Rugged CoreXY

Hi All,

I'm new here and have never built a printer before, but I do earn my living through printing and have been at it almost 5 years now.

I've always been a strong believer in "tinned" software and printing with rafts in ABS, but have been spreading my wings a bit in the last 12 months and want to try something new.

I have a large background in technical fabrication, but none in code/programming (I'm 47 yo, most of us don't).

I'm just starting to formulate a plan for a water jet cut aluminium printer, with a 300 x 300 build volume (probably 300 high as well, but maybe 600) and a corexy gantry utilizing precision rails instead of rods.

I own 8 printers (3x UP Plus, 4x Zortrax M200, 1x Wanhao D5S), and just feel there's a need for a rugged "tool" or "prosumer" machine for people in my position, with open source software but high end reliable components. But it's a fine line with quality/reliability suffering in the name of price.

Here's the components and ideas I've come up with so far:

1. 300 x 300 build envelope.
2. Correct use of precision rails, ie, 2 parallel rails per axis with 2 carriages per rail. According to my contact in Hiwin this is something that is commonly flouted in 3D printers, and much better rigidity can be achieved using their install specs.
3. 4 rail, twin ball screw Z axis (left and right side) which is seen more often these days. I've not seen one in the flesh, but I'm assuming there's a belt driving the 2 ball screws off the single stepper.
4. AC heated bed, specifically one of the Keenovo adhesive silicone mats in 240v (I'm in Australia) temp controlled by the printer in the usual way but powered through a solid state relay via the AC supply. I've seen reports of success with this.
5. Possibly a Smoothieboard.
6. Bondtech direct extruder.
7. E3DV6 hot end. Single head only by the way.
8. Switched mode power supply, 80-240v AC input or whatever it is, possibly something from Mean Well...any suggestions?
9. LCD display with simple jog switch/dial. No touch screen.
10. SD card drive. Don't care about wifi unless it's easy, I'm happy with SD cards but would just want to be able to use any card not just those 2 GB out dated ones.
11. Simple screw/spring manual platform leveling system, but one that holds it's position solidly. I might go with a dial gauge mount.
12. Glass bed mounted on top of heated platform with clips, or even maybe magnets/locating pins.
13. S3D slicer.
14. Enclosed with filtered/temp controlled fan exhaust. I don't want to chase high temps in the cabinet, but I like to keep them constant and warm. I'm assuming my AC heat mat and normal operation heats will give me the 30 degC for PLA and 45 deg C for ABS which I like to print at. I'm not trying to be Stratysis.

It does not need to be fast for me, I'm over that, it needs to be precise. Repeatable and consistent print quality is what I'm looking for. Just thought I'd mention that before some one pulls me on my rails/carriage point.

Some questions would be:

1. What's good for end stops? I've never had a lever switch fail ever, but see people using magnet and optical sensors. It's all about reliable and utilitarian, so I'm all right with switches installed properly unless people say there's a better way.
2. Nema 17 steppers for this size printer?
3. Type K thermocouples instead of thermistors (you can tell I've been into UP's and Zortrax's)? I see a little board for them somewhere that plugs straight onto the Smoothieboard and I can get decent Hanna or Omega probes for not that much more than the ebay cheapies. I'm even considering machining a longer one into the heated build plate for a more accurate reading and less cycling.
4. Any other suggestions?

This will be funded out of my business and I'm not really looking to go cheap. I spend 3-3.5k AUD on an M200 by the time I've finished modding it, and have no issue doing that here.

My programming skills or lack thereof will be an issue, but I'll most likely build something and then ask for help as well as try to learn during the build.

Thank you for any replies in advance, and in the meantime I'll be scouring your wonderful forum!

Edited 2 time(s). Last edit at 05/29/2016 04:25AM by hobbymods.

hiya hobbymods, welcome to the forum. some links to analysis that may help,

• look at what the_digital_dentist did, with son-of-megamax, his printer design is extremely comprehensive and accuracy is extremely good.
• [forums.reprap.org] for an analysis as to why you should use dual rails/rods and triple lead screws. dual screws is not enough. multiplying up the number of rails does not "fix" the problem of bed rotation about [only 2] screws, it just minimises it. triple lead screws is GUARANTEED to stop rotation, the rails (only 2 needed) stop lateral movement. anti-backlash nuts on all 3 lead screws and you're laughing.
• [forums.reprap.org] you're using rails but is there any rotation in the rail carriage part (about the x y *or* z axes) i.e is the linear carriage *perfect*? if not, take a look at rods instead, set up as described in the thread there. experimental idea so YMMV, other people may have better advice here.
• [forums.reprap.org] follow the belt layout outlined here. keep the entirety of each belt in the same plane, offset them slightly (staggered heights). make sure that the belt goes *exactly* through right-angle turns, do *NOT* be tempted to do 89.5 or anything like that. use 2 bolts in the x-ends and stagger the bearings so that the belts come in on *exactly* the same centre line which should be along the centre of gravity of the carriage... and the hotend should be exactly on that centre of gravity.
• above a 300mm span don't use 8mm rods. anywhere. they bend too much.
• i got a meanwell LRS-350-24 prices are very good, they're also quite thin which is nice. but that was for a 200x200 printbed, you may want to get a mains-operated bed heater then the LRS-350-24 would do you fine... yep that's what you're planning. don't bother with 12V any more. make sure the fan from the PSU points *away* from the printbed...
• make sure you use 0.9 degree per step NEMAs, dc42 is extremely knowledgeable about this, you get far better accuracy and can do up to 1000mm/sec with them on a corexy.
• go for a Duet NG, again, dc42 knows what he's on about, REALLY don't bother with RAMPS or anything that uses arduinos.
• put fans pointing at the Duet's stepper ICs.
• consider a Flex3Drive, they are a combination of the best of both the bowden and direct extruder worlds *and then improve on that even more*. mutley3d is an extremely knowledgeable engineer.
• [forums.reprap.org] don't for god's sake bother with a bowden, they're basically shit. i know people say "oh i have no problems at all" - these people have never had a Flex3Drive. the reality is that bowden tubes are not robust, there's far too much to go wrong, they waste filament, and every single one of the problems i listed on that post was something i encountered MULTIPLE times over an extremely intense and very annoying period of 3 weeks. but, it boils down to this: as an engineer, retract of *five* millimetres as compared to 1.5mm for a flex3drive or direct drive extruder really should tell you everything you need to know.
• based on the same analysis as for the z bed i came up with a trick to use 3 springs and 2 fixed screws for bed-levelling. just as in the z-bed analysis the 2 fixed bolts stop rotation and lateral movement, and the 3 springs allow bed-levelling. if you used smooth rods (or part-threaded M3 bolts) and got the drill holes to good tolerances you could ensure that the printbed really really does not move laterally, i have a tiny amount of potential lateral movement because the 2 fixed bolts sit in slightly oversized holes. something for me to fix later.
• enclosed you will not be able to use PLA for parts, they will melt. find a material that has a high enough temperature or just go all-aluminium.... i think you said that's what you planned
• E3Dv6 i recommend going immediately to a volcano upgrade, unless you really really have a need to use 0.3mm nozzles or below.
• i use mechanical endstops and find them to be fine, but i have no experience with anything else yet so cannot make a comparison.

errr that's all i haven't covered everything, as i am learning as i go along, so have not commented on things i don't know about (or have said "i don't know" if i have). you'll like the Duet. it has ethernet and a built-in web interface, and you can get an LCD for it as well.

there's plenty of great people here who are knowledgeable. there's some angry ones, stressed-out ones (including me) and some inexperienced ones too, but that's okay, we're all learning. enjoy

Edited 1 time(s). Last edit at 05/29/2016 05:46AM by lkcl.

---

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

Here's what I made to get high quality prints. I'm working on a core xy design now.

[www.instructables.com]

Like you, I don't care about high print speed, quality is the goal.

I'm not a fan of glass beds, but people seem to like them. I prefer cast tooling plate with a layer of Kapton tape. I used two linear guides on the y axis with ball screw drive, and a single guide rail for the x axis. A single motor in the z axis is definitely the best way to go.

It's refreshing to see someone on this forum who isn't focused only on making the cheapest possible printer. Welcome!

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
lkcl
hiya hobbymods, welcome to the forum. some links to analysis that may help,

• look at what the_digital_dentist did, with son-of-megamax, his printer design is extremely comprehensive and accuracy is extremely good.
• [forums.reprap.org] for an analysis as to why you should use dual rails/rods and triple lead screws. dual screws is not enough. multiplying up the number of rails does not "fix" the problem of bed rotation about [only 2] screws, it just minimises it. triple lead screws is GUARANTEED to stop rotation, the rails (only 2 needed) stop lateral movement. anti-backlash nuts on all 3 lead screws and you're laughing.
• [forums.reprap.org] you're using rails but is there any rotation in the rail carriage part (about the x y *or* z axes) i.e is the linear carriage *perfect*? if not, take a look at rods instead, set up as described in the thread there. experimental idea so YMMV, other people may have better advice here.
• [forums.reprap.org] follow the belt layout outlined here. keep the entirety of each belt in the same plane, offset them slightly (staggered heights). make sure that the belt goes *exactly* through right-angle turns, do *NOT* be tempted to do 89.5 or anything like that. use 2 bolts in the x-ends and stagger the bearings so that the belts come in on *exactly* the same centre line which should be along the centre of gravity of the carriage... and the hotend should be exactly on that centre of gravity.
• above a 300mm span don't use 8mm rods. anywhere. they bend too much.
• i got a meanwell LRS-350-24 prices are very good, they're also quite thin which is nice. but that was for a 200x200 printbed, you may want to get a mains-operated bed heater then the LRS-350-24 would do you fine... yep that's what you're planning. don't bother with 12V any more. make sure the fan from the PSU points *away* from the printbed...
• make sure you use 0.9 degree per step NEMAs, dc42 is extremely knowledgeable about this, you get far better accuracy and can do up to 1000mm/sec with them on a corexy.
• go for a Duet NG, again, dc42 knows what he's on about, REALLY don't bother with RAMPS or anything that uses arduinos.
• put fans pointing at the Duet's stepper ICs.
• consider a Flex3Drive, they are a combination of the best of both the bowden and direct extruder worlds *and then improve on that even more*. mutley3d is an extremely knowledgeable engineer.
• [forums.reprap.org] don't for god's sake bother with a bowden, they're basically shit. i know people say "oh i have no problems at all" - these people have never had a Flex3Drive. the reality is that bowden tubes are not robust, there's far too much to go wrong, they waste filament, and every single one of the problems i listed on that post was something i encountered MULTIPLE times over an extremely intense and very annoying period of 3 weeks. but, it boils down to this: as an engineer, retract of *five* millimetres as compared to 1.5mm for a flex3drive or direct drive extruder really should tell you everything you need to know.
• based on the same analysis as for the z bed i came up with a trick to use 3 springs and 2 fixed screws for bed-levelling. just as in the z-bed analysis the 2 fixed bolts stop rotation and lateral movement, and the 3 springs allow bed-levelling. if you used smooth rods (or part-threaded M3 bolts) and got the drill holes to good tolerances you could ensure that the printbed really really does not move laterally, i have a tiny amount of potential lateral movement because the 2 fixed bolts sit in slightly oversized holes. something for me to fix later.
• enclosed you will not be able to use PLA for parts, they will melt. find a material that has a high enough temperature or just go all-aluminium.... i think you said that's what you planned
• E3Dv6 i recommend going immediately to a volcano upgrade, unless you really really have a need to use 0.3mm nozzles or below.
• i use mechanical endstops and find them to be fine, but i have no experience with anything else yet so cannot make a comparison.

errr that's all i haven't covered everything, as i am learning as i go along, so have not commented on things i don't know about (or have said "i don't know" if i have). you'll like the Duet. it has ethernet and a built-in web interface, and you can get an LCD for it as well.

there's plenty of great people here who are knowledgeable. there's some angry ones, stressed-out ones (including me) and some inexperienced ones too, but that's okay, we're all learning. enjoy

Mate that's some awesome info to go through, cheers for that.

Some of it I don't understand, but I do understand Bowden's being shit....my one dabble in them made me hate them.

Yes of course no PLA in the hot zone. In the case of me using printed parts which I will do my best to minimize, I'll use Zortrax polycarbonate.

I will use rails and understand the importance of getting them aligned correctly, the holes will be cut by waterjet off my DXF files to 0.1mm accuracy, I don't want to use rods on this one.

I will get some further explanation of the pulley alignment as I get into it and have read all the info I can find here.

I've absolutely never heard of the 3 ball screw thing ever, I'll look into it. That's the expensive bit.

I certainly don't consider myself and expert, I just know enough to be a pain in the arse lol.

Cheers again.

Quote
the_digital_dentist
Here's what I made to get high quality prints. I'm working on a core xy design now.

[www.instructables.com]

Like you, I don't care about high print speed, quality is the goal.

I'm not a fan of glass beds, but people seem to like them. I prefer cast tooling plate with a layer of Kapton tape. I used two linear guides on the y axis with ball screw drive, and a single guide rail for the x axis. A single motor in the z axis is definitely the best way to go.

It's refreshing to see someone on this forum who isn't focused only on making the cheapest possible printer. Welcome!

Yes I've looked at this one already, and I loved how you explained that speed is not everything.

I totally get that....why are they always on about it?

Obviously I have to stop rubbing myself over this and get started, and I have some ideas of my own, but I can see I'm in the right place to avoid the pitfalls.

I actually make a (very small) living out of printing, and can assure you that cheap printers are a false economy.

Holes are not what align linear rails (installed according to the specs you mentioned). Hiwin (as well as THK, IKO, SKF and several others) describe proper technique which is a machined lip for the bearings to rest against. It's not just the location of the holes, but the flatness of the surface you need to be concerned about. This is also specified in installation instructions, you can find these documents on all of the websites of the manufacturers I listed earlier.
As to your comments about 2 carriages per rail, if you are serious about "doing it right" work out the loads seen by your printer and make reasonable determinations as to what you need. You don't always need two per rail if you keep the moment load within specs. In the machine I'm currently building I am using 9mm rails for X and Y, with two rails for Y and a single rail and carriage for X. Z is a single 15W sized rail which is 30mm wide, and it is cantilevered and well within moment load spec. Your print area and other machine sizes will be different than mine, so you have to work out what works best in your situation but do a little math, it can save you a lot of cash.

I think switches are fine for X and Y endstops as they don't generally require high precision. The Z axis stop should be very precise. I used an optical switch pulled from an ink jet printer for that one. There's no hysteresis in the optical switch (unless your adjustment mechanism creates it).

I don't think there's any advantage to thermocouples over thermistors unless you intend to go to very high temperatures.

NEMA-17 motors can deliver up to about 100 oz-in of torque. That should be enough depending on how hard you intend to push the printer and the requirements for starting/stopping precision. See Oriental Motor's motor sizing calculators.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
I think switches are fine for X and Y endstops as they don't generally require high precision. The Z axis stop should be very precise. I used an optical switch pulled from an ink jet printer for that one. There's no hysteresis in the optical switch (unless your adjustment mechanism creates it).

that's good logic dd, i should find an optical switch for the z on the sandwich.

Quote


NEMA-17 motors can deliver up to about 100 oz-in of torque. That should be enough depending on how hard you intend to push the printer and the requirements for starting/stopping precision. See Oriental Motor's motor sizing calculators.

dc42 mentioned that the higher torque (higher ampage) motors have a lower resistance, which may end up with a mismatch on some stepper drivers. i made the mistake of getting 2.5A 84oz.in (60Ncm) steppers.... and trying to drive them with a Rumba v1.2g which is only rated up to about 1.4A on its steppers.... got lots of pauses in the middle of printing. even the duet 0.8.5's steppers are only rated up to 2A, you really have to pay attention and make sure the steppers are rated higher than the motor ampage.

Quote
hobbymods

Mate that's some awesome info to go through, cheers for that.

no problem

Quote

Some of it I don't understand, but I do understand Bowden's being shit....my one dabble in them made me hate them.

well it was a quick hit-list and there's links to the discussions where relevant which gives some background and context

Quote

I will get some further explanation of the pulley alignment as I get into it and have read all the info I can find here.

it took me a while to work out corexy, the photos etc. just don't help... because it's 3D. how can you understand something that's 3D and covers 3/4 of a square metre when you need detailed photos of like 10mm or less at each critical point? so yeah do go over as many photos as you can, or just download the Fusebox 3D STL, Alex_F did a complete model containing the entire design, you can load that into meshlab, walk round it. just copy that *but* make sure to use 2 bolts and do that offset-bearings trick.

oh - for god's sake don't do what i did on the sandwich200v2 which was to think that vertical carriage rails are okay. they f******g well aren't. read closely that discussion me and realthor are having about where to place the hotend relative to the bearings and rods: any travel in the linear bearings gets amplified, the further away the hotend is from them. i _really_ screwed up by underestimating that, and have a whopping *ONE MILLIMETRE* of travel at the hotend thanks to the amount of machining tolerance travel in the linear bearings. one millimetre! ain't doing that again...

Quote


I've absolutely never heard of the 3 ball screw thing ever, I'll look into it. That's the expensive bit.

3 ball screw? you mean 3 lead screws? it's just 3x TR8 rods to give triangulation (just like a 3-point bed-levelling system - exactly the same principle), but i used a closed loop belt arrangement, a GT2-16 on the motor and i 3d-printed my own GT2-50tooth pulleys, then looped a 900mm belt around all 4. real simple. most of the time people don't want to use TR8 rods because the pitch is higher (2mm per revolution instead of 1 or 1.25mm per revolution), but with the gearing 50:16 from the pulleys it's *better* than the usual M8 x 1mm pitch threaded bars for resolution.

threaded bars are a pain because as i found out, they often aren't straight, so the nut wobbles as it rotates. that has a *really* obvious effect on the print quality, it actually shifts the printbed sideways, so every 5 or so layers you get an entire line of print shifted over by a noticeable amount.

also, nuts tend to wear out, i actually managed to strip one entirely on the mendel90 i had. i wondered why i couldn't level the damn printbed properly...

they're not hugely expensive, i think it's like $12 per TR8 rod, and $4 per brass TR8 nut. you want brass because it's self-lubricating against steel. oh, and some... what are they... i made my own housings using skating bearings, people mentioned there was a better way to do it... somethingsomething K bearings... argh what were they - they're only $2.50... ah: KFL-08 pillow-block bearings, that was the ones. [www.ebay.com]

someone else, crazy guy, was talking about using triple independent Z motors. sounds mad... until you realise that with 3 independent motors you can do automatic print-bed levelling with a z-probe on the hotend. which is very very cool.

Autoleveling assumes that settings made at the start of the print will be maintained throughout the print. How many i3 people have problems keeping two motors synchronized? Adding a third motor is asking for trouble. Autoleveling is a self-fulfilling concept- if you include it in the machine, you will need it because you included it. The only reason for using multiple motors is to save a few $ on pulleys, bearings, and a closed loop belt.

My printer's Z axis has had pulleys and a closed loop belt with a single motor from day one, and after aligning the X axis perpendicular to Z, I have never had to touch it again. Ever.

I had thumbwheels and springs in my printer's bed leveling system but almost never had to make any adjustments. A few months ago I did some work on my printer's Y axis including removing the thumbwheels and springs from the leveling system. The bed is now bolted tightly in place on the leveling screws. So far I've had no need to adjust the bed leveling, even though I've transported the printer on its side in my car a few times. When I get to the destination, I stand it up and start printing without any adjustments. All it takes is solid frame construction and a flat bed plate.

You guys are thinking in low budget mode. With hobbymods budget, he can easily build the machine solidly enough that the bed will not get out of tram.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
Autoleveling assumes that settings made at the start of the print will be maintained throughout the print. How many i3 people have problems keeping two motors synchronized?

i don't know - but from what i know of i3s it would not surprise me in the least. i had a mendel90, designed by nophead, and had absolutely no problems at all with the two motors staying in sync. absolutely none. i trashed one nut because i switched on z-lift and forgot to grease the damn threads regularly, but that's my fault. i ran that mendel90 for 2 years, printed about 500m of PLA on it - didn't have a single motor-sync problem. nophead gave instructions on how to connect the wires up so that the Z motors were in series. i was to be honest pretty surprised by that, and even more surprised that it worked perfectly.

so *good engineering* means "no problem". the mendel90 has - had - a melzi. single-board. plenty of heat distribution for the stepper ICs. i3's? RAMPs. says it all.

Quote

Adding a third motor is asking for trouble. Autoleveling is a self-fulfilling concept- if you include it in the machine, you will need it because you included it. The only reason for using multiple motors is to save a few $ on pulleys, bearings, and a closed loop belt.

no, the reason for considering to use 3 motors is not to save money, please don't make assumptions that i have to contradict, it's not fun. you could say "did you mean to suggest that as a way to save money?" and i could answer "no" without having to contradict you... which takes efffort. don't make me do it - it's painful (literally) as my RSI is getting worse.

the reason is to be able to do auto-bed-levelling. a single motor cannot do auto-bed-levelling.

note, dd, i did *not* say "let's wire up 3 motors to the exact same stepper IC". dc42 i believe is considering looking at doing an addition to the reprapfirmware so that you can run triple lead screws off of triple motors running on triple stepper ICs. that's *completely* different from the s**t called "i3".

ok.

so.

that's sorted.

i'm interested to hear how easy it is to set up the initial calibration. how would you do the z-calibration on the arrangement that you have, if there are no spring-loaded screws? do you undo the grub screws on the pulleys and turn the lead screws by hand? do you remove the belt, turn the pulleys by hand, then put the belt back? how do you put the belt back if it is under tension?

the reason i ask is because i have a triple lead screw arrangement with a closed loop belt, and putting the belt on is a pain in the neck.

Edited 1 time(s). Last edit at 05/29/2016 05:41PM by lkcl.

Ikcl is building a Rube Goldberg foldable printer
using mainly kitchen utensils as tools,
because he has to move every week.
Exactly opposite of your build.

Don't get screwed by him -- if he doesn't know the capability / precision of a ballscrew
He's a rod man --- averse to proper linear guides.

Z drive --- 3 points determine a plane -- with two ballscrews, what keeps bed from tilting
stiff tooling plate on good frame could be made to work
but if you have three ballscrews driven / syncronized by a belt drive --- would be best

With water jet available -- no plastic printed parts for you.

Digital Dentist the sage of well built printer

Quote
Koko76
Holes are not what align linear rails (installed according to the specs you mentioned). Hiwin (as well as THK, IKO, SKF and several others) describe proper technique which is a machined lip for the bearings to rest against. It's not just the location of the holes, but the flatness of the surface you need to be concerned about. This is also specified in installation instructions, you can find these documents on all of the websites of the manufacturers I listed earlier.
As to your comments about 2 carriages per rail, if you are serious about "doing it right" work out the loads seen by your printer and make reasonable determinations as to what you need. You don't always need two per rail if you keep the moment load within specs. In the machine I'm currently building I am using 9mm rails for X and Y, with two rails for Y and a single rail and carriage for X. Z is a single 15W sized rail which is 30mm wide, and it is cantilevered and well within moment load spec. Your print area and other machine sizes will be different than mine, so you have to work out what works best in your situation but do a little math, it can save you a lot of cash.

Yes I did get all that about the "shoulder" for alignment, and yes I've been over those documents thoroughly. They also go on to instruct what to do when that isn't available.

Of course I am aware of the flatness spec, and will be using a stiffer grade ally (forgot the number at the moment, it's morning here) for the X gantry plate that the rails will be mounted on.

I am serious about "doing it right" which is why I'll be over engineering the hell out of it, and I did post that I wasn't taking budgetary short cuts.

I have already considered the single rail for the X, possibly with a single longer carriage, but I'll be sure after I do the math.

I get genuine Hiwin at a good trade (business) price, and have an existing relationship with the distributor. An extra carriage for the 9mm MG rail costs me bugger all, no need to skimp.

Quote
cozmicray
Ikcl is building a Rube Goldberg foldable printer

and proud of it! there's a concept called "permaculture" which summarises as, "respect your environment and re-use and repurpose what you have around you".

Quote

using mainly kitchen utensils as tools,

spoons and table knives make _great_ tools - been using them since i was 7, on mechano, and to take neighbour's household appliances apart.... you remember the saying, "a bad workman blames his tools"? you ever hear me *complain* about using spoons and table knives?

Quote

because he has to move every week.

and wants a printer that is accurate, fast, and can be packed up into 1/2 the space. yes.

Quote

Exactly opposite of your build.

not quite.

Quote

Don't get screwed by him -- if he doesn't know the capability / precision of a ballscrew
He's a rod man --- averse to proper linear guides.

i've just never used them yet, ray. i did say i didn't know because i've never yet used them. please don't associate words, sentences or assumptions about me when you know nothing about me, when it's physically painful to type and you force me to have contradict what you're saying it's not a lot of fun, so please have some respect and don't do that, ok?

Quote
cozmicray
Ikcl is building a Rube Goldberg foldable printer
using mainly kitchen utensils as tools,
because he has to move every week.
Exactly opposite of your build.

Don't get screwed by him -- if he doesn't know the capability / precision of a ballscrew
He's a rod man --- averse to proper linear guides.

Z drive --- 3 points determine a plane -- with two ballscrews, what keeps bed from tilting
stiff tooling plate on good frame could be made to work
but if you have three ballscrews driven / syncronized by a belt drive --- would be best

With water jet available -- no plastic printed parts for you.

Digital Dentist the sage of well built printer

Ha ha that's funny. Yes, I'm on the same page as the dentist lol. Built to a spec and standard, not a price.

If I want another budget priced piece of shit that I work on as much as I print with, I'll talk to Wanhao.

Let's not take it too much to heart, I'm really just dreaming at the moment and enjoying new knowledge.

Like all these things, everyone has their own approach and there's more than one way to skin a cat.

Edited 1 time(s). Last edit at 05/29/2016 06:26PM by hobbymods.

The point of leveling the bed, either auto or manual, is to get prints to stick to the bed. If they don't stick they fail, so the goal of leveling is to improve reliability of the printer.

The designer has a choice- build the machine solidly so that it can maintain bed level, or build it less solidly and require frequent manual leveling or autoleveling. There is no reason to add autoleveling to a solid machine that maintains the level setting. If you choose the less solid construction approach, usually to reduce cost of the printer's frame and mechanism, you add the complications and potential failure points of autoleveling while compromising print quality.

Ultimately, money spent on autoleveling buys you consistent first layer quality with reduced overall reliability. Money spent on solid construction buys you quality for every layer of the print without reducing (and maybe increasing) reliability.


Adjusting the X axis in my printer is as simple as loosening the grub screw on one pulley and turning the lead screw without turning that pulley. When the X axis is in correct position, tighten the grub screw. You would do the same with a three screw mechanism.

Edited 1 time(s). Last edit at 05/29/2016 06:38PM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
The point of leveling the bed, either auto or manual, is to get prints to stick to the bed. If they don't stick they fail, so the goal of leveling is to improve reliability of the printer.

The designer has a choice- build the machine solidly so that it can maintain bed level, or build it less solidly and require frequent manual leveling or autoleveling. There is no reason to add autoleveling to a solid machine that maintains the level setting. If you choose the less solid construction approach, usually to reduce cost of the printer's frame and mechanism, you add the complications and potential failure points of autoleveling while compromising print quality.

Ultimately, money spent on autoleveling buys you consistent first layer quality with reduced overall reliability. Money spent on solid construction buys you quality for every layer of the print without reducing (and maybe increasing) reliability.


Adjusting the X axis in my printer is as simple as loosening the grub screw on one pulley and turning the lead screw without turning that pulley. When the X axis is in correct position, tighten the grub screw. You would do the same with a three screw mechanism.

That's where I'm at.

You need to own and use a few of the off the shelf cheapies with all these functions to realize this.

I will spend my money on making the axis' as solid as I possibly can, because yes, it's not only all about that first layer. There's thousands more layers where that came from and they look best without axis shift too.

Totally coming around to the Mic 6 ally plate as well, as my waterjet cutting guy can probably supply it at the same price as I can get glass cut anyway.

Edited 2 time(s). Last edit at 05/30/2016 03:14AM by hobbymods.

Quote
DD
Adjusting the X axis in my printer is as simple as loosening the grub screw on one pulley and turning the lead screw without turning that pulley. When the X axis is in correct position, tighten the grub screw

I do that too. The thing I don't like about it is that it means I can't put flats on the lead screws for the set screws to tighten against. It hasn't bit me yet, but I'd be happier with flats, I just can't think of a good way of being able to adjust the relative heights if I do that. Perhaps some sort of adjuster on one of the Z carriages...

The load on the screws is the weight of the X axis, which isn't much. It increases a little when the motor accelerates, but not enough to matter. A single grub screw, even on a round shaft, can easily hold the pulley fast.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
Autoleveling assumes that settings made at the start of the print will be maintained throughout the print. How many i3 people have problems keeping two motors synchronized? Adding a third motor is asking for trouble.

All 3D printers rely on all the axis motors not losing steps from the start of the print through to the finish. Otherwise you get slipped layers. In a delta printer, every movement requires synchronised movement of 3 motors.

Keeping motors in sync is not a problem, unless you power the motors off and on again.

---

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
the_digital_dentist
The point of leveling the bed, either auto or manual, is to get prints to stick to the bed. If they don't stick they fail, so the goal of leveling is to improve reliability of the printer.

agreed.

Quote


The designer has a choice- build the machine solidly so that it can maintain bed level, or build it less solidly and require frequent manual leveling or autoleveling. There is no reason to add autoleveling to a solid machine that maintains the level setting. If you choose the less solid construction approach, usually to reduce cost of the printer's frame and mechanism, you add the complications and potential failure points of autoleveling while compromising print quality.

Ultimately, money spent on autoleveling buys you consistent first layer quality with reduced overall reliability. Money spent on solid construction buys you quality for every layer of the print without reducing (and maybe increasing) reliability.

there is another consideration - beyond that of "money buys you quality". if we may, let's go over a potential scenario with the design that you've created - or let's imagine that someone else follows the advice that you've given of removing the springs and to use the triple-axis lead screws to level the printbed. it's perfectly reasonable to consider that, because, well.... 3-point levelling, i agree absolutely you can just use the lead screws.

so let's imagine that someone has done that, but they are experimenting, so they have some firmware where the hotend hasn't quite been set up properly, or the slicing software hasn't been set up properly yet, and it heats up the hotend and a lot of plastic oozes out. let's imagine then that there is a huge blob of plastic on the hotend which they don't notice is there (or don't realise what's coming next). during the print, that blob then goes onto the printbed (they don't notice it because it's at the back of the hotend, out of their line of sight). that blob ends up on the printbed, and the hotend runs over it.

now, with a spring-loaded printbed, what happens is that the hotend will push that huge blob of plastic down, the springs will go "boingggg" and no damage is done.

HOWEVER...

... in the arrangement that you are recommending, there exists the possibility that, under the circumstances described above (which i believe you would agree are quite reasonable and rational), the printhead would continue to travel over the blob of plastic, and, given that the printbed cannot move downwards out of the way, one or more of the following is, i believe you would agree, quite likely to happen:

• the glass could crack
• the heated bed PCB (which is fibreglass) could be bent and possibly crack
• the carriage could be damaged
• the belts could be broken
• if the belts are strong enough the NEMA17s would stall instead, however the stalling would result in severe juddering, which could potentially cause huge movement spikes, smashing the hotend repeatedly down against the glass
• thus the hotend nozzle could be damaged as well as the glass cracked

and many more other possible ways in which the printer could be damaged, too numerous to list but just as equally severe in consequences.

now, i know that you may feel that the quality of your work, the quality of the prints and the care that you take is sufficient so that it is unlikely *for you* that this scenario could occur. which is why in the above scenario i specifically took *you* out of the equation and gave a scenario which was "someone else copies this idea and does not have as much experience as you".

in other words, the springs act NOT as a "cheapo nasty quality compromise that should not ever be considered in a high-quality design where budget is effectively unlimited", they're actually something to consider as a *really important* safety valve.

i fully appreciate that having 2 sets of 3-point levelling systems sounds initially to be a bit silly (triple lead screws that are adjustable *and* triple spring-loaded bed adjustments), particularly when the tolerances on accuracy for each levelling system are compounded, but can you see that there might be a really good reason to have the springs?

Quote


Adjusting the X axis in my printer is as simple as loosening the grub screw on one pulley and turning the lead screw without turning that pulley. When the X axis is in correct position, tighten the grub screw. You would do the same with a three screw mechanism.

ok great. i like that. my only thing is, i picked a pulley design that has triple grub screws, so i'd have to remove 2 in order to use this idea.

Quote
JamesK
I do that too. The thing I don't like about it is that it means I can't put flats on the lead screws for the set screws to tighten against. It hasn't bit me yet, but I'd be happier with flats, I just can't think of a good way of being able to adjust the relative heights if I do that. Perhaps some sort of adjuster on one of the Z carriages...

*hand-waving and grimacing so on...* yeahhh perhaps... an M3 screw would be involved in that? and we're back to screws and stuff which we were endeavouring to eliminate because the assumption is that they're redundant!

there's a rule which i can't remember where it came from, which is "things should be as simple as they need to be... but *no more*". in other words, if you oversimplify (for perfectly good reasons) you risk leaving something out - an edge-case which may turn out to be high-risk with catastrophic consequences.

i'm glad you mentioned the "flats on the lead screws" thing, JamesK. i was wondering what was slightly bothering me about only having one grub screw, but couldn't think what it was.

the only other thing i could think of was, having a "belt slackening" mechanism. you could adjust the lead screws, get them perfectly in position, then put the belt on and *then* hit the "tighten belt" button / lever. the only thing is, you'd better hope like hell the lead screws don't move whilst you're putting on or tightening the belt. but, that's always solveable by double-checking and redoing the task if necessary.

Quote
cozmicray
Ikcl is building a Rube Goldberg foldable printer

ray, i thought about this (and also have had a bit of a rest so my RSI isn't as bad). although it's funny, what you're saying is that the entire design - 3 months of work - is entertaining but ultimately of absolutely no value whatsoever.

a question for you: have you considered the consequences for yourself of describing other people's time and effort as being worthless, even if it is done in jest?

for the archives - for anyone else reading this now and in the future: ray is an extremely knowledgeable and competent engineer, whose slightly grumpy advice is extremely valuable and appreciated. he reads datasheets and app notes and over the relatively short period i've been on here i've seen that he has provided links to them repeatedly to the benefit of many people here on this forum. along this vein, he offered some help on the subforum for the sandwich200, and, having already done a comprehensive analysis, i was in a position to give a detailed, comprehensive, rational and reasonable explanation as to why his really very helpful advice would not fit the extremely comprehensive design requirements, but i specifically went out of my way to thank him for giving the advice and mentioned that it would likely be extremely useful advice for anyone else who did not have such comprehensive design requirements.

my feeling is, then, that ray has likely never encountered someone before who, because of the comprehensive nature of the requirements that have been set, has been *unable* to follow his advice *not* because they cannot understand them or because they are too inexperienced to follow them rationally, but because there are *genuine* reasons why the advice he gave could not be followed.

as we got into the (extremely comprehensive) chain of logic as to why his advice could not be followed, my feeling is that he began to realise that he was out of his depth. most 3D printer designs are a fairly simple, straightforward process of "divide and conquer". the placement of the PSU generally has absolutely nothing to do with the mechanical design, for example. it's bolted onto the frame, out of the way, and it's "sorted". however, in a *compact* portable design (compact being a completely different requirement from portable) the position of the PSU has to be extremely carefully considered. after months of careful design work, i had *ONE* place left that i could fit a *LOW PROFILE* PSU. and that's just one component.

so when cozmicray describes the sandwich200 as a "rube goldberg" machine, what he is really saying is, "i tried to follow the comprehensive requirements and the extremely lengthy chain of logic but i simply couldn't cope, so everything that lkcl is trying to achieve therefore looks to be completely pointless. entertaining, pretty, but ultimately functionally absolutely useless, because in *my* mind it's clearly far too much for *me* to cope with".

now, what he *could* have said, was, "wow man, i'm sorry, i just can't cope - btw, how in god's name are you managing to deal with all this complexity? why is your brain not melted to jelly?" to which i could have replied, "i'm having ice-cream with that jelly for dinner. ok, i'm not, but complexity is something that i've trained for decades to deal with. thanks for you help and advice btw, i'm sorry i couldn't use it" and we would be cool.

i made a decision over twenty five years ago to tackle the things that other people can't cope with, because they're simply too much. "straightforward" is not a challenge for me, but more than that: why should i take away the joy and sense of achievement that other people get from achieving something that is within *their* capabilities? so i deliberately stay away from "easy" as much as i can.

however in the case of the sandwich200, i didn't have time to do that. i *really wanted* to use a pre-existing 3d printer design. but a comprehensive analysis of about 3-4 days of research showed that no such design (portable, compact, fast, accurate) existed. so from *NO EXPERIENCE* in 3D printer design i have had to learn. very very fast. that leaves huge holes in my knowledge, rather limited time to respond or analyse certain techniques (as to whether they may or may not fit the extremely comprehensive requirements), the potential risk of getting stuck with an improperly-evaluated design decision and so on.

so.

cozmicray does that help you to give you some food for thought? hobbymods i apologise for taking up this sub-forum with these thoughts. i felt it was important to go over them at the time and place where cozmicray has commented, as opposed to a separate subforum where they may not be connected by other people reading these public archives, some time in the past. he could have chosen *not* to say what he said. he could have chosen to contact me privately, or to use the sandwich200 sub-forum for his quite funny (but barbed) comment.... but he didn't.

so. anyway, hobbymods: now at least you see what i meant that there is a huge mixture of experience as well as emotion on this forum... and that we're all learning.as a community from the insights that we have to share with each other - not just technical. but... can we keep it technical? i'd much prefer that.

Regarding bed screws and springs: Unless you do a better job of engineering the spring arrangement than I did, you risk the bed moving horizontally when the head hits a plastic blob. This results in shifted layers.

I had springs on my bed, but due to the slightly shifted layers issue (only occasional, and usually less than 0.1mm), I changed to a solidly mounted bed system. This is still adjustable for levelling, but is locked in position once adjusted.

You all sound like much better engineers than me, so feel free to ignore my findings. But I am quite fussy with my prints, and found spring loaded mechanisms too hard to keep locked in the x-y plane.

I think there's a very large variation in the amount of tension that people use for spring levelling systems. I use stiff springs at close to coil bind, and I'm fairly sure the heat break would bend before the springs moved appreciably during a head crash. I don't think there is much likelihood of any lateral movement of the print bed on the springs at this compression level.

Yeah....I'm getting all the emotion.

Back on track....did a little research and my contacts assure me that 15mm EG/HG rails not only have better features than the miniature rails, but he can do them cheaper.

That's some big rails, but I'm up for it if they fit and aren't too cumbersome. Plentiful used ones on ebay as well for prototyping.

Yes I will have a 3 point spring loaded manual bed leveling system, and yes I agree that springs on the build plate are a necessary insurance policy against busting something really expensive. I just plan to do something better than chinese spec wing nuts and soggy mismatched springs.

I think I've been misunderstood about my Z axis plans....I'm hoping to do something like this monstrosity (check the PRICE!!):

[3dprintersuperstore.com.au]

Only with rails instead of rods, and the 2 ball screws driven by the same motor/belt/pulley system.

While I doubt that POS would even start up let alone print, that's the basic setup I'm looking at.

As for the gantry, this is the material/look I'd go for:

[www.thingiverse.com]

Again, with rails not bronze bushes glued onto things and without all the strange methods of mounting things. Motor mounts would be slotted for belt tensioning in my case, but I like the idea of a nice ally assembly that could be lifted straight out of the top of the printer for servicing.

While with the Z axis I feel that mounting the rails on flat panel alone should suffice (done it before), for the gantry plates it wouldn't be that much more to get a 1-2mm deep groove milled into the plate so as to provide the reference edge correctly pointed out to me in a previous post, for that perfect job.

Definitely for the Z axis and most likely the Y axis I will be using 4 carriages (2 per rail), but for the X axis I should be fine to use only one carriage per rail for the small plate that I'll mount the extruder/hot end on, possibly the longer carriage option in that case. I WILL NOT mount the extruder on a single rail/carriage.

Obviously this will be heavy, but we've covered the fact that I'm not interested in printing at 900mm/second and am going for repeatable high quality prints.

Brings me to the question though...NEMA 23's? Yes/No? 17's OK if cranked right up?

Will a Smoothieboard drive bigger motors?

I'll start sketching stuff up soon and cut up some templates out of some corrugated plastic sheet I've got here before I shell out on ally.

Appreciate all the input!

Quote
nebbian
Regarding bed screws and springs: Unless you do a better job of engineering the spring arrangement than I did, you risk the bed moving horizontally when the head hits a plastic blob. This results in shifted layers.

I had springs on my bed, but due to the slightly shifted layers issue (only occasional, and usually less than 0.1mm), I changed to a solidly mounted bed system. This is still adjustable for levelling, but is locked in position once adjusted.

You all sound like much better engineers than me, so feel free to ignore my findings. But I am quite fussy with my prints, and found spring loaded mechanisms too hard to keep locked in the x-y plane.

If there's a big blob of plastic on the bed your print is finished/failed so there isn't much point talking about how the print is going to look if the extruder bangs into the blob. Forgetting about a blob of plastic on the bed, if the springs are so soft that the nozzle won't have a problem, or the glass won't break (if you still use glass, I don't because it can break), etc. in the event of a head crash, the bed is probably going to wobble while you're printing. Every time the bed starts or stops, some of the springs are going to compress and the bed will tilt downward. That can/will affect print quality. Of course, it shouldn't be a problem in a machine in which the bed moves in the Z axis.

While we're on the subject, have you considered that glass is not the best material to use for a bed plate? It can break, either by the nozzle crashing into it, prints sticking aggressively, tools being dropped, or the plate being dropped (if you're one of those who print so much that you can't wait a few minutes to start the next print so you swap glass plates). Broken glass can cut you. It is a poor thermal conductor- what is the benefit in that? You use metal clips to hold the glass in place. Those clips stand up above the bed surface. If the extruder crashes into them, will it survive unscathed? If you have to swap plates, why not swap aluminum plates?

I didn't recommend that anyone should take the springs out in their printer. I simply said that I did and haven't had any problems or need to adjust the bed leveling. The springs I had in place were so stiff that they would not have protected the nozzle or a glass bed from a head crash. On one occasion, maybe 6-8 months ago, I had some sort of problem, I don't remember the details, but the nozzle crashed into a print. The print was stuck to the bed solidly so the heatbreak bent. SS heatbreaks bend so easily that soft springs or no, you can't protect them except to be careful in operating your printer. If you have an E3D v6 hot-end, I recommend you keep an extra heatbreak handy. You can straighten a bent heatbreak, but you can probably only do that a couple times before it breaks.

Even if you're going to use 3 leads screws to lift the bed in the Z axis, you'll probably still use leveling screws for the bed. If the bed is heated, you will want it stood up off the undercarriage that couples it to the lead screws so you don't end up heating the whole assembly. I suppose you could just use fixed spacers instead of leveling screws and use the lead screws to level the bed, but putting thumbwheels on the leveling screws is probably a lot more convenient. OTOH, if your construction is solid and you don't have to keep adjusting the bed level, you may as well just use the 3 lead screws to do the leveling.

Regarding pulleys and grub screws- I don't understand why the number of grub screws matters. My pulleys have two grub screws each. If your pulley has 3 grub screws, loosen all of them, rotate the lead screw to adjust the tilt of the X axis (or the bed if it moves in the Z axis), then tighten all three grub screws. It ain't rocket science.

Edited 2 time(s). Last edit at 05/30/2016 08:47AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Oh, yes...the build plate.

I will be using mic6 ally plate for all the reasons mentioned above. On very stiff leveling springs.

Plus I can probably get it done cheaper than glass, as I don't have a glass guy, and my water jet guy will just charge for the time on his machine plus the sheet of material, which shouldn't be too much more than standard ally.

Edited 1 time(s). Last edit at 05/30/2016 08:54AM by hobbymods.

Awesome. You're making me very jealous for your contacts. It costs me a fortune to get tooling plate. After a year and a half of working around cheap components I do find myself wishing I could just drop the cash on the good stuff DD is usually right about these things.

Quote
JamesK
Awesome. You're making me very jealous for your contacts. It costs me a fortune to get tooling plate. After a year and a half of working around cheap components I do find myself wishing I could just drop the cash on the good stuff DD is usually right about these things.

I'm in the heart of a mining industry that has collapsed...there's a lot of skills/facilities around here and they're all looking for a gig. Now's the time.

I'm sure he is right, as he has actually built a printer that is doing amazing things with large ABS prints. Proofs in the pudding in this case.

Cheap is so often a false economy when you look at what spending 25% more will get you.

Edited 1 time(s). Last edit at 05/30/2016 09:03AM by hobbymods.