Ultimaker++ (800x800x730mm build env.)

[mattinv 0]

**UPDATE: follow this project on my blog:

http://3dlargeprinter.blogspot.com/

**

Good morning everybody,

I'm working on a large Ultimaker, which I will call the Ultimaker++ for now. The build envelope is approximately 800(X) x 800 (Y) x730mm (Z). Z can easily be increased, but for budget purposes (linear bearings, ball screws, beams, etc...), I was limited to 730mm.

The objectives of this build is as follow:

- Very high printing speed: the X,Y,Z gantry is built using open end and closed belts (I'm working on a self tension mechanism for X & Y, Z can be adjusted by moving the plate that holds the steeper motor). The design is such that accuracy is improved, with no loss of speed. The gantry weight has been reduced to its minimum.

- High accuracy: Extremely rigid frame; Linear bearings on all 4 axis. "double belt / pulley" system for X / Y.

- Adjustable, spring loaded table (25 points). I'm still working on being able to integrate a heated bed...

- 95% of the parts "out of the shelves" or printed parts.

- All custom parts are made of 0.25" (6.35mm) thick aluminum plates (but other materials could be used). All can be laser cut / water jet / CNC.

- Re use as many as possible of the original parts of the Ultimaker.

- Extruder by Geo Hagen - Thank you again for the files (

http://www.thingiverse.com/thing:15897

)

- Fits on a standard 48"x40" pallet for shipping.

I'm finalizing the design and I will post the files on Thingiverse.com as soon as it is complete.

I'm looking for people to print some of the custom parts for me, since I do not have my Ultimaker yet.

I have two options:

Option #1: X & Y -> belt driven, Z -> ball screw

Advantages:

- Only 2 ball screws, X axis is much faster

- Will work with UM electronic

Inconvenient:

- possible backlash due to long belts on X due to the weight (2500g) of the gantry -> loss of accuracy.

Options #2: X -> ball screws (limit of 250 mm/s @ 50 PPS / 3000 RPM), Y -> belt driven, Z ball screws.

This is a little bit more expensive but the envelope for the exact same volume goes to: 800mm (X) x 930 mm (Y) x 800 mm (Z)

Advantages:

- very accurate on all axis.

- bigger build envelope

- fast

Inconvenient:

- limited to 250mm/s on X axis*

- more expensive (2 extra ball screws, bigger stepper motor for X axis)

* I would think We could compensate for that by printing along Y mostly (not sure if the software part can do that...).

I'm currently polishing the design and starting the BOM. I'm not sure what the cost is yet, but I'll post a detailed list including all standard (shelf) parts.

You can also follow the conversion on the google group: http://groups.google.com/group/ultimaker/browse_thread/thread/8556925cbc0db0cc

[ddurant 0]

Wow! That looks pretty cool!!

(sorry - got no answers to your questions)

[zmaster 0]

I'm not an expert in mechanics so i'll give my opinion just on the electronic side:

3. At first sight it should be possible, the controller has the same signals. Some changes to the firmware may be necessary for polarities and of course an adaptation PCB would be needed.

4. This page have many useful indications:

http://www.reprap.org/wiki/Heated_Bed

Designing a single big PCB heater would be quite expensive for the prototypes, if you want to go this way it would be probably better to use an array of 4 or 9 of them.

A couple of considerations: that's an awesome print area but you might consider using a larger nozzle to extrude otherwise it will take ages to print. There's a video on YouTube of something similar and a big print took 3 days to complete. I think the required precision is related to the size of the object so if you print something bigger it needs less accurancy.

The heater will consume lots of power. And this is another reason to try to get fast prints. You should probably close the sides with transparent panels to create an heated chamber so the heat will stay there: this also reduces power consumption for a print as the heater will be powered for less time.

[mattinv 0]

Since you mentioned the electronic, here is what I found.

I looked at the wiring diagram of the UIM24008 Minuature Microstepping driver (

http://www.zappautomation.co.uk/pdf/UIM ... %20_2_.pdf

).

These seems perfect for an upgrade of the stepper driver. Up to 8A for good torque and/or speed.

I will put acrylic side windows. This is just a first sketch, I still need to work the mechanical side of it.

Now, my only concern is the backlash. It was mentioned on the Google thread (

http://groups.google.com/group/ultimake ... 8d55c48ebd

) that it might be an issue with the long belt.

The only solution is to switch to a ball screw for the X axis, which slows down the gantry. My problem is finding a stepper motor, ANY, that will have some torque at 50 rps. If somebody has an option to recommend, I'd like to hear it.

So far, the only one I found is the WO-5718X-15E (

http://www.linengineeringstore.com/prod ... px?proID=3

), that can operate at 50 rps and has a torque of 14.16 oz.in at that speed (@ 20,000 pps).

Is there a pulsation limitation with the Ultimaker micro controller?

Anybody has a better option? I need to control 2 x 950mm ball screws...

Should I use 2 stepper motors and run them from a single stepper driver?

[mm_build 1]

My problem is finding a stepper motor, ANY, that will have some torque at 50 rps. If somebody has an option to recommend, I'd like to hear it.

So far, the only one I found is the WO-5718X-15E (

http://www.linengineeringstore.com/prod ... px?proID=3

), that can operate at 50 rps and has a torque of 14.16 oz.in at that speed (@ 20,000 pps).

 

I had a look in your other thread on google groups. Looks like you have done the math using a ballscrew with a pitch of 5mm. On the same site there are ballscrew configurations with 10mm pitch which would give you twice the travel at half the rpm. According to the chart you provided that would give 40 oz.in (+282%) torque at 10,000pps. If you run the motors in their sweet spot for torque/speed, you should theoretically be able to add the needed gearing afterwards.

If you upgrade to a truly massive NEMA 34 motor (86mm) you can easily achieve ~600 oz.in at 2500pps, which works out to ~150 oz.in when geared to the equivalent speed of the smaller motor running at 10,000pps. I have seen ballscrews with up to 20mm pitch available online, if you used those you would bring your required speed down to 5,000pps, it's only an extra 2:1 gear to get the rest of the way.

Also you were asking about the electronics side which I know next to nothing about, but this

http://reprap.org/wiki/Stepper_motor

might make for some good reading.

[mattinv 0]

Well, here is the dilemma.

I kind of forgot about it, but:

1.8 deg -> 200 step; at 5mm per rotation, the accuracy with a gear ratio of 1:1 is 0.025mm / step.

I'm looking at the Gecko G210X (

http://www.geckodrive.com/g210x-p-51.html

) to see if they might help.

These might be able to compensate.

What do you guys think?

Here are a few other options:

M-3424-6.3[X]

http://imshome.com/products/nema34.html

[mm_build 1]

The first charts listed were in half-stepping mode, which actually doubles the steps per rotation to 400 (and the resolution to 0.0125mm). Also this means they are running at about 70% Torque from if they were in full stepping mode.

http://en.wikipedia.org/wiki/Stepper_mo ... f_stepping

The last chart you listed shows Torque rated in full stepping mode, so for a fair comparison the values would have to be adjusted.

Also for the size of a printer you are going for, I don't know what kind of resolution you want but presumably you would be using a larger nozzle size, etc, so it would be less sensitive.

[mattinv 0]

Another stepper motor that would work (up to 50 rps):

http://amci.com/stepper-motor-control/stepper-motors-size-34.asp

[mattinv 0]

As promise, the second version if we want to avoid too much backlash in the belt for the X axis (the load on Y is extremely light).

Options #2: X -> ball screws (limit of 250 mm/s @ 50 PPS / 3000 RPM), Y -> belt driven, Z ball screws.

This is a little bit more expensive but the envelope for the exact same volume goes to: 800mm (X) x 930 mm (Y) x 800 mm (Z)

I made sure there will be no slippage on X axis pulley:

And I'm still working on the belt tensioner.

Advantages:

- very accurate on all axis.

- Will need an upgrade of the stepper driver for the X axis.

Inconvenient:

- limited to 250mm/s on X axis*

- more expensive (2 extra ball screws, bigger stepper motor for X axis)

* I would think We could compensate for that by printing along Y mostly (not sure if the software part can do that...).

[Daid 306]

That's one long belt for the Y move. Are you sure that won't cause any problems?

The whole thing looks crazy enough that it might even work

[mattinv 0]

Hi Daid,

I'm not too worry about the belt on the Y axis because the weight it is moving on the miniature block bearing (Hiwin MGN 9) is very light: 185g (including the bearing, pulley, etc...). To be precise, it is less than the weight moved along X or Y on the ultimaker, since the weight of some of the bearings and axles is removed. So the backlash will be minor, and since the belt is tensioned, probably less than what you would have on the UM.

Most of the weight is on the X axis, about 2500g (which contains the Y axis), or 4X the weight on the X axis of a UM. This is why it was decided to switch to a ball screw on this one.

By keeping Y extremely fast, if I print mostly on Y and move more slowly along X (which can still go up to 250mm/s, although it could be more but I'll have to try different stepper motors), I should be able to have very precise and fast prints.

I think I'm going to start a blog where I'll post all the details on each part of the machine and the calculations that went along on deflection, stepper motor selection, etc...

Right now, I'm working on the BOM and I will post everything online later today.

[Daid 306]

Btw, 250mm/s is faster then I can even move the Ultimaker during printing. However, if you are lugging 2.5kg around then I'm wondering about max acceleration you can get.

I'm not a mechanical engineer, so if you could do a blog about the math behind the forces on the machine and explain it in a way I would understand, then you would become a hero

[mattinv 0]

I will start a blog on the Ultimaker++. But I want to present 3 options, similar and different in many ways. I will show some of the results for the deflection on each beams of the machine.

I mostly finished the BOM for option #2. Price came at just over $4300 for the build. So I want to try to offer a version under $4K for people to consider. Closer to $3500 would be best.

Option #2 is the "rolls-royce" of all 3 versions, so I think it is not too bad.

Option #3 will be the cheapest. I'm looking at the big items and trying to reduce them to a minimum.

Here is what I'm working on now. I'm converting the inch frame to metric. This should allow for the build to be made anywhere worldwide, plus people can now use some cheap aluminum extrusion from overseas. In the CAD model I have presented so far, I used the series "15" of 80/20. It is expensive, the fastener are not cheap. And it is limited in compatible suppliers. The new revision will be build using series "40" of 80/20, very similar to the series 15 in size.

Next step is the conversion from rail bearing to shaft + ball bushing. The question remain as to the conversion to ball bushing on the X axis. It only saves $105 assuming we purchase the cheap blocks from ebay... So I might keep the original design on this axis at first.

250mm/s is the maximum speed. It will probably take a few cm to reach it.

As for acceleration, I will see if I can find some easy way to calculate them. If anybody has an advice, or a (free) software that would do it, please share.

[mattinv 0]

Hi Everybody,

I finally posted all the modification and a first final version of the Ultimaker ++ on this blog:

http://3dlargeprinter.blogspot.com/

I encourage everybody to check it out and criticize. As soon as I get my Ultimaker, I will start ordering and printing some parts for it, and put it together.

But first, I want to make sure I iron out as many problems as possible. So I would really appreciate any criticism and suggestions.

[owen 19]

That link isn't working for me.

[mattinv 0]

Hi Owen,

Check again and let me know if it worked this time. http://3dlargeprinter.blogspot.com/

I uploaded quite a few images and I will keep adding articles about the designs.

Like I said on google groups, I'd love to get some feedback and suggestions on possible modifications.

There is so much to be said about each part of the assembly, and always room for improvement!

[owen 19]

Yep it came up earlier on a tab I already had open. Haven't had a good look so can't give you any feedback sorry.

[mattinv 0]

When you get some time, check it again. I guess it takes some time to load because I posted a few pictures on each post.

I'm also working on a granule extruder to mount on it, you might me interested in this section too.