Thoughts on building a new CoreXY 3D printer

So after almost two years of printing on a classic Mendel-style machine which I started building as a Mendelmax 2.0 replica but ended up altering the original construction quite a lot, I think that the time has finally come to build a second printer where I invest all the knowledge I have gathered while working on the first printer.

After so much drama with the moire and rippling effects and much research on the topic I have come to the conclusion that a CoreXY machine will be the best bet for me. I definitely don't need a delta printer, the polar printers aren't anywhere near the cartesian ones in terms of print quality, SCARA is pretty much uncharted territory, so CoreXY seems like the last available viable option. It is fast, reliable, fairly simple and the cube-like constructions make it easier to design an enclosure for printing ABS

I was looking at some existing CoreXY designs (no need to reinvent the wheel here) there is this one that caught my eye: [www.youmagine.com]
I like a lot that the heated platform is supported on both sides and is driven by two motors. This adds a lot of stability to the Z axis.

Could anyone suggest other similar designs that I should consider? It would be a great advantage of the original CAD files where published together with the STLs because this would make it easier to modify the construction according to my needs...

PS. If anyone is wondering... I did see this thread but didn't want to bloat it as my question is not exactly the same...

Edited 1 time(s). Last edit at 06/25/2017 12:30PM by icefire.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

I don't know why you are writing off delta printers, because they can perform very well if they are built correctly, although cheap delta kits can be a nightmare. But if you want to go CoreXY, this one [somei3deas.wordpress.com] very good in the videos, has a mixing hot end, and can print at impressive speeds.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
icefire
I was looking at some existing CoreXY designs (no need to reinvent the wheel here) there is this one that caught my eye: [www.youmagine.com]
I like a lot that the heated platform is supported on both sides and is driven by two motors. This adds a lot of stability to the Z axis.

I don't think this printer is really a CoreXY. It uses the cross-rod configuration like the Ultimaker printers. BTW, does anyone know the term commonly used for this type of cross-rod XY mechanism?

Quote
dc42
I don't know why you are writing off delta printers

Mostly out of fear of the unknown. Deltas are more difficult to build and calibrate. I haven't really invested a lot of time reading about them but they seem to be more complex on the mechanical side which also means more potential problem areas. That said, I am sure that a properly built delta printer with the right components will work like a charm but on the other hand, this statement is true for almost all types of printers.

My goal here is to build a machine with a simple and robust kinematic setup.

Quote
LoboCNC
I don't think this printer is really a CoreXY. It uses the cross-rod configuration like the Ultimaker printers.

Yes but this is exactly what I like about it. It seems simple and yet effective.

Edited 1 time(s). Last edit at 06/25/2017 04:46PM by icefire.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Deltas are far simpler mechanically than most other types of printer. The bed is fixed. You build a triangular base frame (which is very easy if you buy the metal vertices), three identical towers with (preferably) linear rails or wheeled carriages, and a triangular frame (again, metal vertices. Also this almost trivial). Then you need 6 rods matched in length, with ball joints at the ends. These are not difficult to make if you build a jig first, or you can buy a set of matched rods with magnet ends and matching ball studs. All this is IMO much simpler than a system with a moving bed.

Calibrating a delta printer can be trivially easy if you have three things: an accurate build, a good Z probe, and an electronics/firmware combination that has a good auto calibration algorithm. You can manage without the Z probe, but then calibration will take longer. The cheap delta printer kits have none of these things, so calibrating them is often difficult.

The disadvantages of the delta are that the circular or hexagonal build area doesn't suit everyone, and that they are not really suited to direct extruders because the mass of the motor attached to the effector would degrade the performance. So they generally use either Bowden or cable-driven extruders, and multiple-extruder deltas are rare.

HTH David

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Thanks for the help. Well, it is not a decision I will make overnight. I will take a closer look at the delta printers.
Meanwhile, I am still looking for a good ultimaker-like design...

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

IMO ultimaker style gantries and I think that's the name for them, are more complex than true corexy (etch-a-sketch mechanism).

I started with a smartrapcore alu and heavily modified it. If I were going to do it again I'd use some of these ideas (I have some of this on my machine)
- a cube or cuboid (since corexy works rectangular as well as square) extruded aluminium frame. Corners joined by tapping the extrusions and drilling for access to screws so no weak corner brackets.
Make the gantry sit below the top of the frame to enable enclosing the top easier the base and sides are easy to enclose.
-pulley holders supported top and bottom
-linear rails
-metal mechanical components, cnc isn't needed for most of this, folded or drilled preformed aluminium can be adapted to do this.
-sliding metal motor mounts for tensioning the a and b belts separately with easy to adjust thumbscrews.
-a two side (minimum) supported, vertical z axis, could be leadscrews or belts to drive it, 3 or 4 sided support even better,no cantilevered beds.
-cast tooling plate bed with mains heater.
-duetwifi controller
-two x carriages (I have a scheme for doing this which is much easier than the corexyu scheme, which is admirable but complex) has to be the best way to do multi-material printing or printing two identical objects simultaneously.

Just some ideas, not a cheap machine but would be pretty awesome. To be fair my corexy with only half of this goodness prints really well, so I'm not upgrading it just yet.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

I agree with pretty much all points, however, if I reuse my current print bed (320x250mm), I am going to support it from two sides with large and stable parts. A three-point-guide-rail-support may be a good but still driven by two motors. The cast tooling plate is also not a must. I currently have one but with an adequate Z probe there is practically no difference.

However, I have been thinking on what

Quote

dc42

said. Maybe a Delta would actually be cheaper to build while still providing good print quality. I still have to think about it.

My priorities for the new build are (in order of importance):
- print accuracy and precision, i.e. when I print a 40mm circle, a 40mm circle comes out
- print quality, i.e. lack of artifacts like moire/rippling, etc.
- build costs
- simplicity and proneness to defects

Speed is not a huge priority.

I still don't know if I am going to keep my current build or reuse some of the good parts.

What are the typical issues with deltas? Like for example, all printers with moving beds have (at least to some extent) a problem with ripples. Are there similar sorts of typical setbacks for deltas?

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

There is an extensive thread about moire/ripples in the delta forum. Basically with all metal construction and decent controller such as duet wifi/ethernet you won't see many artefacts. Stationary bed is a nice thing to have and no moving motors. But build has to be very precise. My kossel Xl frame has 0.1mm max difference between the tops and bottoms of all three towers over 315mm. The carriages, rods, effector etc all have to be very precise, use linear rails not rollers or rods. See this guide [duet3d.com] if that looks doable then by all means build one they are very impressive machines. Expect some headaches, probably slightly more than with corexy, certainly more than with cartesian.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
Expect some headaches, probably slightly more than with corexy, certainly more than with cartesian.

What sorts of headaches? I am not scared of having technical issues because I always end up learning new things from them but I don't know what to expect from a delta...

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

If you have time read my build log for my kossel XL and of course DC42's who inspired me to build mine.

Things like slightly misaligned towers, carriages, rods not exactly the same length, the choice of how to fix the rods to the carriages and effector (ball joints are okay but have play/backlash, magnets are great but have to be done right or they come apart).

I have to say if I were doing it again, I'd make a frame like the Tevo Little Monster, two precut plates for top and bottom horizontals and c-beam extrusions, very simple and if the plates are laser/waterjet cut, then very accurate and rigid. But I'd use Duet controller, piezo probe, genuine v6 hotend, tooling plate bed and mains heater, metal effector and carriages and Haydn Huntley's magnet system. The only issue with 4080 cbeam is that there isnt a channel in the centre for attaching linear rails with t-nuts, but you could drill the extrusion along the central inner aspect, or use a pair of rails per carriage (which would have amazing rigidity, not that you need it with one good rail per tower). The bigger you make them the more sensitive they are to these small changes, the same as any other printer but with a delta three pairs of parallel arms have to precisely hold a lightweight platform level across the whole build area and height. My microdelta was not that precise, easy to build and calibrate (which is easy with duet controller and a good probe - such as a piezo probe) but only prints 100mm diameter. The kossel XL I would take the pepsi challenge for precision with any delta out there, it calibrates to a 0.006mm deviation across its 300mm bed, but was a long slow build requiring a no compromise approach to accuracy.

But build one, deltas are not scary they're awesome.

Edited 2 time(s). Last edit at 06/29/2017 02:47PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions