Design a perfect 3D printer

I know this is a reprap forum...but I'll ask the question: If you can design a 'perfect' 3D printer, what would you do? Which parts would you use? Cost is not a major issue, just as long as I don't get " why don't you just buy an xzy machine" answers. Here's what I have:

1. willingness to learn to do anything new
2. I have a makerbot complete electronics set
3. I have lots of stepper motors size 12 and 23.
4. I have lots of slides from 5" to 6 ft travel distance
5. I don't have to use any of my current stuff, in other words, I'm willing to buy all different parts
6. I can post my build progress if anyone is interested.

My experience in electronic and mechanical design is zero. This is a learning process for me. I know that buying a reprap or makerbot would be quick and easy, but I can never buy anything off the shelf. Besides, this will be a study in design and development for me. My ideal 3D printer:
1. be able to print at least 12" x 12" x 12", bigger if possible
2. as precise as possible
3. as reliable as possible
4. prints using plastic, but other material such as metal would be a plus.

Thanks.

I really don't think you can realistically develop a new printer that's better than what is available without some experience. The learning curve is steep enough without redesigning the machine also. If you already had electronics and mechanical experience, it would be a different matter. I pondered the same question for a while before deciding to get started with a MakerGear Prusa. Plus, I'll have a working printer when I start work on the next machine.

Even with the kit, there is still plenty to be learned.

Since you already have some of the components (which MakerBot electronics, btw?), have a look at all the various RepStrap machines in the wiki to get ideas on how to proceed. Maybe start with the WolfStrap.

My own thoughts on an "ultimate" machine borrow a lot of ideas from Fab in a Box. Plenty of good ideas there. I was thinking about building something along those lines with powerful enough axes to handle a small router head. Use the right materials, and maybe you could mount a laser for sintering, and cutting, or maybe even a MIG gun. Still, with no experience, all this will be over your head. Walk before you run, etc.

May be I'll have to be a bit more specific. I do have some experience with building CNC machines. Building, not operating, since I managed to blow up one stepper motor and destroyed a gecko drive last night. That gecko drive was supposed to be 'indestructible'. I do have a machinist working for me to help with any machining work. Having never seen a 3d printer in real life, I can't make some designed decisions. For instance, my tiny desktop cnc mill is built incredibly robust, but weights a good 100 lbs. When it worked for about 2 hours, there is significant inertia (?) when there is stop and go motion. Despite a robust table, the machine moves, which make me think that if something such as melted plastic is on the table, the movement would knock it over. Do repraps and makerbots move when printing? Perhaps a lighter gantry or a heavier table would solve the problem.

I see that most people use some sort of slide and pulley system. What sort of slides do people use? and why? I have THK square linear slides. These have a decent amount of preload though and may need a strong motor. Then there are round slides all over ebay. THK are the most precise, but they are the heaviest as well. Should I go with the THK square slides? It seems that with the 1.75mm filaments, these machines are getting more and more precise. Would not be long before 1mm or .5mm filaments are available. I want a drive system that anticipates these improvements.

I've been scanning these forums, seems that the makerbot step extruder 6 is the way to go....anything better out there?

I have the latest makerbot electronics package, not sure what the model is, it's on their website.


The Fab in the box design is very interesting. I have thought about a combo design for quite some time. I decided that there's too much compromise in the design. The z axis travel is usually too limited for a tall 3d print. Plus if you are going to cut plastic and wood, there will be too much dust and that would get into the melted plastic. I have a 6 ft by 4 ft CNC build in process, with an open frame design. Perhaps I'll do a combo machine later. My two favorite designs at this point are the Grassroots machine and the Arizobot. My critique of the Arizobot is that the horizontal movements are place on top of the machine. Seem that there would be too much movement when the machine is working. The grassroots machine is likely sturdier. The grassroots machine looks great, but I'm not sure about the choice of leadscrews. Too much backlash. Plus the heated platform is so large that perhaps there will be problems with the bearings. Of course these are just theoretical critics since I've never seen them in real life.

Edited 2 time(s). Last edit at 03/09/2011 07:10PM by HereinCS.

Some of your comments reveal that you really have a lot more to read up on with regard to how these printers work. For example, the filament size has nothing to do with the precision of the printed part.

I really think you should build an existing design or mount an extruder on one of your CNC machines to get familiar with the printing processes before you attempt to design a machine.

You will not improve upon a high accuracy CNC platform with an extruder mounted on it.

If it can mill metal to under +-.001 in it can extrude plastic to at least that.

The height of an extruded 'liquid' plastic bead is so short even the apparently sudden movement on the platform will have no real effect on it's position until it cools. Not to mention the speed is easily controlled.

You are right, I really should look at some live samples prior to doing my own. Too bad there's not much where I live. Perhaps the next time I'm in CA I'll have to look for a local reprap group.

In regards to filament size, if smaller does not equal higher resolution, then what's the advantage? Why are people seemingly moving towards smaller filament size?

I'd suggest a http://reprap.org/wiki/T-Slot bot.

Say, a basic gantry design made from 20 mm t-slot, with a maximum footprint of ~16" x 24", (450 mm x 600 mm). Along with well specified linear slides that aren't a special one-off via ebay.

You will not improve upon a high accuracy CNC platform with an extruder mounted on it.
Oh, certainly. And milling/routing is lovely, and slowly growing under our banner. And a routing head is necessary for engraving and drilling pcbs, and also when we get into cut-and-paste reprapping.
http://reprap.org/wiki/Shape_Deposition_Manufacturing.
To say nothing of self-machined repraps.

But a machine that is good for routing is extremely overbuilt for moving a lightweight extrusion head around.

Obviously, we should do it both ways and do up both machines in the wiki, this being reprap.

And, most importantly, have build instructions and parts lists in the wiki. Unlike the existing body of wiki-hosted t-slot bot research, which is fuzzy and amorphous.

In regards to filament size, if smaller does not equal higher resolution, then what's the advantage? Why are people seemingly moving towards smaller filament size?
Perhaps it can use a lighter head and therefore a less heavily built positioning system?

I've been scanning these forums, seems that the makerbot step extruder 6 is the way to go....anything better out there?
It's an active research area:
http://reprap.org/wiki/Category:Cold_End
http://reprap.org/wiki/Category:Hot_End
I've not seen a feature by feature comparison.

Also, you should join reprap-dev, our secret official mailing list http://reprap.org/pipermail/reprap-dev/
because you're (secretly) an official developer.

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!

I wouldn't mind building a t-slot machine for your wiki. A 16 x 24" machine foot print will likely limit the build area to 12" x 18" at most, or do you mean a machine that can build 16 x 24"?

I'd like to use metal rods as linear guides rather than THKs, since custom THD rails can run quite a bit costly. Can you suggest a reputable linear rod dealer that can cut the rods to custom lengths?

How about timing belts on all the axis? Is that possible? I don't see any belt drives on the z axis.

I'm not sure exactly what you are looking for for the wiki, or how to do a wiki. How about I post my build step by step and you put it in a wiki for me?

I think you are confusing filament size and nozzle diameter.

The diameter of the nozzle determines how fine of a layer can be extruded, not the diameter of filament fed into the nozzle.

The diameter of the filament has so far settled on 1.75mm or 3mm. The smaller filament requires a smaller stepper to achieve the same nozzle pressure, but at the risk of kinked filament.

VXB.com sells very reasonably priced linear rails.

Timing belts can be used for Z, though if the machine loses power while the z axis is at the top, it can plummet into the table (which is bad). The threaded rod provides enough friction for the z axis to remain stationary when power is off.

Thanks. I spoke with vxb..seems that they can't do custom size rods at a reasonable price. Maybe I'll just go with THK linear rails. But I'll have to find some with not too much preload. I'm going to spend this weekend thinking about the size of this machine, then I'll contact my 80/20 distributer to make an appropriate table.

Just so I know what is necessary for this wiki thing, what are the rules? Is my understanding correct that the wiki is there so that people can make a same design? So obviously I can't use any of my existing rails since they are one of a kind and not reproducible without substantial cost. What other info and criteria are necessary?

our wiki is chaotic; but we like it that way (kinda). bottom line is, if it's worthwhile, it'll eventually be made pretty by people who do that sort of stuff.

you can read about how to make a page here: http://www.reprap.org/wiki/Development

I'm not sure exactly what you are looking for for the wiki, or how to do a wiki.
It's pretty straight-forward. Just log in and click "edit" and you'll hit the ground running:
http://reprap.org/wiki/Perfect_T-Slot_Bot

I wouldn't mind building a t-slot machine for your wiki. A 16 x 24" machine foot print will likely limit the build area to 12" x 18" at most, or do you mean a machine that can build 16 x 24"?
It's all good. Folk will adapt your design to their own ideals as soon as you start doing it up in the wiki.

I'm going to spend this weekend thinking about the size of this machine, then I'll contact my 80/20 distributer to make an appropriate table.
Who's that? Folk keep on mentioning:
[www.misumi-ec.com]

I'd like to use metal rods as linear guides rather than THKs, since custom THD rails can run quite a bit costly. Can you suggest a reputable linear rod dealer that can cut the rods to custom lengths? ... I spoke with vxb..seems that they can't do custom size rods at a reasonable price. Maybe I'll just go with THK linear rails. But I'll have to find some with not too much preload. I'm going to spend this weekend thinking about the size of this machine, then I'll contact my 80/20 distributer to make an appropriate table.

We've got a bunch of entrepreneurs who will do up kits including rods and printed or generic bushings / bearings, etc. if the design takes off. And, of course, likewise for rails.

Just so I know what is necessary for this wiki thing, what are the rules?
We tend to make it up as we go along. Technically, I think we're a cryptomeritocracy, only more anarchic.

Is my understanding correct that the wiki is there so that people can make a same design?
Yup. And then extend the docs, translate them into german, improve the design, etc.

So obviously I can't use any of my existing rails since they are one of a kind and not reproducible without substantial cost. What other info and criteria are necessary?

Once you've got a napkin sketch up in the wiki, wave it at reprap-dev, our secret official developers' mailing list http://reprap.org/pipermail/reprap-dev/ since you're secretly an official developer. This way you'll get good crit, and I can ping Alexander Watanabe et. al. and get a status update out of them, and also bring up cut-and-paste reprapping. (Mendel's not really built for it.)

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!

Thanks. Sounds good. I'm going to do this using timing belt. I've never used timing belts before for a drive system, and it's always been a curiosity. Plus belts should be much lighter than ball screws and more accurate than lead screws. I have 3 nema23 mtors coming form Kelinginc either tomorrow or Monday. They will be perfect size for this machine.

Walk me through the timing belt design. I think most people here like SPD/SI for their belts and pulleys. Here are my questions:

1. I like the looks of their TruMotion timing belts....any objections or other suggestions?
2. Should I use a continuous belt or open ended?
3. Is .200" XL pitch the way to go?
4. Obviously the active pulley goes to the stepper motor, but what does the passive pulley go on? My stepper motor's shaft is 1/4", so I figure I'll have to mount the passive pulley on 1/4" shaft of some sort to keep things simple. How do I do this?

Edited 1 time(s). Last edit at 03/10/2011 11:06PM by HereinCS.

Hi HereinCS,

The second pulley is an idler pulley, so it gets mounted to a ball bearing. Depending on the type of pulley, you can either have the ball bearing inside the pulley and have it spin on a fixed axle, or mount the pulley directly to a shaft and have the shaft supported by bearings. The choice of which one will depend on what kinds of parts are available, and what is more convenient for your machine's assembly.

Different timing belts will not make a huge difference in quality; some might last longer than others or run a little more quietly, but for performance it isn't very critical which ones you choose. Ensuring that your pulleys are coaxial with the shaft will make a much larger difference on the precision of the machine than the brand of belt used. So I recommend making your decision based on cost.

I think that the most popular pitch of timing belt here is MXL (0.080"), which reflects the fact that these are small desktop-size machines. I doubt that 0.200" pitch belt will be ideal, because it will require large and heavy pulleys.

In the standard Mendel design, continuous belt is used for the Z axis, and open-ended for the other two. That's because the Z-axis belt is used to link two leadscrews together, and it has to be able to rotate continuously. The X and Y axis belts directly drive the carriage and are used for reciprocating motion (not continuous), so open-ended belts are used. Basically you only need continuous belt if the belt will have to go through a full rotation; otherwise, open-ended belt is easier to work with.