Core XY with dual gantries using a duet wifi
-
So just to be clear for stupid ppl like me, the duetwifi can with the present firmware support 2 moving gantries as X and U as a core XY - IDEX?
No, some firmware work would be needed to define which motors control the U axis.
Um so it would be 2, actually U and V? I sort of assumed the gantries and belts and steppers would be totally independent of each other.
-
I’m currently (slowly) building a dual carriage core-xy printer. The kinematics were described in http://forums.reprap.org/read.php?397,737863 over at reprap forum and David wrote a bit about what would need to be implemented in firmware.
Would love to see the Duet/RepRapFirmware support it! -
I'm struggling to get to grips with this idea. From what I can ascertain, to keep the axes on the same plain, they share a common Y but have different X axes. This means that the X axis must be very wide and /or one would lose a lot of travel. So, the only advantage I can see over simply fitting two hot ends to a single X carriage is that there would be a weight saving on one of the X axes. However, the Y carriage would still have to move both the X axes so the Y axis weight will be the same. Therefore any speed advantage would be limited to pure X moves which is unlikely to have much of an impact on the total time it takes to print something. It seems a lot of complexity for very little practical gain or have I missed something?
-
Well, you do have 4 steppers contributing to z-axis movement and 2 for x-axis. Although you have to move shafts and the two hotends for z-axis movement and only a hotend for x-axis. Differences in z offset of the two hotends should not be a problem using two carriages and you could use two very different hotends if you find that useful.
-
Yes, it would be interesting to support 2 moving gantries. In effect, each tool would be a separate CoreXY machine except that they would share the Z axis. We'd need to map the Y axis to the V axis on the second tool, and define the U and V positions as being defined by the U and V motors in the same way as X and Y. It doesn't sound difficult.
Yes, that is exactly what I'm thinking of! Time to start setting it up in cad then!
-
Ahh.. two gantries, I was thinking one gantry (shared Y and Z and independent X). I’m too preoccupied with what I’m building so I missed that part, sorry
-
Yes, it would be interesting to support 2 moving gantries. In effect, each tool would be a separate CoreXY machine except that they would share the Z axis. We'd need to map the Y axis to the V axis on the second tool, and define the U and V positions as being defined by the U and V motors in the same way as X and Y. It doesn't sound difficult.
Yes, that is exactly what I'm thinking of! Time to start setting it up in cad then!
I still can't get my head around the mechanics of how this would work. I can't see how you could have the gantry's on the same plane without one interfering with the other. Unless, the gantry's are stacked one above the other? But then you'd have to somehow deploy one hot end to move in Z by the thickness of one of the gantry's. I just can't picture it in my mind - be interesting to see a design.
I am in fact working on a stacked gantry design but that's a just a way of trying to mount 3 extruders above a diamond hot end whilst keeping the footprint small and the Bowden tubes short. In this case, both gantries will be active at the same time and the second gantry would probably be passive - i.e. linked to the lower gantry but without any motors or belts of it's own.
-
I presume you would
partpark one gantry at one end of the Y axis while you use the other one. Just as on an IDEX machine you park one X carriage at the end of the X axis while you use the other one. The printable length of the Y axis for each head would be reduced by the amount of space taken up by one gantry, and the printable length accessible to both heads would be reduced by the amount of space taken up by both gantries. -
Yes, it would be interesting to support 2 moving gantries. In effect, each tool would be a separate CoreXY machine except that they would share the Z axis. We'd need to map the Y axis to the V axis on the second tool, and define the U and V positions as being defined by the U and V motors in the same way as X and Y. It doesn't sound difficult.
Yes, that is exactly what I'm thinking of! Time to start setting it up in cad then!
I still can't get my head around the mechanics of how this would work. I can't see how you could have the gantry's on the same plane without one interfering with the other. Unless, the gantry's are stacked one above the other? But then you'd have to somehow deploy one hot end to move in Z by the thickness of one of the gantry's. I just can't picture it in my mind - be interesting to see a design.
I am in fact working on a stacked gantry design but that's a just a way of trying to mount 3 extruders above a diamond hot end whilst keeping the footprint small and the Bowden tubes short. In this case, both gantries will be active at the same time and the second gantry would probably be passive - i.e. linked to the lower gantry but without any motors or belts of it's own.
Like dc42 said, I'm thinking of dual gantries. A bcn3d sigma for example uses one Y-gantry where two X-carriages live. Imagine instead two independent Y-gantries with one X-carriage each. Each gantry+carriage is controlled by its own set of steppers.
-
I presume you would
partpark one gantry at one end of the Y axis while you use the other one. Just as on an IDEX machine you park one X carriage at the end of the X axis while you use the other one. The printable length of the Y axis for each head would be reduced by the amount of space taken up by one gantry, and the printable length accessible to both heads would be reduced by the amount of space taken up by both gantries.Yes that was the idea, ie parking a complete gantry each end was forming in my mind. The original Dbot is a 300(x)x300(Y) corexy printer but its made from openbuilds 2040 so its easy to re-dimension (cut). My original idea was to build something big enough to print my own storm trooper armour parts in one piece (to save glueing) for that I need 500~550mm in one dimension of X or Y and 450~500mm high and then maybe cut it back to something more physically manageable later if need be.
Being lazy the 2040s come in 1500mm lengths so I just chopped up 2 lengths and ended up with 4 x 750mm pieces and did the same for Z and there is the frame https://ibb.co/dtCE35. I just added the HGR15 Linear rail on top yesterday giving an absolute max stoke of 670mm, or around 600mm~630mm NET stroke. So the linear rails came with x2 boggies each anyway so I have added these both to the top rail since two were spare. https://ibb.co/bKvqbQ The second gantry will cost me 75~100mm at most if I hand and overhand them so on Y so I think I'm still OK for 550mm, Worst case I have to buy a bit more 2040 (4 x 1m), about $40US worth, that is not going to break the bank.
So structurally I am set, linear rail I am set, next and biggest nightmare is somehow running 2 sets of belts, hello fusion 360 and some evenings not to hard, its the multi-dimensional curves that drive me nuts. The last part will be the firmware to allow it to work. Would there be anything else Ive missed?
As an aside unlike the wanhao pile of crap mainboard that lasted 3 months (that would have cost me $165US to replace and they are dying like flies due to bad relays) v the duetwifi at $200US has well proven itself, its quiet and the web gui is just nice.
-
Hello im also thinking of dual extruders
i like this ide
http://www.thingiverse.com/thing:1446731
that way you dont need extra steppers motors
only need a way to power 2 magnets in the tool change script
is there a way to send a signal to the magnet to turn on/offfor example the tool script will do something likt this
->move left turn of magnet #1 (park extruder1)
->move right turn on magnet #2 (get extruder2)
->printsee here
https://www.youtube.com/watch?v=FiTmRHkiEUgsounds simple …
-
@Mrblom, you still carry the weight/mass of both from the looks of it, the whole idea is to have as low mass as possible consistent with rigidity etc.
Steppers are very cheap, like $6~11US so to have 4 instead of 2 is no biggee. The real cost is the headache of the design and the mainboard and, well is it worth it?
My original idea was to have a quick swap hotend, I might still do that if I cant get the belt layout sane. A job to investigate over the winter (I am in New Zealand) me thinks.
Thanks for input all…..it looks interesting and possible.
-
hello
it will be only a lite extra mass
it will only carry one extruder at the time
the extra mass will be the center "item" where the belts are connectedcheck this video
https://www.youtube.com/watch?v=FiTmRHkiEUg -
I drew up the original top end in Fusion360 and the only way I can see to make it work is to swap out the 2040 rail for 2080 and the gantries will be tiered slightly, otherwise you just cant get the belts to pass through each other! No biggee and not that hard to print either.
-
hello
it will be only a lite extra mass
it will only carry one extruder at the time
the extra mass will be the center "item" where the belts are connectedcheck this video
https://www.youtube.com/watch?v=FiTmRHkiEUgYou are doing this on a Prusa i3 type printer? in effect the Z moves slowly so mass isnt an issue so for this design architecture having 2 hotends might be OK as Y is the bed.
Extra mass hardly seems small as it would be in the order of 300~500grammes, hardly small unless using a bowden which isnt any good for flexibles.
For a core XY on the other hand I am moving X and Y and rapidly not just X. Everything I read says keep mass so low I just dont see the point of putting 2 complete moving hotends on Y's gantry as Y's mass is then increased which is what I wish to avoid.
With 2 gantries I avoid carrying the 2nd hotend's mass around when I dont need it. In terms of complexity I need 2 more stepper motors and 2 sets of belts and idlers in the order of $30US. I will need to print 4 parts at about 7hours each part. Then the 2080 rail but since I need to buy more 2040 anyway its only a marginal cost to upgrade to the 2080 for me, I can then cut up the 2040 I have for where I am short.
-
Just a thought but 2080 is twice the weight of 2040. Can you not tier the gantries by changing the mount and still use 2040? I still can't picture how this is going to work so maybe you can't but it was just a thought..
-
The gantries are still 2040 they are not changed at all from the original design and yes they are tiered. Its the frame that needs to be 2080 as the double set of belts need to be stacked to pass each other. So this mass does not matter, except the printer's frame at 780mm(Y)x550mm(X)x760mm high is already getting big and heavy. I am trying to see if I can get 2060 to work by having steppers at opposite ends but I suspect not. I am drawing it up in Fusion360 and have the original in front of me….
-
Ah OK. Out of interest, what is the range of movement on each of the 3 axes for that frame size?
-
this is what i mean
https://www.youtube.com/watch?v=3UE9jU6RDP8
(the video is of a early "dry run")
have been running this for 2 weeks now
works grate and no extra mass also make it posible to add more tools -
Any updates on this?
Long term I'm thinking the project at work would be best served by adaptive slicing with a large nozzle for solid infill and a small nozzle for detail. While calibration is easier if the two nozzles are on one head I don't like the idea of a second nozzle dragging over the print. Perhaps I'm just being over cautious there.
