Requirements for a bigger build volume

[jonnybischof 60]

I've been thinking about making a larger version of the UM. (something like 500x500x300 mm build volume)

If it were trivial to do that, the UM would probably already be a bit bigger. But what are the main requirements in order to make a bigger printer?

- The electronics and firmware shouldn't have any problem handling larger spaces?

- Cura doesn't seem to have a problem with a 500x500x300 custom machine.

- The shafts will be longer, and therefore any straightness-issues may become bigger. So you can't use the cheap banana-style shafts. Ok, but good quality industrial shafts come in 2000mm standard lengths, so they should have some degree of usefulness or they wouldn't produce them in that length?

- Longer belts shouldn't be a problem?

- When does build height become a problem?

- The frame needs to be straight. No problem if the right material is used...

- Are there any other problems I'm not seeing?

 

[skint 61]

Great question Jonny! Have been thinking about this too recently. Once you have the basics done how hard is it to upscale?

I seriously doubt that upscaling a current UM by 25% build volume would decrease the stability of components by much? What about the ambient temperature differences, would they effect build quality?

Let`s build!

 

[jonnybischof 60]

Ambient temperature shouldn't have more of an effect than with the standard size, I guess (?)...

If you use the same 6mm plywood, maybe you get some negative effect. Though I have to say the standard UM frame is amazingly sturdy. But wood does have very high, and most importantly uneven, thermal expansion coefficients. Aluminium is not that much better (and it's useless when you want to build a heated chamber), but chrome-nickel steel should have pretty much no measurable thermal expansion in this application.

I was thinking of making a double-layer construction: 3mm plywood over 1.5mm chrome-nickel steel. That makes for a very sturdy construction and provides good thermal insulation at the same time.

Maybe, the stuff I'm using for my new build plate could be useful, too. But it's really expensive... Showstopper for 99% of all hobby builders.

 

[Dim3nsioneer 557]

What about the Bowden tube? It would have to be longer too. Might that give some problems with e.g. retraction?

 

[tommyph1208 55]

While its not a problem to get some open ended belts of ~ 1 m. length for such a size printer, wouldn't it be ALOT harder to keep those belts tight?

Maybe look into some stronger belts or better ways of tightning?

[jonnybischof 60]

Right - longer bowden tube needed. Shouldn't be much of a problem though. But it's a needed part...

You're right, retraction length would probably need to be increased which also increases retraction lag. Could be that increasing the build volume means losing somewhat of printing quality...

I think I'd definetly need bigger diameter pulleys, because the belts must be kept apart from each other. I don't know about tension. Does it really make a difference?

Stronger belts could be an option, if necessary.

 

[drayson 75]

If the whole machine is bigger, what would hinder to place the feeder app. 400mm in the middle above the UM on a gantry, pointing down. So no ~180° bend for the bowden is needed, nearly straight line so less problems with retraction lag...

Yust my two cents... :-)

 

[Dim3nsioneer 557]

If the whole machine is bigger, what would hinder to place the feeder app. 400mm in the middle above the UM on a gantry, pointing down. So no ~180° bend for the bowden is needed, nearly straight line so less problems with retraction lag...

Yust my two cents... :smile:

 

That's almost three cents...

 

[jonnybischof 60]

*Muaha*

I wanted to suggest the very same thing, but then I found a problem with it:

If you do that, then you still have to make the bowden tube long enough to reach into the corners. But when that assembly moves right into the center of the platform, it will develop a lot of (most probably too much?) tension because the bowden will be much too long.

So we'd have to compensate for that...

If the feeder assembly were able to slide up and down, it could compensate the bowden length. But I don't know if it would suffice to have it "move freely on a vertical slider" or if the friction / needed force would still be too great. Also this would take a lot of space upwards (not a real problem, but it means a huge loss in elegance...).

Or you could mount the whole feeder system on rubber bands, being able to be dragged around. That also means a lot of force on the printhead which is bad. I guess even just a little more force would already be a problem because the head is supposed to move very quickly. If you have a large build volume, being able to print fast is very important or you'll have 100+ hour prints...

/edit:

If you say "gantry", do you mean another x-y gantry on bearings / bushings which will move together with the printhead?

Maybe it could work if you add a whole second gantry, with shafts, motors, pulleys & belts that moves together with the printhead. If it were just "free floating" on an x-y shaft gantry, there would probably be too much friction...

 

[drayson 75]

Ok, sorry, my mistake - I ment some sort of a gate where the feeder is fixed/mounted in the center... (so no moving gantry)

 

[Daid 306]

For a machine that big, it might be better just to mount the feeder on the head, as everything is getting heavy compared to the feeder anyhow.

Stiffness of the axis is also very important. I'm not sure if the square gantry of the Ultimaker scales very well to larger sizes.

 

[jonnybischof 60]

For a machine that big, it might be better just to mount the feeder on the head, as everything is getting heavy compared to the feeder anyhow.

Stiffness of the axis is also very important. I'm not sure if the square gantry of the Ultimaker scales very well to larger sizes.

 

That's exactly where the UM excels: feeder not mounted on the head. It makes some things much easier... If the printhead gets heavy, you need a much stronger gantry (or move much slower) etc. etc. ...

So, it is definetly not trivial to make the UM bigger. That's a good first conclusion I think.

I guess I won't attempt to build one. Seems like a lot of work without guarantee that it will actually work well enough.

(Besides, I have enough other projects waiting to be realised o.O)

If you could just lasercut some bigger panels, buy longer belts, shafts and maybe a few other parts, that would be nice. But it seems there would be some major problems (like the bowden length) that can't just be fixed without having to redesign half the machine...

 

[randyinla 14]

I keep wondering how hard it would be to increase the Y length of the UM2. Remove the front panel, cut new, wider side panels. Longer rods/belts wouldn't be that difficult. But then what to do about a longer heated build plate and glass?

Increasing the X seems more challenging due to the placement of the steppers/gears etc.

I'm only talking a few inches. Like, from 8" to 12" so I could print full face masks/helmet or even shoe-sized objects... Aurora race tracks etc.

 

[Dim3nsioneer 557]

Maybe your questions will be answered shortly.... :ph34r:

 

[zoev89 73]

Protospace in Utrecht they have some experience on this subject. I think trying to stick to the same possible resolution might be trap. If your model is bigger you might want to trade in resolution. Otherwise your print time explodes.

[drayson 75]

Ok, Dim3nsioneer seems to know something - related to the next weeks announcements??

[Dim3nsioneer 557]

I can definitively confirm the first part - I know something - otherwise I would not be able to write... :lol:

More serious: The elefant project in the Netherlands recently used an add-on kit for the UM original which allowed them to print several meters high objects. So the technology for extending Ultimaker axes exists...

 

[3dcase 37]

http://www.calumdouglas.ch/3d-printing/ultimaker-3d-printer/vertical-down-filament-feed/

Perhaps you are Looking for something like this Jonny?

Bookemarked it because its on my to do list For a UMorginal you could probably make the bowden tube shorter.

Just add some springs to give the feeder some movement and reduce the weight.

With a standard bowden tube and some springs, you can cover a 400x400 platform.

Edit....

pretty sure the owner of the site is among the forummembers somewhere....

 

[randyinla 14]

The elefant project in the Netherlands recently used an add-on kit for the UM original which allowed them to print several meters high objects. So the technology for extending Ultimaker axes exists...

 

I stumbled across their website with the live cams a few weeks ago. I thought the inverted UM originals and seemingly limitless Z height was AWESOME!!!!!!!!! Looked like they had a floating feeder system too. The feeder seemed to be moving on springs or rubber bands or something. Pretty cool stuff, indeed!

 

[Daid 306]

So, it is definetly not trivial to make the UM bigger. That's a good first conclusion I think.

I guess I won't attempt to build one. Seems like a lot of work without guarantee that it will actually work well enough.

(Besides, I have enough other projects waiting to be realised o.O)

 

One of our interns did a study on a bigger printer. He actually build it (it's big, 1000x1000x1000 I think). It can be done, but you better switch to a CoreXY system. He also had problems getting enough extrusion, even with a 0.8mm nozzle. As you want much faster printing if your printer is that much bigger.

Not sure how many details I may share, but it clearly shows it can be done. He did the math on all the axis and stuff to make sure it's strong enough. The Z stage was one of the tougher nuts to crack, as it gets really flexible when it's that big.

 

[xeno 108]

The two biggest problems I encountered when I enlarged my UM to an UM+,

was a sturdy frame and very sturdy Z-platform.

everything else work very easy, the UM original lends itself very easily for extension

 

[jhertzberg 19]

At some point, do you think it makes sense to move the CoreXY gantry on the z axis to simplify making a level (and hot) bed?

 

One of our interns did a study on a bigger printer. He actually build it (it's big, 1000x1000x1000 I think). It can be done, but you better switch to a CoreXY system. He also had problems getting enough extrusion, even with a 0.8mm nozzle. As you want much faster printing if your printer is that much bigger.

Not sure how many details I may share, but it clearly shows it can be done. He did the math on all the axis and stuff to make sure it's strong enough. The Z stage was one of the tougher nuts to crack, as it gets really flexible when it's that big.

 

[duesentrieb 0]

Yes, makes definitely sense. Core XY seems to be one of the best systems for larger printers. I am not sure where exact the border ist, but I would use core XY when going bigger than 500x500mm and maybe the Ultimaker gantry for smaller ones. But core XY also works for smaller systems. Working on a large one, but gut too many other "small" side projects... So it will take some time. Hope it´s finished by the end of the year.

 

[randyinla 14]

If the build platform gets a little wobbly when it's larger, how about a matching worm gear attached to the front side that is directly linked to the original one in the back? Or in keeping with the three-point alignment, add a Z worm gear to the front left and right corners, again, linked to the back one?

As for assuming faster printing has to occur for bigger builds, I don't agree with that. I am perfectly fine with the speed of the U2 as it is and wouldn't expect it to be faster just because I am printing larger things. It's my decision to print a larger object, I'm fine with the speed. I rarely ever print faster than 50mm per sec so I'm kinda used to it.

 

[jhertzberg 19]

The z screw controls the height and holds the weight. The smooth rods prevent wobble. You may want to either add one linked leadscrew opposite the current one, and add two more smooth rods in the front corners, or put leadscrews in the front corners, and smooth rod front middle. The delux approach would be to run each of three leadscrews on separate steppers, use a sensor, and microstep to automatically level the bed.

 

If the build platform gets a little wobbly when it's larger, how about a matching worm gear attached to the front side that is directly linked to the original one in the back? Or in keeping with the three-point alignment, add a Z worm gear to the front left and right corners, again, linked to the back one?

As for assuming faster printing has to occur for bigger builds, I don't agree with that. I am perfectly fine with the speed of the U2 as it is and wouldn't expect it to be faster just because I am printing larger things. It's my decision to print a larger object, I'm fine with the speed. I rarely ever print faster than 50mm per sec so I'm kinda used to it.