really don't like to use up his machine time since he already loans us our build space. So we try to limit ourselves to just a single waterjet job per season, and it's usually a 1/4" plate job. Though I'm sure he'd do more for us if we asked.

The other thing 3D printing lets me do that waterjet wouldn't is adding the belt flange directly on the 3D printed part. It's not a huge deal though because it can always be added on after the fact. It's just another part I don't have to worry about making.

In the end, I've just made it sound like I need to send our sponsor an e-mail ;)

We've cut numerous gears (12-32DP) on a regular waterjet w/ the taper and it's worked out fine.

I'd recommend 20DP here bc vexpro product. It certainly does add waterjet runtime, but completely eliminates the printed part.

[Kevin Kolodziej]

I have never seen 3D parts used for this before! A couple of questions: Were the 3D printed plates proving any structure for the robot frame? Or were they just mounted onto the frame? Also, how thick did you print them?

I'll have to get some better pictures after Thanksgiving when we start meeting again, but no, the gearbox did not provide any structure for the frame. The output shaft ran through a bearing on the chassis and the motor was clamped to the top rail of the chassis - the plates were only holding the gears. Here is a preliminary version of the gearbox: https://www.chiefdelphi.com/media/photos/38361

They were printed solid ABS, 1/4" thick.

[Michael Hill]

We've cut numerous gears (12-32DP) on a regular waterjet w/ the taper and it's worked out fine.

I'd recommend 20DP here bc vexpro product. It certainly does add waterjet runtime, but completely eliminates the printed part.

I was a bit dissuaded by Cory's post (https://www.chiefdelphi.com/forums/showpost.php?p=1590531&postcount=18), which is why I started looking into 3D printing an alternative.

[Bluman56]

For the past couple of years, 334 has used 3D printing (Ultem specifically) extensively. In this (https://www.chiefdelphi.com/media/img/3ba/3ba05d39818d04d1552d8c6695163ac2_l.jpg) picture you can see their two front drive wheels as well as their shooter wheel/hood adjustment mechanism and turret gear all 3D printed.

[fargus111111111]

We 3D print quite a few components, mostly spacers and battery trays. In 2016 we tried printing the pulleys for our track drive and we did shatter several of them after lots of use, but we did not fill them fully and we were not nice to them. (repeated full speed over the rock wall did them no favors) We love to print with bright orange plastic so we can easily show it off. Our printed rollers on the sliding part of our 2015 arm, which supported the weight of a full stack held out past the frame of the robot, have never had any problems. That being said, I think that 3D printed parts are better used for more constant stress situations rather than high shock situations.

[Why_A_Username?]

SLS is definitely a really cool technology that makes great parts. Unfortunately, most teams will not have access to SLS parts as the technology is considerably more expensive than an FDM printer. FDM parts are plenty robust enough for that application you described.


-Brando

You guys made that roller on an FDM printer? Did you vapour smooth them? That is a super smooth and clean finish on them.

[Munchskull]

How do people model custom HTD 5 pulleys for 3d printing?

[lorykzarr]

How do people model custom HTD 5 pulleys for 3d printing?

You could try using the model of some pulley stock (with the correct tooth count, profile and pitch of course) so that you don't need to draw out the profile in a sketch. Then add whatever bolt holes or bores you need.

[Munchskull]

You could try using the model of some pulley stock (with the correct tooth count, profile and pitch of course) so that you don't need to draw out the profile in a sketch. Then add whatever bolt holes or bores you need.

Thanks, however part of my motivation is learning how pulleys are designed from the ground up.

[troy_dietz]

You guys made that roller on an FDM printer? Did you vapour smooth them? That is a super smooth and clean finish on them.

It looks like it may be a combination of a lower layer height, high quality filament, and possibly the way the picture is taken. (you could also have the layers start on the inside of the roller to prevent Z-scar from showing)

I'm not saying it isn't vapor smoothed ABS, but there are other ways to get that effect.

[Cothron Theiss]

Thanks, however part of my motivation is learning how pulleys are designed from the ground up.

Check out the Gates belts and metals design manual. Gates has lots of information about the design of timing belts and pulleys. Also, Gates refers to their pulleys as "metals." Just a little tip to navigate their site.

[cbale2000]

How do people model custom HTD 5 pulleys for 3d printing?

I can't speak for everyone else, but I do know that Autodesk Inventor has a tool that allows you to generate standard sprockets/gears/pulleys of various sizes and tooth counts.