The quest for the ultimate extrusion screw
In the plans of the current extrusion machine a basic drill is being used in the barrel to push and melt the plastic to the nozzle. This seems to work, however, in the industry a different type of screw is used which compresses the plastic while it’s melting. These are way less accessible than the drill, for obvious reasons. So Dave had the idea of possibly trying to produce these ourselves. Therefore this topic will be dedicated to the research on trying to design/model/produce this.
I was quickly doing some research on this yesterday and it seems to be a world on its own with different designs of screws, different compression ratios and different “zones”. If there’s anybody that might have some experience or knowledge on this that could help out, let us know here.
I turned the basic idea for the screw design into a quick 3D model, to see if I would be able to model it. Now I need to find the right design requirements to turn that into the eventual 3D model, so we can turn this into a physical prototype.
Has anyone explored this possibility to extrude beams of an infinite length ?
I think this method will cause the plug to be pushed by molten plastic as the extrusion goes, though I think the operation would need to be repeated if the extrusion goes quicker than the cooling.
We should make sure there will be always an entirely cooled transversal section of the beams inside the metal part attached to the nozzle.
Another idea is to fuse two beams after extrusion by heating an end of each beam then pushing the ends through a metal tubing of the same shape. working with plastic is not like working with wood you can’t fuse lumber as you could using plastic 🙂
For the compression screw siemenc designed, the diameter of the threaded section is 25.7mm. There are drawings in the download kit.
Barrel: 26mm ID 34 mm OD. Screw (drill): 26mm diameter.
There are no tolerances given. Depending on where you source your tube a 26mm drill may not even fit inside, so you do need a small gap. I think you just have to find what works for you, there doesn’t seem to be a definitive answer to this, but some clearance between screw and barrel is necessary. Too tight and the screw will bind and seize, too loose and it will not extrude with as much pressure, but will probably still work.
So, here is a screenshot of Dave’s video ‘Build the extrusion’..
The question therefore changes to: Isn’t it easier to clean it when it has no sharp end?
Here is a question for those of you who decided to go with an auger wooden drill bit for the screw (photo below).
Do you cut the sharp edges off at the end of the screw (the pointy thing)?
I have done screw design professionally for about 7 years. I’m happy to help. There are a lot of variations in the market, but you’re on the right path to get a basic extrusion screw that will work well in this open source extruder. Here are the basic design parameters I suggest we target for this screw:
Compression ratio : 2.5 or 3:1
Zoning: 1/3 length feed, 1/3 length transition, 1/3 length metering
Pitch: square (lead equals diameter)
Channel length:flight width = 4:1
Hand: right hand. This is industry standard, and the motors should be easy to re-wire to turn in reverse if your gearing setup spins in the wrong direction.
I’d be happy to design a model with these parameters if someone would be willing to specify the shank design, length, diameter, and show where the feed port falls on the screw.
It’s a manual milling machine that I converted to 4-axis CNC. The screw rotates as the cutter moves along it’s length, but as the screw is long and thin it’s very flexible and I had to take numerous shallow cuts.
The screw is designed to have a compression ratio of 2.5:1. The lead is constant and the depth of the thread varies along it’s length. I could have also varied the lead and kept the depth constant, or some combination of the two, I don’t know what real difference it would make, if any.
Another reason I wanted to make my own screw is that drill bits are right-handed, which means to feed material forward you have to rotate them anti-clockwise. I made my screw left-handed so that it can turn clockwise, which is just better from an engineering viewpoint. If you look at the pictures you posted above you’ll see that most of these screws are also left-handed.
I’ve thought about this too. Using a wood auger bit seems a bit basic to me, it’s not designed for this and it’s unlikely to be a precision fit inside the barrel unless you happen across a piece of tube of the ideal bore. Ok, it works acceptably, but I thought I could do better. Below is a picture of a compression screw I just made, it’s not pretty and it took me several days to make on a hobby milling machine, but I’m fairly pleased with it (haven’t been able to test properly yet as I’m still building the injection machine).
I wouldn’t make another screw this way. At the time I considered building a machine to automatically grind these, which is the way I would go if I were going to make more than one. If there’s really interest from people wanting this type of screw I could look into this.
I have a couple of spares for @210USD if someone needs one?
Noted 😉 But I’m not going to be mass producing these by this method. Perhaps if there’s enough interest a larger order from Dave’s supplier would bring the price down?
wow that looks nice @siemenc! We got our screw from china installed in our extrusion machine. It fits perfectly and works super nice. It sin’t necessarily smaller but it does have waaay more pressure! We are trying to make solid tubes with it. Filling up a hollow tube with the extrusion machine, we already managed to make a 40mm x 3M. II have to say it goes beyond our expectations. Super impressed by this screw!
It seems like the machine guy at the fablab is out of the country for another 3 weeks. So no indexer-action until then. In the meantime I might source the material stock for this.
Here’s what I can extract from the original design files. Diameter of the barrel is 26 mm on the inside. The current drill being used has a length of 585 mm in total. There’s a distance of 365 mm between the end of the barrel and the feed of the plastic material. The drill is locked in a pipe in the back with a diameter of 13mm. Is there anything else you need to know?
Do you use a special software for generating these kinds of screws?
Here’s some images from the 3D model I made of the design concept. What I so far understand is that the distance between the turns of the spiral logarithmically build up towards the end of the barrel, while the diameter decreases. I now need to know what the optimal ratios for these design requirements are.
@faridahmadsofizada, that would definitely not be ok. As soon as plastic melts inside it will go anywhere except for the nozzle. I guess it is not that easy to find a proper one, but this combination is not good at all. This question has been asked before in this topic/ forum, you could check it out.
What are the tolerances for the barrel and the screw PLEASE HELP!!!
@siemenc we used a really long bar. We wanted to do what you said (using cooling to get infinite bars). However we found that the newest added material is going totally through the beam untill it gets to the end. So if I use for instance black, green, blue. Blue will be in the total beam in the core and the end of the beam, than the green, than the black. Maybe @davehakkens can show an image of that.
We guess it is because friction and the cool temperature of the tube that the new material is traveling true the core till it reaches the end.
Because of the molten core the plastic has not so much struggle to travel true the tube.
What you mentioned is done in the industry but requires a lot of cooling.
Industry extrusion beams
Btw shooting plastic sound really cool, this is a slow process.
@adyn we use still the same shreds. The screw gets only smaller were the plastic already is molten.
Aha ok gotcha! For a second I thought the plastic shoots like a canon all the way to the end of the tube and fills it up like that. :p Now I understand it just pushes the dried plastic further through the tube, right?
Does that mean that you can create plastic bars with infinite length? By making the length of the tube a bit shorter and perhaps cool it somewhere in the middle and leave the other end of the tube open so the hardened plastic comes out. Perhaps from a certain length it needs some help with feeding the plastic bar out of the tube. Or is that what you did now as well?
exactly what Dave says. Temp 235-220 with extrusion at full speed (@Davehakkens what is the speed). Filling took almost an hour. Then let it cool down and it is really easy to slide out. We were really surprised, first time something goes in one way as we hoped.
Later, we put a small wooden block in the tube to build a bit of pressure to get smoother surfaces. I will show a small stool, hopefully, next week.
Yeah it is hard to set it up square and parallel. But what I actually meant was that the front and the back part of the screw can be made with a straight milling bit fairly fast. But for the taper you will get small steps with the straight milling bit. And setting up that part in the cam-software is less straightforward than we expected. I can provide images to explain better if you don’t get what I mean.
It is possible to mill steel at the fablab in the center of Oslo, but we haven’t tested it on the machine in the fablab outside of Oslo yet. We’re first going to try in aluminium so we can check whether we can make that work. After that we can potentially dig into milling steel on the indexer.
It took a lot longer than I expected, over a week. The problem was the small motor I used for the milling attachment would overheat after 10-15 minutes and then I had to leave it for 1/2 hour to cool down.
The material is bright mild steel (EN32B).
Not sure what you mean about the tapered part, is is just hard to set the machine up square and parallel?
The problem I found with milling it that way is that to do steel, you need a very rigid setup which is why I went with the lathe. Can the Fablab machine mill steel?
That looks pretty sweet! @andyn Pretty crazy lathe skills there. How long did this take? Which material did you use?
In the mean time we got the indexer running at the Fablab in Moss. It’s been a very slow process of setting everything up. The indexer was now just standing on the machine with not positioning whatsoever so we had to spend some time fixing that (and to be honest I assume there will be some deviation on the positioning as well, as we were manually trying to put it in the correct place with the tip of a V-bit as an alignment tool on the y-axis).
We did a test run in wood which seemed to work quite well but looks horrible because of the wood chipping off from the wood grain. The tricky part with this setup is machining the conical part. One reason being the big distance between the clamping on the side, the other reason because of the geometry.
I decided the screw I made above wasn’t good enough. Though it worked well with injection moulding pellets, the channel in the screw was too small to feed shredded plastic consistently, even if it was screened quite small. Making a bigger screw was a challenge and I had to build a couple of things to do it. The new screw is a lot closer to Siemen’s design, 25mm diameter, 25mm pitch, it’s even right-handed!
Haven’t tested it yet as I have to rebuild the whole machine around it. A couple of comparison pics below and here’s a video of how I made it: https://youtu.be/cf4dQ2xw4OM
In principal, screw extruders are using shear flow and a pressure gradient to push polymer forward and a lot of the shearing and viscous dissipation imparted by the screw is what’s melting the polymer.
The single screw and twin screws are both operating on the same principles but twin screws are more difficult to model. Additionally for twin screws: the two screws can rotate in the same or opposite direction.
Honestly, it seems like compression molding might be worth looking into as well because you can get a decent variety of molded parts, it’s simple, and it’s least expensive.
You must be logged in to reply to this topic.