Extruder Rate (request for crackpot ideas)

.

Edited 1 time(s). Last edit at 12/08/2007 09:34PM by The Guy.

Oh, I forgot to mention: this design would have almost no moving parts. I.E. no framework, no threaded rods, no end-stops or gauge wheels. Just those components that moved materials into the extruder head. Everything else would be achieved by magnetic or electrostatic steering.

---

The Guy

That is some amazing, but crazy tech. I can't believe that my tv could make itself!!!!

Holy Shit?

No - What I want as a finished product is the ability to walk up to it and say "Tea, Earl Gray, Hot" and get a cup of hot earl gray tea in a cup. Welcome to Star Trek.
Now all of the technologies exist in very crude form. All we have to do is refine the primary results to a finished product. Here comes hopefully less that 300 years of testing.

Bob Teeter

TheGuy,
Your idea sounds great. Perhaps that will be our Crick&Watson model.

Hey! What happened to my original post!? somebody chopped it out. I don't want to have to type that all in again!
[ whoops! my bad, it got moved to page 1, sorry ]

Glad everyone agrees that it is a sufficiently crazy idea. Those are usually the tastiest. I can't imagine someone mobilizing all the individual components to make the thing though. Would cost a pretty penny but work great. Interestingly all the design and driver software and most of the wiring from the current reprap system would transfer directly over. One would only need to add the E-M coils or electrodes and appropriate electrospinning equipment and then do a few calibration runs.

Edited 2 time(s). Last edit at 12/11/2007 09:05PM by The Guy.

---

The Guy

Hi Bob,

... for 3D-fabbing "Tea, Earl Gray, hot" you mustn't wait 300 years:

Imagine a Darwin-two-head-extruder in a refrigerator-housing - one head extrudes a cooled ceramic-paste for the cup-slices and the other extrudes frozen Earl-Grey to form the 'Tea'-slices ...

When ready after some hours put it in a microwave-oven, heat it until the ceramic sinters and the tea boils - voil

Viktor - An I can purchase it where????? Do we need a third extruder for the lemon wedge?


Bob Teeter

... and one for sugar and another for milk too ...

Viktor

I originally designed a large, fast repstrap machine that I named Godzilla. There were so many issues that came up on ramping up the extruder and especially the cartesian machine, inertia being the biggest, that I eventually decided that the easiest way to ramp up reprap production was simply to build more repraps.

Consider, the Stratasys machine that Adrian and Ed has, with all its additional expense and sophistication, prints about 10 cm^3/hour. If you work back from that figure you get their extruder head operating at around 15-20 mm/sec. Keep in mind that that is in a controlled temperature enclosure kept at just a few degrees below the melting point of the plastic being extruded. Is it reasonable to suspect that we're going to be able to ramp up the Mk II to match that printing, as it does, at ambient temperatures.

It's better, I think, to simply build more of our cheap repraps than trying to kick up their speed. The challenge there is to build more intelligence into the control boards so that one PC can operate several repraps at the same time.

Hi Forrest,

... i remember a thread with a copy-mill - someone tried to run two extruders in parallel already?

It's maybe an idea to build a longer bed and 8 extruders in parallel (or even more), so you can fab 8 (or less with malfunctions) equal objects in the same time, as a single extruder needs for one ...

Viktor

Forrest Higgs wrote:
>Is it reasonable to suspect that we're going to be able to ramp up the Mk II to match that printing.

No need to ramp up, the extruder already comfortably extrudes 0.5mm filament at 16mm/s which I make to be 11 cm^3/hour.

One simple way I think we could increase it would be to reduce the friction in the filament holder side of the pump. Perhaps a PTFE insert or a series of pulley shaped rollers.

---

[www.hydraraptor.blogspot.com]

Spent my spare 3.5 minutes today tinkering with polymer mono filament and an electrostatic module. Easily got 7 inches of deflection from 7.5kvolts at 80hz. So the extruder head I envision could cover at least that much area with minimal expense. Picture 2 electrodes at right angles (or 3 at 120 degrees) connected to such an E-S source via some of those high voltage transistors. Or alternately, 2 such sources could be controlled individually by a lower power set of transistors. The rapidly flowing stream of polymer could then be controllably steered by the same base hardware/software.

The E-S module I'm using is from a negative ion generator, model# SW750 cost me about 11$

---

The Guy

TheGuy Wrote:
-------------------------------------------------------
> Spent my spare 3.5 minutes today tinkering with
> polymer mono filament and an electrostatic module.
> Easily got 7 inches of deflection from 7.5kvolts
> at 80hz. So the extruder head I envision could
> cover at least that much area with minimal
> expense. Picture 2 electrodes at right angles (or
> 3 at 120 degrees) connected to such an E-S source
> via some of those high voltage transistors. Or
> alternately, 2 such sources could be controlled
> individually by a lower power set of transistors.
> The rapidly flowing stream of polymer could then
> be controllably steered by the same base
> hardware/software.
>
> The E-S module I'm using is from a negative ion
> generator, model# SW750 cost me about 11$

That's a very interesting idea you have!
I see it working better with a flow of matter that has little cohesion, as in a gas. Using a high viscosity liquid you would encounter problems with the inertia of the filament, right? I fear that the system would start an oscillatory movement, now what I mean?

Also, you made this ES experiment on what type of filament? It didn't need to have a charge before? That's pretty neat! Was it HDPE?

Hi Fernando, how's the work o UV curing going?

Oscillatory movement is an intentional part of electrospinning but I see no reason it would exist undamped in a controlled circumstance. The inertia of the filament is also a usable property although my thoughts on using the ES steering method with plasma spraying instead of electrospinning would leverage flow properties in the manner you suggest.

I've experimented with several types of filaments now from cotton, nylon to HDPE. Since electrostatics is a surface effect they all respond in the same manner with the only variations seeming to have to do with stiffness and mass. If one is employing repulsive steering then the filaments must first be charged the same as the electrodes but attractive steering operates on both opposite and neutral charges.

In regards to using UV curing, the liquid mer could be sprayed or streamed into a chamber lit with a UV source, the stream or particles would begin to undergo polymerization on the way to the work piece while still remaining sticky and liquid enough to join the work piece as polymerization completes. Tricky timing there.

I'm thinking that a grid of addressable metal pins could be placed underneath the workpiece's base plate in order to apply varying static repulsion/attraction to different regions. This might be one method of overcoming the overhang issue.

---

The Guy

Only in part.

The ultimate overhang issue, in my mind, is a bell with fabbed-in clapper. The bell, sitting upright, has a clapper that hangs free underneath. Even if you don't require the clapper be hinged relative to the bell, you have the issue of how do you start fabbing the clapper out in the middle of nowhere.

It works!
chemical engineers at Princeton have produced exactly the effect I was thinking of. 100 nanometer line widths!
Check it out:
[www.physorg.com]

---

The Guy

Nophead said:
"My experience with HDPE is that you can run the axes faster than the filament flow to get a finer thread. E.g. I can comfortably extrude a 0.5mm filament at 16mm/s through a 0.5mm hole."

In general, the farther the drop from the extruder head to the surface, the more the filament thins out. So what could be done is start with a larger hole and extrude with the extruder head farther above the surface, so the filament is still thin when it lands.

This is not a free lunch, since the filament will now land somewhere behind where it was extruded; so for example, if the extruder travels a sharp corner, the filament will come down in a round corner. However, the software could model where the filament would land and make a path that would compensate for the lag. This is a tricky software problem, but it is solvable because it has already been done in the case of laying undersea cables where the drop is sometimes kilometers and the cable still has to be layed down accurately in order to avoid undersea hills and rifts.

There will be a limit to software modeling and compensation, so there will be height above which the filament can not be guided accurately. However, a nozzle area doubling is probably possible and the software, while difficult, only has to be written once.

>In general, the farther the drop from the extruder head to the surface, the more the filament thins out.

Not in my experience. As it leaves the nozzle it widens and then sets. If the head is far above the the work surface the filament will not stick.

---

[www.hydraraptor.blogspot.com]

I just thought of a wierd idea... what if the extruder printed a hollow tube, rather than a solid thread and could manage the gas pressure within that tube...

This should stop, or at least seriously reduce shrinkage.

The print should be able to be run a lot quicker.

The extruder could spray CMYK mixed pigment inside the tube during the print for colouring if the material used is transparent.

The strength to weight ratio should also go up by a reasonable margin, dependent on the wall thickness of the tube.

To finish or start a line, the extrusion nozzle would twist to seal the ends of the tube and reduce straggly bits.

If the gas is hot, it should mimic some of the benefits of a heated chamber.

You'd use less material per print.

You might even be able to vary the gas pressure while printing to add an extra degree of resolution.

-------------------

Does this at all sound like a feasible plan, or should I go and invest in a couple of rolls of rubber wallpaper and a no-smoking jacket...?

-------------------

*edit* - I have now made up a rough piccy of the idea, so I thought I'd add it to the post..

note - I have made no attempt at doing any sort of accurate scaling on this pic..

-------------------

*2nd edit* - if metal powder was another pigment, it may be able to print an insulated conductor, also, if florescent pigments were used in conjuntion with the right arrangement of conductor, lights could also be possible.. haven't really looked into this much, but it might just work... which would be pretty cool

Edited 2 time(s). Last edit at 02/10/2008 01:52PM by deadgenome.

Hi deadgenome,

... the idea isn't so bad - but if your filament is fluid enough, to stick on the surface, it isn't stable enough to hold the tube-form against pressure (it'll blow up or collapse partially), so you can't control the diameter with pressure-changes.

I think you should evolve your idea with two different materials - the core is from a stronger and denser material and the hull from a softer and more adhesive one, so the tray would stick to the surface and the thermal contractions of both materials could interfere to a non-contraction compund ...

Viktor

Cheers for your input Viktor... I was thinking of the gas as being very, very close to atmospheric pressure... but I take your point. Another possible variation on the theme could be having the middle being foamed plastic, that might provide enough structure to prevent collapse, or it might just be introducing unneeded complexity into the extruder, though I'm assuming that reducing the distortion through shrinkage is worth at least a little bit of extra complexity.

Um, actually both nophead and I have overclocked the extruders. I got it up to about 20 mm/sec in an extrusion-only and I know nophead has actually printed things at up to 16 mm/sec.

16 mm/sec will print a max of about 30 cubic centimeters/hour assuming a .8 mm diameter extrusion thread (0.5 mm^2 cross sectional area). That's nearly 20 kg of printing maximum per month. That's a LOT of product.

I think that there are other practical problems that begin to get in the way when you start overclocking reprap machines. The one I keep running into is noise. Using threaded rods on Tommelise at over about 8 mm/sec tends to be make a considerable racket. I'm trying to quiet that down by switching over to nylon threaded rods, but the jury is still out on that one.

I've gathered from some things that Vik has said that Darwin isn't exactly silent either, but I don't know for sure what kinds of noise levels that he's actually getting. I think I heard him say that he's averaging 6-7 mm/sec.

Forrest I would take a serious look at using a lead screw it's more expensive than rod but is designed for the job, my CNC router runs at 40mm/sec and motor noise is all you hear.

---

Ian
[www.bitsfrombytes.com]

I may well get to that. There's even a little company that manufactures them to micron precision down the coast from me that I've talked to. The cost of the lead screw isn't what bothers me nearly so much as the cost of the thrust collars. Those things cost about $30 each and are just a piece of plastic with a few wires in. :-s

Been thinking about The Guy's electrospinning printer idea & have been wondering if using continuous inkjet technology ( apparently patented in 1867 by Lord Kelvin, according to Wikipedia anyway - [en.wikipedia.org] ) in the same way might work. So instead of firing a constant stream of liquid, you are firing loads of individual drops at a very high rate towards a pretty much infinite range of points along a line which you then scan back and forth across the total build area. To get around the issue of the drops cooling, you could either make sure that they are very, very hot before firing them, or focus a laser onto the area that you know a drop is landing, so it is nice and sticky when the drop arrives. A benefit of doing things this way round is that the amount of deflection is controlled by the charge given to each individual drop, while the magnetic field itself stays constant. This means that you can be firing loads of drops at the same time, in a range of materials, thus enabling an extremely rapid build.

----------------------

*edit* - on the extruder idea I posted further up the page... if I kept the configuration the same, but made the pigment also be the higher melting point plastic that Viktor suggested, then this should mean that the idea of varying the width of line using gas pressure might still work. The other thought I had on that today, is that a very complex internal structure to the print line could be attained by having a large diameter barrel in which the filament's cross section is arranged, which then narrows to the diameter required for the actual printed filament, much like the way that a stripy toothpaste dispenser works.

Edited 1 time(s). Last edit at 02/11/2008 11:39AM by deadgenome.

I have been following the RepRap for about a month now and I am becoming more interested as time goes on. But this is my first post (so Hi to all the RepRappers).

One idea that I though of is to have a 2 dimensional array of heads with the heads evenly spread across the array and all attached together so they move as one object.

For instance, if you used a 3 by 3 array with a work space of 30cm, you would have heads at the X=5cm, X=15cm and X=25cm on the y=5cm line, then this is repeated again on the Y=15cm line and Y=25cm line.

This way each head is only needing to cover an area of 10cm by 10cm on each layer. Where a line continues within one layer from one cell (the 10cm by 10cm area each head covers) the head in the other cell will have filled that part in.

Because each cell is small (and you can add more heads if you have the space), then it might be feasible to just scan the heads over the 10cm by 10cm area and only turn the head on (extrude) where you need the material to be deposited.

The old dot matrix printers used a system like this (except that the head continued on and started a new set of cells further along the page). The difference is that unlike a dot matrix printer, the print heads cover the entire print surface (the base of the print box).

These are certainly interesting notions.

IMHO, one of the key things to reprap success is getting extrusion to be repeatable and controllable. (I'd rather get a good part slowly, than a deformed one quickly. Though for some applications, quick/dirty may be OK.)

Unfortunately, the extrusion process is not simple: plastic is flowing, cooling, (hopefully adhering) and may be being stretched/deformed by the motion of the head -- all at the same time. Right now, we control temperature (of only one point in the extruder), and roughly control the flow rate of plastic via the temp and extruder-motor duty cycle. At this point, the motion of the extruder is likely far better controlled than the other process variables.

The temperature (both in and probably just outside) of the nozzle influences the flowrate. We may want to do something to measure (and actively control?) the temperature *profile* outside the nozzle, to control how the filament solidifies and bonds to the underlayer. Perhaps a little heatlamp attached to the extruder head?

Similarly, it would probably be helpful to be able to measure the pressure of the melted plastic (that is about to transit the orifice.) Sensing that would probably facilitate controlling the flow rate.
However, I don't know how to measure the pressure of hot, viscose melted plastic? (Especially without significantly complicating the extruder. I'd be interested in any ideas on good way to measure this. Similarly, ideas for valves that work with melted plastic (and can return to operation after yesteday's plastic has solidified) are an attractive notion, but I don't yet see how to make them work reliably. Ideas?

-- Larry

-- Larry

Quote

One idea that I though of is to have a 2 dimensional array of heads with the heads evenly spread across the array and all attached together so they move as one object.

This rather sidesteps the issue of making a faster or otherwise better extruder though

It would only work with large parts, or multiple small identical parts, so it lacks a certain flexibility.

You've not only got the expense of extra non-reprappable extruder components (screw, barrel, insulator, nozzle, bearings, motor, rotary encoder), the need for extra controlling hardware (1 PWM channel, 1 DC motor channel, 2 lines for the encoder? 1 line for the nozzle valve), but you have to make a tougher and more rigid frame to bear the extra weight and the motors to push it around, and then deliver that much more power to the final device to keep those big motors going and all those extruders hot...

Further in the future, what do you do when you have exchangeable toolheads? Make three of everything and have big tool carousels to hold them all?

This wasn't intended to sound mean or confrontational, by the way. I've had a bad day