My CoreXYU Printer project

First a few goals I wanted to achieve with this project:

• Large print volume for printing big objects.
• Print PLA, ABS, TPU and other materials.
• A minimum of two colours/materials in one print.
• Layer height: 0.20+ (0.1 should not be a problem but takes too long to print)
• Print speeds: 60-100mm/s with decent quality
• Travel speed: 250-300mm/s

Quick summary of build:

• Dual extruder CoreXYU IDEX (two carriages share same gantry).
• Outer dimensions without top: 750x750x1200mm
• Build area: 500x500x750mm
• 8mm milled aluminium heated (1400W) bed.
• 40x40 aluminium profiles for sturdy frame.
• 20mm steel linear rods with igus iglide sleeve bearing, 12mm aluminium for gantry, cheap and readily available.
• Heated enclosure (up to 70°C), planning to use PC sheets. Electronics and steppers are located at bottom of printer to keep out of enclosure.
• Three-point auto levelling bed that can tilt at end of print for easy access to bed.
• Air pump part cooling.
• E3D v6 hotends (might consider water cooled hotends later).
• 24V PSU.
• Duet Wifi + duex 5 as controller.
• Remote direct drives for lightweight carriages (Zesty nifty).
• 4 * Neam 17 0.9deg 2A 46Ncm 17HM19-2004S steppers for corexyu.
• 3 * Nema 17 1.8deg 2A 59Ncm 17HS19-2004S1 steppers for bed.
• 2 * Nema 17 1.8deg 2A 59Ncm 17HS19-2004S1 steppers for extruders.
• Quite a lot of printed parts as I do not have access to an aluminium cnc mill (yes, I know printed parts is not ideal for print quality but lightweight carriages might mitigate this problem somewhat).

Originally, I considered a Ultimaker style gantry but got inspired by prot0typ1cal dual corexy project Hlidskjalf and decided to go for that kind of kinematics.
This project is my first 3d-printer build so the design probably has quite a few flaws I’m sure you guys will point out
The end goal is not a finished printer that can be mass produced but a one of printer I can continue to improve and try new ideas on.

I have built the basic frame, installed the heated bed, extruders, hotends, a z probe and part cooling air pumps.
I have implemented corexyu in the dc42 version of reprap firmware for the duet wifi.

Some recordings from the build.
Testing corexy kinetics: [youtu.be]

The first prints, sliced with cure 2.6 using RepRapFirmware 1.19 beta 5 with my corexyu patch and changed M104 and M109 to make them set both active and standby temperature. To my knowledge, the first prints with a CoreXYU printer.

First test of corexyu firmware support: [youtu.be]

First benchy, need a ooze shield and part cooling!


Second benchy with cooling and shield, better but still need tweaking.


Third benchy, removed the purge tower and increased the speed and tweaked some settings. The cooling did reduce ooze from standby extruder a lot.



Third benchy printing: [youtu.be]
Need to decrease temperature on the white PLA a bit me think.

Not the prettiest printer out there yet


Edit: Managed to inline the pictures.

Edited 8 time(s). Last edit at 06/29/2017 12:20AM by lars.arvidson.

Cable management usually last on the check list.
Waiting for my printer to work so I can print a control case.

3D printed parts can be as strong as aluminum, just need to be bigger and add more fasteners.
Your printer looks pretty sturdy, with a massive build volume.

Well done.
Glad I could be of inspiration.
Only wish I could have finished mine 1st, LOL

Well done!
I thought one of the advantages of CoreXYU is to park the unused hotend in a place where the nozzle opening is blocked by a piece of sheet metal to prevent oozing? ( One at either end of the gantry)

Yep, I look into that although it’s not a priority atm. You will still have quite a long travel from parked when hotend is at active temperature. Atm part cooling is doing a decent job limiting the ooze when parked at standby temperature. I have to study it a bit more before deciding what to do.

The other way to prevent ooze is to use set the standby temperature lower than the active temperature and retract a significant amount of filament when a hot end goes to standby. This works extremely well on my dual-nozzle Ormerod, using a standby temperature of 150C and 10mm retraction (less retraction might work but I haven't tried). However, all-metal hot ends don't take kindly to large amounts of retraction, so you would probably need to use en E3D Lite 6 with this technique rather than the E3DV6.

---

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].

David, do you see a problem including the change of M104 and M109 in you version of the firmware? All I did was to set standby as well as active temperature when these commands are run.

Edit: Btw, I used standby temp of 185 and retraction of 16mm on tool change printing the third benchy.

Edited 2 time(s). Last edit at 06/29/2017 11:24AM by lars.arvidson.

Quote
lars.arvidson
David, do you see a problem including the change of M104 and M109 in you version of the firmware? All I did was to set standby as well as active temperature when these commands are run.

Edit: Btw, I used standby temp of 185 and retraction of 16mm on tool change printing the third benchy.

Please explain what you mean by "set standby as well as active temperature when these commands are run".

---

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].

I'm no expert on Cura 2.6 but it looks like it uses M104 & M109 to set temperatures both for active and inactive tool.
It start of the print with setting both tools to active (printing) temp:

M104 S210
M104 T1 S210
M109 S210
M109 T1 S210

At start of first layer it sets inactive tool T1 to standby temp (standby temp as defined in cura):

;LAYER:0
M107
M104 T1 S185

Some time before switch to T1 it sets the inactive tool to the active temperature (without wait). I guess this is done to minimize wait for temp at actual tool switch.

G1 F1500 X226.097 Y246.462 E29.0372
G1 X226.167 Y246.532
G0 F6000 X225.745 Y246.675
M104 T1 S210
G1 F1500 X224.845 Y245.775 E29.06258
G1 X224.775 Y245.705
G0 F6000 X224.845 Y246.341

Some time before switch to T1 it sets active tool T0 to "Final Printing Temp" 195

G1 F1500 X261.642 Y251.46 E32.75277
G1 X261.712 Y251.53
M104 S195
G0 F6000 X261.397 Y251.78
G1 F1500 X260.401 Y250.785 E32.78084

Tool change looks like this:

T1
G92 E0
M109 S195
M104 T0 S185
G1 F1200 E14
G0 F18000 X233.8 Y255.038 Z8.1

My change to the code in bold:

void GCodes:etToolHeaters(Tool *tool, float temperature)
{
	if (tool == NULL)
	{
		platform.Message(GENERIC_MESSAGE, "Setting temperature: no tool selected.\n");
		return;
	}

	float standby[HEATERS];
	float active[HEATERS];
	tool->GetVariables(standby, active);
	for (size_t h = 0; h < tool->HeaterCount(); h++)
	{
		active[h] = temperature;
		standby[h] = temperature;
	}
	tool->SetVariables(standby, active);
}

After this change Cura 2.6 gcode works like a charm!

Edited 2 time(s). Last edit at 06/29/2017 06:48PM by lars.arvidson.

I can see why that change works for Cura. But with slicers that don't manage active and standby temperatures like that, it would prevent you from using the automatic management of active and standby temperatures provided by RRF because the standby temperatures would get overritten.

Edited 2 time(s). Last edit at 06/30/2017 06:06PM by dc42.

---

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].

Hmm... Maybe there could be a config command to disable standby temp and only use active temp for both active and inactive tool? That would support a slicer like Cura but not changing anything for someone that what to explicitly set standby temp with G10.

Edit: Maybe G10 R-1 could disable standby temp...

Edited 2 time(s). Last edit at 07/01/2017 01:17AM by lars.arvidson.

I quick progress report.

I switched to Cura 2.6 from s3d. The dual extruded support in s3d appears to be a bit lacking. I wanted to use a skirt but it put it around the part from one extruders right were the other extruder would print, not around the entire model. I have had similar problems with single extruder where the skirt was put under support. I found Cura 2.6 to work well although on a bit slower than s3d. I did a firmware hack so both standby and active temperature was set when M104and M109 was issued to take advantage of Curas excellent temperature management.

I remodelled the carriages in Fusion 360 (originally done in OpenSCAD) and made a few changes. I had placed the BLTouch mount in the path of a belt to the other carriage, don’t know how I could have missed that. Moved the mount a bit and added a cable management feature. I also simplified the hotend fitting. It’s now only a slot in the carriage. Turns out that when I printed the two carriages one was a perfect fit but the other was a bit too loose. Will have figure out a new design. Maybe one where I don’t depend so much on dimensions but use bolts to secure it. Second problem is the air tubes. After a bit of fiddling the holes get too big, need to be able to secure the tubes without using cable ties.



I had a bit of z wobble that I can see when printing in white pla, mostly on a y-direction. I figure the problem was that there was that there was too much play (y-direction) in the bushing on the two front guide rails. The back guide rails don’t stabilize the bed in y-direction (bed tilt feature , I can lower the front to get easy access to the print once it’s done). I put in screws to squeeze the slop out of the bearing on y direction. X direction should be good as the two opposing front guide rails takes care of that. This removed a lot of the banding but there a little bit of banding left. One of the leadscrews are not straight, will cut a new one and replace it, might help a bit…


I also tested the 1.19 beat8 version of the duet reprap firmware with CoreXYU support but there are some bugs affecting IDEX printers in it so I’m back to using my own fork.

I have made my first G29 mesh probing with the BLTouch, worked like a charm. 21x21 probe points. Takes some time to preform! Need to tweak it to get it flatter!


Edited 1 time(s). Last edit at 07/04/2017 03:09PM by lars.arvidson.

Printed my first item that used more than a tiny part of the bed. An arm to hold the cables going to the carriage. First few tries were failures. I’m a bit embarrassed I did not figure out what went wrong first time it happened. I had the filament spool on a bolt in one of the corners feeding the filament directly to the extruder instead of using the PFTE tube running with the other cables. When the extruder did long fast moves away from the spool the filament sometimes jammed itself on the spool stopping the carriage from moving and offsetting the layer.
Once I used the PFTE tube running with the other cables it worked better…
It’s not too pretty, the infill shows through the walls. I’m happy with the air pump cooling, although the pump is too loud. The rounded tip and holes came out close to perfect without support.

Quote
lars.arvidson
...
I also tested the 1.19 beat8 version of the duet reprap firmware with CoreXYU support but there are some bugs affecting IDEX printers in it so I’m back to using my own fork.
...

Please elaborate, otherwise they are unlikely to get fixed unless someone else reports them. I don't have an IDEX printer to test on.

---

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].

Hi David!

I did post about it on the duet forum, [www.duet3d.com], maybe you missed it...

Hi Lars,

Did this [www.duet3d.com] work around the issue? (i.e. use S2 on the G1 command in tfree.g that parks the old tool).

I'll put a fix in beta9.

---

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].

Just G1 S2? No X?

Normally G1 S2 don't work well with CoreXYU (at least as I remember it using it by mistake instead of S1 some time ago ), moving one stepper out of synch with the others is not optimal...

Edited 1 time(s). Last edit at 07/08/2017 03:23PM by lars.arvidson.

I mean use S2 as well as the X parameter.

I've put a provisional fix in the source code and if I finish other work (support for bed levelling using multiple Z motors) then I will release that version tomorrow.

---

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].

Ok, I'll give the new version a try tomorrow then!

I have had a bit of problem with the tension block tilting slightly resulting in the belt wearing on the idler flanges (or worst case slipping of). It appears it’s mainly the block titling a bit and not the idlers themselves. Ideally the idlers bolt should be supported above and below instead of just screwed into the block but I’ll think I’ll leave that as it is for now and redesign that when and if they start to tilt.

I made a piece that supports the rods that the block slides on so they are held horizontal. I added insert rivet nuts secure the idlers a bit better. I made the holes for the rods slightly smaller and printed the block with enough outlines to make it solid.


So far it has worked well.

I modified the design of the carriage so that the important dimensions and offsets are multiples of the layer height. The hotend fits good now. Added a grub screws to keep the airpipes in place and slightly redesigned the endstop holder. The endstop can now be inserted and removed with the wires soldered on to it.
I got me some spiral cable management sleeving and combined with the arms I printed keeps the cables tidy.


I also tested the 1.19 beta 9 firmware. The “Bed levelling using multiple independent Z motors” feature look promising. Had to do a small fix to get it to work with CoreXYU. Its currently only reporting adjustment, not actually doing it, but I hope David enables that in the next beta...
I also tracked down a IDEX tool change bug that was very annoying. The firmware did not properly update the second tools position when changing tools. If you did a relative move after the tool change to T1, say G1 X-1, it would use the X position of T0 and move T1 to this pos – 1 resulting into the tools crashing together if T0 was in a position not outside T1s min/max.

All in all, I happy with the progress. I will print and replace the other carriage next.
This far I have printed all the parts in PLA as I have a lot of it and I got it cheap but I’m thinking I’ll start replacing the stepper mounts with parts printed in ABS soon. The steppers are pretty hot and I think the PLA is getting soft.
I might use PC-MAX to replace the parts that will be inside the enclosure but I’ll wait until I’m happy with the design as it’s expensive. Might go for ABS first and later switch to PC-MAX.

Lars, did you buy the pipes from makerhive or did you make you own? Can you post pictures of the pipes and how you bent them? I like the effectiveness of the berdAir (a system like you have) but the pipes that come with it i find are difficult to bend just right..

Hi Qdeathstar,
The pipes are 4mm x 3mm 304 Stainless Steel Capillary Tube Length 250mm Stainless Pipe I bought from banggood. They are pretty hard to bend. I put an electrical wire that’s a bit over 2mm into it and apply a lot of force with a 20mm steel shaft, simple and effective, not very precis but you can cut it down to the length you need.



Edited 2 time(s). Last edit at 07/16/2017 09:25AM by lars.arvidson.

And a picture of the carriage with the air tubes.

which air pump did you use?

I got two (24V version) of theses: [www.aliexpress.com]

They are too loud if you run them at full speed. Atm I'm using a SSR and it cant do PWM frequency high enough to run them on less than 100%. I'm planing to connect them directly to my duet wifis fan output once I get a current limiting NTC thermistor and should then be able to control the speed.

Anyone got a air pump that generate acceptable level of noise (or even perceived as quiet)?

Edited 2 time(s). Last edit at 07/17/2017 05:36AM by lars.arvidson.

Lars,
What you want is a rotary vane, not a diaphragm pump.
Rotary vanes are much quieter, more efficient, and last longer.

[www.ebay.com]

Oh yea, and more expensive :p

Quote
prot0typ1cal
Lars,
What you want is a rotary vane, not a diaphragm pump.
Rotary vanes are much quieter, more efficient, and last longer.

[www.ebay.com]

Oh yea, and more expensive :p

Interesting, I'll look into it.
Thanks prot0typ1cal!

This rotary vane principle isn't far from a peristaltic pump.
I wonder, if we could print one?

Peristaltic pump

with all the diy membrane pumps on the internet i am surprised no one thought of printing one... u just need a dc motor and some surgical glove silicone.. i dont see y it has not been done..

Quote
o_lampe
This rotary vane principle isn't far from a peristaltic pump.
I wonder, if we could print one?

Peristaltic pump

That would be an interesting project, I wonder how many L/min air you could pump and how durable it would be...
The pumps I’m using now is rated at 13.0 L/min. It might be a bit more than is needed, maybe 6L/min is enough?