3D Printed derailleur.

Derailleur arm*

So, about a month ago, my Norco XFR3 decide to take it to the top rope, salute the crowd, and drop an elbow on my roommate's Giant. What really happened was a 3D printed bike hanger broke (that's what I get for being cheap with filament) and dumped the bike on the ground from about 5 feet.

The damage was a bent derailleur arm and a bent rear wheel. Lovely. Remind me to hang the bike facing the other way next time.

The wheel can maybe be repaired for about $20 and, if not, the LBS has 700c's for about $100. Ouch but not the end of the world for a new wheel.

What I CANNOT abide is a $34 dollar derailleur hanger. $34 dollars for a couple hundred grams of Chinesium with some threads cut into it. So, I pulled the hanger and started trying to straighten it as best I could. A little work on the vice and a little bending and pulling on the bike and it was shifting decently. I'm fairly certain any vibration is now coming from the bent wheel but we'll (heh) have to wait and see. 

While I was doing all this, I couldn't help but think of how well these hangers COULD do as 3D printed parts. They're not the hardest metal in the world and there's probably some thermoplastics out there that are harder. Those plastics are probably brittle enough to still break away and prevent major damage to the derailleur, though, should they need to. And the hanger is tiny. Probably less than a dollar to print one out of solid PLA (not advisable) and not much more to print one from something like PETG (probably better). 

A quick YT search netted this:

A pretty compelling prototype, if you ask me. And this bike is a commuter hybrid, not a mountain bike. 

I copied the gentleman's method and I'm here as of now:

Plan is to make a very lightweight prototype, check fitment, tweak, then reinforce in a similar way to how our friend in the video did. Once I've got something that works, I'll mount it, check operation, and test ride. If it survives my normal commute at least once, I'll print a few more and throw them in my roadside repair kit. If it works better than the bent hanger, I'll use one as a permanent replacement (knock on wood). 

Only real concern is I ride around campus like a horse's ass and it's hard on my bike. I don't see that stressing the plastic in such a way that it would break but still. Should be fun to see how it goes.  

Tasty.

Still needs threads but I'm pretty happy with that. I'll do the milled portions tomorrow and then print. 

This is a cool idea.

1 point though...  the hangers job is to protect the frame, not the derailleur.  In the old days, you could break off the hanger from the frame and have a ruined frame.

In reply to wvumtnbkr :

Ah, I see. I can see how they would be the case, looking at it again. Why in the world are they so expensive? I've seen some out there that are forged but even then. Parts shortage? 

Small run machining costs more than you think. $34 for a part like that when the vendor and machinist both want to stay in business? Plus some anodizing? Quite plausible. There is more cost in a part than just the value of the raw material. And yes, bikes and parts are in short supply right now.

The 3D printed one is interesting. It has to be stiff to keep precise shifting but able to bend/break sacrificially in case of an impact. Could make for some interesting design work. 

I straightened so many derailleur hangers at my first job in a bike shop. We had the special Park Tools thing that screwed right into the hanger that made it easy. 

In reply to Keith Tanner :

The small parts run makes more sense. I'm guessing they make about a billion and that lasts them until the end of the model's life out in the market/ aftermarket. 

In reply to thatsnowinnebago :

I have that special tool as well. The hangers on my Rocky Mountain are made of cheese. I can tell when it's bent by the shift behavior. 

Printing time. I'm still novice enough on SW to not know when it's acting weird and boy, did that extruded cut take forever to do. 

Anyways.

I'm thinking this guy is gonna be a little off. The drawing I used to trace it wasn't the highest resolution. I might just trace the one I have, upload it, and use it as a template if this one isn't jiving with reality. 

Here's the original for reference:

That's the worst pair of calipers in the world, by the way. Spend more than $20 on your calipers, guys.

Print in wax or PLA and cast yours part out of aluminum. It works a treat. 

It you have the ability to print carbon fiber impregnated nylon it works a treat. We are using some of it in the lab but have not finished the strength tests yet.  

https://www.makerbot.com/3d-printers/materials/method-nylon-carbon-fiber/

I thought about doing a possible casting at some point. If I can get it to  fit really well, I may do one just to try casting.

I really like carbon-impregnated PLA. There's some kind that can be baked after printing to increased hardness quite a bit. I've used it before and it's surprisingly strong. Even cut threads into it for a shift knob.

Do ya'll vent your printers that use nylon? I'm super interested in it but I want to be careful about printing it in the living room. If it needs an enclosure, I'd build one but I don't have one now. 

Also, we're off:

We use carbon/nylon at work on our Makeforges. Wonderful material. 

I have a method X at work and we keep ours in a chemical hood just to be safe. Hood and instillation was like 10x the cost of the printer but no chemists around to use it so I sneaked my printer in. 

It's not too expensive or difficult to build a hood at home. There's about a trillion takes on making one out of an IKEA end table. 

I HATE using PETG, I should probably just step up and build a hood. Nylon seems to be the only other easily accessible/easy to work with material that gives good strength and temp resistance. If we exclude carbon PLA, which we should, cause that stuff is a pain to work with and it's damn expensive.  

Maybe a dumb question, but can you print the threads into your hanger? Or do you need to tap the hole after it's printed?

In reply to thatsnowinnebago :

You can do either but the material you print with and the program you use the CAD the threads are going to determine how successful you are. Just like any metal or any other plastic, the threads are only as strong as what they're cut into. A steel bolt in PLA threads can pretty easily wreck the threads with not a lot of torque. I've cut threads into carbon impregnated PLA with good success. I'm CAD-ing threads into this guy, we'll see how it goes. I would prefer to cut them with a tap or die, I think, since you're not as behold to the drafting software or the printer to get good threads. You just have to make sure the hole you draw is small enough to cut good threads in to. 

All that said, there's some pretty hard filaments out there and I've seen both methods work. Even solid PLA will do a decent job at holding a thread. I'd trust it with this job if PLA weren't so bad in the sun.  

This is take number 2 for the bike stand. The center piece that holds the pedal is what broke last time. It's printed at 100% infill this time, so.........stand under your arch.

I actually got the bike shifting decently yesterday. That's a new rear wheel; good luck finding anything in 700c, disc brakes, quick release these days. All the local bike shops had next to nothing in stock. I got the rear hanger straight enough to tune the derailleur okay-ish. 

The derailleur is printing as I type this. I'll throw it on there and see how well I can get it to shift. If it's smoother and more reliable, I'll keep it on there and keep the metal one as a spare. Then to get the brakes to stop squeaking. 

Turbo_Rev said:

In reply to thatsnowinnebago :

You can do either but the material you print with and the program you use the CAD the threads are going to determine how successful you are. Just like any metal or any other plastic, the threads are only as strong as what they're cut into. A steel bolt in PLA threads can pretty easily wreck the threads with not a lot of torque. I've cut threads into carbon impregnated PLA with good success. I'm CAD-ing threads into this guy, we'll see how it goes. I would prefer to cut them with a tap or die, I think, since you're not as behold to the drafting software or the printer to get good threads. You just have to make sure the hole you draw is small enough to cut good threads in to. 

All that said, there's some pretty hard filaments out there and I've seen both methods work. Even solid PLA will do a decent job at holding a thread. I'd trust it with this job if PLA weren't so bad in the sun.  

If I think printed threads are appropriate I model them by hand. It takes me about 3 minutes and most of that time is looking up the thread profile in the machinery handbook.

In reply to APEowner :

Any experience with SolidWorks's threads? I just went with their auto-generated threads and was wondering how it would go.

Turbo_Rev said:

In reply to APEowner :

Any experience with SolidWorks's threads? I just went with their auto-generated threads and was wondering how it would go.

That is all I use. 90% of the time the metric option cover me and I have a part number for some of the more common bolts for assemblies. 

Cool beans. Hopefully I won't have to finagle too much with the threads, then. Worst case, I'm going to print blank holes and cut threads but I'd rather avoid it.

Look at this awful crap:

WAY too big. Not sure what happened with the scaling but it is not correct. Weirdly the 10mm hole for the derailleur was the right size. The smaller holes were too small.

I cut threads into the small holes and they held the frame hardware just fine. The larger hole's threads held the derailleur's fastener just as well, with plenty of torque.

I think the photo I used, shot in glorious 144p, is to blame. It might not even be the right derailleur. Or maybe it had some weird resolution? 

Anyways, I traced the derailleur I have and scanned it. I'm going to check units, trace another one, print, and fit. Stay tuned.  

P.S. - got this other project put together. Anyone know ROS?

All right, so...

Traced the hanger on a sheet of paper and scanned it. Got a much better result.

That doesn't LOOK too close and that's because it isn't. But it's lightspeed closer than that first print, jeez loweez. The bottom contour of this hanger actually fits (mostly) correctly in the frame.

Everything needed to be shifted towards the front of the bike and down about 2mm. You can see where the axle crosses through the frame and hanger has a big ol' lip getting in the way. Who knows how that happened, probably and little fudging with the paper trace that got carried forward. One of the smaller mounting holes was dead on, so I just made some tweaks in SW and printed another:

 Voila! This one actually mounts to the frame correctly and the arch where the axle passes through is flush with the frame. I also managed to sort out functional threads with this one (the hole on the last one was bigger than the max diameter for M10x1.0 and the derailleur mounting hardware just passed through it). You can see the obvious gap between the frame and the cut in the hanger, which will hopefully be the last tweak before I print at 100% infill. I don't *think* that the hanger needs to touch the frame for any kind of support (the factory one doesn't) but what the heck; nature abhors a gap. 

With the second hanger, the bike is pedaling and shifting okay-ish. Worse than with the metal arm but I'm pretty sure that just tells me the obvious: arm was bent. Also, the brake disk that was rubbing yesterday is rubbing again; so, pretty sure everything shifted a little with the straight arm. All the disk needs is for the caliper to be moved over a little, so no biggy. One more tweak, fit, print a final at 100%, install, and tune shifting.  

P.S.- each prototype takes about 25 minutes to print at 5% infill, so that's nice.

If you do a drawing of your model and print it on paper 1:1 you can check fits in one dimension very quickly.

In reply to APEowner :

I'll have to use that next time. I think it was a combination of the pencil tracing under the bevel of the original metal part and then a little more error generated while trying to trace in SW over some fairly thick pencil lead.