My 3d printer build

RobTC

Member
Well usually filament will have a tolerance of about 0.05mm or better (diameter), and that's for 1.75mm filament.

Yes, you can get nozzles quite small, E3D has 0.25mm nozzles. I've heard of 0.1mm. 0.4mm is a good middle point between speed and detail, it's probably the most common overall size. Obviously a smaller tip will greatly increase your print time but I assume it will increase your quality by allowing you to print thinner layers. Now the most common way to increase detail is too simple print thinner layers, but I'm not sure how limited I am with this 0.4mm nozzle, so far I've only got done to 0.1mm layers. I just found a good read on filament tolerance: LINK
I haven't gone through all the math of that page yet, but it does say that the error with 3mm filament is lower. Which is obvious from a statistical analysis point of view, but it doesn't seem like it's necessarily the case in practice once you take into account the cross-sectional area ratio problems, the angle limitations of your extrusion motor and backlash on your extruder gearing.

The only thing I'm aware of is the ability to change layer thickness during slicing.
Fair enough. Looks like I can look forward to some incredibly long print times, then!

Actually I started out thinking the same. But ABS is nice! I really like the texture, and of course the flexibility. I would almost prefer it over PLA, but that's just my personal preference.

Actually it's the cooling that causes warping. ABS can expand up to 1% when melted, so it's the uneven cooling of layers that causes warping. As long as you can keep the print above it's glass transition temperature (around 105 degrees C). Now I've been heating my heatbed to 105C, but obviously only the bottom layers are hot enough, so the taller you build, the quicker the layers cool. So an inclosed chamber is really the only option. Unless you use brute force to hold it down from warping away from your platform (ABS dissolved in acetone applied to my glass).
Yeah, I've only heard good things about ABS, that may be the direction I go in. I said that because I'm not intending, by and large, to be producing functional parts with it, so the physical properties of the plastic I use are irrelevant outside of the print area. But since ABS seems to behave better in the printer as well as making tougher parts, that may be the direction I end up going anyway.

Yeah, I misspoke a little, I meant slump, not warp, but it doesn't seem like 105C would be enough to cause that if it's extruding at 200-something.. Definitely sounds like a heated enclosure is the way to go.

I ordered my hotend from Filastruder. $3 shipping via USPS. Thats the best way to get it if you're in the US. Filastruder is the official US supplier.
Yeah, I discovered that later, it's pretty handy! Are you thinking about those dual extruders for support material yet? I'm not sure if I shouldn't build a little extra X-direction to take the wider head, even if I don't use it at first.

Yes the sain smart kit came with, I wanna say 2 thermistors, but they didn't have any documentation on their resistance or rating (100K ohm I assume). So I went ahead and used one from a set I bought, for the PCB. As for the hotend, it's supplied with a nice one. But it's only rated for 300C max, so I would have to get an alternative thermocouple to achieve high temps (I may have to in order to print PC, it flows at around 300C).
Oh, ok. Yeah, I read thermistor and thought thermocouple, for some reason. I think I want to try that, since PC and nylon are super useful if I do end up producing end-products (rather than parts of mouldable masters). Gearing and mounts may be useful for other stuff i have planned (mini diode laser cutter!) and being able to print in PC would be insane. Wonder what its strnegth is like compared to standard (extruded? rolled?) PC sheet.

I bought a kit off ebay for $30. Really nice quality! Theres a lot of people that have kits already printed and ready to ship, for common printer models. I haven't look into shapeway much, but I'm sure this is the most cost effective way.

Again, Ebay is your friend here. Most people get their parts from there. I bought just about everything for this project from ebay (except fasteners & threaded rods).
Oh, ok. I'll check it out, though I'm concerned about proper retraction from printed straight spur gears. Would it be better to try to get helical ones? Or even some cheap brass gears, maybe, Amazon does some reasonably priced ones. Do you have any problems with backlash causing ooze using the "eBay specials"?

I have a sort-of relationship with the guys at OpenBuilds, and I like their hardware, though I do buy plenty of general purpose hardware on eBay, it's my go-to for sure!

Actually that's how it used to be. Nowadays it cost the exact same to run both 1.75 and 3mm. Just check filament suppliers by you (they're the same in the US). And everything seems to be focusing more on 1.75 lately. Personally I would stick with 1.75 because it's easier to manage, and it seems logical that it's easier to control for the extruder (doesn't have to move so slow to be extruded). Also there's a lot of talk about the back pressure on 3mm vs 1.75mm.

Really it boils down to the cost of filament.
Interesting. I'll bear that in mind, and look into the back-pressure stuff.
 

Master Builder

Jr Member
Yeah, I discovered that later, it's pretty handy! Are you thinking about those dual extruders for support material yet? I'm not sure if I shouldn't build a little extra X-direction to take the wider head, even if I don't use it at first.
At first I was really thinking heavily about it, but know I've figured the only advantage is if you can get filament to use as support material at a much lower cost, or if you had the need for dissolvable support (PVA cost more than your primary filament most of the time. There is HIPS that will dissolve in some other solvent that does not affect ABS, and the price for that is pretty good). Or if you wanted to print in multiple colors.

So right now I'm really not going in that direction anymore. Printing support material with your primary filament works great for me. And you don't have to be carrying around all that extra weight of a second extruder.

Wonder what its strnegth is like compared to standard (extruded? rolled?) PC sheet.
I'm not sure, I know it's probably the strongest/most durable thing you can print on it's own (There are other filaments reinforced with fiberglass or carbon fiber). But from what I've heard/read is that it's really hard to print, almost to the point where it's not worth it. So I think you would only print it for certain parts that need PC's properties. I'm sure you wouldn't use it as your everyday filament. Even then, there are other filament options that are a lot easier to print, but still have a lot of the properties you'd be looking for. I'll have to find out what filament I'm thinking about again.

Oh, ok. I'll check it out, though I'm concerned about proper retraction from printed straight spur gears. Would it be better to try to get helical ones? Or even some cheap brass gears, maybe, Amazon does some reasonably priced ones. Do you have any problems with backlash causing ooze using the "eBay specials"?
Mine works great! I have though about the amount of slop it has, as to if it's affecting my prints, but I haven't seen any signs of that whatsoever. When you put more thought into it, it's always going forward, so slop wouldn't affect it in that mater. Even when it retracts, say 2mm, it would revert back to the exact same position before moving on, so there can't be any room for error there.

Now what I have been wondering is how the gears mesh, and if that could not accurately rotate the second gear exactly in unison while going through it's teeth (if you figure out what I'm trying to say). Because I'm not exactly happy with how my gears mesh together. And if you're printing really low layer heights like 0.1mm and lower, you can barely see the gears rotate, so I was just wondering how all that could be affected.

Now if you're running into ooze problems, you simply have to retract more.
 

RobTC

Member
At first I was really thinking heavily about it, but know I've figured the only advantage is if you can get filament to use as support material at a much lower cost, or if you had the need for dissolvable support (PVA cost more than your primary filament most of the time. There is HIPS that will dissolve in some other solvent that does not affect ABS, and the price for that is pretty good). Or if you wanted to print in multiple colors.

So right now I'm really not going in that direction anymore. Printing support material with your primary filament works great for me. And you don't have to be carrying around all that extra weight of a second extruder.
Hmm, interesting. I was thinking about it more as a method of using ideal orientation, regardless of overhangs. Or even internal overhangs that aren't tool-accessible. Seems like the higher cost would be offset by the relatively low usage (I assume)

I see PVA and HIPS (dissolves in limonene) for not-too-terrible prices, I dunno if they're available in pellet form to Filastrude though.

The head weight is something I've been wondering about. I've seen a lot of people say that too much head weight is detrimental, but at the same time, CNC mills and routers have far higher head weights and maintain proper stability and acceleration at higher feed rates. Do they have better firmware, more heavy-duty construction, or something else that keeps them on track? Seems odd to me that it's a 3D printer killer but fine for anything else.

I'm not sure, I know it's probably the strongest/most durable thing you can print on it's own (There are other filaments reinforced with fiberglass or carbon fiber). But from what I've heard/read is that it's really hard to print, almost to the point where it's not worth it. So I think you would only print it for certain parts that need PC's properties. I'm sure you wouldn't use it as your everyday filament. Even then, there are other filament options that are a lot easier to print, but still have a lot of the properties you'd be looking for. I'll have to find out what filament I'm thinking about again.
Yeah, that makes sense. Probably just for heavily used mounts or mechanisms. Might need some experimentation if it's that difficult though. I did just find this that said it's not that bad as long as you set up properly beforehand, it's not exactly plug'n'play: http://www.protoparadigm.com/news-updates/printing-polycarbonate/ - Price-wise it's still not an everyday material, but it does seem useable for some reprap/machine-building work. Interesting that they had a different experience, I wonder why.

Mine works great! I have though about the amount of slop it has, as to if it's affecting my prints, but I haven't seen any signs of that whatsoever. When you put more thought into it, it's always going forward, so slop wouldn't affect it in that mater. Even when it retracts, say 2mm, it would revert back to the exact same position before moving on, so there can't be any room for error there.

Now what I have been wondering is how the gears mesh, and if that could not accurately rotate the second gear exactly in unison while going through it's teeth (if you figure out what I'm trying to say). Because I'm not exactly happy with how my gears mesh together. And if you're printing really low layer heights like 0.1mm and lower, you can barely see the gears rotate, so I was just wondering how all that could be affected.

Now if you're running into ooze problems, you simply have to retract more.
Yeah, that's true. Even with say, 0.2mm of backlash, you could retract it 2mm, it would actually only retract 1.8mm, but when you restart forward 2mm, the first 0.2mm is simply re-taking-up that backlash. So it cancels out. Hmm, ok.

Yeah, I'm looking at tiny layer heights, there are hotends out there that can do a few microns. Inconsistent gear meshing could definitely cause problems there. I've seen dual-motor direct-drive extruders, might be a better option for precision work. I'll see if I can grab some of the stuff I've been looking at, since Firefox just forgot all my tabs again.
 

Master Builder

Jr Member
The head weight is something I've been wondering about. I've seen a lot of people say that too much head weight is detrimental, but at the same time, CNC mills and routers have far higher head weights and maintain proper stability and acceleration at higher feed rates. Do they have better firmware, more heavy-duty construction, or something else that keeps them on track? Seems odd to me that it's a 3D printer killer but fine for anything else.
Well obviously CNC routers/mills are build a lot tougher, and heavier (not sure how they compare on a scale ratio though). But I think the real separation is that you're running a printer at much faster speeds. I could run my printer all day at speeds like 20mm/s and it wouldn't miss a beet. It's when you try to achieve crazy fast speeds that you will run into problems.

Another thing, I think CNC routers/mills run by means of lead screws. Now using those cuts down tremendously on the maximum speed you can achieve, but it increases in everything else, accuracy, torque, and next to no backlash. Now stepper motors can only run so fast, you can max them out quite easy, so whatever mechanism you use for axis drive have to be quite direct (no gearing, or stepping down) if you want to achieve any speeds. Now belts are the most used, but those also introduce backlash and accuracy by small amounts. Another alternative to belts that a lot of people have be using is spectra line (fishing line) wrapped very tightly around pulleys, it's supposed to eliminate almost all the backlash/and meshing issues that belts have.

Now if you were thinking of dual extruders, I would go with a bowden design. You wouldn't have any issues there as to weigh. I'll definitely be going this route on my next printer.

Direct drive extruders are used to achieve higher speeds and overall simplicity, not necessary more control, if anything they probably have less. You also need a high torque motor to drive them.

Yeah, I'm looking at tiny layer heights, there are hotends out there that can do a few microns. Inconsistent gear meshing could definitely cause problems there. I've seen dual-motor direct-drive extruders, might be a better option for precision work. I'll see if I can grab some of the stuff I've been looking at, since Firefox just forgot all my tabs again.
Actually, filament diameter consistency would pose a much greater problem than gear meshing. Overall, I don't think the gear meshing poses that big of a threat.
 

RobTC

Member
Well obviously CNC routers/mills are build a lot tougher, and heavier (not sure how they compare on a scale ratio though). But I think the real separation is that you're running a printer at much faster speeds. I could run my printer all day at speeds like 20mm/s and it wouldn't miss a beet. It's when you try to achieve crazy fast speeds that you will run into problems.

Another thing, I think CNC routers/mills run by means of lead screws. Now using those cuts down tremendously on the maximum speed you can achieve, but it increases in everything else, accuracy, torque, and next to no backlash. Now stepper motors can only run so fast, you can max them out quite easy, so whatever mechanism you use for axis drive have to be quite direct (no gearing, or stepping down) if you want to achieve any speeds. Now belts are the most used, but those also introduce backlash and accuracy by small amounts. Another alternative to belts that a lot of people have be using is spectra line (fishing line) wrapped very tightly around pulleys, it's supposed to eliminate almost all the backlash/and meshing issues that belts have.
I'll be building mine with similar stock to basic table routers, which are intended to run with 2-3lb on them, so I'm fairly confident in its strength and rigidity. Over-engineered slightly, but I still have no real concrete extruder direction and with anything CNC its better to go way too heavy than not.

When they're built properly (expensively) they use ball screws, lead screws tend to wear excessively on high torque implementations like that, but I want to build the whole thing for the price of one ball screw, so I'm planning on just going with lead screws. I have fairly limited interest in speed- if it can go 20 hours or whatever without me having to think about it, that's ideal. The 20-50mm/s range seems fairly reasonable to me, if that's the price of accuracy. I may change my tune on that with sub-0.1mm traces, of course.

Now if you were thinking of dual extruders, I would go with a bowden design. You wouldn't have any issues there as to weigh. I'll definitely be going this route on my next printer.

Direct drive extruders are used to achieve higher speeds and overall simplicity, not necessary more control, if anything they probably have less. You also need a high torque motor to drive them.
Yeah, I got that general idea from what I saw too, though I've seen others saying that direct drive is better in some other ways. So... not sure. Fortunately I have time to think about it before getting around to buying anything.

Actually, filament diameter consistency would pose a much greater problem than gear meshing. Overall, I don't think the gear meshing poses that big of a threat.
That's what I thought but I've done a bunch of research onto extruders and it seems to be an area of active interest due to its direct impact on print quality- not so much on individual traces themselves as on larger features as a whole. Not sure yet, but it's definitely something I want to look more into. I saw that others had also used brass gears for higher accuracy and compactness, but there's also a weight issue there. I dunno, it may just be a case of picking a compromise.

When I can get to it- busy mouldmaking at the minute- I'll re-find what I've seen. Some of it appears to be contradictory, so it'd be useful to get another opinion from someone with known good results if possible.
 

Master Builder

Jr Member
Some more prints!

A bear!
bear.jpg

Some new frames for my sisters glasses she broke. (4th revision)
glasses 1.jpgglasses 2.jpgglasses 3.jpg

Some thank-you tokens for my firefighter class instructors.
hydrant and barrel.jpghydrant.jpg

A few days ago I managed to almost eliminate my peskie z-banding issue. I'll have to upload some better pictures so you can actually see the detail (cameras charging). With that solved, you can expect to see much better prints.
banding solved.jpg

Stay tuned, I will be announcing a helmet project here soon.
 
Top