Looking for someone who can replicate a complex gear for a toy from the 1980's

avo_jeknavorian

Hello! I have a plastic gear that I am trying to have made so I can print it using my 3d printer. This is a complex gear with various windows and posts.
It's for a toy from the 1980's called the Armatron by Radio Shack / Tandy. It's a robotic arm. I enjoy tinkering and restoring these things, and this gear is something I'd love to have as a file to print.
Is there anyone in the Boston area that would like to attempt this? I have 1 intact gear.

Thank you!
Avo

nick_papageorge073

Is it a spur gear? There is a spur gear featurscript that can help a lot.

Injection molded gears are quite specialized, and you sometimes need the gear pair to replicate the one gear, due to specialized tooth modification made per gear pair. All that said, for a fun time fixing an old toy, you can probably ignore those modifications.

dirk_van_der_vaart

https://www.stlfinder.com/3dmodels/radio-shack-armatron-robotic-arm/

avo_jeknavorian

Thanks for the tips. I will keep researching! It’s not the gear on stlfinder. This one is very complex.

EvanReese

Can you post an image of the gear in question?

dirk_van_der_vaart

This is useful?
https://www.scribd.com/document/401753119/The-Armatron#

avo_jeknavorian

Here are a bunch of pics. It’s a complex gear. It has posts, holes, and windows for other gear access. You’ll see a pic of it installed on the main arm.mplex heg

S1mon

Depending on the 3D printer technology used, printing a gear like this will only get you so far. Even if you have the exact CAD, the tolerances and surface finish of 3D prints are not great for gears, especially at smaller scales like this. You can probably make something that will work for a while, but it will be noisier, and have more backlash than the original. If you can get it CNC’d in delrin (POM) that would be better.

If I had to do this, I would find someone with an optical measuring machine or a CMM to help get the locations of the centers. If you can’t arrange that, I would set up a good camera with a long lens (to reduce distortion) and take a very careful orthogonal view and use that as a picture reference.

EvanReese

Oh cool. I agree with Simon that CMM would be the most accurate way to do it, but I suspect the photo reference, plus careful caliper work, and perhaps print iteration would be plenty to get it functional, and I expect a resin printer like formlabs (there are a lot of other hobby-grade ones too) could handle the detail well enough. Longevity of that print is more of a mystery to me. You could also print a master, to make a silicone mold and cast them out of resins to make copies and open your material library up.

If I were retired I'd take this on for you, but you probably don't want to wait that long.

avo_jeknavorian

Here are a bunch of pics. It’s a complex gear. It has posts, holes, and windows for other gear access. You’ll see a pic of it installed on the main arm.mplex heg

avo_jeknavorian

Thank you for all of the great advice and comments!

nick_papageorge073

Nice gear. It doesn't look too hard to model fundamentally, but like the other's said, if the center to center distances of the internal shafts aren't right, it won't work. Plus, the material if 3D printed will be all wrong. That gear is either POM or Nylon. 3D printed FDM will mesh like crap, and it also usually prints oversize and will be tight. I used to design gears for transmissions for infant swings that rocked babies to sleep. We could never prototype the gears well. We had the best RP machines, Objet, SLS, SLA, none of them could make a gear worth anything. Even machined POM was difficult if it was a compound gear, and also very expensive.

I think your best bet is to see if you can find the same used toy on ebay, and steal the gear out of it.

Or if you want to do it yourself to see if it can work good enough, first you have to find the module of the gear. If the toy was made in the 80's, its probably from Hong Kong or ROC. PROC had just opened their borders for export a few years prior, so might not be made there yet. So its probably metric module gears. Unless, Radio Shack designed it in the USA, and only had it manufactured in HK or ROC but to their design. In that case, it might be imperial based. You can figure that out with gear charts, calipers, and counting the teeth. Then plug that into the onshape spur gear featurescript. Print that out and adjust until it matches your gear exactly. Pressure angle is most likely 20 deg or 14 deg. I don't see any profile shifting by eye, so the featurscript should build you the right shape of the teeth. Then, build the rest of the gear with the bosses for the shafts of the little gears. I'm sure that will take a lot of iterations of measuring, printing, testing, and repeating, to get to fit and mesh right. Especially if all you have to measure it are pin gauges and calipers.

If you pay someone to do it, I think you are looking at 1000 USD minimum, and maybe even no-quotes. Its more if you can find someone to do it as a favor to you for free that has equipment they can measure the gear with, and then model it. An optical comparator can probably also get you the boss locations.

nick_papageorge073

I like Evan's suggestion of a silicon mold. You can use your intact gear to build the mold. Then you would not have to model anything in CAD. The molded parts usually come out right on the money without shrinking. You can get about 10 parts from a mold. The part would be a lot stronger, and smoother, than FDM printing.

Still a complicated thing to do though, and it will help finding someone with experience building silicon molds, with pins to use for boss ID diameters.

The male pattern part usually survives the mold building process, but there is a chance it could be scrapped.

roman_jurt190

Maybe ask the guys at MITs Center for Bits and Atoms if you can use their Artec 3D-Scanner (0.5mm precision)...  with the scan imported to onshape.. the reverseengineering would be fairly easy... 
http://cba.mit.edu/tools/index.html
http://cba.mit.edu/tools/index.html#:~:text=Cutting metal stock-,Instrumentation,-Artec Leo

an even higher-res scanner would probaby be better...
from my experience with a hi-res 3D-Scanner (Einscan Pro 2x, 0.04mm res.) it would take not more than 20 min...