Quote:
Originally Posted by artdutra04
Also, remember that objects that are 3D printed via the FDM process (where ABS/PLC plastic spool is melted and extruded into the printed object) are weak in between the vertical layers. There are two ways to deal with this: 1) accept this and design/layout the parts with this in mind; or 2) design hybrid parts that use other parts or hardware to reinforce the 3D printed parts.
|
Art's First point is probably one of the most important things to keep in mind while printing parts. Depending on the exact geometry of the part, printing in the strongest orientation (the layers do not want to pull apart from one another) is not always the orientation that will most accurately produce the part. There are some interesting ways around this, either splitting the part into multiple parts where each sub-part is optimized for geometry or strength, or as Art said, you can make a hybrid part where some other component reinforces the weakened area.
Quote:
Originally Posted by Oblarg
I pulled the CAD of bbman's website, with the hope that the profile is correct (or close enough to correct that it'll work, the fact that their CADs have different profiles for HTD and GT2 pulleys seems to indicate that they're not just using a single erroneous profile for all of them).
I had heard that other teams had successfully machined pulleys from CADs obtained this way, so it seems worth a shot.
If it doesn't work, I'll check out the parametric pulley link nicholsjj posted (thanks for that, by the way) - though I don't have OpenSCAD on my computer right now.
|
I used your method exactly in Decemeber-ish of last year to make a prototype pulley (ignore the versa keys...) and the profile was pretty darn close to perfect.
https://db.tt/WT0HnkDQ
https://db.tt/Wd92gCkg
From what I saw, the printer ran a few thousandths over while making the teeth, so if I were to make that exact pulley again, I'd probably remove a couple thou from the tooth profile and see if it cleans up with tooth engagement at all. All in all, wasn't a bad final product, IIRC, the teeth being a touch big was only noticeable once the belt wrapped more than 120* or so.
Also, as far as I can remember, the pulley was surprisingly strong, we clamped the belt in a vise and beat the pulley up a bit and there was little to no permanent damage. It's worth noting that the pulley was printed from Ultem (polyethermide) which is something like 2.5-2.75 times stronger than ABS in the same machine. Odds are, ABS
should hold up just fine, but if I were using ABS, I'd make sure that at least one wheel was directly driven from the transmission or something.
While we're on the subject, here's a picture of one of the larger print runs done for last season's robot:
https://db.tt/tqhDMaaS
If I remember correctly, that build took somewhere around 11 Hours to print from beginning to end, including warm up time. Starting from the left most side:
-(4) parts with rectangular holes were servo brackets that were printed a touch too thin to hold up to handling before the servo was mounted,
-(2) spacers to offset a sprocket from our shoulder joint - they had the standard 1.875" hole circle printed into them, along with a 1.125" OD protrusion on one side to pilot into a bearing bore, and on the other there was a 1.125" ID bearing bore to hold an FR8 Bearing,
-The large clam-shaped parts on the right side were a spacer for our shoulder's pivot. There were (6) Hexagonal pockets printed into the spacer to hold 10-32 threaded inserts along with some neat cutaway geometry to allow #25 chain runs through the pivot to our climber. I stood on one of the shoulder spacers to see what would happen, and it didn't move a bit.
Also, here's one of the encoder mounts made over last season - initially, this was the 'standard' mount used on both our drive and shoulder although the shoulder's final encoder configuration was a bit more complicated (cooler) than this:
https://db.tt/kJqK0kE3