Are there any drill bit screw extruder heads out there?

I just realized this in the last few days: people extrude filament (like the filabot) using screw extrusion. Why then do we use filament + a hobbled screw or gear? Would it be possible to use a drill bit in a tiny hole and perhaps add a continuous servo for an ultra cheap extruder head? The way this would work may or may not include filament, but it may be ideal for an ultra-inexpensive head that uses raw pellets.... [0.02]

---

Measure once, Cut twice, Print 3 times.

I think measuring exactly the amount of plastic extruded would be difficult, and the output quality of printed part relies on knowing it. Also it might get difficult to extrude fast enough.

This said, feel free to experiment, it would be nice to prove me wrong here.

---

Most of my technical comments should be correct, but is THIS one ?
Anyway, as a rule of thumb, always double check what people write.

Check out Granule Extruder and the associated links.

Looks like they switched from pellets in 1.1 to filament in 2.0.... no explanation why, it seems like it was on its way to working.

Also, can anyone tell me why their design is on wiki as a "standard" extruder? Do they have some sort of clout/officer authority in the reprap community or are they just updating wiki's as they deem fit?

Yes, Simba, the author of those wiki articles has special clout around here.

Simba Wrote:
-------------------------------------------------------
> Also, can anyone tell me why their design is on
> wiki as a "standard" extruder? Do they have some
> sort of clout/officer authority in the reprap
> community or are they just updating wiki's as they
> deem fit?

Maybe because the author is the founder of the RepRap project?

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

Until about two years ago, attempts were made to concentrate on a single design. It didn't work out very well, as for example access to the central repository was very limited and version management best practices weren't applied. Claiming to be a "standard" is probably a remaining from these days and I think it's OK to change the wiki towards showing the plentitude of designs we have now.

Oh, and expect a heated discussion when you try to choose another design to be "the best" or "the most widely used" or whatever you consider to be the most important point on a design. IMHO, the wiki should always show many options, still give good advice for newbies. Even if this advice is "they all work the same, so it doesn't matter what you choose".

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

you will never get a working grannulate extruder that mounts to your x-carrige. Grannulate extruders only work when built like the filabot, a stand-alone filament-extruder period.

This is not an open challenge, but the time is you'rs to waste.

Sounds like you know how to motivate people. Why do you think it is so impossible? Its sounds really straightforwards if you keep the part light by having a conveyor belt of sorts for plastic pellet.

This idea still seems obvious. Am I missing anything? I've built and started testing granulate extruders now in two sizes (1/4" and 3/4" with 192 and 450 watts power respectively). I have a 300 deg C temperature limit based on new thermocouples and kaptop-wire-copper-furnace cement only designs.

I'm not seeing why the extruder couldn't work attached to a gantry, but I could certainly see may be inconsistent as far and controlled extrusion amounts if poorly designed.

Quote

I've built and started testing granulate extruders now in two sizes (1/4" and 3/4" with 192 and 450 watts power respectively)

Have you performed a process capability study or similar to see how tightly you can control filament diameter?

It is easy to extrude plastic into a filament shape, it is difficult to do so with the level of consistency that our 3D printers require.

What level of consistency is needed? I remember reading a link online for a recent home made extruder (http://forums.reprap.org/read.php?1,166442). The filament was like 1.78 mm +- 0.02 mm. I can bet you that the rest of a reprap system doesn't have enough accuracy (i.e., 0.02 mm) at the hot-end level, + hobbed bolt, + heat flow differences, for this to matter. I think it would be a problem if there was more than 5% variation along the filament and a major problem perhaps about 10%.

Also, reprap itself is really a crude filament maker. It takes 3 mm or 1.75 mm (which has error itself) and tries to extrude it to 0.5 mm. I think the output quality of the reprap nozzles are representative of what any crude home made filament maker or auger could do... My guess is they are accurate to about 5% too.

As a rule (I do a lot of experimental engineering research) I've find that most systems, experiments, biological functions, etc. can easily absorb 10% variations in condition without it being easily noticed. Luckily I don't manage a nuclear power plant or anything : D

I will add that the places I've seen 2% differences matter are with bridging, and with "pimples" forming around the perimeter (i.e., 98% versus 100% extrusion ratios). But this would mean that <2% variation is needed to control factors that are essential to print surface quality, not the ability to make 3D prints that work in general.

Edited 2 time(s). Last edit at 11/28/2012 01:35PM by Simba.

The extrusion feedrate generated by the slicer is based on the cross sectional area of the filament, which is calculated based on the inputted filament diameter. The cross sectional area is is A=1/4*PI*D^2, therefore any variation in the filament diameter has a very large impact in the cross sectional area of the filament. For example, a 5% variation in the filament diameter results in a 10.2% variation in the cross sectional area.

In order to to keep the filament cross sectional area within 2% (to use your number) of the value the slicer is using, the filament diameter must vary by no more than 1% (eg, for 2.85mm filament, the diameter must be between 2.8356mm and 2.8643mm). This is, to put it lightly, very difficult to achieve.

You're right that the overall system is tolerant of fairly large variations in filament diameter, if you don't care about surface finish and final part appearance. However, many people do care about such things, and high quality output is expected of modern RepRaps. It is no longer good enough to be able to produce parts that "more or less" look like the CAD file.

Personally, when I inspect new filament, I consider +/- 0.03mm or smaller to be acceptable. Any variation larger than +/- 0.05mm means I probably will not order from that supplier again. I take at least 8 measurements over the length of the spool in multiple orientations and average the result.

For a process capability study to validate the output of an experimental filament making extruder, I would suggest extruding at least 100 feet of filament and taking diameter measurements in 2 orthogonal directions every foot. This will give you enough data to have statistical significance, and generate a nice bell curve of data points.

Thanks for the input. I will do my best to address these when I extrude my own filament. What would you say your findings are for the best filament out there - is is .03 mm for 3 mm diameter? Is it gaussian?

I suppose for a granule extruder, the tolerances have to be much different. If you could help me address this in your next response that would be great: What is the combination of error due to the input (filament) and output (nozzle) of a typical reprap? If the filament has to have .03 mm (i.e., 1%) error, and lets say the extruder has 2.5% error, than the combined error of the system is some non-linear function of the two. So, a granule extruder with 3% error might even produce higher quality prints overall if the error is dampened, i.e., the hysteresis in the error is more consistent, or less "amplified" as large variations caused by small variations in input material.

Also, I'm a bit more cavalier than most with my approach, though I can value high quality prints, I would choose functionality to focus on for enabling features as a firststep/prelude to future revisions that enable better print quality. Sometimes I feel like the community as a whole has a "Mr. No" mentality when it comes to small changes, while I think the overall field has gotten here by embracing rather ballsy left-field ideas. Now the field is starting to mature and it seems like converging on more identical rather than different solutions. That's good for end performance and compatibility, but bad for wild new ideas.

Quote

Is it gaussian?

Heh, never heard it called that before. I think you are asking if the distribution is normal. I've never done a full set of measurements on a sample of filament - I've never needed to since I don't make my own. The reason I suggested it for you is that you are trying to demonstrate that a new approach works well, and in my mind that requires a higher standard of evidence than a supplier who is making filament the conventional way (big molding machine, special die, special cooling system, etc) would be expected to provide. That said, I would expect a normal distribution unless something was not working right.

As for extruder error, I'm not sure I'm following you. What sources of error does an extruder have, and how would we expect to detect and quantify them? Are you going to measure free air extrusion diameter?

Once those questions are answered, then you have to figure out if you can separate the extruder error from filament (input) error.

All that aside...
In my mind, the biggest conceptual stumbling block is you are going to have to rethink how to modulate and shut off the flow of molten plastic from the hot end. Since you have a greater volume of molten plastic and less precise control of the melt chamber pressure, stopping and reversing the feed screw will probably not give satisfactory control over the output.

The second biggest issue is you have now substantially increased the moving mass of the extruder, which will do bad things to the print speed and output quality.

My opinion is there's not theoretical reason that a granule extruder couldn't be used. But there are many practical reasons why filament extruders are superior, even given the price premium we pay for filament over raw pellets. With enough effort these limitations can likely be overcome, but IMO this effort is better spent on other aspects of RepRaps at the moment.

Thanks. I want to hear from you and others how granule extruders are so difficult so I can address each issue in my attempt. Lets all (readers, not you persay) be honest and figure out what the issues are. Are we resistant to change because the potential cost savings (and ability to use novel materials like elastomers) is not as important as having something tried and true that is working today? Do we want to avoid change for sake of consistency? Or are there real fundamental reasons the carriage mounted extruder can't work with ease?

I should point out, I'm getting close to finishing my own. I can see how the added weight will limit acceleration (not speed so much) because of perhaps adding 50% or 100% more weight. I think most people assume the weight would increase dramatically, but there is not good reason to expect more than a 100% increase in weight. That could also be easily offset by using 24V instead of 12V or by using two motors or a twice current/force motor. If the Y carriage and Z are connected, same problem, but usually this would only affect the X.

How would the added weight affect print quality in any way - even if it lowers print speed? I don't see an argument here, unless the weight somehow causes the belt to slip or the entire frame to bend somehow.

My philosophical argument for bringing up the existing filament extruder quality was just an example. You never measured the accuracy of the output of your hot end directly (you measure the indirect output once it is squished into layers for calibration). But you would want someone like me producing new filament to be extra cautious. So my argument is that if a relatively crude extruder works to our satisfaction, one that takes 3 mm filament, scours into it with a hobbed bolt for grip, and then pushes it out a fine nozzle exit, why is there any reason to believe home made pellets extruders would be extruded any worse? I see strong evidence a home made extruder will work well. So I must not understand the issues

I believe it will be harder to control ooze. Moreover, a lot of attempts have failed without explanation. For example, right now I know the Up! Engineers are trying to make one carriage mounted pellet extruder. I suspect the reason their approaches failed is because a stepper can not provide the torque necessary to push pellets through with an auger. I have a solution I am working on that also solve the other issues in question but I will reveal it if it works!

Edited 1 time(s). Last edit at 11/28/2012 11:40PM by Simba.

Simba Wrote:
-------------------------------------------------------
> Or are there real
> fundamental reasons the carriage mounted extruder
> can't work with ease?

I think that the auger designs are fundamentally wrong. Here's why:

The drive screw is not a positive displacement pump. It doesn't have a constant relation from input rotation to output volume like progressive cavity pumps, which physically force the material forward at a known rate. Instead the action is more like a boat propeller in a shroud, where the rotation produces some thrust but the "pumped" material is also free to just rotate along with the screw or even flow backwards.

If I was developing a granulate extruder, I would start by building a positive displacement pump of some type that can handle molten plastic. With that I would be able to extrude accurately regardless of the pressure before the nozzle or other conditions.

Industrial extruders use augers, but their size allows them to work differently. They are horizontal and the input end works as a traditional screw conveyor, with the tightening screw pitch compressing and melting the pellets towards the output end. Scaled to reprap-size, the whole drive screw would heat to the extrusion temperature, preventing the conveyor action. The viscosity of plastic at that size would cause the plastic to stick to the screw and make releasing air bubbles hard. It would have enormous backflow if you tried to reach the speeds at the nozzle diameters that the filament-fed extruders are capable of.

Fantastic. I believe I can address (and have in theory) 100% of these potential challenges.

Currently my filament (from 1.5 mm drill bit) is coming out as the following over 1 foot:
1.43, 1.44, 1.40, 1.42, 1.43, 1.38, 1.49, 1.43 (mm). Considering it was a very *very* early run totally uncontrolled (no temperature or force or speed control) and unoptimized, I think we have a grand new direction to go into here.

Std: 0.031959796
Mean: 1.4275

Hello,

Is there any final outcome, can we make extruder from drill bit and barrel ?

i am interested to make one.

can you please guide me ?

Hey, you definitely want to see this then,
Not a drill bit, this is more of an engineered solution.
[youtu.be]

Drill bits are not the right shape

Yes you need to move material from one end to the other via a thread, but the depth of the thread needs to taper away while maintaining the outer diameter


Just like the one in the video above.