CoreBoxx Rewinder Parametric TPU Tread Idler & Rim Rollers

BACKGROUND:   (OR WHY DOES THIS EXIST?)

There's been a whole lot of interest in the Filamentalist derived rewinders for Voxel3D's CoreBoxx.

One issue with the original parts release and design is the use of rubber bands around the “tread” part of the roller wheels.   Unfortunately the  BluRolls kit supplied rubber bands don't cover the entire surface of the “tread” and don't fit particularly well.  

Other folks have provided models for “addon” TPU “covers” to facilitate better traction, but I wanted to do it as a single multi-material print, mainly to keep the facing surface at the original dimensions to minimize the impact to the gearing ratio when the rewinders are being driven.

My though was just to recess the existing “tread” face by some number of single wall print widths to make room for the TPU material.

However, the supplied fusion file (in addition to being really hard to navigate as nothing is labeled at all) wasn't set up for parametric use whatsoever.    Changing the setback of the “tread” face by modifying the values on the sketches totally broke the model at multiple steps downstream as did most of the other values.

This is simply because the way the sketches were put together there was no way for them to work parametrically between the Revolve and following steps.   

After fiddling with the file for an hour it became real apparent if I wanted to achieve my goal a full on from scratch remodel was in order, and this is the result.

As I may have need for something similar down the road, I built this file to be completely parametric.

WHY?

The problem with the base unit that a number of users have found is if your spool isn't in the right range of widths, the rubber bands may not hit the spool flanges at all, and without traction to the spool flange, the rewinder doesn't work, and your filament change will result in a mess of “buffer less” filament.

Folks have found alternative rubber bands, hand painted on “tacky glue” and a host of other fixes.

Being as I have an XL (and you “could" do this via the MMU3 as well), and can print these in a single pass without having put the TPU part on after the face, this approach worked best for me.

WHATS PARAMETRIC?

In the original source F3D?  Pretty much nothing.   Most any change to any of the sketches will break in several places downstream.

I this from scratch rework?  Pretty much everything.  In the photos is a portion of the settable parameters for the file (the ones you would likely need to change), along with examples of a 3 vs 4 layer TPU setback.   

Everything is fully parametric, with a single caveat - changing the TPU_Thickness_Idler (and roller) will change the edges for the bottom chamfer, so you'll need to reselect those.   The step will highlight red, edit the feature and select the changed edges as shown in the example SS in the photos.   

The Modeling for Printability in CAD notion

Wherever I can I set all my values based on layer heights (first and normal as they aren't always the same), as well as single wall thicknesses, which are in the user parameters as Layer_Height_First and Layer_Height_Normal and Single_Wall so if I change the print layer height or nozzle size everything in the model that's based of that will still be correct if those values change.

I changed the way I CAD stuff with printability in mind several years ago and it's eliminated a whole ton of issues, and a lot of that is based in modeling things with the actual values that will be used at print time.

If I later need to change the scale of the model, go to a different axle diameter, need more accuracy (detail) and drop to a small nozzle or speed up the print and go to a larger nozzle, everything that's based on those values at print time are built into the print so I don't even up wasting tons of print time printing 2 ½ perimeters because I stuck an arbitrary value in for how wide something its.  It'll be modeled in at 2 or 3 perimeter widths from the get go.     

By building stuff truly parametric from the sketch on, you can change pretty much any aspect of it months or years down the road (long after you've forgotten why you did that in the first place) and not have to spend a half a day reworking the model to be able to get a usable export because you needed to change something.

Sometimes that's in knowing what your cad package has as limitations.   For example, fusion sucks at revolves that have tangents in the profile that are not locked to a quadrant of the curve.   It'll look ok but the step export isn't water tight, as fusion will attempt to treat the whole area as a circle and the step export will not be a closed poly surface but an open polysurface with a revolved “partial” circular face.   

You could spend a whole ton of time trying to get it to work (and fail) or just accept that there's some things that fusion is just absolutely miserable at (a lot of them actually), and fix it post step export elsewhere (I use Rhino for that and a lot of other stuff).

Other Changes:

While I was in there I fixed the small roller (Idler) as it didn't have enough material for the M3 screws to cut threads for.   Also fixed the hole sizes to accommodate the more common Capstan type M3 Screws.  

I got rid of the 3 sets of holes on larger Rim rollers as well and fixed its hole sizes as well.  If you want to change the “lock” screw sizes you can easily via the user parameters.  For example you “could” use actual set screws vs shouldered capstan screws and as a result make the through holes much smaller.   That was one of my original ideas when it became apparent I was going to have remodel the whole thing, but I didn't have 20 of the same diameter and length setscrews on hand so went with the traditional M3's.  

I also modeled in a brim as I printed the whole thing out of ASA and needed the brims to preclude lifting.  

It's a separate body in the step file so if you don't need it you can just delete it in your slicer.