Dualwire-Gantry (DW-G)

1,495 views
Skip to first unread message

Kulitorum

unread,
Jun 14, 2014, 5:00:41 AM6/14/14
to h-bot-and-core...@googlegroups.com
I'm working on a printer using a new principle for seperate XY movement of the printhead - the Dualwire-Gantry (DW-G) method.

http://www.reprap.org/wiki/Dualwire-Gantry_(DW-G)

I am working on a 3d printer with this principle. The XY axies are moving, the rest of the machine (endstops, z-axis) is still begin worked on.


Does anybody have an idea for a better z-axis then what makerbot and ultimaker uses?

Kulitorum

Ryan Carlyle

unread,
Jun 14, 2014, 2:42:38 PM6/14/14
to h-bot-and-core...@googlegroups.com
The "best" z-stage in my opinion is a three-leadscrew design belted off one stepper. Check the F306. Cantilevered Z-stages are fine for small bots but the longer you make the Y axis, the worse they get. 

That's an interesting gantry design. 
  • The Y axis is kinematically equivalent to a standard Replicator-style Y gantry -- the wire crossover serves the same function as a Y sync torsion rod. Kind of slick in that regard since it allows substituting several expensive parts (toothed pulleys, shaft couplers, etc) for cheaper parts. Could be made with belts too if you want. Just put the stepper in one of the corners rather than mid-line.
  • The X axis has some H-bot similarities. For example it looks like it has the same racking problem that H-bots do. But you can invert the belt path on one side of the gantry to eliminate that. It will require a couple more idlers but will balance torsional forces during X moves. See pic.


Eric Hsu

unread,
Jun 16, 2014, 1:51:42 AM6/16/14
to h-bot-and-core...@googlegroups.com
That's a really interesting method of XY movement. I think ill have to investigate it further. Might even implement it in my next printer.
I'm currently working on a scissor mechanism to keep the build platform level while using one leadscrew or belt drive to move the platform in the Z direction. I'm thinking of having the leadscrew/belt drive orientated vertically and applying the force on one side of the platform and wondering if the scissor mechanism will keep it level.
Maybe the picture will make more sense. 
DSC_0045~2.jpg

Kulitorum

unread,
Jun 16, 2014, 9:01:25 AM6/16/14
to h-bot-and-core...@googlegroups.com
Hi Eric,

Thanks.

I do like your z-axis idea - could be very stable, if it works :) Have you considered putting the stepper on the bottom of the assembly, so it just pulls the "legs" together, to raise the bed?

I think I will try to implement your z-axis, just to have the whole printer use new methods for everything.

Thanks.

Ryan Carlyle

unread,
Jun 16, 2014, 10:35:38 AM6/16/14
to h-bot-and-core...@googlegroups.com
Can't put the leadscrew horizontal at the bottom because it won't have a linear steps/position relationship. 

Ryan Carlyle

unread,
Jun 16, 2014, 10:40:05 AM6/16/14
to h-bot-and-core...@googlegroups.com
I like the scissor-lift idea, but I think it'll have to be perfectly parallel in order to work right. Might be difficult to build. Could work really well though. 

Eric Hsu

unread,
Jun 16, 2014, 12:32:45 PM6/16/14
to h-bot-and-core...@googlegroups.com
Glad you guys like the idea.

I like the scissor-lift idea, but I think it'll have to be perfectly parallel in order to work right. Might be difficult to build. Could work really well though. 
 
Yeah, I realized that after I drew it out. I was thinking maybe I'll put the leadscrew nut on some kind of swivel so the stage won't bind if things aren't built exactly to spec. or maybe put turnbuckles on the scissor so the build platform can be adjusted square with the leadscrew.

Marc Lee

unread,
Feb 4, 2015, 8:45:08 PM2/4/15
to h-bot-and-core...@googlegroups.com
Ryan, please can you explain why you think the original x-axis referenced by Kulitorum has racking problems ... and hence why you think your modification overcomes that racking? Thanks.

Enrico Da Re

unread,
Feb 5, 2015, 10:08:41 AM2/5/15
to h-bot-and-core...@googlegroups.com

Marc Lee

unread,
Feb 5, 2015, 9:44:42 PM2/5/15
to h-bot-and-core...@googlegroups.com
Thanks Enrico - yes I had seen that and I am very interested in the concepts behind Richrap's Slic3r, in particular the use of spectra with DW-G movement for X/Y, and the use of spectra for Z too.

I'm after a 400-500 mm cubish print volume, an all metal design (with no printed parts) to suit a ~80 C enclosure and all movement running on V-slot/wheels.

So I'm thinking of extending the spectra Z to the 4 corners to cope with a larger/heavier ~400 mm square heated bed too.

And I'm looking for others building similar printers to bounce ideas around with - anyone out there?

Unfortunately, this forum seems a bit dead ...... or is it "just resting" like in the Monty Python dead parrot sketch!

Ryan Carlyle

unread,
Feb 6, 2015, 9:50:03 PM2/6/15
to h-bot-and-core...@googlegroups.com
It's pretty dead.

Does spectra creep at 80C?

Marc Lee

unread,
Feb 7, 2015, 12:56:36 AM2/7/15
to h-bot-and-core...@googlegroups.com
Honeywell Spectra says this:
Advised long duration temperature limit 70° C
Advised short duration temperature limit (non-constrained fiber) 145° C
Advised short duration temperature limit (constrained fiber) 130° C

Is all Spectra from Honeywell or are people using different sources?

Simon Merrett

unread,
Feb 15, 2015, 5:00:03 PM2/15/15
to h-bot-and-core...@googlegroups.com
Hi Ryan (or anyone else who cares to answer),

Please could you explain where the extra couple of idlers are in your schematic, compared to the original  (non-inverted) X axis wire route? I've been staring at the two drawings for a while and can't work it out. The only difference I can see is the different anchor point and stepper placement.

On a general note, I think there is potential to get much more efficient conversion of rod length to build volume using this system. I'm a novice, having only build a QUBD TwoUp kit, but I'm intrigued about the possibility to recycle its parts into a much larger printer, with build volume circa 500x500x500. Obviously increasing printing speed is important for being able to use that kind of volume in a realistic way, so I'm intrigued by this method. Combined with a Bowden extruder and a powerful hot end, large volume printing could be quite accessible.

Thanks

Simon Merrett

unread,
Feb 15, 2015, 7:19:29 PM2/15/15
to h-bot-and-core...@googlegroups.com
Marc,

I'm very interested in making a printer in this way. Please update us on any progress you make on this thread and I'll do the same.

Simon

Marc Lee

unread,
Feb 15, 2015, 9:35:39 PM2/15/15
to h-bot-and-core...@googlegroups.com

Marc Lee

unread,
Feb 16, 2015, 7:51:24 PM2/16/15
to h-bot-and-core...@googlegroups.com
....and another topic on Spectra/wire http://forums.reprap.org/read.php?2,377858,page=3 (9 Feb on)

Marc Lee

unread,
Feb 16, 2015, 8:02:50 PM2/16/15
to h-bot-and-core...@googlegroups.com
Simon - I don't think there are two more idlers in Ryan's schematic compared to the original DW-G X-axis. And, in any case, I don't believe Ryan still thinks that there is a significant racking issue that needs to be overcome by using his schematic - correct me if I'm wrong please Ryan - see https://groups.google.com/forum/#!profile/h-bot-and-corexy-3d-printers/APn2wQcBTyqw_DhoupJAmgrZB4KRYVgj1BhlwI5UCphh9O8nbb0KYcTAxdENz5Xz6VSiFmkO5b6H/3dprintertipstricksreviews/qJoy7QfikUI/T2gEKHeLKewJ

Simon Merrett

unread,
Feb 16, 2015, 8:10:29 PM2/16/15
to h-bot-and-core...@googlegroups.com
Thanks for your updated comments. Here's the text of my last email to you for others to look at if they wish:

I'm afraid I have no knowledge of spectra and I plan to use much lower temperatures than that. I would want to check that the motors and electronics (including insulation) are either shielded from the heat or won't get damaged if they are trying to transfer their own heat of operation into an unforgiving 80*C.

Today I stumbled upon v-slot, which you can get metal wheels for, so that could be an option for your frame material. I certainly am considering using a single 2020 section of it for my X gantry and two sections of it for my Y rails (but with delrin/plastic wheels to keep the wear on the wheels, which are replaceable, and not on the rails). Having looked at the challenges of scaling 3D printer axes up, two significant issues appear to be the mass of the moving parts and the sag in the rods/rails. The 2020 v-slot appears to be better than 16mm dia steel rod from some uniform distributed and point load deflection calcs I have done, for a 600mm span. This is especially valuable as you can use only one of the sections to achieve what you'd need two 16mm steel rods for in the X gantry. This then drops the mass of the gantry rail from over 1kg to about 300g. I'm not saying that the wheels/mounting plate won't eat up some of that saving, compared to a 3D printed gantry end housing a LM16LUU bearing but I think it's definitely worth a shot and initial costs don't look significantly larger (quick survey of Aliexpress). Does anyone know a good supplier of V slot to UK? The vslot-europe website says all parts are only available for preorder.

Moving on to the mass of the moving parts a bit more, this seems important to me because we are probably going to rely on a system which appears more prone to the potential for slipping against the wire, compared to timing belts. If accelerations are too great we might lose accuracy in such a drastic way that the print becomes untenable. Another consideration for my planned large Y travel is that the wire is not infinitely thin and so it spirals up and down the spool/bobbin/pulley on the stepper motor. I'd have to do some calcs to check that my envisaged range of movement wasn't going to demand a stupidly tall pulley to deal with all those coils winding their way up it as it takes in and pays out wire. My backup mechanism is to use a HTD 3 belt in a U shape, similar to the DW-G Y layout. If I get to this point, I might get a larger (NEMA 23) stepper to try and accelerate the larger mass in line with the X axis ability. Of course I'm just pondering out loud here - I haven't even checked my Printrboard can serve a NEMA 23!

Finally, on Z-axis, my thoughts are to go with a 3 No. 8mm dia trapezoidal lead screw with brass nuts. You can get a set on Aliexpress for about $27. I plan to mount these vertically in 608zz bearings at both ends and either the top or bottom of one of them will have the Z stepper coupled to it. I will try and print a combined HTD pulley, stepper shaft AND 8mm lead screw coupler to connect them. The remaining two screws will be attached (in the same plane as the Z stepper) to their bearings with a combined HTD pulley, 8mm lead screw AND 608 bearing coupler. I have found a great pulley design software on Thingiverse  which I will explain to you if you want a head start on how to work it out (it took me a while but I'm not going to tar you with the same brush!). I will use a tensioner bearing too and this whole setup will hopefully allow me to do a Z axis with only one stepper. I saw one post where a novice (as I am) asked for help to scale up a printer (he didn't own one yet) and he was told he was ridiculous for not considering the weight of his massive parts and how he would spec a stepper powerful enough to lift them. The poor bloke was so new he didn't realise that the Z axis never needs to lift the part, only the Xand/orY axes OR LOWER the bed. So I'm hopeful that even if the bed gets heavy (about 2kg for a 500x500x3mm sheet of mirror glass) we won't have too much work for the NEMA 17 Z stepper to do, as the print will be "falling with style" as Buzz Lightyear might say.

Simon Merrett

unread,
Feb 16, 2015, 8:32:56 PM2/16/15
to h-bot-and-core...@googlegroups.com
Have you considered kevlar yarn?
Message has been deleted

Alex Borro

unread,
Feb 17, 2015, 8:11:12 AM2/17/15
to Marc Lee, h-bot-and-core...@googlegroups.com
What GT2 belt are you guys using??

I have about 20 meters at home and it weights about 10g/m not 96g/m.. and they are good quality one.. 
Yes, it weights more than cables, but my CoreXY bot with 320mm x 320mm area uses about 4m of GT2.. this is awesomes 40g os running belts.. I can get 450mm/s of travel speed, 20t pulleys, 4.8Kg motors and 1/16th microstep running marlin. By the way, I have just fixed some CoreXY in Marlin dev branch - recommended to folks running such system.

Cheers.

Alex.

2015-02-17 9:20 GMT-02:00 Marc Lee <m....@shinfin.com>:
By my calcs both Kevlar yarn and steel wire give significant weight savings over GT2 belts.

0.5 mm diameter Kevlar yarn weighs a very small ~0.3 g/m and stretches about the same as a GT2 belt which weighs about 96 g/m
0.5 mm diameter steel wire weighs a very small ~1.54 g/m and stretches about half as much as a GT2 belt which weighs about 96 g/m

So yes, whilst Kevlar is lighter than steel, this is negligible for 500 mm cubish 3D printers which will have an XY belt system of about 5m length - as steel wire will still be under 10g and stretches half as much (which may be a benefit). Whereas 5m of GT2 belt weighs about 500g which is very significant.

I haven't tried Kevlar yarn or steel wire. But factors other than weight will drive the decision, like ease of termination, wear on the pulleys, minimum pulley radius (hence pulley mass).

--
You received this message because you are subscribed to the Google Groups "H-bot and CoreXY 3d printers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to h-bot-and-corexy-3d-...@googlegroups.com.
Visit this group at http://groups.google.com/group/h-bot-and-corexy-3d-printers.



--
"Não é o mais forte da espécie que sobrevive, nem o mais inteligente. É aquele que se adapta melhor as mudanças"  ( Charles Darwin )

Alex Borro

Ryan Carlyle

unread,
Feb 17, 2015, 12:33:04 PM2/17/15
to h-bot-and-core...@googlegroups.com
Spectra DWG with light, low-friction pulleys should have negligible racking force. If you use a heavy belt with big pulleys (eg to respect belt bend radius at reverse bends on the X carriage) it will have some racking force during Y axis accelerations. Whether that impacts print quality will depend on the Y motion stage drivetrain stiffness and rigidity of the Y carriage / X rail interface.

Ryan Carlyle

unread,
Feb 17, 2015, 12:34:44 PM2/17/15
to h-bot-and-core...@googlegroups.com
As for steel wire -- I think when you figure out the minimum bend radius to hit a reasonable bend cycle fatigue life, it's going to be larger than you want to use. There's a reason we all use fiberglass or aramid for belts and lines.

Marc Lee

unread,
Feb 17, 2015, 10:51:43 PM2/17/15
to h-bot-and-core...@googlegroups.com, m....@shinfin.com
@ Alex Borro - many thanks for picking up my error in GT2 belt weight. I apologise for that. I've deleted my old post and rewritten it here....

By my calcs both Kevlar yarn and steel wire give some weight savings over GT2 belts.

0.5 mm diameter Kevlar yarn weighs a very small ~0.3 g/m and stretches about the same as a GT2 belt which weighs about 10 g/m
0.5 mm diameter steel wire weighs a very small ~1.54 g/m and stretches about half as much as a GT2 belt which weighs about 10 g/m

So yes, whilst Kevlar is lighter than steel, this is negligible for 500 mm cubish 3D printers which will have an XY belt system of about 5m length - as steel wire will still be under 10g and stretches half as much (which may be a benefit). Whereas 5m of GT2 belt weighs about 50g which is still low (but might perhaps have some impact on printing - I don't know).

I haven't tried Kevlar yarn or steel wire. But factors other than weight will drive the decision, like ease of termination, wear on the pulleys, minimum pulley radius (hence pulley mass).
To unsubscribe from this group and stop receiving emails from it, send an email to h-bot-and-corexy-3d-printers+unsub...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages