Foldable, portable CoreXY (Fusebox derivative)

flex3drive arrived last night (thank you to mutley3d), fitted it together, installed the E3Dv6 hotend with the volcano upgrade and 0.6mm nozzle. space is so tight i'm having to use 25mm fans. the bottom one i will need to redo the fan-holder and turn the bottom fan through 45 degrees to point downwards at the hotend. i'll make some sort of cone later to direct airflow towards the nozzle.



this one shows the flex3drive in place - the extruder motor is *sideways*.. yes really .... so that when the box closes the flex3drive holder mount doesn't stick up through the box. yay. the actual flexible driveshaft is 750mm long, which is something like 200mm too long, but we'll deal with that later. the other end of the flex3drive has the top half of the extruder holder. that whole assembly is, again, way too high to fit in the box, so it's removable. turns out that friction's pretty strong and it's braced in 3 places, so i don't think i'm going to need screws (even though i arranged 5 screwholes).



ok onwards.... nearly there. microswitches are in place. motors in place. wiring to do - and i can do some tests...

Edited 4 time(s). Last edit at 05/13/2016 06:18AM by lkcl.

first print was successful! nothing blew up, or broke. z-screw wasn't calibrated correctly (25% out) so the object printed which should have been 15mm came out a bit blobby and was 12.5mm high. corrected that and did a 2nd print, and it's good.

i can see clearly that there are inaccuracies from belt-dragging (photos / update tomorrow). compared to the v1, the belt seems a lot stiffer. i really feel i'm going to have to use that "unnamed" plotter-style cartesian belt arrangement instead of corexy. v3 already!

but, the flex3drive works really well. the 3 lead screws provide a rock-solid bed. i took the opportunity to drill an extra hole in the acrylic front bed-rest-plate, to convert to a 3-point bed-levelling system (the alu MK3 hotplate has a 3-point hole). but due to the way i did the bed springs, the bed was was wobbling sideways. long story will document it tomorrow, but basically i put two upright (fixed) guide screws where the front two bed-levelling springs _used_ to be. ideally they should be 3mm diam smooth rods (because the threads catch on the acrylic) but we'll sort that another time.

whew, yay!

[update] image:



[update] video uploading [youtu.be]

ok so the one on the left was the 2nd print, the 2 on the right the 1st. 1st was incorrectly calibrated z height and printbed height. 2nd print came out ok, looks like the corexy belts are a bit loose though

Edited 3 time(s). Last edit at 05/14/2016 07:17AM by lkcl.

---

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

i'm adding a series of new videos here [reprap.org] which are some calibration tests. some of them successful, some of them not - the hilarious ones are me nearly getting smacked in the head by the flex3drive as the carriage moves at 500mm/sec from one side to the other...

here's where i printed at 450mm/sec, a 150mm-long part, 0.3mm layer height, 0.6mm nozzle on the volcano e3d - cura is reporting a whopping 90 (NINETY!) mm3/sec flow rate and there's no zagging! i'm just amazed and delighted.



the blobbing on the bottom edge is because i hadn't got the printbed height calibrated perfectly. there is, just in the middle, an area around 40-60mm, where the filament is no longer "shiny" that tells me that the limit (at a whopping 90mm/sec) is just nearby. however there's no zagging (breaking up of the filament), which is what was happening before.

the corner-bulges on this (fast, long, narrow) part, however are enormous: something like 0.5 mm (!) which didn't happen on the much slower-printed parts in the first and second runs. honestly i feel that i should be sorting that out by not using cura: both repsnapper and skeinforge have options to "round" the corners, i'm surprised this isn't something that's by default in all gcode-generating software.

ok now to find some sort of voronoi object to try

What is your acceleration set to?

Quote
Koko76
What is your acceleration set to?

only 6,000 mm/s^2 at the moment. in the video of the 150mm/sec long part it's a noticeable slowdown at the ends. hmmm i should up the acceleration a bit and try again. [update] 9,000 mm/s^2 i'm getting occasional skipping, but this is with the (5 year old) 1.68A zappautomation NEMA17s so i've set the current limit to 1.4A. i've got some 2.5A motors here but i'll need to set up some fans for the duet, first, to be able to approach the duet 0.8.5 driver's 2A current limit.

[2nd update] 8,000mm/s^2 and increasing the current limit to 1.5A works fine - no skipping, and quality's ok as well. 9,000 was a bit much for these rather battered steppers they've been in an original reprap, and absolutely hammered. the original reprap, if you recall, uses bits of plastic around non-precision-machined plain steel rods (!) so the motors were pretty much maxed out, poor things. anyway they're doing fine at 1.5A.

Edited 3 time(s). Last edit at 05/14/2016 12:26PM by lkcl.

a voroni test - bear in mind this is really small! - that's only 65mm high, wall thickness is only 2mm, the layer height's 0.3mm and it's a 0.6mm nozzle - no underside support, i think this is not bad an achievement. i will try the 0.4mm nozzle some time, then i can get the layer height down to 0.15.



and whoops this is the 9000mm/s^2 test with only a 1.4A current limit, skipping occurred several times. printing this part is very very challenging - it's 230mm long, 2.5mm thick and 10mm high. the photo below is of the worst part, but on the main straight it's actually amazingly clear. zagging only occurs where the print clearly went badly wrong because of the dog's dinner at the end from the skipping.



Edited 1 time(s). Last edit at 05/14/2016 12:26PM by lkcl.

Mm/sec^2 is the unit you are looking for. To each their own, but I consider none of that to be acceptable print quality.

Quote
Koko76
Mm/sec^2 is the unit you are looking for.

whoops, thanks, will correct that.

Quote

To each their own, but I consider none of that to be acceptable print quality.

i'm not done yet! i'm pushing things quite hard, deliberately, to find the limits. when i do production-quality stuff i put the 0.4mm nozzle back in, set the layer height to 0.15, and the speed down to 200mm/sec or less. this is an entirely new from-scratch concept printer that's only been running for 16 hours - gimme a little credit, man

Quote
Koko76
To each their own, but I consider none of that to be acceptable print quality.

koko76, can i ask you a favour? i don't have anything to compare against (this design's pretty unique) - do you happen to have a 0.6mm nozzle (or, as a 2nd preference a 0.4mm), and could you print this out at 0.65 scale, layer height 0.3mm, and post a picture so i can compare? "good" or "bad", i'm not judging: i really do need to know if i'm either just going too far beyond what's possible, or if there's something i should be investigating. [www.thingiverse.com]

so just to reiterate and make absolutely clear: i'm asking out of a need to do comparative engineering quality control analysis, *not* to pass judgement in any way.

I have no desire to do work for you in order for you to perform a "comparative engineering quality control analysis". I don't print at those sizes or layer heights. If parts that my printer made looked like that, I would not be claiming that those speeds were valid. Makes zero difference to me what your design is, I just look at the output. Your opinion is different, it's your machine and you make the output to whatever standards you want, do whatever you like.

Quote
Koko76
I have no desire to do work for you in order for you to perform a "comparative engineering quality control analysis". I don't print at those sizes or layer heights. If parts that my printer made looked like that, I would not be claiming that those speeds were valid. Makes zero difference to me what your design is, I just look at the output. Your opinion is different, it's your machine and you make the output to whatever standards you want, do whatever you like.

koko76, you seem to be quite upset about something, but i appreciate and respect what you've said and how you've said it: i often use this same phrasing myself on this forum. as you can see i've documented this process in some detail, so that others can learn from it (mistakes and all). if by chance you viewed my request as anything other than an *invitation* to collaborate and improve on the general field of knowledge and community expertise, for *other* people's benefit not just my own but yours included, that was in no way the intent. so, to be clear, i respect your decision to decline the invitation.

anyway - jason (mutley3d) has taken a look at that voronoi test and we're working out how to improve it. personally i believe it's just not practical to use a 0.6mm nozzle and 0.3mm layer heights for such small detail, especially with considerable overhangs - the challenge now is to find out what _can_ be done.

Why not just do a search for Voronoi LED Egg tealight shade images
and use those for comparison.

Is the Voronoi LED Egg tealight shade a good test?

CntlVTest -- is a good cal / comparison test

[www.thingiverse.com]

I don't know what is a good test for 0.6mm nozzle and 5000 mm/sec speed?

Quote
cozmicray
Why not just do a search for Voronoi LED Egg tealight shade images
and use those for comparison.

sounds like a good idea so i tried it... unfortunately this particular object i arbitrarily picked without thinking it through in advance is sufficiently new that the only available pictures are 500x500 (ish), with the object in the centre so the resolution is poor, and, because it's a tealight, the lighting is also poor. i'll try expanding the size a bit, decrease the layer height, see how it goes.

Quote

Is the Voronoi LED Egg tealight shade a good test?

don't know. let me think. there's too many unknowns here. my reverse-engineering experience says, "find good baseline and eliminate all unknowns down to QTY 1 (one) then do binary-search on that one unknown".

Quote

CntlVTest -- is a good cal / comparison test

[www.thingiverse.com]

_great_. ok let's take a look.. hmmm they appear to be (different versions, 1 2 3) for 0.4 and 0.48mm nozzles... i'll need to put in the 0.4mm nozzle. i've got a couple of parts to do first, then i'll swap over.

Quote

I don't know what is a good test for 0.6mm nozzle and 5000 mm/sec speed?

danged if i know! hmmm.... out with the 0.6mm nozzle as an unknown then. and out with the... weeelll... it's 450mm/s not 5000...

Quote
Koko76
Mm/sec^2 is the unit you are looking for. To each their own, but I consider none of that to be acceptable print quality.

koko76 (and anyone else who may be following these tests now and in the future), i just re-ran the voronoi test again, scaling up to 1.2x and reducing the layer height to 0.15 in order to compensate for the larger nozzle.... and i just spotted cura 15.04 doing something rather dumb / interesting.

... basically, cura 15.04 tries to make the nozzle extrude in mid-air. rather than go "oh look we have a bit of plastic where we can start from to make sure the plastic doesn't just splut out and drop under gravity".

sooo that would explain quite a lot! on rather a lot of the other prints i've been doing, as well... hmmm.

so, i think i have to see if i can get repsnapper up and running again. it has some different kinds of flaws but i used it for 18 months and got used to its quirks: generally (not always) it does a good job.

okaaaay... right, this is scaled up to 1.2x the object original size, layer height reduced to 0.15, acceleration is 8000mm/s^2 (correct units, yay!) - and it's using cura 15.04 which we now know can't design underside gcode properly - it tries to start an extrude in mid-air. i stopped this print after around 18-20mm as it's sufficient for a test.

first the top side, which looks pretty good given the conditions:



and the underside, which doesn't:



first thing to note, is, there was a "skip" somewhere around 8mm up. so, that indicates "reduce accel, increase current". done - keep an eye on this. must put fans pointing at the controller stepper ICs. second thing: the "drooping" is *not* down to the sandwich200v2, it's because cura's trying to use air as "support" material.

on the inside photo, the buildup of the voronoi holes is really quite nice. the main circle however, you can see it's not really very smooth. i think this may be down to belt judder, i can hear some sort of resonant frequency amplification somewhere, like when you get an old fan bearing going horribly wrong and screaming. some of the idler bearings can clearly be seen to catch the belt (causing it to bow outwards more than it should) - i need to disassemble the belt to fix that, and i'll probably swap over to the newer 2.5A NEMA17s at the same time.

also, i've spotted that there's quite a lot of rotation of the carriage, about the two vertical x-rods. it's about 0.5mm. i'm not very happy about that. it means the bearings LM8LUUs) are not particularly good and/or the 8mm rods are not according to spec. humph. well, you live and learn. no more vertical x-rods for me. i'm working on a redesign to use the "unnamed" plotter design, keeping the motors where they're currently positioned.

harumph.

ok so i've begun mounting the PSU and the Duet 0.8.5, both rather oddly-mounted as they go on the (single) piece of extrusion at the bottom and then through the bolt-hinge at the top:



the PSU is in brown, the top PSU holder is in yellow and i've since added bracing across it. the bolt-holder (green) on the corexy back for the z-rod top mounts (yellow) i had to redesign in order for the PSU holder to fit that 50mm bolt - that wasn't the original plan, and i had to destroy the three pieces of plastic holding the z-rod top mount. they'd been designed to be sturdy and to wrap around the extrusion by slotting into the grooves.... unfortunately on *two sides* meaning that the only way to take them off was to slide them off... and the motor mounts are all in the way! it would have been a disassembly nightmare job.

sooo *sigh* instead that lesson's been learnt now (don't do it) - i created a part that drops down onto the two bolts that are just above the z-rod holder, but left a groove to slot over the other M3 bolt in the extrusion which is at right-angles to the other two (underneath, not visible in the picture). so in this way the part can be secured with 3 bolts on 2 separate sides of the extrusion, but without having to slide both the part *and* the bolts off the rails.

fitting the PSU was therefore quite a big, risky disassembly job, involving *cutting* away the previous parts with clippers. nothing got damaged in the process, except whilst the printer was disassembled i decided to wrap sound-deadening and protective material on the outside, as well as cut away the protruding screws. cutting the screws (why 12mm? because that's what's available from the local hardware store) down to 10mm caused them to heat up and melt the plastic corners (argh....) so those are now loose (argh).

*fortunately* on reassembly the top CoreXY parts that slot into those corners strengthened and stiffened them up, so that they can't rotate or move. also fortunately, they're still attached to enough bolts such that the frame is stiff... as long as it's assembled.

what a dog's dinner. trouble is, i'm running out of time to get this fixed. we've set a (first) moving date - plane tickets for the UK have been booked, so we are definitely leaving on the 31st may - 2 weeks from now. the printer *has* to be ready and fully operational - including its stowing capability - by then. and i also have to get... creative about the kinds of materials that i use: i.e. what's available in my immediate environment. more on this later

Edited 1 time(s). Last edit at 05/17/2016 03:52PM by lkcl.

Quote
pen25
looks good. congrats on getting this to work. how well does it maintain print calibration when you fold it move it and set it back up to print?

hiya pen25, coming back to this for the 2nd revision - not at all i did however get a "fluke" where i only had to make a tiny adjustment to only one of the 3-point levelling points. however in the v2, the front z-axis lead screws are separate assemblies that have to be removed and stowed, and the belt is loose and must be reattached. so, apologies: it's something that i can live with, but some people may not. ehhh can't have everything.

[youtu.be]

ok so after getting the PSU and Duet into place (in a bit of a rush) i need to move on to more critical matters: replacing the CoreXY with the unnamed "plotter" arrangement. above is a video link to a proof-of-concept 1-in-2 model that shows that yes, it's possible to fit a drop-in replacement based on a modified version of the "unnamed" plotter arrangement [forums.reprap.org] first introduced in 3D printing with the "Sli3DR".

here's a photo of the proof-of-concept (1-in-2 scale) which took only a couple of hours to mock up, and, importantly, allowed things to be "moved" so that they could be verified as having the right amount of travel and clearance:



i marked up the belts (750mm cotton) in green and red for clarity. i just realised that the "bearings" are 50% too small but i believe that is not a problem, i'll re-print the "bearings" and try again later, but i think it still works. i could have done this as 3D, but i wanted my 7-year-old daughter to be able to help out, and also see what i'm doing, explain it to her, as well as dramatically cut the time to be confident that this concept will work.

Quote
lkcl
here's a photo of the proof-of-concept (1-in-2 scale) which took only a couple of hours to mock up, and, importantly, allowed things to be "moved" so that they could be verified as having the right amount of travel and clearance:

A concern I have with that photo is that the belts are only going round the motor pulleys for about 90 degrees. Is that really enough, especially if you will be using 16-tooth pulleys? I would have aimed for 180 degrees by moving the idler pulleys.

---

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].

Quote
dc42

Quote
lkcl
here's a photo of the proof-of-concept (1-in-2 scale) which took only a couple of hours to mock up, and, importantly, allowed things to be "moved" so that they could be verified as having the right amount of travel and clearance:

A concern I have with that photo is that the belts are only going round the motor pulleys for about 90 degrees. Is that really enough, especially if you will be using 16-tooth pulleys? I would have aimed for 180 degrees by moving the idler pulleys.

Or add a few more idlers - there's only 19 plus 2 motor pulleys.

Quote
deckingman

Or add a few more idlers - there's only 19 plus 2 motor pulleys.

nono, there's actually 6 more than that - another 3 are in v-rollers per axis amazingly it's actually not as many as are in the current sandwich200 v1, because i had 6 in the printbed cantilevered arms, and the corexy motors are inboard so required 4 more idlers than most people would have in a CoreXY. a trade-off between compactness and number-of-idlers that i was happy to make a decision on. oh, and four LM8UUs.

hey it's nowhere as many as the_digital_dentist uses - he doubles-up the 625 bearings and puts 8mm belts on!

i'll try and reorganise it tomorrow, but with only 1.5 metres of belt on both X and Y i'm pushing the limits to within millimetres of those pre-set limits. mmmm.... fuuun...

Quote
dc42

Quote
lkcl
here's a photo of the proof-of-concept (1-in-2 scale) which took only a couple of hours to mock up, and, importantly, allowed things to be "moved" so that they could be verified as having the right amount of travel and clearance:

A concern I have with that photo is that the belts are only going round the motor pulleys for about 90 degrees. Is that really enough, especially if you will be using 16-tooth pulleys? I would have aimed for 180 degrees by moving the idler pulleys.

yyyeah... it might be doable. i'll reprint some 8mm representations of the bearings and give it a shot. i have to miss the screw-hole points for the motors, but also, the extruder is level with the X and Y motors (sitting sideways!) let me do the equivalent photo to the above, of the current sandwich200v2....



that's a flex3drive coming off the extruder motor, arcing up to dominate the photo then going back down to the extruder. so the green belt has to miss the left of the extruder motor and the red the right (!) i can't move the motors further left, they wouldn't fit in the box when stowed....

.... oh god i just thought of something horrible - rotate the extruder motor (cock it up on one side) by 20 to 30 degrees around its axis of rotation, so that the green belt clears the *underside*, and the red belt, which would be slightly higher up, would go *over* it.... blerrrrgh. well i will do that only as a last resort.

i can't move the extruder motor further down (to where the z-motor is, for example) because there's not enough space. thank you to everyone for letting me think out loud here, btw.

Edited 1 time(s). Last edit at 05/20/2016 03:31AM by lkcl.

Quote
dc42

Quote
lkcl
here's a photo of the proof-of-concept (1-in-2 scale) which took only a couple of hours to mock up, and, importantly, allowed things to be "moved" so that they could be verified as having the right amount of travel and clearance:

A concern I have with that photo is that the belts are only going round the motor pulleys for about 90 degrees. Is that really enough, especially if you will be using 16-tooth pulleys? I would have aimed for 180 degrees by moving the idler pulleys.

ok so put in 8mm circles to represent 1-in-2 scale 16mm 625 bearings, redid the "belt" - it's rriiiight on the limit of what's possible within the pre-existing constraints. 180 degrees isn't possible, the pair of idlers being closer together means the path is longer by 20-30mm. 150 degrees or so should be okay. plus that means the idler's bolt wouldn't be right on top of one of the motor bolt-holes...



the mess for the other belt... gaah. it would actually be better to turn the belt round so that the teeth are facing outwards... wait... frick! which bleedin way are the teeth??? ok ok, on this purple one, they're already outwards, which is why the carriage-idlers are teethed... yeah ok whew that's the right way.

so, for the other belt with the dog's dinner mess of bearings (you can see what's planned in the photo, in that top left corner, belt isn't in yet, and one bearing is missing) it would be better to turn it round so the teeth are *inwards*... that would mean putting 2 smooth idlers on the carriage and toothed-idlers everywhere else.... argh i only bought 4 toothed idlers, i haven't time or funds to get more (only 8 days remaining in the country), i'd have to go with teeth-to-idlers when i'm planning to run this at 500mm/sec and above... argh... i could drop them in later i suppose if i plan for it now *sigh*.

normally on that plotter, the other motor would be on the top edge (as seen in the photos), it's that i have to get them into such a tight pre-defined space that's the issue. they can't go "inboard" because there's a 2nd piece of extrusion along the edges - the frame - which the top assembly sits on.

yaargh stress! i'll work something out.

I have lost track of pre-existing constraints

Maybe like this --- motors at corners
motors up, shafts pointing down

Use of a spring tension idler near motor,
gives you good contact arc on drive pulley




I am crazy for continuing this insanity!!!!!

Quote
cozmicray
I have lost track of pre-existing constraints

Maybe like this --- motors at corners
motors up, shafts pointing down

Use of a spring tension idler near motor,
gives you good contact arc on drive pulley

[attachment 78428 Plotter_Belt_2_j.jpg]


I am crazy for continuing this insanity!!!!!

i am definitely running my health into the ground with the stress and long hours, that's for sure.

i really appreciate the effort and the clear drawing, cozmicray, and apart from one technical error it would actually work very well in any other design. couple of specific reasons why i can't use the approach you describe as a retrofit in this specific design:

the size of the motors relative to the available space (which is 312 x 358mm and cannot be changed without a total redesign) is much larger than you're normally used to, so the scale of the motors you've put in there is off by something like.... a factor of 1.5 maybe even 1.8. so, the frame comes literally in a hard line down about 2mm from that bottom left yellow bearing, and chops off the left-hand side of the motor, likewise on the right side. i have *six* millimetre clearance either side of the 300mm rods to play with, on either side.

so this is why i've put in the same 2 extra idlers per belt that the other guy who did that amazing all-aluminium printer did, the one where he even made his own extruder, and was getting an insane 9,000mm/sec travel speeds.

also, because of the space constraints, the left-side spring-loaded idler would, unfortunately, need to occupy the same space as various things such as the static linear rods/rails (in particular, the horizontal one at the top between the motors), and the belt would be likely to hit various things as well.

the motors cannot point upwards because, all told, they would be approximately 60-80mm above the level of the frame: thus, when the printer is folded they would stick out about 30-50mm through the side of the box.

the shape of 312x358 is a longer rectangle than you've drawn, so the top two moving (direction-change) idlers would be on a rail that actually goes *through* (into) the motor. the rail needs to be about.... 30mm lower than it's been drawn, because the distance from top to bottom, as drawn, is something like 330mm, which is outside of the constraint of being exactly 300mm so as to be able to use a 300mm 8mm rod.

also, the distance between the two motors would be such that those two direction-change idler bearings and their associated carriage would crash into the motors, the spring-loaded idler, the fixed belt-end holder.

also, with the belt going along the top, there, it's in the way of the place where the extruder motor sits.

so although in principle it's a great idea, ray, which would work extremely well for any other circumstances, it would work well for a 3d printer that has around a 400 x 400mm top for example... but not here. if you're going to do 400x400mm, then you might as well do what the original plotter design does, which is put the motors 1/2 way down each side.

those are the main things. there is one design flaw however in the diagram, which is independent of the extremely tight space constraints: the top right spring-loaded idler should be on the *top* static-length red belt, not the variable-length right-side belt. i've made the correction, here:



just as on a corexy design it's absolutely critically important that those variable-length belt portions be lined up and remain in an absolute dead-straight line. the top-right spring-loaded idler is now sitting directly over the static belt-end holder, which is fine, it's at a different height. however..... it would be necessary to arrange clearance for the entire rail assembly that's going to come along, so that top-right motor might need to move a bit further up.

and that, almost certainly, would take the belt length beyond the 1.5 metres that i've already cut the 3.0m belt into. which, again, is yet another constraint that i cannot change within the (now) 8 days remaining until i have to get out of this country.

so, those constraints are - as hard-limits:

• 7 days to finish a 100% working design with the available materials and parts to hand (28 625 bearings, LM8LUUs, LM8UUs, 4 geared idlers, GT2-16s, 300mm x 8mm chrome rods, 300 x 8 x 8 *square* rods)
• 1.5 metre belt length per axis
• nothing above the extrusion must be higher than 40mm above the top of the 15x15 extrusion
• the extruder motor must be directly inline between (same y coordinate as) the two x and y motors i.e. nothing must occupy the space between the x and y motors. you can however go above or below that space - just that space exactly 44mm wide to fit an extruder motor has to be clear.
• the width must remain exactly at 312mm
• the depth must remain exactly at 358mm

ridiculously-tall-order but i think i can do it... with some compromises such as putting belt-teeth against smooth idlers. yuck.

Edited 3 time(s). Last edit at 05/21/2016 04:45AM by lkcl.

Quote
cozmicray
Use of a spring tension idler near motor,

occurs to me, you could as easily use a strong spring on the belt-end holders, but i feel that the same d-shape arrangement that nophead uses on the mendel90 x-carriage belt adjustment would do as well, it's a 2-to-1 gear ratio on the M3 screw, as it turns out that the D-shape is a pulley-system.



Edited 1 time(s). Last edit at 05/21/2016 09:59AM by lkcl.

My sketch was not to scale at all, and the size of all components
is arbitrary.

Your requirements / constraints way beyond insanity or practicality.
You left out assembly only requires kitchen utensils and a Ninja blender

Are you building 10 -15 coreXY mechanisms to swap out on top?

7 days to finish? I think it has been 7 months

Beyond insanity

I believe these requirements will be used for just ONE printer

Quote
cozmicray
My sketch was not to scale at all, and the size of all components
is arbitrary.

understood. i rescaled it in my head, placed actual parts where they'd need to be.... sadly it didn't work. but for other people (with that one small correction) the idea would work very well.

Quote

Your requirements / constraints way beyond insanity or practicality.

well, it is what it is. no use getting stressed about it, although that's exactly what's happening

Quote

Are you building 10 -15 coreXY mechanisms to swap out on top?

can't. i don't have enough spare extrusion to make a replacement. and only the one printer available. so i have to do it as a single "design, verify-by-hand, print, disassemble, rebuild, pray" cycle. i still have the v1 frame (minus the printbed) - i suppose as a last resort i could part-disassemble the v2, reassemble the v1, and use that - so the extrusion that's in it i do *not* want to cut it up.

so, that's why i have to use what i've got, stick with what i've got, and absolute minimise and restrict the changes to the top assembly.

Quote

7 days to finish? I think it has been 7 months

under three, and that's two revisions. it's why i'm not going to revise the frame (again) as i simply don't have time.

Quote

Beyond insanity

gaaah tell me about it. actually, insanity is defined as "trying the same thing over and over, expecting each time a different outcome". as it's never been attempted even *once* it has to be described instead as "totally bonkers and completely out on the edge of being risky beyond belief".

yeah. i'd settle for that as a good description i am reaaallly looking forward to a nice calm time when i don't have to focus on this quite so hard...

Edited 1 time(s). Last edit at 05/21/2016 03:20PM by lkcl.

Quote
cozmicray

I believe these requirements will be used for just ONE printer

yeahh i definitely need to do a v3 to clean things up. ehh the whole idea is, i go through this stuff, experimenting and trying new things, working out how to do it, so that other people don't have to. they can just buy a kit. or copy the end-resultant successful ideas.

ok so it's now been three *days* doing nothing but try out successive variations on the plotter layout. the height has had to increase because each of the key parts (linear X and Y Rods, X and Y belt bearings) all had to be on *different* levels so as not to overlap or run through or over each other. it's incredibly tight, there is literally a millimetre clearance in many places and i am still not sure that the belts (pre-cut to 1.5m) will be okay: it's just too close to call. i've done a video walkthrough of the CAD to explain it. [youtu.be]

Edited 1 time(s). Last edit at 05/24/2016 04:24PM by lkcl.