Full color 3D printer? I guess they forgot something...

[MadModder]

https://www.kickstarter.com/projects/nixtek/tomato-full-color-process-desktop-3d-printer
Now, where did that white filament go? 
Sure, you can load white filament into it instead of black, but then you can't do really dark colors.

[MadModder]

No, they're not fake.
However, you can't do CMYK without a white background.
The printed objects with white in them has no dark colors, so those are CMY+white. Black filament switched out.

He clearly says in the video that there are four colors used, cyan, magenta, yellow, and black. (And there are only four ports to feed filament into.) And while he declares the four base colors, the video shows that there is white filament in one of the four tubes...

Theoretically, you would be able to do black from CMY, but in reality it just becomes brown.

[tom66]

Is 3D printining additive or subtractive with colours? I want to say it's additive (RGBW would work best), but I'm not certain.

[donotdespisethesnake]

Definitely subtractive. It's no different to mixing inks or dyes.

I was going to say "in 2D" , but of course ink on paper is actually 3D because the ink is composed of molecules. The third dimension might be very small, but it is non-zero.

[free_electron]

50 micron resolution my ass. 50 micron layer thickness maybe.  but side to side ? just look a their printed tomato. it's got 'fuzz ' growing all over it.

when is someone going to make a 3d printer that can print a polished outer surface ?

[Mechatrommer]

50 micron resolution my ass. 50 micron layer thickness maybe.  but side to side ? just look a their printed tomato. it's got 'fuzz ' growing all over it.

its theoritically possible on all axis when you calculate it by hand. but there are many things getting in the way, such as vibration and resonance, friction, backlash, missed steps, temp fluctuation and the precision of the amount of material squirted from the nozzle, too much you get blobs everywhere, too few you got empty spaces, too hot you got ooze strings etc. the tomato you saw probably not printed with 50micron height resolution. and the 3d file used to generate gcode apparently was not set detailed enough face resolution thats why you see "not round" tomato, also current 3d printers do not support native arc command, so they have to approximates with few number of lines.

when is someone going to make a 3d printer that can print a polished outer surface ?

when there is technology such as 0.1 degree stepper resolution, when there is motor/stepper strong enough per weight and watt to squirt plastic at massive volumes per unit time, nozzle assembly that can hold such massive forces. because at current technology, if you want to print 50 micron resolution esp on z-axis, you'd rather to not want it. say compared to printing at 300 micron (0.3 mm), printing time at 50 micron can be more than 6X multiples. sounds not so much of a deal? yes if a multiple is a matter of seconds or minutes. a multiple in 3d printing can be hours or days. to add pain to the insult... when there is power outage, the feeling is like want to kill a creature.

[free_electron]

when there is technology such as 0.1 degree stepper resolution,

steppers ? when will the 3d printing world switch to servo drives ? steppers are stoneage junk. pick and place machines, milling machines have long switched to dc drive motors with optical encoders. way higher resolution, way faster speeds.

sorry if i sound ascerbic but i have grown very skeptical about all these things. every kid and his uncle jumps this 3d printing stuff but nobody makes true improvements to the systems. they are all invariable contraptions made form duct tape, plywood and rusty screws driven by recycled floppy drive motors.  i am still waiting to see the first REAL , affordable ( read: +/- 1000US ), easy to use, doesn't need 14 days a month in adjustments to print a small thingie ) 3d printer.

i got one of those printrbots now ( the 600$ one ) and that is about as good as it gets. i've managed to get some very good results using PLA as material. I've sent the same files to people with those 'delta' printers  , makerbots and others and their results are all crap compared to the printrbot . curled edges, don't fit right ,  ragged edges, layer shift, one even clerly had the stepper skip a step somewhere.. there was a clear 'shift'.. not good if you print two pieces that have to mate perfectly.

still. the printrbot output cannot compete with the file i got printed on an objet , in nylon.

so yeah, i'm very ascerbic when i see the umptiets '3d printer' kickstart.... as i'm waiting for the one that really brings and OUTPUT QUALITY improvement, or plug-n-print capability.  I don;t need to 'adjust' my laser or inkjet printer. it just works.

when can we have such a 3d machine ?

[helius]

all of those requirements are met today if you just use SLA instead of FDM technology.

[Mechatrommer]

steppers ? when will the 3d printing world switch to servo drives ? steppers are stoneage junk. pick and place machines, milling machines have long switched to dc drive motors with optical encoders. way higher resolution, way faster speeds.

i agree with this point. probably they dont want to mess with firmware built from scratch, they dont want to risk having a heart attack for coding kinematics and momentum conservation law into the software in different dedicated mcu etc. tracking 4-8 encoders in one single arduino can be a task you know? so the easy and faster way is not reinvent the wheel, use the work of many, usually reprap community. and really, imho the stepper resolution alone is not the bottleneck, the bottleneck is in the extrusion rate, that tiny motor in the printhead, it needs to be powerful and small. i can ramp my 3dprinter to superfast, but printhead motor and gear wont cope. so what i meant is if you want superfine printout, you need higher resolution steppers, or dc motor whatever, but only feasible when the extrusion rate can be ramp up too. so for now, switching from stepper to encoded dc motor seems impractical. + with mechanical difficulties i've stated earlier.

sorry if i sound ascerbic

well... you are free electrom arent you? who doesnt know you?

i got one of those printrbots now ( the 600$ one ) and that is about as good as it gets. i've managed to get some very good results using PLA as material. I've sent the same files to people with those 'delta' printers  , makerbots and others and their results are all crap compared to the printrbot . curled edges, don't fit right ,  ragged edges, layer shift, one even clerly had the stepper skip a step somewhere.. there was a clear 'shift'.. not good if you print two pieces that have to mate perfectly.

different people with different result, for now this cheap 3d print market, its more towards the "owner skills". PLA is easy to print but it has low melt temp not suited for prolonged outdoor use. ABS is much resilient and preferable, but harder to print correct. you need enclosed heat regulated chamber, but many are afraid to breach that stratasys patent.

when can we have such a 3d machine ?

we both, and many others are waiting... cheers.

[Mechatrommer]

all of those requirements are met today if you just use SLA instead of FDM technology.

and at what cost/print volume? maintaining liquids seems not too fun. i'm still sceptical on the resolution or "print smoothness" thereis, i guess the quality is not so much different from FDM, i'm yet to see a promising report on SLA print output compared to FDM.

[free_electron]

all of those requirements are met today if you just use SLA instead of FDM technology.

yes, bat not at my 1000$ pricepoint !

[Mechatrommer]

all of those requirements are met today if you just use SLA instead of FDM technology.

yes, bat not at my 1000$ pricepoint !

there are within your budget, bat there can only print 4cm toys.

[ralphrmartin]

50 micron resolution my ass. 50 micron layer thickness maybe.  but side to side ? just look a their printed tomato. it's got 'fuzz ' growing all over it.

when is someone going to make a 3d printer that can print a polished outer surface ?

Océ make one for fine art reproduction with 2 micron resolution. They dont sell it though, it's only for in house use and they seel the output at 25k per print...

[Corporate666]

50 micron resolution my ass. 50 micron layer thickness maybe.  but side to side ? just look a their printed tomato. it's got 'fuzz ' growing all over it.

when is someone going to make a 3d printer that can print a polished outer surface ?

They are now claiming 10 micron resolution. And they are not claiming it can position that accurately, they say "Capable of printing 10 micron resolution".

They are totally full of shit!  It can't even position to 100 times 10 micron.  A micron is 0.000039 of an inch.

A modern high end CNC milling machine can produce results accurate to perhaps 0.001" if the part is held carefully and the programming is good.  If you tell an advanced machine shop you need 0.0005" results, it is probably not possible to make the part, unless you grind it after the fact to tolerance (but you can't do that on curved surfaces).

These guys are claiming 0.00039" printing accuracy, which is dramatically better than you would get with a $500k CNC machine and it would be very difficult to achieve that accuracy on the CNC without lots of trial and error, temperature compensation and post-process grinding.

It's too bad people aren't held accountable for their claims.  Bunch of scammers.

[ruffy91]

A micron is 0.001mm. Mikron has CNC machines that are capable of positioning with 1-2 micron accuracy reproducible. Why shouldn't it be possible? Glass scales are commonly available with 0.5 micron resolution.

[Corporate666]

A micron is 0.001mm. Mikron has CNC machines that are capable of positioning with 1-2 micron accuracy reproducible. Why shouldn't it be possible? Glass scales are commonly available with 0.5 micron resolution.

Not sure if this is a serious post... but a Mikron CNC machine has absolutely nothing in common with this Tomato 3D printer anymore than a folded paper plane has anything in common with a Boeing 787.

Which Mikron machine can reliably position within 1-2 microns?  The best I have seen is +/- 2 microns, which is - as best - double what you wrote, but also a specific metric for a specific function (table positioning accuracy) of a purpose built machine that costs more money than most people earn in their entire lifetime.  And even if a machine can position to +/- 2 microns, that doesn't mean it can machine something to a 4 micron tolerance - it can't.

The Tomato people are claiming their machine can actually print at 10 micron resolution - which is absolute and total bullshit.  Like me claiming my toy rocket from the hobby store can fly to the moon, land an astronaut, and bring him home again.

Let them print a plus sign and show it's within 10 microns of the size in the CAD program and geometically accurate (parallel/perpendicular).  Let them print a circle, cylinder and sphere and show that they are within 10 micron.

The only way this machine is hitting 10 micron anything is if it prints a blob of material and leaves some chaff hanging off the end of the blob - I'm sure that, at some point, the thickness of that stringy leftover will be 10 microns within the intended size.  But that is an accident, or the nature of the machine... not a design functioning as claimed to produce 10 micron results.

[cavemaneca]

As someone who actually owns a decent quality 3D printer (though admittedly it was self constructed from mostly kit components), I can see where they would get their 10 micron spec from.

While I haven't tested it recently, I first did a large number of calibration tests with a 20mm cube. The absolute MINIMUM layer size I can reliably print at is 500 microns. In testing this means a variance of +/- ~200 microns for the finished part. For the X and Y axis, the accuracy is much better, at least on flat surfaces. The end result was around +/- 70 microns, with the largest variance being ~90 microns from bottom to top of the cube.

So, if they made a simple short object, measured the "smoothest" faces on the opposite sides of a rectangle, and had a more stable machine than my own, they could measure a +/- 10 micron variance. More than like though your average print would be closer to +/- 50 microns on the X and Y axis. The Z axis would likely not be under +/- 100 micron variance unless you specifically chose a object with a layer size that evenly divides into the object height.



For a final note I'd like to give my own opinion on their project.

The only original part I can see is a custom extrusion head. Nearly everything else is basically a copy of the Ultimaker design that is already a few years old. Additionally they are selling at a price point that is absolutely ridiculous for the market considering the likely cost of that single custom part. The at cost value of their product is probably under $500. Either their inept and paying way more than they should for the parts they're getting, or they are greedy and trying to expand their price point to fit their profit margins.

I would not recommend this Kickstarter over getting a proven Ultimaker II. Heck, if you want the extra color buy the U2 second extruder addon(or build your own) and their software handles it quite well. Their multi-chromatic selling point is mostly BS and their wonderful examples mostly show them changing individual layer colors instead of advanced color objects. I doubt their software can do anything more than adjust colors on specific layers, whereas you would want this to be able to actively change colors mid layer.



Overall score : 3/10