[idea] CoreXY "+" - Alternative to "H" arrangement, keeping CoreXY belt

ok so i noticed that corexy is based on the "H" principle, but i haven't seen any discussion of alternative arrangements whilst also keeping the corexy belt layout. i'd therefore like to ask people what they think of the above idea (ignoring that it's drawn by hand with "gimp"...)

to repeat and make absolutely clear: belt arrangement here is *exactly* as per corexy.

there are three linear bearings required per axis (pink for y, green for x), they are slightly offset in height, so that they go through the carriage (light orange). min/max/both endstops can be placed in the corners (statically) which is nice - i find it a royal pain to have to have the endstops moving on the carriage (or the x-ends).

in theory you could have doubled-up rods and bearings, either stacked vertically or horizontally, per axis, but this would be the minimum arrangement.

what is the motivation for considering this?

my primary motivation is that i see in the H design the weight of the carriage is exclusively placed onto the x axis, asymmetrically. i.e. when moving in the y direction, the carriage *and* the rails (and bearings) have to be moved, but when moving in the x direction, only the carriage has to move.

this bothers me. with the above arrangement, the weight distribution is balanced evenly between x and y. when a normal H-style corexy moves diagonally at 45 degrees, due to the "add/subtract" design of corexy, that's when the maximum load is placed on a single stepper: only one motor will be used to move both the carriage *and* the rails.

i feel then that this "plus" arrangement is a much more evenly-distributed load, and also, it avoids the whole "rotating" thing inherent in vertical rods "H" style corexy designs. and it also avoids the "whoops i have to make the rods really far apart in order to fit the hotend in between them" issue of the horizontal-rods variant of "H" corexy designs....

so, it would be a compact design, and have some advantages. what are the disadvantages (apart from needing a minimum of 6 linear bearings as opposed to four in an "H" style corexy).

Edited 1 time(s). Last edit at 04/27/2016 01:48PM by lkcl.

---

-- sandwich200: compact portable folding corexy printer [reprap.org]

hi folks it's been over a week, i'm kinda surprised that there's been no "hmmm..." reaction to this one, at all: i'm really interested to keep innovating on the corexy theme. i'm going to run it by the "mechanical" forum, see if it grabs anyone's attention there, as i wasn't aware the mechanical subforum was there when i first posted this one.

Hi,

You are making some difficult/strange questions, i not a mechanical engineer/expert and not sure if i will tell next is correct or not...

but from your gimp image the belt have the "normal" corexy setup and you have mixed with some of the idea of the ultimaker design (added a axis one the X).
I think this second axis will add more rigidity and not allow the carriage to twist, but this twist problem is related with the h-bot design and not the corexy.

So resuming.. for me this design only make sense if you use a h-bot belt setup and add the second X axis to remove the twiste problem.
but it is a lot more easy to simply upgrade the linear movement, the second axis add to much complexity on the carriage...

If you're willing to add a second gantry, you can just use the Ultimaker's setup, where each gantry is controlled by a single motor through a driveshaft and a pair of rods:

[cdn.thingiverse.com]

(Note that here, the linear rods double as driveshafts; this is not necessary.)

This setup should be simpler and should introduce less elasticity, as well as allowing the use of non-Core-XY-compatible firmware.

Quote
epicepee
If you're willing to add a second gantry, you can just use the Ultimaker's setup, where each gantry is controlled by a single motor through a driveshaft and a pair of rods:

appreciated.... but i really want to evaluate a corexy belt arrangement with this "plus" (cross system). i like it oh - in the "mechanical" section someone pointed me at the quadrap, where people are using square-section carbon fiber carriage-support rods. i *really* like that idea especially as it appears that carbon fibre (solid rods) actually have more structural strength than steel, which i was really surprised by. delrin double-v rollers...

so basically, it might potentially be possible to compete with a delta printer for carriage weight, and still retain the corexy dynamics, *and* have a balanced x-y weight distribution (unlike the H CoreXY arrangement). hmmmm....

Quote
filipeCampos
Hi,

You are making some difficult/strange questions, i not a mechanical engineer/expert and not sure if i will tell next is correct or not...

well that in itself helps, filipe - discussion is clarification

Quote

but from your gimp image the belt have the "normal" corexy setup and you have mixed with some of the idea of the ultimaker design (added a axis one the X)

and, it turns out, from something called the quadrap - yes. except the ultimaker (as epicepee points out) uses rods and belts (or 2 motors on X and 2 motors on Y), and the quadrap uses wire (fascinating design).

Quote

.
I think this second axis will add more rigidity and not allow the carriage to twist, but this twist problem is related with the h-bot design and not the corexy.

yes, exactly. in the h-bot design you could add 3 parallel rods, to stop carriage twisting.

in this design however, i am *not* planning to put in two belts at each end of the rods (like in the ultimaker and the quadrap). i was thinking of doubling-up the Y rods (like in an H-Bot "horizontal" arrangement). the X-Ends (and now the Y-Ends) would be "free-standing" (ok, free-moving). belts would go through the X-Ends, but not through the Y-Ends.

Quote

So resuming.. for me this design only make sense if you use a h-bot belt setup and add the second X axis to remove the twiste problem.
but it is a lot more easy to simply upgrade the linear movement, the second axis add to much complexity on the carriage...

i admit it's complex... well, i think the right english word would be "comprehensive". it's a *lot* to think about, "hmm how would this remain stable and rigid?", and i like thinking those kinds of things through

To be honest i think this is only a good theoretical exercise, but not make much sense in practical. You can simply take the extra cost of adding the second axis and buy a good guide linear rail, problem solved! This solution is adding way to much complexity to solve a problem that can be easy solved with a good linear movement, normally the simplest solution is the best.

Quote
filipeCampos
To be honest i think this is only a good theoretical exercise, but not make much sense in practical. You can simply take the extra cost of adding the second axis and buy a good guide linear rail, problem solved! This solution is adding way to much complexity to solve a problem that can be easy solved with a good linear movement, normally the simplest solution is the best.

how heavy is a metal linear guide rail compared to carbon fibre rods?

let's have a look... [www.bucks-composites.com] 96 grams per 1 metre (that's a solid rod) - so 28.8 grams for a 300mm length. or, you could go for this [www.hobbyking.com] which is a square tube, that's 19 grams for a 300mm length.

let's leave the carriage itself (and bearings) out of the equation for now - what's the weight on a linear guide rail? i take it you're suggesting to have that on the x-ends, so that the entire rail is moving.... [www.robotdigg.com] that's 100 grams!

so we're talking about maybe 40 grams for two carbon rods (one on x, one on y) - the idea i had of doing 3 rods (2 on y, one on x) would be 60 grams, if you went for the quadrap approach which is only 2 rods, that's a whopping 60% weight reduction.

... you start to see why i am giving this idea some serious consideration? it's approaching delta-printer weight territory.

Edited 1 time(s). Last edit at 05/05/2016 04:02PM by lkcl.

Well.. you are probably right about the use of carbon fibre, looks like to be a great material to use on the build of a printer. But this is not a new topic?

And you call the quadrap a fascinating design. To me is more like "build a printer with what you have at home...", is a big mess..

I my opinion you are going the wrong way, your are trying to fix a problem adding more complexity on the system.
What you need to do is the opposite, simplify the system. In this particular case, i will suggest to solve this using 12mm linear guide rail and use a simple h-bot mechanism.
This heavy setup will have enough rigidity and precision to not have the twist problem and is weight will not matter.

But this is only my opinion, nothing more.. You are trying a new design and if you believe it will work than go for it..

Quote
filipeCampos
Well.. you are probably right about the use of carbon fibre, looks like to be a great material to use on the build of a printer. But this is not a new topic?

it's new to me - i'm used to repraps (cartesian), big metal rods n stuff. ah. do you happen to know of any discussions, analysis or actual printers that use carbon fibre rods for the carriage support? i'd like to do some research. found the quadrap (ok, quadarc) but nothing else yet.

Quote

And you call the quadrap a fascinating design. To me is more like "build a printer with what you have at home...", is a big mess..

that's why it's fascinating! the guy who designed it i think is based in the middle east, where, clearly, he has difficulty getting hold of some of the materials that we take for granted here in the west. and he's made it an entire cubic metre, which is just stunning. he... ah.... printed out a naked figurine and put it in a display case, it looks awesome (not just because it's a naked lady!) but because it's 800mm high and looks really good quality.

i just love the ingenuity of people doing 3d printing.

Quote

I my opinion you are going the wrong way, your are trying to fix a problem adding more complexity on the system.
What you need to do is the opposite, simplify the system. In this particular case, i will suggest to solve this using 12mm linear guide rail and use a simple h-bot mechanism.

i don't want an h-bot system - everybody's got those (including me, with the sandwich200v1 and v2) i want to experiment, to do something that people *haven't* done - that's what i find interesting. i already have an H-based corexy system, and i want to push to 500mm/sec and above printing speeds and see what happens.... just because i can.

if i use 12mm guide rails, with the extra weight chances are that i'll start to run into GT2 timing belt stretching issues at those kinds of speeds.

Quote

This heavy setup will have enough rigidity and precision to not have the twist problem and is weight will not matter.

because of the planned experimentation speeds, a heavy setup is exactly what i am looking to avoid - and also to get a balanced weight distribution between x movement and y movement (which you *don't* get on an H-bot or corexy H-based 3d printer)

Quote

But this is only my opinion, nothing more.. You are trying a new design and if you believe it will work than go for it..

i have no idea - that's what i think i am going to have to find out! not for any reason other than "because it would be cool if it worked", and for me there's no better reason. i just... i'd like sooome degree of analysis help (and checking) before wildly plunging ahead

Edited 2 time(s). Last edit at 05/05/2016 05:22PM by lkcl.

---

-- sandwich200: compact portable folding corexy printer [reprap.org]

Ref your OP. I think I see where you are coming from with this, trying to spread the load more evenly by adding central support to the carriages. However, as I understand it, one of the primary advantages of the corexy design is that it offers the potential for greater speed because all the motors are static and not fixed to any moving axis. Therefore the moving mass is reduced allowing for greater acceleration and overall speed. The only thing that you are trying to move around is the hot end with any associated cooling fan so there shouldn't be much of a mass. By adding the central bearings, you are increasing the mass that has to be moved around so potentially negating some of the speed advantage of the corexy design. Whilst it would be more rigid, I personally think that supports at either end will be rigid enough and so adding extra supports in the centre will only increase the mass and therefore reduce the maximum speed potential of the machine. That's just my opinion of course.

Quote
deckingman
Ref your OP. I think I see where you are coming from with this, trying to spread the load more evenly by adding central support to the carriages. However, as I understand it, one of the primary advantages of the corexy design is that it offers the potential for greater speed because all the motors are static and not fixed to any moving axis.

yes. that's why i like it. i especially like that the belts, if the tension is equal, produce equal and opposite cancelling-out forces that do not "torque" the X carriage bars. in fact, i believe that as long as the belts are not stressed out too much (too fast acceleration) they will actively support the X carriage bars and keep them parallel to the frame.

so, to clarify: you (and i) like the corexy design because *compared to the reprap* and other designs where the motors are not static, it offers a reduction in inertia. that having been stated and accepted, i am looking for a *greater* reduction in inertia whilst retaining the clear design advantages of corexy that those of us who have looked at it closely know exist.

Quote

Therefore the moving mass is reduced allowing for greater acceleration and overall speed. The only thing that you are trying to move around is the hot end with any associated cooling fan so there shouldn't be much of a mass.

not quite: there is the weight of the x-ends, the x-end linear bearings, the idler bearings on the x-ends, the two x-rods and... what else... the bolts holding the x-end idler bearings.... the main (huge) contribution to weight is those x-end round steel bars. a quick google search for a metal weight calculator shows that each 8mm x 300mm round bar is about 110 grams, meaning you will have nearly... what... 240 maybe even 250 grams (0.25kg!) worth of x-end carriage to move about. that's a *lot*. [update: actually more than that because the LM8LUUs are 20g each, or if you use 4 LM8UUs those work out to be heavier]

obviously, when moving in the X direction, it's the carriage and hot-end you only need to be concerned about - hotend plus bearings, taking a rough guess here... 100 grams?

... but in the Y direction you suddenly have an extra 250 grams to move about - 350 grams possibly even more, vs 100 grams. that's a 3.5:1 ratio imbalance between moving in the x and moving in the y directions.

so with the original CoreXY design that HUGE weight imbalance between x and y is what has me concerned, and it's something that i believe can be addressed, here, with what i'm naming "CoreXY+".

Quote

By adding the central bearings, you are increasing the mass that has to be moved around so potentially negating some of the speed advantage of the corexy design. Whilst it would be more rigid, I personally think that supports at either end will be rigid enough and so adding extra supports in the centre will only increase the mass and therefore reduce the maximum speed potential of the machine. That's just my opinion of course.

hey no problem, all respectful discussion welcome.

ok, so let's work through it. i'm going to be testing this out with solid carbon-fibre square rods, 1 for the x axis and 2 for y, because the x-axis one is supported to some extent by the belts, but i want to stop possible rotation in Y. the x-axis will have x-ends with idler bearings for the belt (just like original CoreXY). also i'll try it with delrin double-v rollers and 625 bearings (just like in the quadrap) except i will use plastic (IGUS) LM8UU bearings as they have been highly recommended, as being quieter, lighter, and less prone to wear.

so, in the x-axis direction you have:

• QTY 4 idler 625 bearings (20g @ 5g ea)
• QTY 1 10x10x300mm solid carbon rod (30g)
• QTY 2 x-end plastic (10g @ 5g ea?)
• QTY 2 IGUS LM8UU plastic linear bearings (4g @ 2g ea)
• bolts (1g or so)

total x weight to be moved: appx 65 grams

and in the y-axis direction you have:

• QTY 2 10x10x300 solid carbon rods (60g @ 30g ea)
• QTY 4 IGUS LM8UU plastic linear bearings (8g @ 2g ea)
• QTY 2 y-end plastic (10g @ 5g ea?)

total y weight to be moved: 78 grams.

which is roughly balanced.

total for x and y: 143 grams (as opposed to nearly 250 for a standard CoreXY design)

for the carriage, if using Delrin Double-V-roller arrangement

• QTY 9 625 bearings (45g @ 5g ea)
• QTY 1 plastic carriage and Delrin double-v rollers (10g appx?)
• hotend (50g?)
• fans (5g?)
• x and y endstops (5g?)

lots of approximations / estimates there, but apart from the 9 (!) 625 bearings there's not a lot extra that you'd normally expect. compare that to a pair of LM8LUU bearings and they're 20g each (!) so you're looking at a 40g comparable weight. you need 9 625 bearings because there are 3 rods, you clamp 2 double-v-rollers underneath and 1 on top: voila, you have a sandwich that's not going to move.

basically what i'm saying, here, is, for a rod-based linear-bearing design the total weight of the carriage is pretty much comparable so can be excluded from any comparative advantage / disadvantage over a standard CoreXY vs the proposed CoreXY+ design.

question, though: i note in the openbuilds delrin double-v-rollers they have *two* 625 bearings per roller. is that absolutely strictly necessary? would tight friction press-fit over a double-v roller using a single deep groove 625 bearing be okay? if not *sigh* that would mean 18 625 bearings.

so, i know it's a long way down now from the original question, but to reiterate, i think relying on there only being one y-rod is not sensible, although i am aware that the quadrap does that, but they (and the ultimaker) both have belt arrangements that support the moving cross-rods at both ends. the corexy belts would support the x-rod, but the y-rod would be free and could move off of perpendicular. two y rods would stop that happening, although.. yeah, it's quite a lot more 625 bearings. hmm.....

Edited 4 time(s). Last edit at 05/07/2016 09:44AM by lkcl.

Quote
lkcl
question, though: i note in the openbuilds delrin double-v-rollers they have *two* 625 bearings per roller. is that absolutely strictly necessary? would tight friction press-fit over a double-v roller using a single deep groove 625 bearing be okay? if not *sigh* that would mean 18 625 bearings.

lkcl: The openbuilds rollers have two bearings and a 1mm spacer to make the roller. The spacer ensures the bearing sides don't touch at the outside when smashed together. There is a ridge in the middle of the roller that wouldn't allow the roller to be used with just one bearing unless you machined out that ridge.

-os3dp

Quote
os3dp

Quote
lkcl
question, though: i note in the openbuilds delrin double-v-rollers they have *two* 625 bearings per roller. is that absolutely strictly necessary? would tight friction press-fit over a double-v roller using a single deep groove 625 bearing be okay? if not *sigh* that would mean 18 625 bearings.

lkcl: The openbuilds rollers have two bearings and a 1mm spacer to make the roller. The spacer ensures the bearing sides don't touch at the outside when smashed together. There is a ridge in the middle of the roller that wouldn't allow the roller to be used with just one bearing unless you machined out that ridge.

-os3dp

thanks os3dp - as a first experiment i was just going to 3d print something up (and then file the contact-surface smooth if needed, with key files). yeah i saw the 1mm spacers on one of the photos of the web site - nice idea, those go against the bearing's centre hub which doesn't rotate, keeps the bearing from scraping against its support.

it was mainly, why do you need 2 625 bearings at all, would one alone do the job? my concern would be axial movement of the bearing hub (learned that word from cozmicray recently: axial ) and i would imagine 2 bearings if pushed together would cancel out any axial movement... but if they're deep groove bearings and high quality there shouldn't be any axial movement, should there?

if that's true, i feel it would be possible to get away with just one 625 bearing, as long as the double-v-rollers were supported from both sides and there was no "wobble" or stress anticipated, such as the square bar undergoing axial rotation for example. that _would_ be bad... but it's square bar, mounted securely at both ends, it ain't gonna be rotating.

still unsure....

Quote
lkcl

so, i know it's a long way down now from the original question, but to reiterate, i think relying on there only being one y-rod is not sensible, although i am aware that the quadrap does that, but they (and the ultimaker) both have belt arrangements that support the moving cross-rods at both ends. the corexy belts would support the x-rod, but the y-rod would be free and could move off of perpendicular. two y rods would stop that happening, although.. yeah, it's quite a lot more 625 bearings. hmm.....

I agree with you that a single Y rod is not the best way to go. I'm planning a bit of a monster and it'll use a Diamond hot end which is quite a cumbersome beast. IMO the only sensible way to mount it is to have it suspended between 2 parallel rails. I'm thinking along the lines of using 2 lengths of 2020 V slot extrusion with printed carriages and Polycarbonate wheels. I'm trying to find out how the weight of 2020 aluminium compares to the weight of solid steel rods (anyone know?). I'm thinking that for what I'm planning, X and Y will be in region of 500mm so I'm guessing that I ought to look at 10mm rods as a minimum.

Of course, the other factor as well as mass is friction. If one used vslot with a single polycarbonate or steel wheel at the top taking most of the weight, and two wheels underneath keeping it all from twisting or rotating, the contact area with the rail would be tiny but then you would also have the friction of the wheel bearings. Would that translate into less overall friction compared to say a linear bearing? Certainly as a general rule, rolling friction is less than sliding friction so in theory it should be better than plastic type linear bearings. Maybe someone cleverer than I can answer that. I guess the only way to test it would be to build two systems, one with rod and linear bearings, the other with Vslot and wheels each having a carriage of the same mass. Then measure the force required to a) get it moving (static friction), and b) keep it moving (rolling friction - or in the case of a plastic linear bearing, sliding friction).

Lots to think about. The big question is, how much does any of it matter? Do I really care about getting an extra (say) 5% gain in speed? I guess it comes down to what you are trying to print. I'm toying with the idea of something that would potentially have a build volume around 400mm x 400mm x 800mm. So, 5% speed increase will translate in saving hours of print time. Therefore I do care - just answered my own question - sorry for the ramble.

Quote
deckingman
Lots to think about. The big question is, how much does any of it matter? Do I really care about getting an extra (say) 5% gain in speed? I guess it comes down to what you are trying to print. I'm toying with the idea of something that would potentially have a build volume around 400mm x 400mm x 800mm. So, 5% speed increase will translate in saving hours of print time. Therefore I do care - just answered my own question - sorry for the ramble.

well the original designer of the quadrap made a monster double the x and y volume of what you're planning - 800x800x800 build area - i'm just astounded at that. and i believe he uses 10mm aluminium square tube, which strikes me as a bit too flimsy for that kind of span, but clearly it actually worked as he printed that huge 750mm awesome naked statue. in blue

no i'm not going for a 5% increase in speed: this is to experiment at 500mm/sec *stable* speeds, and i will be experimenting at faster than that. a Flex3Drive is arriving soon, and mutley3d uses those on the mendelpro 3d printers he upgrades, and runs *reliably* at 350mm/sec with many of the ringing problems, corner bulges, all that stuff... *gone*. so with a reduction in weight, and keeping to a 200x200 printbed size, i'm only spanning 300mm.

at a 400x400 build volume... mmm... honestly i don't know. i haven't enough experience at that, but i do know, from experimenting with spaghetti under load when i was 16 (don't laugh! it was proper science!) that for double the distance, you *quadruple* the flex. in other words it's a square law. so if you want to keep stability and reduce flex, then i guess you would need to double the rod thickness (if using rods) - that would mean a 16mm rod minimum which is 4 times the weight. that's a hell of a lot. you'd be throwing around 1kg of steel, at least. so yeah, solid steel rods on a 400mm x 400mm span.... not so hot

so i think you can probably see why people go for aluminium tubing, or extrusion, or those linear rails. personally if i was doing a monster 500mm span like that, i'd look at putting in linear rails, but have 2 per axis - one vertical, one horizontal per axis. or perhaps use three linear rails - two vertical, one horizontal (to create an I-beam in effect, but each line of the "I" is a linear rail). that would completely stop any twisting. the only thing is, that's still a hell of a lot of metal to be shifting around.

so this is why i'm thinking about this design, not just for speed, but also so that if you go to a larger span, you can balance out the weight distribution across x and y... especially if you have a heavy hot-end like the diamond.

so what do you think, deckingman - two pairs of aluminium square-section (2 for x, 2 for y), 12 or even 15mm square, you'd need 12 bearings to support the carriage on double-v-wheels. don't be tempted to make it less, i know people have tried to put the upper v-wheels on the x and the lower v-wheels on the y, then use a small amount of bend in the bars to keep the carriage stuck to the bars.... please don't do that

1m of aluminium box-section @ 15mm is like... $10 from a hardware store. 1mm wall thickness, a 500mm length is 75 grams (or so). 4 of those is 300 grams. 12 bearings @ 5g.... 60g. the hotend... god knows what that is... 200g? reprap wiki says 250 WOW you're a loony, deckingman ok so 250g there.... so you're up to 610 already... yeoowch.

... i take it you'll be using say.... 1mm nozzles, and running this at no more than... 100mm/sec?

about the bearings (static, dynamic friction) - honestly if bearing friction is a problem, you're doing something wrong: pick larger diameter bearings: 608s not 625s. of course that's extra weight....

Quote
deckingman
I'm trying to find out how the weight of 2020 aluminium compares to the weight of solid steel rods (anyone know?). I'm thinking that for what I'm planning, X and Y will be in region of 500mm so I'm guessing that I ought to look at 10mm rods as a minimum.

500mm 2020 V-slot weighs in at 222g

-os3dp

Quote
os3dp

Quote
deckingman
I'm trying to find out how the weight of 2020 aluminium compares to the weight of solid steel rods (anyone know?). I'm thinking that for what I'm planning, X and Y will be in region of 500mm so I'm guessing that I ought to look at 10mm rods as a minimum.

500mm 2020 V-slot weighs in at 222g

-os3dp

cool, ok, so 2 of those in a plus, with 6 delrin rollers, assume openbuilds double-v standard rollers 2 bearings each @ 5g = 12, that's 222*2 + 12*5 = 282g which is less than what i was suggesting above (four 15mm box-sections)... some form of linear rod or more delrin rollers at each end... yeah that'd work. it'd be a monster but it'd work.

or you could go with them in a straight H arrangement but then i feel that that's an awful lot of imbalanced weight.

Quote
lkcl


so what do you think, deckingman - two pairs of aluminium square-section (2 for x, 2 for y), 12 or even 15mm square, you'd need 12 bearings to support the carriage on double-v-wheels. don't be tempted to make it less, i know people have tried to put the upper v-wheels on the x and the lower v-wheels on the y, then use a small amount of bend in the bars to keep the carriage stuck to the bars.... please don't do that

The plan at the moment (and it's only a plan) is this.

3 or 4 off 2020 vertical Vslot as linear guides for the Z. Each guide will have a carriage (gantry as open builds call them). That'll take care of any sideways movement of the build platform. To lift it, I'm planning on 3off Z rods but as they don't need to provide any stability as that is taken care of by the linear guides, they can be allowed "float" and I reckon that 5mm threaded rod will be fine (but I could be wrong). I'm thinking that lead screws would be an unnecessary expense with the linear rails, and 5mm rods will give much better resolution due to the smaller pitch ( as well as being a hell of a lot cheaper. The rods will be two at each of the front corners at one in the centre if the back giving 3 point lift. Initially the rods will be connected buy a belt and driven from a single stepper but the longer term plan is to have each one driven by it's own stepper when DC42 gets around to incorporating bed levelling by controlling each individual Z rod. Of course. this will, mean having 3 linear guides and not 4. Still undecided on that because having 4 rails means 2 carriages at each side so a length of 2020 between them, makes it easy to make something for the bed plate to sit on but won't allow the z rods to be used for automatic bed levelling.

The top of the frame will be 20x40 or 20x60 V slot which will also serves at the linear rails for either the X or Y. These will each have carriages to take 2 off 2020 Vslot as linear rails at right angles for the other axis. They'd be spaced with a gap of at least 50mm between them to allow the Diamond hot end to drop in between. Then the hot end, fan and z probe will be mounted on another carriage which will slide on the two 2020 Vslot rails.

So, yes lots of wheels and aluminium extrusion but no cylindrical linear bearings smooth rods or lead screws.

This may never get off the ground but on the other hand it just might.

Quote
deckingman
The rods will be two at each of the front corners at one in the centre if the back giving 3 point lift. Initially the rods will be connected buy a belt and driven from a single stepper but the longer term plan is to have each one driven by it's own stepper when DC42 gets around to incorporating bed levelling by controlling each individual Z rod.

niiice. ok you don't need dc42 to do the bed levelling, you can just copy the deltaprinter homeall.g file and it's pretty much done. *but*.... what you *will* need - in the firmware - is to have Z0 to mean "do all 3 z motors" yet when you want to control individual ones you would refer to them as "Z1, Z2 and Z3" respectively. that _will_ need to be coded up in c++ only thing i can say is, though, aw poo! the duet 0.8.5 can't handle that on its own - i'd need to get the expansion board. oh well - at least it would be possible with the expansion board.

Quote
lkcl

Quote
deckingman
The rods will be two at each of the front corners at one in the centre if the back giving 3 point lift. Initially the rods will be connected buy a belt and driven from a single stepper but the longer term plan is to have each one driven by it's own stepper when DC42 gets around to incorporating bed levelling by controlling each individual Z rod.

niiice. ok you don't need dc42 to do the bed levelling, you can just copy the deltaprinter homeall.g file and it's pretty much done. *but*.... what you *will* need - in the firmware - is to have Z0 to mean "do all 3 z motors" yet when you want to control individual ones you would refer to them as "Z1, Z2 and Z3" respectively. that _will_ need to be coded up in c++ only thing i can say is, though, aw poo! the duet 0.8.5 can't handle that on its own - i'd need to get the expansion board. oh well - at least it would be possible with the expansion board.

Yup. With 3 extruders as well, I'll be using 8 steppers. Duet 0.8.5 does 5 and the expansion board gives another 4 so I'll have one spare channel.

Quote
deckingman

Yup. With 3 extruders as well, I'll be using 8 steppers. Duet 0.8.5 does 5 and the expansion board gives another 4 so I'll have one spare channel.

ye gods! i think that's not good enough. i think you should use the spare one as an automatic loading mechanism on a built-in trebuchet or other siege engine