Welcome! Log In Create A New Profile

Advanced

Hlidskjalf, Dual CoreXY

Posted by prot0typ1cal 
Re: Hlidskjalf, Dual CoreXY
October 16, 2017 05:27PM
bobbyd,

Nice that you've decided on CoreXYU smiling smiley

Might I suggest some 2060 posts for the corners, that way you can attach the linear guide to the center T-slot, similar to what I did on Hlidskjalf, and avoid all those brackets :p
Or just point the corner 2020 in at 45 and work out trick corner brackets...
Re: Hlidskjalf, Dual CoreXY
October 16, 2017 06:06PM
Quote
prot0typ1cal
bobbyd,

Nice that you've decided on CoreXYU smiling smiley

Might I suggest some 2060 posts for the corners, that way you can attach the linear guide to the center T-slot, similar to what I did on Hlidskjalf, and avoid all those brackets :p
Or just point the corner 2020 in at 45 and work out trick corner brackets...
Good suggestions but I think I am fine with the current design. I already bought the extrusions and i kinda like having all of the brackets I think it makes it look more interesting. And I have tons and tons of plastic that I need to do something with anyways. I am curious however as to why o_lampe said my mechanics would fail.

I finished my BOM I plan to have all of the parts bought by April. So I hope to have my printer up and running by like June. Everything should cost me around $2600. I was really surprised, I was able to source almost all of the parts from the USA for just as cheap as some of the stuff on aliexpress after some digging.
Re: Hlidskjalf, Dual CoreXY
October 16, 2017 07:30PM
Quote
bobbyd
Quote
o_lampe
4 ballscrews and 4 linear rails will fail. This is overconstraint twice ( is that the right word? )
IMHO two rails and 3 ballscrews will be better.
Also, I wouldn't put bearings on top of the ballscrews, but selfaligning (thrust-) bearings on the bottom end. Use pulleys and closed loop belts to drive them.

I'm not sure what you mean by over constraint. And are you saying the linear rail in conjunction with the ball screw is over constraint or the aluminum rods connecting the four corners. Also I'm super OCD so having 3 ball screws with 2 linear rails is very displeasing to me. Its why I did 4 rails and ballscrews. And I'm fine with the motor directly connected to the ballscrew with a bearing holder on top. I want to make it as simple as possible. And as symmetrical as possible...

Using 4 rails and screws makes exactly the same non-sense that using 4 "leveling" screws for the bed does. Here are some of the problems with it:

1) Keeping two motors synchronized can be difficult. Keeping four synchronized is nearly impossible.
2) Aligning and maintaining the alignment of two widely separated guide rails parallel to each other can be difficult. Aligning 4 parallel to each other is much more difficult.
3) Any minor misalignment will cause the mechanism to walk or bind, and it will always have minor misalignments.

Two guide rails provide all the required guidance and resist lateral forces. 3 screws provides a stable lifting platform for the bed support the same way that a table with 3 legs never rocks.

You can turn multiple screws by either driving them with a single motor and a looped belt or by using three independent motors with their own drivers and cables and a Duet controller board (plus an expansion?) that can drive them independently to auto level the mechanism each time it powers up.


Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Hlidskjalf, Dual CoreXY
October 16, 2017 11:22PM
What DD said.

Whenever you use more than one of anything, alignment becomes an issue. Add more, the more everything has to be aligned to prevent stiction, or worse, jams.
Which is why you see a lot of 2 guides (rods) and one screw. That and it's cheaper. In theory you could make a 3d printer Z with a single linear guide and threaded rod.

I'm using 2 ballscrews with independent drivers, independent home switches, that require both ballscrews to home during power up (Z and W).
Only I'm using RADDS/Due combo, though more or less the same as a Duet, in any case, running RRF with an expansion board.
Using 4x ballscrews mean you will absolutely need auto level, as there's not enough home switch IO to go around, especially with CoreXYU as you intrinsically already need 4 (X, Y, Z, and U).

Of course, if all 4 ballscrews are belted together, than disregard all of the above, except the alignment stuff :p
Re: Hlidskjalf, Dual CoreXY
October 17, 2017 04:59AM
Back to the drawing board then I guess
Re: Hlidskjalf, Dual CoreXY
October 17, 2017 12:53PM
My printer Z table has 4 ACME screws on a common belt and 2 linear rails. Aligning everything was a real chore. What I would up doing is making four collars that bolted to the Z table that held 4 8mm short rods in the same place that the ACME nuts would go. Using that I could align the guide bearings top and bottom with the rails so all six items are exactly parallel.

Also had a binding problem with the rails that no amount of aligning would get rid of. Don't know if it is a problem with the extrusions not being exactly parallel or with the cheap Chinese rails. Followed Jetguy's method of aligning with no joy. What I would up doing is making two brackets with spherical bearings that allows the bearing trucks 3 degrees of freedom in rotation. Z-travel is smooth now.

Thought that I would have to do something to keep the 4 attachment points held down so the table wouldn't rock. Turns out with this big lift table there is just enough flex that all four corners are bearing weight and there is no movement. The build surface is another thick plate of aluminium above the lift table that is supported at 3 points, 2 in front and one in back.

Advantage of this system is the large Z lift table is held stable by the 4 screws. Also, because the screws are past the ends of the build surface the printable surface area is very large for the outline of the machine. Downside, of course, is this has been a complicated thing to figure out and build. Still might need to swap out the Nema 17 for a Nema 23 on the Z axis.

Working out of town now so still no plastic through the contraption to prove it out.

FWIW,

Mark Napier
Re: Hlidskjalf, Dual CoreXY
October 17, 2017 02:59PM
My printer's Z axis is made with it's own U-shaped "subframe" that allows the whole thing to be removed from the printer as a unit. With just two linear guides, alignment was extremely easy and it just worked form the first time I bolted it together. The 4040 frame members are bolted directly to each other, then the corner plates were added to hold the drive pulleys and bearings. Squareness of the frame was verified by measuring the diagonals. The guide rails sit in a flat slot I milled into the face of the frame members (not entirely necessary- the video linked below was made before I did the milling).

UMMD Z axis test 1

It's belt lifted so there are no issues with screws causing z wobble, and though it still requires careful alignment of the pulleys and belts, it's a lot more forgiving that screws.

Lifting 4 kg

You can see some of the mechanism in final form here. The other side of the bed is identical except for the limit switch.

Zmax limit switch operation

The bed is only 300 mm square, so if you're doing something much larger, you may need to do something more to ensure stability.

Edited 1 time(s). Last edit at 10/17/2017 03:04PM by the_digital_dentist.

Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Hlidskjalf, Dual CoreXY
October 21, 2017 09:23PM
Modified the PanelDue display, just because smiling smiley


Re: Hlidskjalf, Dual CoreXY
October 22, 2017 04:29AM
Nice, I like it!
Re: Hlidskjalf, Dual CoreXY
October 22, 2017 08:26AM
Very nice!

You seem to have a lot of axes. What is the W axis used for? You can hide the later axes (V and W?) if they don't need to be visible to the user, see the P parameter in the M584 command.


Delta printer calibration calculator, mini IR Z probe, and colour touch screen control panel: [escher3d.com]

Large delta printer, and other 3D printer blog postings: [miscsolutions.wordpress.com]

Disclosure: I have a financial interest in sales of the Panel Due, Mini IR height sensor, and Duet WiFi/Duet Ethernet [www.duet3d.com].
Re: Hlidskjalf, Dual CoreXY
October 22, 2017 09:06AM
Thanks smiling smiley

dc42, yea created a StaticBMP4Field (4 bit bitmap with a custom palette) to paint the pic. Plus had to limit maxHeaters to remove the boxes that get painted the background color to hide unused heater icons.
Sweet that PanelDue is open source !

Regarding RRF, the V and W had problems getting the dual ballscrew to home with them hidden. Have home switches on both the Z and W (Z screws). Works fine with the shown, but fails otherwise.

From my config.g
; Axis and motor configuration
M584 X0 Y1 Z2:5 U3 V4 W8 E6:7      ; Set up steppers for x, y, u, v, z and e, must not hide W for dual ballscrew
M574 X1 Y2 Z2 U2 W2 S0			   ; set endstop configuration. X Z and W endstops at min, Y and U max, all NO active low (S0)

home.g
; homez.g
; called to home the Z axis
G91

; Split Z into Z and W
M584 Z2 W5

; Move quickly to Z axis endstop and stop there (first pass)
G1 Z500 W500 F1800 S1

; Go back a couple mm on Z
G1 Z-2 W-2 F600

; Move slowly to Z axis endstop once more (second pass)
G1 Z4 W4 F300 S1

;Add offset by changing 0 to some distance to level bed
G1 Z0 W-0.83

; Combine Z and W
M584 Z2:5 W8

; Set Z position value
G92 Z415
G90
Sorry, only registered users may post in this forum.

Click here to login