Loading
mcmaven

Filament Loader for Multiple Mechanism Auto-Rewind Spool Holder

by mcmaven Jul 12, 2019
Download All Files

Thing Apps Enabled

Please Login to Comment

I want to first thank you for trying out my rewinder design and to let you know I am glad to hear it is working well for you. Please feel free to offer any suggestions you might have based on your experience.

now, I would like to talk about your design.......

wow!

this is a very cool solution to a difficult problem. Is there any chance you could post a video of it in operation?

Today, I put the spool on the shaft with the cap and nut, but do not initially place the spool on the stand. I first feed the filament into the PTFE tubing and through to the MMU a few inches at a time with my fingers. Once complete, I put the spool on the stand and move to the next filament.

This is the best option I have been able to come up with for using the fitting integrated into the stand.

I think the problem is much easier if one of the optional arms is used. The spool can be placed on the stand after loading onto the cap/shaft/nut since the PTFE entry is not blocked by the spool. But, the filament must be pushed through a tiny bit at a time by hand.

I do have a few questions...

It appears that the gears are used to get the crank handle where you want it and not for mechanical advantage. Is that correct?

Did you consider trying to get this to work with the existing stand rather than requiring a custom stand? I think the "base rewinder main" assembly could be implemented as an "accessory" that attached using the existing arm mount location on the stand. There is probably a way to make the crank and gears work as well. If not, I would be happy to make modifications to the stand to enable your filament loader as it is really interesting.

I am wondering if there could be a universal solution that worked across all (or most) of the configurations that people are deploying. I do not see how such a thing could be done at the moment. But, you have really got me thinking about the filament loading problem. Perhaps the situation can be improved for other configurations like you have done for this one.

I was using the same technique for loading the filament, but because the spool wasn't on the stand and the filament comes off the bottom, I had to keep unwinding as I pushed the filament up the PTFE tube. I need 3 hands; one to unwind the filament, one to push the filament up the tube, and another to hold the stand while I did it. This approach lets you push an inch or two into the feeder, clamp it, put the spool on the stand, drop the rack in, then crank.

The gears give a little mechanical advantage, but you are correct, they are mostly there to get the crank high enough so you don't bang your knuckles.

I designed my own stand for two reasons: to add screw holes for adding my extruder and crank parts, and to reduces the print time.
My print of the existing stand and its components took almost 24 hours, and to be honest, I felt that the existing stand was designed a little heavier than it needed to be. I reduced the cross section of most of the beams, and created separate parts to allow them to be printed flat, to reduce print time and the amount of material needed. And to test my theory that parts printed laying down are stronger than the same cross section part printed standing up, because the grain of the fuses aligns with the length of the beams, rather than being perpendicular to them.

The down side is that I could print all the parts for the existing stand in one print session, but mine requires two print sessions because everything is laid down.

Both the crank assembly and the extruder are independent of the stand, and I believe that they could be glued to (or screwed to the existing stand if it had some screw holes in it). I used screws to minimize the print time if I had to redesign the extruder assembly, which I did a couple of times. Often when I'm happy with a design and I'm not concerned about the cosmetics, I will glue parts together and use the screws to align and clamp them, and after the glue sets, I take the screws out and use them for another project.

But while I got your attention... There is a 2mm flange around the Cap gear that was initially frustrated me when I loaded a spool. If the flange rests on the teeth of the Large Gear, the spool looks like it is loaded, but the teeth don't engage. I learned that the only way to reliably drop in the spool is to insert the Cap end first, then swing the Nut end into position. Could you chamfer the flange???

And I will post a video.

no need to post a video....it is easy to visualize what you are describing. I can certainly add a small chamfer to the edge of the cap gear flange.

The other contributing factor might also be the overall length of the shaft+cap. I set the length of the shaft based on "average" tightening of it into the cap to get the overall length to fit the stand. If yours are slightly looser than my "guesstimated average", the assembly would be long and harder to insert. If the nut side lead, there would be a tendency for the cap gear flange to hit the large gear. I'll also look into shortening the shaft a tiny bit as well as my assumption of "average tightness" might be too aggressive.

I initially investigated a multi-piece stand, but eventually went with a monolithic stand for two main reasons.

The first is that I wanted to get everything on one plate so a complete rewinder would come from a single print. As you already observed, that is not possible with a three part stand.

The second is that the channel seemed to come out smoother when printed vertically than any option I could come up with for printing it laying on its side. I will be printing one of yours this weekend and am interested to see how your track channel comes out.

I will admit that the stand is beefier than it needs to be. But, my efforts to lean it out have not really resulted in a significant print time reduction. Will look more closely at the dimensions of your stand components to see if you made some better choices.

Can you tell me what slicer and slicer setting you use? 24 hours seems like a long print time for the base stl files and it would be great to understand how you ended up there. As a reference point, the gcode I post on the Prusa Community site prints in about 11.25 hours and seems to produce an acceptable print quality for proper functioning.

Finally, what glues do you use? I've made many things that are single piece prints and many things that needed fasteners. But, I have never designed anything where I intended to use adhesives. So, I am ignorant on the topic and would like to know more....

I didn't want to deal with digging support material out of the channel, or the resulting rough finish inside the channel, so I split the channel Rack_Side_B into two STLs; a and b. I print the both STLs with supports and the inside of the channel facing up, which results in a
great inside surface finish for the rack to slide on. Then glue them to together.

I'm using the latest Prusa slicer.

As a simple test I created 2 parts: a 6x6x120mm rectangular STL (volume = 4320 cubic mm) and an 8x8x120mm rectangular STL (7680 cubic mm or ~44% larger). The estimated print times with all default settings (i.e. 0.15mm layers,15% fill, etc.) and no support, the slicer estimates 56 minutes to print the 6x6 standing up, and 22 minutes to print it laying down, and 80 minutes to print the 8x8 standing up, and 30 minutes to print it laying down. The addition volume cause about a 30% increase in print time in both orientations, with the height being the big difference. Needless to say, the height contribution to print time drops is minimized when there is more material printed on a layer.

But my rule is that generally; going high, will take longer than going wide. As a result I try to layout my parts on the print bed with the overall minimum height, given the appropriate surface quality and support trade-offs, to minimize the print time.

The 24 hour print time was with the same slicer, all default settings (0.15mm layers) and no supports. I probably should have set the layer height to 0.2 mm.

I use Methylene Clorhide to glue PLA. It works in seconds (like superglue), except that it's a solvent, and it actually melts two parts together. It's great for fixing broken parts because the resulting joint is stronger than the original layer bonds. When I was building my Gyrotourbillon that I broke the balance spring when I was removing the support material, and tried gluing it because I didn't want to wait another 5 hours to print another spring before testing it. To my surprise it is still working great.

This is what I use:
https://www.tapplastics.com/product/repair_products/plastic_adhesives/tap_acrylic_cement/130
with the Thin BD-25/2 applicator:
https://www.tapplastics.com/product/repair_products/plastic_adhesives/hypo_type_solvent_cement_applicator/409

Beware, it is nasty stuff, and extremely volatile. But it's fast (sometimes too fast) and the bond is great.
I keep the can in a capped coffee can, because the caps on the containers sometimes don't seal that well, and you can find that in a couple of months a half full can will be empty. Also I only load a what I need for a job into the applicator, and you normally don't need to use too much. 1/8 or an inch in the bottom applicator is a huge amount.

A friend told me that Acetone works just as well on PLA, but I haven't tried it yet.

I split parts and glue regularly to avoid using supports, if I'm concerned about surface quality. I attached a picture of a model where some parts were split into as many as 5 pieces to avoid the funky surfaces that supports create.

As soon as i downloaded the stls, i figured out what you had done. Agree that supports in the track would have been horrific. Your method is likely as smooth as a channel can get as the track will not even encounter layer boundaries as it slides along.

Also agree that any given volume can be printed faster horizontally than vertically.

Despite both advantages, I will probably need to stick with my current approach. There has been too much positive feedback about the system being a single overnight print and I cannot assume users have access to specialized adhesives. The good thing is that your stand design is out there for people that do.

It might even be worth releasing another remix of just the optimized stand components without the filament feeder. That could be interesting to people who use the optional PTFE arms and thus are not candidates for your feeding mechanism.

I am familiar with solvent welding from PVC pipe. But, I did not know there was a solvent that will weld PLA. There are a number of projects I have started as I could not imagine a workable printing solution. I was about to spend a ton for a roll of BVOH. Now that I know there is a good solvent weld, I will try to segment the parts for assembly instead.

Is there also a solvent weld product for PETG?

I have worked with some high VOC solvents before, so no worries there. I have an outdoor space, respirator, gloves, etc.

Acetone works on ABS both as a solvent adhesive and for vapor smoothing. But, it did not work on PLA for me. (neither did MEK)

I really want to print a tourbillon and have a couple of candidates in my "cool stuff to make" collection. Which one did you get to work?

I literally just came here to say... "multi-media unit"?
Lol

Don't laugh. But I admit that I still haven't gotten mine working right yet. I assembled the printer and it worked flawlessly immediately. I'm afraid that I can't say the same for the MMU. And everyone that I know who has installed it, are having some issues with it. A flawless 15 hour multi-color print has been rare.

did you get a MMU2, or a MMU2S/MK3S?

I gave up on the original MMU2 after literally months of fighting with it.

But, the "S" upgrade was thus far been nearly bulletproof.

We need to talk.

am happy to help, but am moving the discussion to a PM.