Thinking of starting a project, thoughts?

Hi! I'm considering building a largeish-format, high-speed, high-precision RepRap using a few aluminum rods I have laying around. I'm on a pretty tight budget, but I have quite a bit of time. Before I start, I figured I'd ask your opinions on the project and how best to approach it.

The aluminum rods, which will be the basis of the build, are 24" long and colored gray/brownish. Six have 5/8" diameters, the seventh hsa a 1" diameter. I think they're made by igus, as they came with a bag of igus bushings -- some for the large rods, some for the small, all slightly oversized. The rods are also slightly undersized -- too much to use bushings or bearings with decent tolerance, unfortunately. I have a few solutions, though.

I also have a working expanded Kossel Mini, some scrap aluminum angle-iron, lots of tools, and access to a Home Depot.

My goals for the printer are the following: I want the printer to be stable, precise, reliable, and fairly quick. I'll probably get a pretty big build volume -- a full 24" cube would be great -- but that's secondary. I plan to make the frame out of metal, ideally welded steel, and to make as many parts out of aluminum or steel as possible. So I guess this'll be more of a RepStrap than a RepRap in that respect.



I'd like to make it roughly Cartesian (not delta or polar), and I'm debating between two mechanical setups. Here are the pros and cons I've got so far:

One is CoreXY. This would require four rods arranged in something shaped like the letter H with two rods for the "crossbar."
Pros: Uses fewer rods, quicker motion in one direction (I can reorient parts so that's the dimension that's used more), requires only two belts and two pulleys.
Cons: Not symmetric (more inertia in one direction than in the other), lots of belt length to stretch, lots of idlers.

The other mechanism is used in Ultimakers, Tantillus, possibly a few others -- I haven't seen a good name for it. Each axis consists of an H-shaped set of rods, and the two axes are mounted above each other, with the effector sliding on the crossbars of both H's. Like CoreXY, it gives 2D motion without requiring one motor to move another. For this printer, I'll use separate driveshafts from the linear rods, unlike Ultimaker. This would use all six thin rods.
Pros: Equal moving mass among the two axes, simple effector (no belts attached), easier to use a heated chamber (no belts on the gantries), relatively simple construction (because the axes are the same).
Cons: Uses all six thin rods, four belts and eight pulleys instead of two and two, requires more (printed or machined) parts.

I'm also curious as to which is more rigid and which is more precise. Assume that all the belts are properly tensionsed.



My second question regards the frame. As I said, I'd like to weld it if possible, but I have zero welding experience and don't have a welder (though a cousin does). How practical would this be, and would it offer better stiffness/toughness than a bolted-together steel or aluminum frame? Also, how close to orthogonal could I get the axes?

Another crazy idea I had was concrete. Could I mold the frame from cement or concrete, in order to get a super-rigid frame? It would sort of limit my options for transport if it weighed as much as an American (or two Europeans), but it would be both reliable and a cool proof of concept.



Thirdly, linear motion. The igus bushings I have are too loose and wobbly to be much use, so I tried cutting a slit in one and squeezing it. It worked quite well, though it does require some grease to slide properly. So, my plan is to print circular clamps, tightenable by a machine screw, into the various sliding parts, and to clamp them around the bushings. How good of tolerance could I achieve this way? Is there a better way (PLA/ABS bushings, homemade Delrin blocks)? And is there anywhere where I could buy factory-undersized bushings?

Two other options for linear motion are hardware-store square aluminum tube (with 608s or something riding on them) and Openbuilds V-slot. I also have the Openbeam 1515 and nine Delrin wheels that my Kossel uses, and I'm pretty sure it's willing to sacrifice itself for the greater good (I'll ask it to be sure). What do you think of those options if I want more than seven linear components?



Fourthly, the Z axis. I plan to use a thick cast-aluminum plate, with heaters and PEI or BuildTak. It'll be quite heavy, especially if it's a full 24" square, since it'll have to be both larger and thicker than a standard plate in order to be flat enough. If it's a 24" square and 3/8" thick, the plate alone will weigh about 9.6kg, requiring a pretty hardcore Z axis. I have a few ideas.

I like the looks of a mechanism using two Z rods and three leadscrews, like the Fusion3 printer. The bed slides along the rods with only one bushing on each -- not cantilevered, there's no torque -- so they provide resistance only to horizontal motion and rotation around the Z axis. The leadscrews prevent the bed from moving vertically (of course) or rotating around the X or Y axes. This would require three screws and three anti-backlash nuts, which would cost quite a bit (around $30 per screw/nut pair), but it would also give quite a rigid platform. If I use the Ultimaker linkage, I would also need to find another rod, since I would have only the thick rod left.

Another option is a cantilevered bed, like a Replicator 1/2 or Ultimaker. This would put a lot of force on the Z rods, so I might replace them with some square aluminum tubing from Home Depot and have ball bearings riding on them. This would lose some precision, since the square tubes aren't super straight, but it would be cheap and it would require only one Z screw (save a lot of cost and complexity). I could also do that using linear guides of some sort, either ball bearings or Openbuilds V-slot, costing a bit more but gaining precision, rigidity and simplicity.



As for the motors, I plan to use NEMA 23's for all three axes (I'll deal with the extruder seperately). I found a kit on eBay consisting of three 270 oz-in NEMA 23's, drivers to drive them, and a PSU to run it all, for $150. Does that sound like a good idea? Could I get away with reusing the NEMA 17s and DRV8825s from my Kossel, or would they be too slow to haul around the aluminum rods? If I do use more powerful motors, would printed mounts be good enough or should I try to machine them from aluminum? I don't have a milling machine, so I'd be doing it with just a smallish drill press and cutoff saw.



Lastly, do you guys think this is a good idea? I'm a senior in high school, and I leave for college in about a year. I'll be able to bring some hardware with me, but I definitely won't have the space for two printers, and my current one's pretty decent if you can convince it to work. I'm doing this primarily for the experience, not for the result, but I do want to come out with a working printer, because a college kid can't ever have too many plastic trinkets. How likely do you guys think that is? Or should I go take up shuffleboard or something easier?

[EDIT: 24" rods, not 48". I'm not that insane.]

Edited 1 time(s). Last edit at 09/07/2015 03:11PM by epicepee.

1. I have some experience with the Ultimaker-layout, A problem is that it is hard to tell if the x or the y rod will define the z plane. My Ultimaker copy shift between them.
2. The only thing you can be sure of when welding is that things will bend. Very hard to keep the accurate measures needed for a printer.
3. Steel against aluminum is a bad idea and NEVER kill one printer before the other is up and running.
4. For a large printer the print head and the mechanism that move it will be lighter then the print bed.

Forget about this and learn 3d-modeling, Texas hold'em-poker, dancing, beer-brewing or something other useful for college (or all of them).

Edited 2 time(s). Last edit at 09/07/2015 12:48PM by SlowFoot.

I'm not sure that large format, aluminum rods, and high precision all belong in the same sentence, and I'm not sure you could build a printer that way.

I suggest you consider something more appropriately sized for a dorm room, and focus on print quality. Any dope can build a sloppy printer (look how many of them are out there), but it takes some study, understanding, and effort to build something that reliably produces quality prints. Start with a rigid frame and go from there. Don't use threaded rods for anything, especially not linear motion.

Think about the definition of "reliable" and what sort of things make a printer unreliable. How do prints fail? What are the causes of print failure? What can be done to eliminate those causes? What sort of things degrade print quality? What can be done to maximize quality? Disregard your budget and whatever parts you have on hand (unless you have something especially good) entirely for the planning stage. What is the best way to do everything that has to be done? If you're not sure what the best way is, do some research and ask questions. Don't assume that what you already know about is all there is out there. And ferchrissakes, don't assume that what you see in a $300 kit is anything close to the best way or even a good way to do anything. Once you have an idea of what it is you need to do, start looking at the parts and see how you can get them without having to pay full price.

Spend some time at the local scrap yards and get to know the people there. Tell them the sort of stuff you're looking for and they'll either show you where it is or can set it aside for you when it comes in. Likewise, scour ebay for quality (Japanese, German, Swiss, etc.) used machine parts. Have patience- that stuff turns over fast and you won't have to wait too long for something you need to show up at a good price. Be prepared to pay a little more for quality. I'd rather take my chances on a quality used part then get a cheapo new part. Figure out where you can keep costs down so that you have more money to buy the quality parts that cost more.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote

I'm not sure that large format, aluminum rods, and high precision all belong in the same sentence, and I'm not sure you could build a printer that way.

You're right, I'd trade them for steel in a heartbeat. But they're pretty thick, even for aluminum, so I do think they could be used to high precision if I'm careful. If I need to, I could double them up, and use linear guides or something for the other axes.

Quote

I suggest you consider something more appropriately sized for a dorm room, and focus on print quality.

Thank you for being rational if I can make the whole thing self-contained enough, like SoM, I could probably just stick it under a desk. I've toured a few colleges and my 750mm-tall delta would have fit in any of them. I also don't think I could bring myself to cut down these beautiful two-foot rods...

Quote

Think about the definition of "reliable" and what sort of things make a printer unreliable. How do prints fail? What are the causes of print failure? What can be done to eliminate those causes? What sort of things degrade print quality? What can be done to maximize quality? Disregard your budget and whatever parts you have on hand (unless you have something especially good) entirely for the planning stage. What is the best way to do everything that has to be done? If you're not sure what the best way is, do some research and ask questions. Don't assume that what you already know about is all there is out there. And ferchrissakes, don't assume that what you see in a $300 kit is anything close to the best way or even a good way to do anything. Once you have an idea of what it is you need to do, start looking at the parts and see how you can get them without having to pay full price.

I really like that approach in general, and after reading your SoM Instructable I spent a few hours brainstorming and thinking of how I would do this all if I could. Unfortunately, it seems I'm going to have to design it around what I have, as opposed to the other way around, since I have some nice rods and not much money. I do plan to eventually build a CNC aluminum router, sometime during or after college, and I'll probably take that approach then, but right now it just isn't practical.

Quote

Spend some time at the local scrap yards and get to know the people there. Tell them the sort of stuff you're looking for and they'll either show you where it is or can set it aside for you when it comes in. Likewise, scour ebay for quality (Japanese, German, Swiss, etc.) used machine parts. Have patience- that stuff turns over fast and you won't have to wait too long for something you need to show up at a good price. Be prepared to pay a little more for quality. I'd rather take my chances on a quality used part then get a cheapo new part. Figure out where you can keep costs down so that you have more money to buy the quality parts that cost more.

Sounds like a good idea. I'll have to find some scrapyards around here, but eBay and such do sound like good ideas. One issue with that is shipping -- even a brand-new Acme rod costs as much to ship as to pay for on eBay. So, how would you recommend looking for good-quality, cheap parts?

And thanks for actually reading through that novel I posted

Edited 1 time(s). Last edit at 09/07/2015 04:22PM by epicepee.

Get your parents, particularly your father, interested in what you're doing. Point out that you won't be blowing the money on weed, tattoos, piercings, beer, a loud stereo for your car, or a girl with weed, tattoos, piercings, and a loud stereo in her car. You may find that some unaffordable part suddenly becomes affordable. If they ask want you want for Xmas/birthday, tell them "I really want this Japanese 15 mm linear guide I saw on ebay..."

When you're on ebay, look for "linear positioner" or "linear guide" or "ground ball screw" or similar terms, then skip almost everything that ships from China. Look for brands like Thomson, NSK, Parker, THK, etc.

Edited 2 time(s). Last edit at 09/07/2015 05:32PM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

5/8" aluminium rods 24" long (that don't give you a 24" cube build volume, but less) should be roughly equivalent to 0.48" steel rods with respect to deflection calculated for supported ends, if wikipedia Young's moduli are correct in your case. It depends of course on many factors, including how they are attached to the structure, but I would bet that even ignoring all the other factors affecting accuracy and precision you will not be able to exploit the whole size of the bed for a big print, also because the weight of the suspended parts for such a big printer will be significant I guess. Moreover, you should first check that the rods are really straight (we are talking about 0.01" precision over 24"). All this if I did not do any stupid conversion mistake (I am a metric guy, sorry).

I also wanted to build a good "big" printer (12" cube build volume, so eight times smaller than yours) but I found quite difficult to do it even going well beyond a $300 budget. So I am now sourcing the parts to build a super-precise small printer with a build volume 4" cube.

In any case, good luck with your project!

Quote
the_digital_dentist
If they ask want you want for Xmas/birthday, tell them "I really want this Japanese 15 mm linear guide I saw on ebay..."

That worked brilliantly last year, that's how I built my first printer. Hopefully it won't be necessary this time...