Baking gears

In the quest for alternative engineering, and apparently having to much time this weekend, I have set myself the past few days to try to cast (or mould) some gears. I had bought a small gear for the Z belt (attached picture, the red one on the left), but it wasn't cheap and I only have 2 and need some more - 2 for the Z and 4 for the Y axis of Darwin/Richard. So I figured I needed to create some myself.

Initially, I used a mouldable epoxy based product (something called Pattex Repair Express, introduced in 2000, [www.pattex.com]) to create a mould for the cast. This worked well, and I was able to cast a gear using the EZ water (see one of my previous posts) which wnt well. But the EZ water is to brittle and soft, so when drilling a hole it broke.

Then I figured I could try to use parts of an old printer. I took a piece made from POM - this site has some info on that and other thermoplastics [www.azom.com]. POM melts at approx. 210 degrees C. I used a hollow aluminium tube that I filled with pieces cut from the printer part, stuck a rod in so I could press it out and heated it on my gas stove - approximating 200C . When melted, I pressed it in the mould. This did work, but the mould was cold and thus the gear not nicely shaped (see second gear from the left on the attached picture). It was very tough though, and I could easily drill and tap it. Then I applied some heat to the mould (the repair express can be heated up to 150C, so I threw it in the oven) and tried it one more time. Now the gear got stuck and when prying it out of the mould, it fell apart. Guess it became to warm and started to disintegrate. The resulting gear is the third one from the left in the attached image.

The final approach was, that I used some ordinary plaster bought from the hardware shop and made a new mould. Plaster can be heated up to high temperatures. I again melted some POM using the same *very* accurate method as described above, and pressed it in this mould. Additionally, I placed it in the oven at 250C, so it would set nicely. I found out that thermoplastics have a very high viscosity (i.e. tend not to flow very well - guess that's why the do injection moulding under high pressure) and the gear that game out looked better, but still wasn't functional (#4 in the picture). Also, the mould was "through" (i.e. plastic also came out at the bottom) so I had to break it to get the gear out.

The last attempt was a new mould from plaster, not going through, heating it up in the oven at 250, melting the POM and pressing it in the mould (see grey shaped object in the picture). Then I put it back in the oven for a few minutes, at the same time applying some pressure using a small piece of aluminium. The final gear did look like a gear (it is the most right one in the picture) but the mould did suffer quite a bit and can not be reused.

Conclusions? Well, I ordered some CAPA and will get some pre-fab belts to try the approach as described on the Cartesian Assembly page. I do not think making gears this way works, but maybe when an extruder is used, it could be possible to do this. Definitely something I will try later when the machine is working. On the positive side: with sufficient scrap plastic around, re-using that makes things definitely interesting! Maybe someone else has had more success?


Joost

i think vik and adrian actually successfully made gears w/ CAPA that way. that seems to be the most promising route (the bolts give you the pressure required. i'm definitely going to go out and get the required parts this weekend and give it a try.

also, it might be worth looking into either getting pre-fab gears or checking out the gear cutting machine forrest found.

Seems alot of work when you could use this stuff...

[www.smoothonsecure.com]

Yup, it is. But is was interesting from a learning curve perspective - and fun to try a very low-tech approach. But definitely not efficient considering the end results.

Joost

that looks like some very good stuff. i seriously dislike the fact that we are having to make our own gears. i think one of the top priorities for a post-darwin release is to find a way for the machine to make its own gears. either by making the gears and teeth bigger, or creating a reprappable machine that cuts gears from blanks you print on the printer.

RTV silicone rubber is the material you need to make the mold. Cures withing a day, very durable, very heat resisant, and if properly made will replicate all details of the original item..

Another cheap, but much more time-intensive method is liquid latex, which you might be able to find in a craft store or hobby shop. You would want to put a flat side of the gear in the bottom of a box (preferably a box with a glossy surface) then you paint liquid latex on the gear and box floor, a layer or two a day, so that each layer dries partially, but not completely (liquid latex does not bond well to fully-dried latex) until the stuff builds up to a fairly durable thickness. Then let it dry thouroghly then pour plaster of paris over the whole thing. When the plaster cures, you have a floppy latex mold of all the surface features of the gear, and a plaster mother-mold that will support the rubber molt to preserve the overall shape of the gear. You can easily remove the original gear from the mold, then pour in two-part epoxy resin to make a nearly perfect copy of the original gear.

For the large (Z) gears I made a simple press from plywood (3 plates, bottom had 8mm hole, middle large 50 mm hole, top 8mm hole, alle centered) and used a similar method as described on the wiki: place a piece of the timing belt inside the large hole, put warm capa in it and then press the 3 plates together with an m8 stud though the center holes. Worked very well. Small corrections to the gears (i.e. where the timing belt overlaps in the mould) can be done by heating only that part in some hot water and press some belt on the "bad" teeth to have them correclty aligned. For the small gears, I made a mold with a 14 mm hole and through 8mm hole, put some capa in it and pressed a pin through it (as described on the wiki as well). These are not soo nice as the big ones, but they do work.

Joost

I've been looking for an explicit definition, but haven't found one. When you say "CAPA" are you referring to any polycaprolactone polymer in general, or to some particular commercial product?

Rick.Berry Wrote:
-------------------------------------------------------
> I've been looking for an explicit definition, but
> haven't found one. When you say "CAPA" are you
> referring to any polycaprolactone polymer in
> general, or to some particular commercial product?

CAPA is the trade name for polycapralactone used by Solvay. Solvay, to the best of my knowledge, is the only manufacturer of polycapralactone since Dow shut down its production in Port Charles, Louisiana last year.

Add a few drops of glycerine to the RTV builders silicone if you wish to cast deep sections; I recommend mixing thoroughly. We used clear silicone with a trace of acrylic to see if it had bee mixed all the way through. The smell is acetic acid (vinegar).
A wide diameter pulley with a thread on it will work with cable similar to a toothed belt and gear system. Similar to a draughtsmans table. This is suited for precision low force applications and is used in the Modela mini CNC mill.
Also - tickling my fancy is the stick on magnetic strip here:

[www.mutr.co.uk]

Instant precision gearwheel! not particularly RepRappable admittedly

> Solvay produces CAPA.

Actually I found out that "On January 31, 2008 Solvay sold the CAPA