Making a 3D Printer part 2
Assembling and adapting as we go
In part 1 I outlined why/what/how I was going to build a 3d printer. I’ve been working on it slowly for many months but I have finally got to phase 2, codenamed “just make the damn thing go ffs”.
The extruder and hot end
I had planned on using a dual bowden head on the Chimera but when I realised that the logic board only had 5 stepper controllers (MKS 1.4) this made me think long and hard about it and I came to the conclusion that I have no need for dual-headed printing… so there’s no point doing it.
It became obvious that the main point of this new printer was to move plastic like no-one’s business, so I scrapped the dual Chimera head and went for an E3D volcano hot end running a 0.6mm nozzle.
Then, all of a sudden there was a youtube video where e3d showed off their new design…
For a number of reasons this was what I’d been looking for, so I held off. Until the Hemera was released.
Why the change from a bowden tube to a direct drive? I tested the bowden tubes a lot and I just didn’t like the outcome.
Retractions and hops weren’t as predictable and flex filaments, especially the really soft TPU I have simply doesn’t work well… Considering my main reason for going with the bowden was for speed and the new nozzle size essentially means I am going to print slower, it just seemed to not matter anymore.
The build : The frame
All of the frame was made of 3030 t-slot which is very common in 3d printers. A lot of people online are using 2020 t-slot but 3030 is more rigid and this printer is going to be a bit big and move fast so longer sections will flex more than I’d like if I used 2020.
After toing and froing quite a bit, I settled on a frame size of 530mm wide, 500mm deep and 600mm high, this was after a great deal of printing trial parts and guessing about clearances.
The core size I was working to was the 400mm print-bed, but I will only be using 360x360mm of it. I tend to avoid the 10–15mm edge of the bed normally so this doesn’t worry me.
About max Height
Because I want to enclose the top of the printer to retain heat, there will need to be a big gap at the top of the printer for the wiring to reach the print head without it being obstructed by the top of the case, this will probably be done with an extension to the top of the printer rather than within the frame, just so I can squeeze as many MM out of the Z axis as possible.
Parts list
- 4 x 600mm t-slot for the vertical corner rails (Z)
- 4x 500mm t-slot for width rails (X)
- 6x 470mm t-slot for the depth rails (Y)
- 8x 3030 tslot corner mounts
- a pile of 90 degree 3030 tslot corners with m6 threads
- nuts and bolts
Next — CoreXY Movement
The CoreXY setup is a bit interesting, it uses 2 stepper motors just like other Cartesian setups but they have a very different layout and they move weird… The two steppers work in unison to achieve the same motion but in a completely different way.
In the drawings below I’ve illustrated the motors as dark blue and the belts as orange and green. The direction of the green and orange belts are illustrated in yellow and blue on the outside, the diagonal blue/yellow arrows show the direction that the belt pulls the carriage, the purple line shows the direction of movement of the carriage.
Why use CoreXY?
Stepper motors are a little bit heavy and removing this weight from the carriage reduces the start/stop acceleration “ringing” (explanation here) you get from changing the motion of the stepper motor mass.
Stepper motors also generate quite a bit of heat and don’t operate well when they are really hot. With the CoreXY layout, the stepper motors can be outside of the heated chamber — that I plan to build.
Le problem? it’s all 6mm belts
All of the CoreXY designs I found use 6mm GT2 Belts — these belts are fibreglass backed and have quite a lot of stretch in them, this stretch is more dramatic on a CoreXY because the belts are a lot longer. I decided to use 10mm belts to get around this stretch problem but it meant that I had to hack/design my own parts.
I wasn’t sure where to start so I used the parts from this thingiverse project as a base for the XY arrangement to design my own parts.
Extruder carriage
The extruder carriage has some particular challenges. It needs to be light, rigid and it needs to have all the extruder bits involved. More over all of the parts and indeed the wiring need to be a little adjustable and I need to be able to remove and replace parts within a reasonable amount of time. To get maximum print area I decided on a tall/thin carriage shape.
I started this build while the e3d Hemera was being developed and I got one of the first to land in Australia. This changes things!
Components of the extruder carriage:
- End-stop microswitch (X axis)
- E3D Hemera
- E3D Volcano hot end
- BL-Touch probe and bracket
- 24v blower fan and ducting to the hot end
Unfortunately I couldn’t find a match for my requirements anywhere so I designed my own and printed them on my trusty Wombot Drafter.
Multi-part
Most printer carriages I have seen are all in one piece, this has some weight benefits but make it a bit harder to work on. Because of all the moving parts and adjusters all over the place I figured I’d split up this carriage.
Main Carriage
The main carriage is where the belts mount and where everything that makes it a printer lives. It’s got some long bearings and a heap of bolt holes!
Tensioners
The CoreXY movement requires the belts to be the same length and tension, if they’re not, the motion skews in favour of the tightener/shorter side, this leads to squares being parallelograms or worse; loud banging and the sounds of the printer breaking….
To put tensioners on a standard movement is easy, you put a spring retainer on them:
This doesn’t work with the CoreXY because the belt moves further than the longest distance between any of the bearings, the only place that made sense to me was where the belts joined the printer, on the carriage head.
I put some rivnuts in there with a smooth-edged block and used an m3 screw as a tensioner on each side, this means that I can adjust both belts on both sides using varying lengths of m3 screw if required. There’s only about 10–15mm of adjustment here which should be plenty but it’s going to be annoying regardless
The tensioners can be tightened from both ends of both of the belts.
Extruder Plate
The extruder plate is where everything bolts in that relates to melting plastic. This went through about 15 iterations and a heap more that didn’t make it to the printing out phase.
BL-Touch Mount
The extruder starts with the BL-Touch mount (which had a lot of tweaking that I’m not showing):
Cooling Duct
The most interesting part of this for me was making the part cooling duct. For a long time my prints have been cooled from one side, I can tell this especially in PETG prints as they are more shiny on one side than the other — this also means that the duct wasn’t perfectly focused which is my bad. I wanted to make the ducting come from two sides in a “sharp” formation (more information about cooling ducting types here).
The point of the cooling duct is to direct air onto the part without cooling the hot-end or the heated bed. After some experimentation I developed a slightly adjustable version that looks like this:
Cabling bits
The wiring for all the parts on the carriage needs to mount somewhere, for now I will just mount the wiring at the top and hope it doesn’t collide with anything. I made a simple bracket that slotted into the space where the X-microswitch goes and mounts on panel mount.
In future I plan on using a drag chain which is a rather nifty piece of kit that is designed for this purpose, it will bolt to both the extruder plate and the frame:
Ta dah! the extruder carriage
It looks a little messy but I’m a little bit proud anyways.
Heat bed carriage and Z-Axis
Considering the size of the print area, I thought it wise to use two z-axis motors, there’s plenty of people who complain that the steppers become out of alignment easily but I suspect a lot of those people move their Z steppers by hand so the way to fix this is to join them with a belt!
For the drive, I’m using 4x440mm 8mm rails and dual 8mm lead screws, mounted to the frame with metal mounts I procured on Ebay.
Because of the size of the linear bearings and the height of the extruder and carriage, I am limited to about 360mm build height which is plenty big enough.
I hard mounted some lead screws to brackets and the brackets to the frame, this means I can drive the vertical lift from two steppers with belts and the weight is held by the shaft brackets.
I made up a frame out of 3030 t-slot and made some brackets for the print-bed frame and assembled it all together.
Then added the heatbed and wiring.
Wiring
Most of the wiring had all been cut to the correct lengths for not-my-printer, which ended up being varying amounts of not-quite-right for this printer but I made due with what I had and extended the wires that weren’t long enough.
Landing a big fish — marlin setup
Getting Marlin setup for this printer was a learning experience, there was no how-to for this printer because all of the dimensions and therefore settings were of my own making.
What’s great about Marlin is that the config files were very well documented so getting it all to work was pretty easy.
Finishing up
Early test print!
I got excited when I had everything together and I just hit print and ended up with some pretty messed up prints.
It was at this point that I began to fully comprehend the complexity of molten plastic and just how much I have relied on the sensible defaults of all those settings in Simplify3D of course this printer is totally new and very new to me. After a some time I managed to get the settings better aligned.
Making something nice
What most people don’t know about 3D printer firmware and tuning is that it’s hard. This is because the manufacturer does it for you. Have you ever looked at something and thought “there’s not enough materials in that to justify it’s price”? Sure some things are just overpriced but a lot of things take a great deal of time, patience and skill to get them right.
After about 5 hours of testing I managed to get PLA dialled in for this printer. But it didn’t quite look right…
But then I really tuned it with some really new PPLA from www.aurarum.com.au this has to print a little hotter and similar to PETG it needs some relaxed pressure (normally PLA can just be mashed on there any which way). This is a test spool so the colour isn’t perfect but the material is very good.
The relaxed pressure means any wobbles are made more obvious.
The left is Spike3d, the right is my Wombot Drafter, with equivalent settings — the smaller printer has a lower-power heater so it needs a slightly higher temperature..
That wobble is what happens when the belts aren’t the right tension.
Final verdict
CoreXY: I don’t like it
After all this fun and interesting creative engineering I don’t like the CoreXY movement. The main reason I don’t like it is because the belts have to be very tight to get an accurate movement.
The tension on the printed plastics means that either the plastic or the belts will stretch eventually.
Hemera is awesome!
I put a lot of time getting the Hemera extruder working for this printer and it really payed off — it prints flexible filament beautifully and fast.
It’s smaller than I expected...
The print bed is 400x400mm, but because of the frame size and the carriage, I’m only able to use 330x360 of that… this means I’ll need all
Next steps
I’m still going to go forward with most of the parts and the build envelope but I will move to a more-standard cartesian movement. I’m entertaining something similar to the Ultimaker movement which has some of the benefits of the CoreXY but without so much tension.
I still need to add in a more powerful controller, case sides and some internal lighting.
