Welcome! Log In Create A New Profile

Advanced

Building a Mendel in 2021. What would you choose?

Posted by ComeWhatThey 
Building a Mendel in 2021. What would you choose?
October 12, 2021 10:46AM
Howdy! I've long been a fan of the reprap project, I remember first reading about Darwin when I was in high school and was inspired, even if actually building one was way out of reach for me. But now, some 13 years later, I find myself in a position where doing something like that actually does seem to make sense...

The why

I want to get some experience of DIY building, with an eye to eventually making something with a significantly larger build area for automotive parts, as well as multi material stuff and laser/CNC. I’ve learned that it’s often best to start small and build up, to get an idea of the sorts of issues you might run into with scaleability, and so that when you do get there you can concentrate on that rather than trying to learn the basics at the same time.

So for now I’m just looking for something on the smaller side, more of a low pressure project to play around with and get comfortable with. And while I’m aware that stuff has come along significantly in the past decade, I actually feel like going back to the beginning (to an extent) and getting that first hand look at how some of those developments occurred and what fuelled them could be really interesting and valuable.

I actually already have access to a 3D printer, my partner has an flsun delta, so I have no trouble getting printed parts. But also because of that, build area isn't too much of a concern right now; if there's anything larger I wanted to print I could always use that. Making my own is more about getting that build experience, and also having a bit more freedom to get to grips with the software side of things, and to experiment with different hardware and settings.

Speaking more philosophically for a second, I actually feel like reprap could be just as relevant now as ever, which makes it particularly sad to see the wiki in such a sorry state. The fact that home 3D printers are much more available and widespread than a decade ago, the fact that so many of the components have reduced in cost as a result, that should play directly into the hands of an open source DIY option, not against it.

Obviously not everyone can or should build their own, and the community is better off for there being so many affordable options to get straight into it. There’s more to print, more people to ask for help, more resources and easy to follow youtube videos. But it’s also way easier and cheaper to get the 3D printed parts you need to put your own together if you have a friend or family member you can ask.

I know there are practical and structural reasons why this isn’t the case, but a lot of the early hype around home 3D printers seemed to be around how they could “end mass production” of cheaply made, disposable parts that get shipped all over the world with massive environmental impact, just because it’s cheaper than producing things as they’re needed more locally. I can’t help but feel some sense of loss that home 3D printers themselves have largely become part of that system, with everyone seemingly rushing to cash in and produce the “next big thing” that you absolutely must buy from them, rather than taking a more sustainable approach to design.

As I say, I think the wider community is in a pretty healthy place, as much as can be expected. And I think this is more or less how things will continue unless and until there is a much bigger push to enshrine right to repair in law. I just wish that reprap builds and documentation had kept up more, to make the DIY option as easy as possible for those who may consider it. Anyway...

The what

I want to build a Mendel, or one of its close derivatives, for a few reasons. I’m aware that there’s been much discussion about the drawbacks of the design and there are good reasons why more modern designs have largely moved away from trapezoidal shapes. But the idea of putting one together using stuff from the hardware store has always appealed to me, and let’s not ignore the most important factor, they just look cool! xD

Despite all the drawbacks, clearly many people have managed to make well performing and reliable printers from these designs, and the process of getting to that point seems like a great opportunity for learning. I do think it strikes a good balance for my (admittedly niche) use case. They can be built fairly cheaply, heavily modified, have parts re-used and iterated, and because of the threaded rod construction, scaleability is fairly easy. I’m intrigued by the notes in the wiki about scaling to use standard lengths of 13” silver steel for the axis rods, and I guess that could apply just as much to scaling down a Mendel as it does to scaling up a Huxley.

As fun as it would be to build my own Gen7 board, for now I will be using a cheap RAMPS kit for the electronics. I can also find Nema 17 motors much cheaper and more available than Nema 14s, I assume there was a time when they were priced according to their capabilities but the economics of mass production seem to have changed that.

In the interests of keeping things simple I also want to start out with a geared extruder and without a heated print bed, as these are things I could fairly easily iterate later on. So, which design to build?

Sells Mendel

Controversy time! In the process of my research I came across this old comment on reddit:

[www.reddit.com]

I’m sure much of it is subjective, debatable or just incorrect. But something about this outlook really speaks to me. Obviously a 12 year old design will have downsides compared to a more modern printer. But this would be the epitome of a “back to basics” approach, allowing me to see where they all came from, what improvements were made, and what, if anything, was lost.

They are notoriously difficult to assemble and square, but I am experienced enough with complex builds that that doesn’t particularly turn me away. If anything, the main downside is the cost.

While many parts of a 3D printer have dramatically reduced in cost over the past decade, fasteners are not one of them. And these make up a significant proportion of the materials cost for the printer, £42 out of around £180 total build cost. Is this just throwing money away relative to a design optimised to use less fasteners?

Huxley

These appeal to me largely because, as I say, I’m not too concerned about print area. Even with the reduced height that comes with using a geared extruder rather than a Bowden. Smaller shapes are generally stronger, and requiring less desk/storage space is always good. Most designs would require some modification to work with Nema 17s, which could be both a plus and a minus. I could build one for near half the cost of a Sells Mendel.

But unfortunately, the documentation seems a bit lacking in comparison. There’s a few different models and it’s not clear which one is being talked about at any given point on the wiki, or what the differences really are. Some use different electronics, some use push vs pull configurations for the Z axis, some of the printed parts seem slightly different. I was only able to find a BOM to use for pricing up by going to the page for the emaker model, on the original page the link is broken.

I’m sure I could piece one together, but given all this I wonder if it might be easier to just modify something else to be smaller if that’s what I want. It’d likely be over-engineered with 8mm rod in a smaller frame, but that’s rarely a bad thing...

Prusa Mendel i2

Sort of the other end of the scale to the Sells, and seemingly about as far as development got before moving to significantly different designs and custom fabricated frames. Drastically fewer fasteners than a Sells, and overall about 2/3rds of the cost. Some of the improvements make a lot of sense, and it’s better documented even than the Sells it seems.

Easier to assemble, less to go wrong, but also less to adjust and fine tune. Obviously this could be fairly easily converted into a Samuel if that’s where I decided to go.

I’ve really gone back and forth a few times on what would be best to build, so I welcome any input! And thanks a lot for reading smiling smiley
VDX
Re: Building a Mendel in 2021. What would you choose?
October 12, 2021 11:21AM
... good luck with your project! smileys with beer

I'm following RepRap since 2007 -- and every now try to "inject" some new/alternative options and ideas, which were partially accepted, but some not so (or waiting).

For now I'm focussed on Laser-applications and some aspects of "laser assisted 3D-fabbing" too, but not so meant for common use, but more for high-tech "niches" aimed on accuracy/resolution and mixed materials ... will still need some years until my "target" will reach maturity state, but actually already near to some partly aspects, like materials, laser-types and -powers smoking smiley


Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]
Re: Building a Mendel in 2021. What would you choose?
October 12, 2021 06:26PM
A machine that moves the bed in Y will not scale up beyond about 300x300 mm. As the bed gets larger it becomes heavier and harder to throw back and forth at print speed. Deltas don't scale up well, either, so stick with a square/cartesian machine of some sort. I suggest that instead of one of the old bed flinger designs, you start with a design that will scale well such as a core XY in which the bed moves in Z or is fixed and the XY mechanism moves in Z. There are some differences between making a large one and making a small one, but the differences are relatively easy to deal with. The XY mechanism scales up easily as does the Z axis. Whatever you learn from the smaller version will be applicable to a larger machine.

If you intend to dismantle the smaller machine to reuse parts for the eventual larger build, don't cheap out on the controller, extruder, or power supply, unless you want to learn about what a PITA it is to deal with crappy parts. Invest in quality up front and you won't have to buy again and again. Use a 24V power supply.

Don't bother using a DC powered bed heater - go right to line power switched by an SSR. 12 or even 24V heaters are usually underpowered and you'll have to spend extra money on a bigger power supply to operate it. You won't be able to use a DC powered bed in a large printer, so don't waste time, money, and effort on a small DC powered bed.

Don't expect 20 mm t-slot that you use to build the frame for a 150x150mm bed machine to be adequate for a 500x500 mm bed machine. You have to scale up the frame members in all dimensions, not just length.

And always remember this: PLA sucks. Never use it to print anything that might get a little warm in a sunny window, near any type of heater (including a stepper motor), or might accidently get left in a hot car. In short, if you care enough about it to bother designing and printing it, it deserves better than PLA.


Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
VDX
Re: Building a Mendel in 2021. What would you choose?
October 12, 2021 07:05PM
... one of the R&D-projects, I was involved in, was started to develop a "all-in-one" machine for CNC, plasma, laser and 3D-printing (my part was the 3D-printing and some aspects of high-power IR-diode-lasers).

The machine-base was designed for 1x1m with 4 combined heat-beds, 500x500mm each with cooling and heating by fluid through bores below the print-bed surface.

This project was progressing until roughly 50%, when the corona lockdown crumbled down the involved companies enough, so two resigned and the project was canceled (I've got most of "my components" for free, so tinkering with them in my basement now).

Don't expect an easy way, when developing an "oversized" 3D-printer -- the related part in the project was calculated for around 150k€ (but too, combined with other aspects, needed for plasma- and laser-welding/cutting) ... if building/developing this on your own, a 500x500mm printbed and the enclosure for lets say 60degC ambient temps will need some thousand USD in components and "experience" (or wrecked/incapable prototypes), until you'll get first results (have a friend in a company with this experience).

But again -- no need to give up! - good luck with your own aims!! smileys with beer


Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]
Re: Building a Mendel in 2021. What would you choose?
October 12, 2021 07:41PM
For clarity, I'm not looking to build a larger machine now (or even re-use parts for one later), just something small to get to grips with. But thanks for the tips regardless! For the sorts of parts I'm thinking about that would need a larger printer, I do have the advantage of only really needing it to be particularly large in one axis rather than all three.
Re: Building a Mendel in 2021. What would you choose?
October 12, 2021 09:00PM
Well... if you want to spend many, many hours building and tweaking... Huxley I found more frustrating.

I2 I ended up with something quite good that could do Abs, if not fast then at least consistent.

Now I have a delta with autolevel, I'm way, way past either of them. Too much faff and not enough print.
VDX
Re: Building a Mendel in 2021. What would you choose?
October 13, 2021 04:43AM
... if for delta - one of my smaller prototypes, presented 2007 in the RepRap builder blogs (a pity, the thread was lost in a server failure or such) was a pretty small delta (building envelope around 50x50x50mm), meant for high precision pick'n'place/microassembly -- it was made from only 12 parts and partially "self-aligning" with the ball-magnets.

Here some of the related images, I could still find - the first sketch and the assembled prototype:





Here a "mature" version (not from me) with more common ball-joints:





And here a "benchmark" for a "one-sheet-delta" (milled from a sheet of PEEK) sitting on a PZT-actuator-arrangement:



Edited 3 time(s). Last edit at 10/13/2021 04:48AM by VDX.


Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]
Re: Building a Mendel in 2021. What would you choose?
October 14, 2021 08:42AM
Quote
ComeWhatThey
I’m aware that there’s been much discussion about the drawbacks of the design and there are good reasons why more modern designs have largely moved away from trapezoidal shapes.

Triangular designs are still the best way to get stiffness. These "good reasons" are mostly simplicity/cheapness. But there's a reason why some of the rectangular designs got nicknamed "Wobblebot". Rectangles have no structural stiffness, they rely on stiff corners and stiff beams.

The Mendel/Huxley design is also one of the few which don't rely on accurately shaped parts. Look at VDX' design shown above: these vertical beams/pillars have to be precisely the same length and cut at exactly 90°, else the tower wouldn't fit together. A Mendel is better here, rods can be a few millimeters longer or shorter, parts still fit. It even works with not-so-straight rods. That said, accurately cut parts are easily made in modern workshops, so most printer designs count on that.

While the initial idea of RepRap was to build replicable printers, it soon became a race for the cheapest design. Don't expect engineering brilliance everywhere :-)


Generation 7 Electronics Teacup Firmware RepRap DIY
     
VDX
Re: Building a Mendel in 2021. What would you choose?
October 14, 2021 10:44AM
... one of my other easier to manufacature/print designs was about "morphing fingers" and "hands", made from 3D-dispensed silicone "bodies" with embedded capialry-networks, supported by pressure/vacuum from an oszillating pressure/vacuum-pump beneath and controlled by a raster of ventiles in a plate --- here a related thread (or search for posts from me with "morphing" and "all dates" set):

[reprap.org]


I was talking about much more complex/sophisticated designs too, in the shape of freshwater-polypes or squids, which can be easily printed/dispensed with silicone or UV-cured polymers ... but the software to precise control this types of "morphing structures" is way more complex than the commonly used cartesian/noncartesian variants ... think more on "reinforced learning AI-software" to control them ...


Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]
Re: Building a Mendel in 2021. What would you choose?
October 15, 2021 09:20PM
I wasn't turned off from earlier designs by all the people (here and elsewhere) who've been telling me it'd be a pain and there's no point. I knew it wouldn't be the best printer and mainly wanted to learn by doing, keep it simple. And the recommendations for i3 style kits, while perfectly valid, didn't exactly inspire me. But looking into what else I could do for very similar money is what has really got me thinking.

It may well be because I was relying on the wiki too much that I was missing some of the options that do follow a similar philosophy to what I was looking for - Voron, BLV, HEVO, Smartrapcore... Hopefully when I'm a little better versed in modern trends I can put aside some time to help clean it up a bit.

Changing my mind a bit on the electronics too, I was a bit overwhelmed with the choice of more modern boards that I thought it was best to keep it simple. But there's lots of stories around of the cheap RAMPS boards being badly made or even failing, so definitely more research to do on that front.

Especially when something like an MKS Robin can be had for around the same money as a full RAMPS kit, which is 32 bit and has integrated TMC2209s. Klipper compatible too. A lot more capable, but still cheap enough that it doesn't feel like I'm committing to it for life.

I have been looking into the possibility of building something inspired by the Voron Legacy, repstrap style. Not exactly to spec, but I think it could be done for a lot cheaper using some of the old reprap hacks. Keep the belt layout and a lot of the 3D parts, use drill rod for the linear shaft, printed PLA LM8UU bushings. Wades/J6 print head. Unheated bed to start out with. Maybe even a wooden frame.

Seems like a good mix of achievable, capable and upgradeable.
Re: Building a Mendel in 2021. What would you choose?
October 16, 2021 10:08AM
It's great that you're going to try to design your own machine, even if it is based on someone else's design. One of your goals was to learn from the process, and you will most certainly do that.

I can make a couple recommendations regarding that design. It appears to use a stacked belt corexy mechanism. Notice that right at the motors the belts bend 90 degrees and at the opposite corners the belts simply loop 180 degrees around the pulleys. If you count up all the pulleys that each belt bends around, you'll find 7 including the drive pulley. That's a LOT of pulleys, each of which adds to the load on the motor. Why not eliminate the bend in the belt right at the motor mount and simply put the motor at the opposite corner where the belt is making a 180 degree bend around that pulley anyway? That reduces the pulley count to 5. 5 pulleys vs 7, what's the big deal? If you ever get a chance to manually push a corexy mechanism around by hand, try moving the extruder carriage in the Y direction. You'll find it pretty easy. Now try moving it in X. It is much harder because that motion requires the belts to move at all the pulleys. That difference represents additional load on the motors, and will ultimately be one of the limits to speed/jerk/acceleration of the mechanism. Adding 4 more pulleys will have an effect.

The front of the frame of that design is open - it has no cross member between the two end pulleys. That will lead to a less rigid structure and problems enclosing the machine. There's no good reason to do that.

The belts/motors are all inside the frame in X and Y but stand up over it in Z. That will make it easy to put side panels on the machine, but will leave it difficult to fully enclose. How are you going to cover the top with the mechanism standing up like that? I suggest building so the vertical members of the frame stand up well above the XY mechanism to allow a flat panel to be added to the top to fully enclose the machine.

It looks like the belts are tensioned by sliding the motors a few mm in the slots in the motor mount (I'm going off the few photos here), but if the motors are mounted at the two end pulleys as I suggested above, belt tensioning can be accomplished simply by sliding the whole motor mount along the frame member and tightening it down. I built a couple sand table mechanisms that incorporate some of the ideas suggested above. You can see the mechanism in this video. I used the t-slot as the Y axis guide rails with sliding PTFE bearings, but I would not recommend doing that for a 3D printer. Stick with round rails or linear guides for the high accuracy and precision that a 3D printer requires. The point is that the mechanism uses the minimal number of pulleys, has no projections outside the XY fame, the vertical corner members could be made tall to allow full enclosure in a 3D printer built this way, and belt tension is easily set by sliding the motor mounts along the t-slot to which they are bolted without the added load of extra pulleys.

The pretty covers on the extruder are nice, but you'll find, especially if you go cheap on the extruder, that those covers are going to be in the way most of the time. The extruder is the least reliable part of the machine and you're likely to have to work on it frequently. The covers make it look nice, but add pointless mass to the extruder carriage and get in the way when it's time to clear a clog out of the hot-end, etc.

I can only see two screws used to lift the Z axis, which is fine, but hopefully they are driven by a single motor. If there's a motor on each screw, you'll have to keep releveling the bed because the motors will jump a bit each time power is cycled. That will lead to the bed tilting and you'll have to do something to untilt it. Using a single motor driving both screws will keep them permanently in sync so you'll never have to tweak that part of the bed leveling. If you want to make it super reliable, use a flat plate for the bed (cast aluminum tooling plate) and put it on a 3 point kinematic mount. You will not need autoleveling at all because it will stay level after the initial, simple, manual setup.

Don't even think about using PLA to print the motor mounts or extruder carriage. Use ABS or PETG. PLA does not belong anywhere near anything that gets warm or hot.

Edited 3 time(s). Last edit at 10/16/2021 10:17AM by the_digital_dentist.


Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Building a Mendel in 2021. What would you choose?
October 16, 2021 06:07PM
Yeah thanks for that! I can't say I know too much about belt layouts, I guess other corexy printers do seem to do it the way you describe? Looking at Hypercube/Smartrapcore. I did wonder why it uses pairs of flanged bearings too, instead of toothless idler pulleys. But those wouldn't fit within the printed parts as they're designed.

The Legacy actually does have a front cross member, it's just moved down slightly to make it level with the print bed in the zero position. I had assumed this was done to avoid interference or have better airflow or something, but looking closer I guess it's so that you get a better view of the print and/or can take cool videos. The 0 doesn't have one at all, but clearly it's small enough that it doesn't particularly matter - at the speeds those things print at, any instability would definitely have been picked up. And the 2.4 is a standard cube, but then it has the entire XY mechanism on a gantry and the print bed is just static.

As I understand it the Legacy isn't specifically designed to be enclosed, but there's a few ways it could be done. Could add a "top hat" style enclosure or put it in something larger. Either way there's no real need for it to be structural, so it could be added later. I will consider lengthening the frame but will have to see if it works out, from what I can see the cheapest 2020 I can get is in packs of 1m lengths.

Thanks for that vid, 40s in is a highlight winking smiley

I certainly wasn't planning on using the afterburner print head, for cost reasons as much as anything.It actually does have two motors for the Y axis, but there's an inductive proximity sensor on the print head that lets you run a macro in firmware to sort out the tilt for you. You can see it here at 20 seconds. Over complicated? Possibly. But again, these machines seem to be designed for speed.

And yes, we have a roll of PETG on the way which should hopefully do most of the printed parts, don't worry about that spinning smiley sticking its tongue out
Re: Building a Mendel in 2021. What would you choose?
October 16, 2021 06:35PM
Flanged bearings are used because the toothless pulleys sold for 3D printers have bearings that are too tiny and will wear out very quickly. Flanged bearings are better because the balls in the bearings are much larger and should last a lot longer. This is the kind of small stuff that you don't have to learn by doing - you can rely on other peoples' experience and save yourself a lot of trouble, time, and some money.

Belt layout is critical for proper operation of the mechanism! The #1 rule is that the belt has to be parallel to the movement it is creating, exactly the same as a screw drive.

The autoleveling stuff makes for cool youtube videos, but it adds a lot of unnecessary complication to the machine. It also means that if there's any problem with the sensor, or the drivers, or the cables, you won't be able to print. Reliability comes from simplicity, not complication! Building a mechanism that stays level is not difficult if you understand how it works.


Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Building a Mendel in 2021. What would you choose?
October 18, 2021 11:24PM
Quote
I can make a couple recommendations regarding that design. It appears to use a stacked belt corexy mechanism. Notice that right at the motors the belts bend 90 degrees and at the opposite corners the belts simply loop 180 degrees around the pulleys. If you count up all the pulleys that each belt bends around, you'll find 7 including the drive pulley. That's a LOT of pulleys, each of which adds to the load on the motor. Why not eliminate the bend in the belt right at the motor mount and simply put the motor at the opposite corner where the belt is making a 180 degree bend around that pulley anyway? That reduces the pulley count to 5. 5 pulleys vs 7, what's the big deal? If you ever get a chance to manually push a corexy mechanism around by hand, try moving the extruder carriage in the Y direction. You'll find it pretty easy. Now try moving it in X. It is much harder because that motion requires the belts to move at all the pulleys. That difference represents additional load on the motors, and will ultimately be one of the limits to speed/jerk/acceleration of the mechanism. Adding 4 more pulleys will have an effect.

So I asked around about this and it seems the motor placement on the Voron serves two functions. Firstly it removes the need to have a twist in the belt to avoid the teeth contacting one of the pulleys. And secondly it helps to maximise the available space in the frame by having all of the things that take up space within the frame on the same side of the printer - motors, electronics, power supply. I did ask why those things couldn't just be moved to the other side, and you'd turn the printer 180 degrees to get an unobstructed view of the print, but haven't recieved an answer.

There may well be other reasons, possibly related to parts of the design being borrowed from the other printers in the range, but I think you're right in that the layout you described can be both simpler and cheaper.
Re: Building a Mendel in 2021. What would you choose?
October 19, 2021 08:07AM
Twists in the belts are only needed if the pulleys are small or you want quiet operation at very high speeds. The stacked F625s used in that design are still a little small to have the belt teeth hitting them- they could cause print quality problems (as well as making zipping noise at high speed). Gates recommends a minimum of 9 teeth in contact with a smooth pulley surface, and since the belts are wrapped at 90 degrees on the pulleys, that would be a pulley diameter equivalent of 36 teeth. Belt pitch is 2 mm (a little less when bent around a pulley), so pulley circumference of 72 mm (approximately). 72/ pi = 23 mm (approximately) in diameter. F608 bearings are 22 mm in diameter (close enough!). I used stacked F608 bearings and 9mm wide belts in my corexy machine and prints are great.



This is the minimal pulley count belt layout for corexy::



I used larger pulleys in the corners to provide extra clearance for the belt when the X axis is near the corner pulleys. It isn't strictly necessary. The most important thing about this diagram is that belt segments labeled A-H must all be positioned parallel to the guide rails (in every corexy mechanism, no exceptions!). Otherwise belt tension will vary as the extruder carriage moves and prints will be distorted. There are a lot of bad designs out there! If you see a clothespin spring belt tensioner, beware! The layout is probably wrong.

See: [drmrehorst.blogspot.com]

Edited 5 time(s). Last edit at 10/19/2021 11:12AM by the_digital_dentist.


Son of MegaMax 3D printer: [www.instructables.com]
Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Building a Mendel in 2021. What would you choose?
October 21, 2021 01:21AM
It will be one of your life's great joys to design and build your own machine and I hope that you decide to do it.

One fun thing is getting to decide which parts you'll focus your attention on. Since there are existing designs which work well, you can even decide to focus your attention on nothing in particular except following someone else's instructions, like I did when I built my Mendel90, and you'll still get a working machine to be proud of. The standard good DIY design looks like a Prusa or is like an LM8UU/CoreXY belts/Off-the-shelf hotend/cantilever falling bed design.

But if you decide to let your mind come up with ways to change some part of the printer, there are many temporarily-adoptable interests for you to find before and during the construction of it. Many aspects could be theoretically improved on if you decided to try something new, which I think is one of the best parts. I'll give you two examples. DigitalDentist here built a kinematic mount for his heated bed to allow it to be removed and put back on in the same spot every time (the style of mount he chose is repeatable to nanometers), and his design also makes the bed move during thermal expansion in a predictable line. That's an aspect of the heated bed aspect that was improved on by someone who took a temporary interest in it for a few months and he came up with something. Another place to improve the printer is the Z-Axis. I posted about a very repeatable and simple-to-build Z-Axis mechanism (that uses the same theory DigitalDentist used) that is both more accurate, and scales to larger and heavier bed designs than the standard cantilever Z-Axis (here). But that's two of like 10 different aspects of a 3D printer.
Re: Building a Mendel in 2021. What would you choose?
October 26, 2021 12:05PM
If you are building a CoreXY, you might want to consider the MaybeCube. It's a linear rails design that supports a high level of customisation. It's also fully enclosable. I took a lot of steps to keep the costs down (for example using standard extrusion lengths, so no custom cutting required, hidden joints are used as they are the most rigid (and self-align) and also keep the fastener count down). It also addresses most of the points that the_digital_dentist raises, since I read most of his blog while I was designing it (it does however use a 7 pulley CoreXY design, since using a 5 pulley variant would have significantly reduced the build volume for a given extrusion length).



Or if you are looking for something smaller, you might want to consider the BabyCube

Re: Building a Mendel in 2021. What would you choose?
November 10, 2021 01:48PM
Quote
MartinJB
If you are building a CoreXY, you might want to consider the MaybeCube

I really like that design, thanks for sharing! Do you have any issues with the unsupported X rail?


Help improve the RepRap wiki!
Just click "Edit" in the top-right corner of the page and start typing.
Anyone can edit the wiki!
Re: Building a Mendel in 2021. What would you choose?
November 21, 2021 11:03AM
I did originally have provision in my design for X-axis support, but have found it unnecessary. X-axis flex is low because:

1. Torsion about the X axis is low because of the close pulley spacing on the Y-carriages
2. The forces on the X axis generated by the CoreXY system are generally in directions that don't flex the X-axis (apart from the torsion mentioned above)
3. MGN rails are hardened steel and very rigid
4. The frame is extremely rigid, so the X-axis is not subject to frame flex.
Sorry, only registered users may post in this forum.

Click here to login