MAVB3D
Posted by pugzor
|
MAVB3D June 24, 2017 09:30PM |
Hi all,
I want to pick up my build which sort-of started maybe 18-24 months ago now. Long story short, I found a girl (we're getting married in September) but this is something I got a damn good start on and would like to finish.
Anyway, I've had a few printers starting with a home-made Cherry Pi IIIS delta. Was a silly way to get introduced to 3D printing and I really should have done something easier like a Prusa i3. Probably 8 hours building and 40 hours debugging! I did get it printing awesomely though so I'm confident to go back to something a little more conventional but scaled up. Also acquired a M3D Micro from Kickstarter, which is cool, but the lack of a heated bed really makes it pretty limited. Actually considering backing the Obsidian too when it goes up in the next few days but we'll see.
Here's how well I got my delta honed in:
Basically the idea for this was to build a massive volume 3D printer that could do its thing for days on end without supervision. I forget what the acronym for MAVB3D is, but it was cool at the time so I'm going to keep it as that for the project name.
Design is mostly built around those Vulcanus Max (CoreXY) printers that were floating around a few years ago. Admittedly they were the inspiration although I have deviated from the design a bit. [www.instructables.com] End goal is to have a print envelope of around 500x500x500mm with maybe a 50 micron print resolution.
I printed most of the plastics required as per here:
Main thing I haven't printed yet is the extruder carriage as I want to go a bit custom with that. First iteration will be just a single direct drive (was considering Bowden for a while, but that's going to be unusable with the massive tubes required) and I'll try to make it dual-head one day. Hopefully I can use 23mm high NEMA17s to keep down the weight of this carriage. Although it's being made to support massive objects I'll be using standard 1.75mm filament with a 0.4mm nozzle - I do want good quality prints despite the size.
I figured taking a triangular approach to mounting the print bed was going to be the easiest way to scale this up. It'll be driven by a single NEMA23 as discussed in this topic: [forums.reprap.org]. The bed will have three leadscrews driving vertically (two on the sides, close to the front, and one at the rear in the centre) and four smooth rods at the four corners (on the sides). Lead screws will sit on these castor-type bearings that I found (can't remember their technical name):
Also made these nice little brackets to hold the leadscrews in place with standard 608 bearings:
Main drive for the leadscrews should be 16 tooth at the motor and 40 tooth at the base of the rods. Got plenty of everything though:
Hoping to manually adjust bed level, again using triangular technique, using valve springs from a 1998 Toyota Camry. Funnily enough they are brilliant for this application - they're made to withstand extreme heat, are made to have a large amount of pressure, and were absolutely free. Two of these will be on the back edge of the print bed (at either corner) with a third at the centre middle of the front edge. Once I hone these in I probably will never need to adjust them ever again.
I've got a nice big chunk of flatness-milled aluminium for the bed. It's 8mm thick if I recall and should have a flatness of 0.38mm over 6ft (again, if memory serves). Might throw a very thin piece of float glass over the top to protect the bed too. Bed is already ~10kg. Did get a nice big silicone mat to heat it up too which will stick on the bottom.
Already have built the frame. It's large! Base is 750x750mm on the outer and barely fits through doorways. It's mounted on some chipboard too for extra stability. Unlike the Vulcanus Max I used 20x40 pieces of rail to give it more rigidity in most areas, and my corner posts are 40x40mm. Should be bloody rigid.
For electronics I was going to go one of those cheap AliExpress MKS SBase boards (seriously, dat price!), but I've since decided I'll pay 10x more and go for a Duet WiFi + Duex5. Mostly made this decided this for the support factor (David and those involved are priceless) and less to troubleshoot without having to worry about the electronics being dodgy. Might go a PanelDue and IR Probe too but we'll see how money comes.
Decided to run the thing on 24v due to the amount of power needed. Ended up being more economical when it came to power supplies. Have a few other things I want it to handle (like LEDs) so may have a small 12v PSU to handle a few independent systems. We'll see how it goes though.
First thing I need to do is perform a stocktake, see what I really do and don't have, and figure out exactly how I'm going to execute some of the more creative bits and pieces. I'll hopefully make a post or two on here as things progress! No doubt there'll be requests for troubleshooting come through.
I want to pick up my build which sort-of started maybe 18-24 months ago now. Long story short, I found a girl (we're getting married in September) but this is something I got a damn good start on and would like to finish.
Anyway, I've had a few printers starting with a home-made Cherry Pi IIIS delta. Was a silly way to get introduced to 3D printing and I really should have done something easier like a Prusa i3. Probably 8 hours building and 40 hours debugging! I did get it printing awesomely though so I'm confident to go back to something a little more conventional but scaled up. Also acquired a M3D Micro from Kickstarter, which is cool, but the lack of a heated bed really makes it pretty limited. Actually considering backing the Obsidian too when it goes up in the next few days but we'll see.
Here's how well I got my delta honed in:
Basically the idea for this was to build a massive volume 3D printer that could do its thing for days on end without supervision. I forget what the acronym for MAVB3D is, but it was cool at the time so I'm going to keep it as that for the project name.
Design is mostly built around those Vulcanus Max (CoreXY) printers that were floating around a few years ago. Admittedly they were the inspiration although I have deviated from the design a bit. [www.instructables.com] End goal is to have a print envelope of around 500x500x500mm with maybe a 50 micron print resolution.
I printed most of the plastics required as per here:
Main thing I haven't printed yet is the extruder carriage as I want to go a bit custom with that. First iteration will be just a single direct drive (was considering Bowden for a while, but that's going to be unusable with the massive tubes required) and I'll try to make it dual-head one day. Hopefully I can use 23mm high NEMA17s to keep down the weight of this carriage. Although it's being made to support massive objects I'll be using standard 1.75mm filament with a 0.4mm nozzle - I do want good quality prints despite the size.
I figured taking a triangular approach to mounting the print bed was going to be the easiest way to scale this up. It'll be driven by a single NEMA23 as discussed in this topic: [forums.reprap.org]. The bed will have three leadscrews driving vertically (two on the sides, close to the front, and one at the rear in the centre) and four smooth rods at the four corners (on the sides). Lead screws will sit on these castor-type bearings that I found (can't remember their technical name):
Also made these nice little brackets to hold the leadscrews in place with standard 608 bearings:
Main drive for the leadscrews should be 16 tooth at the motor and 40 tooth at the base of the rods. Got plenty of everything though:
Hoping to manually adjust bed level, again using triangular technique, using valve springs from a 1998 Toyota Camry. Funnily enough they are brilliant for this application - they're made to withstand extreme heat, are made to have a large amount of pressure, and were absolutely free. Two of these will be on the back edge of the print bed (at either corner) with a third at the centre middle of the front edge. Once I hone these in I probably will never need to adjust them ever again.
I've got a nice big chunk of flatness-milled aluminium for the bed. It's 8mm thick if I recall and should have a flatness of 0.38mm over 6ft (again, if memory serves). Might throw a very thin piece of float glass over the top to protect the bed too. Bed is already ~10kg. Did get a nice big silicone mat to heat it up too which will stick on the bottom.
Already have built the frame. It's large! Base is 750x750mm on the outer and barely fits through doorways. It's mounted on some chipboard too for extra stability. Unlike the Vulcanus Max I used 20x40 pieces of rail to give it more rigidity in most areas, and my corner posts are 40x40mm. Should be bloody rigid.
For electronics I was going to go one of those cheap AliExpress MKS SBase boards (seriously, dat price!), but I've since decided I'll pay 10x more and go for a Duet WiFi + Duex5. Mostly made this decided this for the support factor (David and those involved are priceless) and less to troubleshoot without having to worry about the electronics being dodgy. Might go a PanelDue and IR Probe too but we'll see how money comes.
Decided to run the thing on 24v due to the amount of power needed. Ended up being more economical when it came to power supplies. Have a few other things I want it to handle (like LEDs) so may have a small 12v PSU to handle a few independent systems. We'll see how it goes though.
First thing I need to do is perform a stocktake, see what I really do and don't have, and figure out exactly how I'm going to execute some of the more creative bits and pieces. I'll hopefully make a post or two on here as things progress! No doubt there'll be requests for troubleshooting come through.
|
Re: MAVB3D June 25, 2017 12:01AM |
"End goal is to have a print envelope of around 500x500x500mm with maybe a 50 micron print resolution....I printed most of the plastics required as per here: "
I'm not sure 500x500x500 mm, 50 um resolution, and printed plastic parts all belong in the same sentence, let alone the same printer. For that kind of size and resolution, rigidity in every component will be critical. Plastic part use, especially printed plastic parts, should be minimized.
If you have a flat aluminum bed you don't need to put glass on it. It's better to use a material like PEI that prints will stick to without slopping hairspray, glue, or root beer, on it.
Using four rails is asking for trouble. If there's even a tiny misalignment (and there will be) the whole thing will bind or walk instead of moving smoothly. You shouldn't need more than two rails, preferably located close to the screws. For the sort of resolution you're after, I wouldn't use end-supported rails, either. Use fully supported rails or linear guides.
When you heat the 500 mm bed to 100C is is going to expand by almost 1 mm (about 0.5 mm when heated to 70C). Your bed support system may not be too happy about that, and you may have problems with the bed lifting or the bed or support structure flexing. I designed a three point support mechanism for my latest printer that allows the bed to expand without creating any lateral forces on the leveling screws. I support the bed from below using a rigid support structure in the form of a "tee" made from 40x40mm 8020 extrusion (probably overkill for my printer, but probably about right for yours). The bed plate sits on 3 screws, two of which have spherical heads. There is a reference screw that is typically not adjusted, whose spherical head sits in a conical hole in the bottom of the bed plate. The conical hole prevents any lateral motion at that location, but since the screw head is spherical, the bed can swivel on that screw when leveling. On the opposite side of the bed there is another spherical support screw, the pitch adjuster, but this one sits in a conical slot (made with a countersink bit on a mill). The slot allows the bed to expand in the X direction but prevents any motion in the Y direction. The spherical screw head allows the bed to pivot for leveling. The third screw, the roll adjuster, is simply a screw that contacts the flat bottom of the bed plate. That allows the bed to expand in X and Y and to pivot for leveling adjustments. At each of the three screws there are springs that hold the plate down onto the screws.
Here's the general idea, viewed from the bottom:
Here's what it looks like in practice (top view):
The screw on the left is the reference that sits in a conical hole. The screw on the right is the X axis pitch adjuster. The screw at the front center is the roll adjuster and has a thumbwheel under the assembly. The reference and pitch adjusters are spherical head screws used in RC cars. They have 2.5 mm hex sockets on the top so they are adjusted from the top side of the bed. All three screws are threaded into teflon or torlon blocks that don't mind the temperature and grip the screws tightly so the bed can't wobble.
You may find that with only 3 supports, a bed that size may sag under it's own weight and not provide the flatness you need to print in 50 um layers. You may have to do something a little more complex to support the corners on the side with the roll adjuster. This is a situation where autoleveling that maps the bed may be useful.
Setting aside the difficulty and expense of building a large printer with the sort of resolution you're after, have you considered how long it will take to print a large object in 50 um layers? You certainly can do it, but people usually match the printer size to the object size. Large printers usually print large objects in thick layers so they'll finish in a reasonable amount of time. Things that require very high resolution are usually small and usually printed on smaller printers that are much easier to build to the required rigidity and resolution. Just something to think about...
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
I'm not sure 500x500x500 mm, 50 um resolution, and printed plastic parts all belong in the same sentence, let alone the same printer. For that kind of size and resolution, rigidity in every component will be critical. Plastic part use, especially printed plastic parts, should be minimized.
If you have a flat aluminum bed you don't need to put glass on it. It's better to use a material like PEI that prints will stick to without slopping hairspray, glue, or root beer, on it.
Using four rails is asking for trouble. If there's even a tiny misalignment (and there will be) the whole thing will bind or walk instead of moving smoothly. You shouldn't need more than two rails, preferably located close to the screws. For the sort of resolution you're after, I wouldn't use end-supported rails, either. Use fully supported rails or linear guides.
When you heat the 500 mm bed to 100C is is going to expand by almost 1 mm (about 0.5 mm when heated to 70C). Your bed support system may not be too happy about that, and you may have problems with the bed lifting or the bed or support structure flexing. I designed a three point support mechanism for my latest printer that allows the bed to expand without creating any lateral forces on the leveling screws. I support the bed from below using a rigid support structure in the form of a "tee" made from 40x40mm 8020 extrusion (probably overkill for my printer, but probably about right for yours). The bed plate sits on 3 screws, two of which have spherical heads. There is a reference screw that is typically not adjusted, whose spherical head sits in a conical hole in the bottom of the bed plate. The conical hole prevents any lateral motion at that location, but since the screw head is spherical, the bed can swivel on that screw when leveling. On the opposite side of the bed there is another spherical support screw, the pitch adjuster, but this one sits in a conical slot (made with a countersink bit on a mill). The slot allows the bed to expand in the X direction but prevents any motion in the Y direction. The spherical screw head allows the bed to pivot for leveling. The third screw, the roll adjuster, is simply a screw that contacts the flat bottom of the bed plate. That allows the bed to expand in X and Y and to pivot for leveling adjustments. At each of the three screws there are springs that hold the plate down onto the screws.
Here's the general idea, viewed from the bottom:
Here's what it looks like in practice (top view):
The screw on the left is the reference that sits in a conical hole. The screw on the right is the X axis pitch adjuster. The screw at the front center is the roll adjuster and has a thumbwheel under the assembly. The reference and pitch adjusters are spherical head screws used in RC cars. They have 2.5 mm hex sockets on the top so they are adjusted from the top side of the bed. All three screws are threaded into teflon or torlon blocks that don't mind the temperature and grip the screws tightly so the bed can't wobble.
You may find that with only 3 supports, a bed that size may sag under it's own weight and not provide the flatness you need to print in 50 um layers. You may have to do something a little more complex to support the corners on the side with the roll adjuster. This is a situation where autoleveling that maps the bed may be useful.
Setting aside the difficulty and expense of building a large printer with the sort of resolution you're after, have you considered how long it will take to print a large object in 50 um layers? You certainly can do it, but people usually match the printer size to the object size. Large printers usually print large objects in thick layers so they'll finish in a reasonable amount of time. Things that require very high resolution are usually small and usually printed on smaller printers that are much easier to build to the required rigidity and resolution. Just something to think about...
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
|
Re: MAVB3D June 27, 2017 11:07PM |
@Mark, I've been thinking pretty heavily about some of the things you've said, mostly around the rails. Originally I thought you were referring to how the Z axis moves but I'm now worried about the gantry (X&Y axes). Smooth rods were planned but I suspect these would be quite prone to sagging too over the ~750mm they need to cover.
Based on the layout I'd originally been working with, changing these to fully supported linear rails isn't going to be fun as it'll require a lot of the gantry parts to be redesigned. Eg, motor mounts, carriages, belt routing, etc. I don't mind doing that (some rework now will save frustration with crappy prints, resulting in needing to do it anyway!), but I wanted your opinion - should dual smooth rods be fine for the carriage which holds the extruder/s (keeping in mind it'll be direct drive)? The side rails I'm fine to convert to supported linear rails but I'm worried about the additional weight/space required for the cross-rails.
There are a few brackets and mounting plates which I'll just make using metal parts too, I think.
What method would you recommend for auto bed levelling too on a large printer? Not sure whether to go PEI (hard to get large enough sheets here) or something else, but won't go glass. Probably Kapton tape. Original specs called for an inductive sensor but given I'd be going a Duet WiFi + Duex5, is there a better option?
@newbob I have no idea to be honest about the speed but think it was conservative... maybe 0.4mm nozzle at 35mm/s? Could have been faster. Was using extremely cheap Flashforge filament too so very happy with the results. I'll have to start up the PC with S3D on it again and see what the settings were. Extruder was a e3d v6 clone which performed quite well. Originally had a jhead but never quite had much luck with it. Admittedly I've been quite impressed with a lot of the knock-off gear I've had.
Based on the layout I'd originally been working with, changing these to fully supported linear rails isn't going to be fun as it'll require a lot of the gantry parts to be redesigned. Eg, motor mounts, carriages, belt routing, etc. I don't mind doing that (some rework now will save frustration with crappy prints, resulting in needing to do it anyway!), but I wanted your opinion - should dual smooth rods be fine for the carriage which holds the extruder/s (keeping in mind it'll be direct drive)? The side rails I'm fine to convert to supported linear rails but I'm worried about the additional weight/space required for the cross-rails.
There are a few brackets and mounting plates which I'll just make using metal parts too, I think.
What method would you recommend for auto bed levelling too on a large printer? Not sure whether to go PEI (hard to get large enough sheets here) or something else, but won't go glass. Probably Kapton tape. Original specs called for an inductive sensor but given I'd be going a Duet WiFi + Duex5, is there a better option?
@newbob I have no idea to be honest about the speed but think it was conservative... maybe 0.4mm nozzle at 35mm/s? Could have been faster. Was using extremely cheap Flashforge filament too so very happy with the results. I'll have to start up the PC with S3D on it again and see what the settings were. Extruder was a e3d v6 clone which performed quite well. Originally had a jhead but never quite had much luck with it. Admittedly I've been quite impressed with a lot of the knock-off gear I've had.
Sorry, only registered users may post in this forum.