Hugh Lyman won the Desktop Factory Competition

It's official, Hugh Lymans' filament extruder won the $40,000 prize in the Desktop Factory Competition.

[techland.time.com]

[www.thingiverse.com]

Any one know when extruder kits will start being sold?

Edited 1 time(s). Last edit at 03/04/2013 05:08PM by PeteD.

I am still skeptical about the capability of this machine to make usable filament. As far as I can tell Lyman hasn't published any information on tolerance and process capability (ie how consistent the diameter is).

In his original BOM he does publish the filament diameter measurements on a 10 foot length at 12 inch intervals. The ten readings averaged 1.73mm with a +0.02mm, -0.04mm variance.

Edited 1 time(s). Last edit at 03/04/2013 05:44PM by PeteD.

Also the speed involved. 8 inches per minute? That would take forever to go through a batch of pellets. It would take just as long to make the filament as it would to print anything with it!

Or power consumption. At 8 IPM

Let's assume that 36" of 3mm ABS filament weighs 6g (according to my scale). So 1lb of filament is right around 225 ft. (1'=2g, 454g/2g=226') Therefore it'll take a little over 5.5 hours to extrude a pound of filament (226*12/8/60).


What are the running costs per hour? The extruder could be insulated to minimize heat dissipation and increase efficiency.

I'm not so concerned about tolerance but consistency - how uniform is the filament along it's full length? Nevermind 10'. What about the results over 1000'?

Questions to be answered...

Hopefully we'll like the data.

Edited 1 time(s). Last edit at 03/04/2013 05:51PM by akhlut.

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

I get 12"/minute with the Filastruder, I can do 1kg letting it run overnight.


PomeroyB Wrote:
-------------------------------------------------------
> Also the speed involved. 8 inches per minute? That
> would take forever to go through a batch of
> pellets. It would take just as long to make the
> filament as it would to print anything with it!

Not sure about Lyman's, but with the Filastruder I use 50w, and it takes 16 hours to do 1 kg. That's 0.8kwh, or about 10 cents of electricty where I live. Plenty cheap.



akhlut Wrote:
-------------------------------------------------------
> Or power consumption. At 8 IPM
>
> Let's assume that 36" of 3mm ABS filament weighs
> 6g (according to my scale). So 1lb of filament is
> right around 225 ft. (1'=2g, 454g/2g=226')
> Therefore it'll take a little over 5.5 hours to
> extrude a pound of filament (226*12/8/60).
>
>
> What are the running costs per hour? The extruder
> could be insulated to minimize heat dissipation
> and increase efficiency.

greenman,

That's pretty good. What tolerance and variance data do you have? And do you have a 3mm die?

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

The die is 1.59 (#52 drill bit) which produces 1.73 filament at around 175C (measured at the die, the heater would be hotter). With mine, I'm finding that the feed rate can vary by as much as an in/min. It seems that variances in back pressure in the extruder changes the speed of extrusion while keeping the diameter in the +- .03-.05mm range. I've tried drawing down the plastic from a larger die using Lyman's spool winder. I replaced the latex rollers with a Wade's extruder to make certain the the pulling speed was absolutely consistent. I couldn't keep the tolerance tight enough however. When stretched straight to the puller, the plastic quickly cools and becomes a mostly rigid rod. When the melted plastic tries to come out of the extruder faster, it gets bunched up behind the hardened plastic which is moving at a fixed rate.

I've managed to extrude and wind filament with a .05 tolerance by letting it drop into a loop on the way to the spooler -

akhlut Wrote:
-------------------------------------------------------
> greenman,
>
> That's pretty good. What tolerance and variance
> data do you have? And do you have a 3mm die?


I've measured ~50 times over the course of a kg of extrusion, and I see +/- 0.03 90% of the time, with a few +/-0.07. I can upload each individual measurement later this week, or even send you a sample if you'd like.

No experience with 3mm, as I don't use 3mm printers, but I can give it a shot later this week!

If it's not clear from above, Ian Johnson is also using a Filastruder.

Edited 1 time(s). Last edit at 03/04/2013 06:35PM by greenman100.

What's the fastest speed anyone's been able to run these at reliably? 1 foot per minute is still pretty darned slow...

greenman,

That looks like quite a good setup.

When are you bringing a 3mm to market? You're out of the beta stage, yes?

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

... if you can make even thinner filament - 0.3 to 0.1mm thick wires would be perfect ... can be +/- 0.05 off too without problems ...

I'm developing an alternative FFF methode, that uses thin wires directly, without melting them in a reservoir

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

PomeroyB Wrote:
-------------------------------------------------------
> What's the fastest speed anyone's been able to run
> these at reliably? 1 foot per minute is still
> pretty darned slow...

Meh... It's faster than what I print at, so I consider it a win.

I haven't tested a 3mm die personally, but changing the die might be all that's necessary. It's possible I'll also need to spec out a faster motor. I'll try to test one this week and let you know.


akhlut Wrote:
-------------------------------------------------------
> greenman,
>
> That looks like quite a good setup.
>
> When are you bringing a 3mm to market? You're out
> of the beta stage, yes?

PomeroyB Wrote:
-------------------------------------------------------
> What's the fastest speed anyone's been able to run
> these at reliably? 1 foot per minute is still
> pretty darned slow...


The issue is that motor power scales with the cube of the linear extrusion speed, roughly speaking. Going faster would require a much pricier and much larger motor. I cannot print as fast as I can produce filament. I've started selling rolls to friends because I've stockpiled so much of it.

This seems quite possible. I have some .4mm - 1.0mm drill bits, so I'll give it a shot.

Tim


VDX Wrote:
-------------------------------------------------------
> ... if you can make even thinner filament - 0.3 to
> 0.1mm thick wires would be perfect ... can be +/-
> 0.05 off too without problems ...
>
> I'm developing an alternative FFF methode, that
> uses thin wires directly, without melting them in
> a reservoir

VDX, I can give it a try with the puller. Something that thin would probably work better being stretched down to diameter. Filament that thin may be flexible enough to bend when the extruder speeds up rather than change diameter.

3mm filament would probably need a faster motor or bigger barrel. It feeds really slowly with the Filastruder, since compared to the 1.6mm die it's like taking your thumb away from the garden hose so it doesn't squirt anymore. It can be tricky to guide 1.75 filament where you want because of the stiffness. Any attempt to bend it can push it sideways at the die, and if you don't straighten it quick enough it sets into a kink. I imagine 3mm would be even stiffer and therefore more finicky.

An interesting thing in Lyman's video. Watch the chain drive at about 1:20 in the video. Notice the chain riding up the sprocket on top? Either he has some really cheap grade chain and sprocket, or the chain has been stretched, which is why you see it riding up because the pitch of the chain doesn't match the sprocket. So a lot of his power could be lost in the chain drive, and both the chain and sprocket will wear quickly.

Gary H. Lucas

Quote

The issue is that motor power scales with the cube of the linear extrusion speed, roughly speaking

Why is that? I thought flow rate was proportional to pressure so power should be a square law.

---

[www.hydraraptor.blogspot.com]

VDX Wrote:
-------------------------------------------------------

> I'm developing an alternative FFF methode, that
> uses thin wires directly, without melting them in
> a reservoir

Let me guess: you want to melt it by a laser.

---

Detlef

Excalibur Hotend

reprapclean32.png

reprapzone.blogspot.de

Correct, flow rate is roughly proportional to deltaP across the die, but in my experiments I'm seeing somewhere between a quadratic and cubic relationship. Not sure if there's other factors at play, like turbulence or varying motor/gearbox efficiencies in my admittedly limited sample size of motors (~ 6 different types).

Even if the relationship is quadratic, you're looking at something like 100x larger motor to achieve 10x higher flowrate. If you can find a gearmotor that is rated at half a horsepower, 10RPM, ~85 ft-lbs torque for a reasonable price, let me know! Unfortunately this will mean uprating all the fittings as well, as operating pressures will go from ~150PSI to ~1500PSI.

You'd also need a heater capable of ~500w instead of ~100w, which doesn't exist in the band heater form factor I've been using, certainly not at 15v.

It's just more practical and easier to wait overnight to make a spool of filament. Doing it faster gets expensive, fast.


nophead Wrote:
-------------------------------------------------------
> The issue is that motor power scales with the cube
> of the linear extrusion speed, roughly speaking
>
> Why is that? I thought flow rate was proportional
> to pressure so power should be a square law.

Edited 1 time(s). Last edit at 03/04/2013 08:17PM by greenman100.

I imagine cooling will also be a problem with high speed filament extrusion. As it stands now, most desktop extruders look like they just allow time and room temperature to cool the filament. That may not work well if the filament is coming out at a higher speed.

theodleif Wrote:
-------------------------------------------------------
> VDX Wrote:
>
> > I'm developing an alternative FFF methode, that
> > uses thin wires directly, without melting them in a reservoir
>
> Let me guess: you want to melt it by a laser.

... right guess

For the CO2-laser any plastic material (even fishing line) would be sufficient, but for the much easier to handle diode-lasers it should be a dark (good absorbing) plastic ... best black or dark green ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

For .3mm filament it would probably be best to go in two stages. Make 1.75ish filament, and then run that through a 3D printer extruder/hotend turned sideways. I have a 6mm bolt threaded into my die so I was able to screw on a .5 nozzle I had around. The output was a bit wiggly, and eventually enough pressure built up that I was concerned about burning out the motor. With the 1.59 die I was able to draw down to .6mm running at almost 1meter/minute but I couldn't get the Wade's to go any faster. It would probably be fine with a 1mm die.

Do you think melting the filament with a laser as it comes out would blend the layers together more and prevent ridges? Or is it more about achieving higher extrusion speeds since there is no backpressure from melted plastic inside the nozzle?

... it's more about feeding the thin wire with minimum force and high-speed - as fast as you can melt/fuse it.

I've done this with thin metal wires and kW-lasers, but should be even simpler/easier with plastic wire and a diode-laser head ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Problem is that to get good speeds without needing too much power you need either a conical shorter auger or one with a variable pitch (the latter is how injection molding machines works, the former is used in drawing extruders). This way the back part of the screw is under less pressure and so friction

One idea I had was to use the augers from a meat grinder (usually conical auger, but some have both cone and variable pitch), in a 2 steps system to get full venting (very important to not have bubbles imho), should work and be cheap enough.

for reference, an industrial extruder for plastic injection of 1.5 kw (screaw+heater) pass through 3.5 kg/h of ABS pellets.

PomeroyB Wrote:
-------------------------------------------------------
> What's the fastest speed anyone's been able to run
> these at reliably? 1 foot per minute is still
> pretty darned slow...

Compared to what?

I would guess that you'd have to consider the flow rate of the plastic, and what would happen to its integrety if you heated it up enough to be extruded quickly. I wonder if the usability of the plastic would suffer? 60 feet an hour seems pretty reasonble to me, unless you're looking at this from a filiment supplier angle, which would kind of go against the whole point of one of these machines...

Ah, you got me there. I guess "slow" is based on one's frame of reference.
If the math akhlut did above is correct, 1 pound of filament is approximately 225 feet. 2.2 pounds (1kg) seems to be a "standard" spool size, and that is approximately 495 feet.
So at 60 feet per hour, it will take 8.25 hours to finish 1 spool.
If you can crank it up to 8 feet per minute, you can do just under 1 spool (480 feet) in just one hour.
I'm not sure why others don't feel that being able to create an entire spool of filament that fast is an incentive to improve these machines.

As for usability and integrity, I'm not sure what you mean. There's no reason for there to be a difference in the temperature of the plastic when you are extruding fast, rather than extruding slow... Velocity does not affect melting points. Heat will be needed to be added to the system faster, since it will be leaving faster, but that should not affect the integrity of the plastic.

... with higher speeds and then needed 'sharper' temp-gradients the controlling of the filament diameter after extruding and until solidifying is much more ambient sensitive and should be tested first ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]