corexy with 2 or 3 direct drive extruders

Quote
alan richard
Forget CoreXY go with cartesian, or maybe even something of the prusa style.

I say this from the perspective of having successfully built a coreXY.

I think I too was 'seduced' by the idea of having a high speed, low mass printhead churning out super fast prints.

I don't think moving mass is the limit to speed; It is partly the ability to extrude plastic quick enough and mainly the desire by most people to have a half decent looking print at the end of it and turning the speed down.

They are a pain to set up. You have to have 10 pulleys (including the motors) bang on square in 3 planes. If they're not parallel there are issues of belts riding off idlers or binding on pulley edges, If they are not perfectly square to the moving elements then the equations of motion above are not correct.

Compared to 4 pulleys in cartesian where each pair (or axis) of which can be considered and assembled as separate entities, then you need only get the 2 sub assemblies at 90 degrees.

I reckon they are also more prone to generate noise and resonance with so many points of contact between belt idlers/ motors and framework. Maybe that's just my cheap bearings. You can't skimp on components. The moving axis needs to be fairly substantial to cope with the torsion of setting up belt tensions (tho' it is equal when they are set up)

I used v-slot to build mine, I know using that for a cartesian build would have been a doddle in comparison, with so many easy ways to drive it (check out 'fixed belt in the slot like rack and pinion' - You can't get it wrong and nothing needs alignment)

CoreXY isn't fully and correctly implemented in Marlin (and probably others) firmware. I think there are bugs. Some features are not available - babystepping for one.

If using Ramps and Marlin, there is a problem with driving them at 1/32 step - there is an absolute limit of 125mm/s or 7500mm/min. feedrate on X or Y axis.
Set it to 7501mm/min and it does some horrible things to the sync of the motors. Print with it at real world sub 100mm/s type speed with 1/32 and there are occasional noises that don't inspire confidence.
Diagnosing motor issues is awkward even if you set it to cartesian to move only one motor, you're still dragging the other belt about and its hard to measure diagonal moves.

I mention Prusa's - Beauty is in the eye of the beholder they say, but I think that the one thing that Core XY and cartesians do have in common is they end up looking ugly big cubes that must contain the entire print volume, whether they're printing or not.
At least with a Mendel or Prusa style the print volume is out in fresh air that can be shared by other domestic appliances when they're not printing. And they can be prettied up a bit. And there's a lot of know how behind them.

Sure, the pro's make boxed in printers of the Core/ Cart variety -the motion control is better but would a home made effort be any better at printing than a well established reprap mendel style design?
There's a reason Stratasys spends millions of $ and thousands of man hours developing a printer - 'cos that's what it takes.

Anyway, enough moaning, here's mine if you want to see it. Just don't bother making one.

d091aab0-4652-47f2-8661-272374a8bf73_zpswfp764i9.jpg

fac50791-6fb0-4271-9fc8-5a5d27e29f43_zpsynrykd60.jpg

70b_zps0lqw0prh.jpg

e122f4b2-5f07-47d4-9168-3b7ed861602b_zpskarzx3k4.jpg

Love it really. Prints great. 304x214x220 print area in a box 420x480x450. Just waiting for the covers to arrive and its done.

-a

Quote
alan richard
Hi Andrew. You are quite right, its not optimal.

It does print ok so far with PLA up to 80mm/s but something rougher like Timberfill wouldn't go up there.
I expect the friction on the filament will vary depending where the print head is which may show itself on bigger prints.

I don't usually do; 'form over function' but that was one case on this build to try and keep it all within the box.
I'll probably get some more ptfe tube tube to loop up and over in a conventional way.

(I could have taken 60 or 70mm off that cantilevered Z axis too - another sop to aesthetics to have the hotend visible from the 'front window' of the machine.)


@J.A.C. - 'Scuse me having an anti-CoreXY rant on your thread, especially here on the CoreXY forum! I was a bit tired and cross at the time.

CoreXY is a lovely quirky mechanism, set up right it works fine. Its a nice engineering challenge.

I just feel that in the 3D printing subculture there is unfair bias towards it being the answer to every problem without much focus on some if its negatives.

My own conclusion from doing this is that it is better suited to either mass production, fast non 3D printing applications, or a printer that has a huge span where every spare gram needs trimming off.

I am a printing newbie but familiar with machines. As a one-off 'build it in your shed' reprap, I think you could make a printer that's just as capable but a lot easier to put together using other methods and you did say 'simple is better' in your first post.

That's a handy looking stack of extrusion in your first picture. I used 20x20 standard extrusion and bought a couple of metres of Openbuilds V slot 40x20 for linear motion. I already had the round bar and SC10UU bearings (It started out as a Mendelmax 2 lookie like and evolved!)

With hindsight, a cartesian using this method

v%20slot%20belt%20and%20pinion_zpsy8j1gsyq.jpg

would have gone together easier, saved a lot of workshop time, cost less, run quieter and I'd feel more comfortable about firmware.
I can't really think of any downsides now I understand the printing side better.

Good luck whatever you build.

Alan

Quote
LarsK
Hey, well, it is better to buy an overall good printer then to buy a few very good components and go cheap on the rest. But it is a lot of money for a very expensive hotend and then I think the mechanics should follow.

I think you should spend some money on your existing printer first - In my opinion you need it to print ABS if you wanna print parts for your new printer.

Quote
the_digital_dentist
Do you realize that the corner cubes are unnecessary? T-slot extrusions are easy to connect together by tapping the center hole and using the screw size that the slot is designed to fit.

All you need are a tap and a drill. No precision is required- hand tools are fine. Use the tap to thread the center hole and use the drill to make a tool-access hole to tighten the screw. 1.5" square 8020 is designed for 5/16-18 tap and button head screws and carriage bolts. Check the appropriate sizes for whatever you're using.

I'm a big fan of rigid frames, but I doubt corner gussets do anything for rigidity. My printer's frame is big, and the machine is heavy, but it has neither corner cubes nor gussets, yet is very rigid. The money you don't spend on corner cubes and gussets can be put into using larger size T slot extrusions. THAT will get you more rigidity.

15.jpg

Quote
Edvardas
...
I have built a Vulcanus V1. Its direct drive extruder X-carriage weights 650 grams. What surprised me is how printing quality does not suffer by high accelerations. Parts printed at 30mm/s and 100mm/s accelerations look just as good.
Now I am turning my Vulcanus X-carriage into a one that has 3 nozzles runned by a single stepper motor. X carriage should weight around 1000 grams so around the same as dual direct drive extruder carriage. I will let you know in a couple of the days on how printer copes with an extra weight.

BTW I have used metal corner brackets on my Vulcanus V1. I do not like the threaded rods on the Z-axis but the X-Y axis runned by TMC2100 stepper drivers work like a piece of art.

Quote
hucilu

That would be great! I wil be eagerly waiting for your answer To be sure,you are building it with smooth rods?
I also considered using alu corner brackets simultaneiously with those corner cubes to get best possible results.
And i want use lead 2 lead screws connected with belt to stepper motor on the bottom of frame.

Quote
Edvardas

Quote
hucilu

That would be great! I wil be eagerly waiting for your answer To be sure,you are building it with smooth rods?
I also considered using alu corner brackets simultaneiously with those corner cubes to get best possible results.
And i want use lead 2 lead screws connected with belt to stepper motor on the bottom of frame.

Yes, I use 8mm rods all around. I wanted to get the printer running before I start modifying things.

My X carriage mechanics are finished and it weights 1300 grams (twice as the one on Vulcanus V1).
Tomorrow I will finish electronics and update the firmware and then it will be a testing time

I can see you are really strugling with a way to connect 2020 profiles. I guess you are not from Europe?

Quote
hucilu
Yes iam from Europě.We have localy avaiable 3030 profiles but i wasnt sure how get sturdiest construction/connection.

Quote
Edvardas

Quote
hucilu
Yes iam from Europě.We have localy avaiable 3030 profiles but i wasnt sure how get sturdiest construction/connection.

Well then why not use Motedis?
I have been using these corner brackets [www.motedis.com] They do make a sturdy construction but make it hard to make things square. Now I would use these cubes for the corners of the frame and corners of the bed: [www.motedis.com]

And as for sturdiness... Vulcanus has extra 2020 beams where Z-axis motors mount and where corexy mechanism mounts. This makes it really stiff (look how some people make corexy printers by just having cube without any such bracing beams). If I had to build a printer again I would stay with 2020 profiles.

Oh and do yourself a favor and go with T-nuts instead of filing normal nuts. These without guidance work great with corners brackets [www.motedis.com] and these with guidance works well in mounting printed parts to the 2020 profile [www.motedis.com]

Quote
realthor
Beware of the deflection on smooth rods caused by heavy carriages. I see big numbers here.
I have attempted to calculate the maximum weight of something at the middle of smooth rods (static weight that is) and i got much smaller numbers being safe for deflection-free (or within reason) system. Please read that post I have linked. There are other guys around that confirmed the numbers with different approaches for calculation.

I believe it's true that CoreXYs should go linear guide route and leave rods behind but there is room for carefully designed printers with rods. You just have to know a lot more and be twice as attentive with your design. I wouldn't dismiss supported rods rails for a better/cheaper approach.

Regards.

Quote
DjDemonD
This seems to cover it [blog.helixlinear.com]

Quote
Koko76
If you can get both for he same price, and the ballscrew is of passable quality go ballscrew. Better efficiency and typically better precision. Either way, if using for x and y you will need to deal with backlash inherent in both. Frequently this is done with two nuts that are spring loaded against each other.