Back in May of 2020 I declared that the new big printer I designed was done. Well, it was mostly done. From the initial design, and build out it was, but there were a number of small design issues I just pushed forward with, with the idea I would revisit them down the road. Since my initial design in 2019, a lot has changed.
- The parts for the printer were all printed on my heavily modified Anet A8 with a Titan Aero extruder. This meant they were printed in pretty variable environmental conditions for especially finicky filaments. The accuracy of that printer was good but not nearly what the new printer is.
- My knowledge of slicing 3D files and best print settings for various filaments has evolved over the last 2 years. I’ve really dialed in the right settings for some CF and Glass loaded Nylons, meaning I can print lighter parts that are stronger.
- My knowledge of 3D design has evolved and gotten progressively better. I was able to go back into my old designs, clean them up, make them easier to modify, make them symmetrical if they weren’t and should have been, tighten tolerances, etc…
- New Extruders and technology coming that requires new mounts for the X and U axis. The X axis carriages for the Extruders were heavily modified from some draft designs. I need new clean sheet designs going forward.
- Visually misaligned belt paths, and other “rushed” work has lead to bearings failing faster than expected in the drivelines. This caused me to keep working back as I was going to design parts that ended up with me redesigning quite a bit of the X axis.
Misalignments and rebuilding the entire X Axis.
When I originally designed the X Axis, I worked around some other fairly clumsy design aspects and spacing of the belts. Using a 16T Pulley on the stepper means the belt spacing is so tight under the MGN linear shuttle the extruders ride on, that the heads of 3mm screws may interfere with one or two of the belts. My solution before was to have the front set of belts higher than the rear. This also gave more spacing for the Titan Aero on one of the mounts.
Well this meant the adjusters on both sides of the X axis needed to be unique per side, and it caused some other odd interference on the Y Axis hubs and the Hall sensors. I wanted everything to be symmetrical now, so only one adjuster was needed that could be used on either side, and to keep the hole spacing in the X/U axis mounts consistent. Below are the design changes I went through.
The Left and Right (E0/E1) Extruder mounts are clean sheet designs. I used the old ones for some guidance, but otherwise every single dimension is changed. They are simpler, stronger, lighter, and more compact in key ways. The hole spacing and sizing is easier to print, and not as finicky. Also now that I have the design in all its build up steps I can iterate it for other extruders easily.
Final E0/E1 design of what I am calling the V3 changes to the printer:
Main changes for the E0 and E1 mounts:
- 2mm thicker motor mount plate, with adjustable motor mount holes to help dial in a perfect consistent head height to the table from E0 and E1. 1.5 mm of total adjustment on each side.
- Wider but thinner top mount (1mm thinner). Wider for better stability and connection to the side reinforcements
- Single central belt duct, aligned to the natural angle of the belts due to the 20T Idler and the 16T Pulley. Sized to keep it as small as possible but not interfere with the belts
- Belt spacing and height optimized for consistency from each hot end.
- Form fitting bolt holes to mount to the MGN12 shuttles, the old ones were ovals to use small and long MGN. Since I am just using Long style ones I can get more rigidity from using tighter mounting holes.
- Better and more exact tolerances for all the fittings and screws
- Stronger side reinforcement, fewer holes and tighter design
- Optimized hot end wire routing
- Improved belt locking setup
- Printed in new/tested batches of 3DXtech CarbonX PA6 filament. Including the fan duct, Hall sensor mounts, and fan nozzle – those used to be PETG on the past iteration.
- Optimized the size, angle, and mount for the part cooling nozzle.
- Optimized print settings – finer layers, lower infill with higher infill extrusion, and one more wall layer. Also printing hotter for better adhesion in a preheated chamber.
One unexpected improvement turned out to be weight! I wasn’t expecting a full 20gm drop in weight from each extruder mount. 40gm from moving inertia total on the Y-axis is not bad. These mounts are also very strong, I tried to test one to failure and needed a pliers to do it the V2 model I was able to break in my hand.
X Axis Belt Tensioner update
With the original design, the left and right belt tensioners were unique in their alignment. The one mounted to the left side of the X Axis was higher and spaced further out, and the one on the right was lower and more centered. For the last 2 years the one on the right has never been properly aligned and I think was the cause of multiple bearings failing and some unusual ringing and artifacts in the print from the E0 extruder (minor but noticeable on PETG).
I also incorporated some new style of shoulder screws that are normally used in foldable knives. This lowered weight, but also made the design easier and allows me to tighten the tensioner but not put any side loading on the pully bearings. You can find these style of screws here.
I printed the tensioners in CarbonX as well which let me print them lighter and stronger as well.
Once the parts were printed, test fit with as much of hardware as I could I was comfortable taking apart the printer. I only own one printer, so if I screwed up on a critical dimension with the belt, I would take apart my printer only need to put the old parts back on. So this was a leap of faith on my evolved design concepts and processes.
Below is the swap over work over about 5 hours. Note – parts are blurred because I am beta testing some hardware for a company and cant share images of the parts yet. It was a pain to blur these out so I would rather not have had to do that.
The final look of things is quite nice:
From some of the other renderings and pics above, you can see that there are cut outs for the Titan Hob gear, and the BL Touch on E0 to avoid contacting E1. This lets them get as close as possible together with no interference. E1 in this design is also moved forward about 3.5mm from the past setup. This lets me make the mounts stronger, and lets me get more usable workspace for mirror printing.
The change in wiring cutouts and ziptie locations make running wires much easier.
Was the change worth it?
Yes! these changes resulted in belts run centered, level, and symmetrical from E0 and E1. Besides for it for scratching that itch of knowing something wasn’t designed well initially. Print quality went up immediately. Ringing on the X axis is less than it was before at high speed (a testament to the better belt layout), and almost no ringing on the Y axis (a testament to the much stronger and stable mounts). I don’t have before/after pics of tests, but from some calibration cube prints in PETG it was clearly awesome.
It took about 2 weeks of dabbling in redesign due to the incredibly tight tolerances, but everything worked in the end. An Example: the E0 Tool (left side) has clearance for the front most belt by .2mm vertically and .5mm of the face of the tool! Knowing my design assumptions were that accurate makes me very happy.
Passive Chamber Heating RepRap Macro
One of the tasks I have been pondering for a long time, was using the 700W heated bed to heat the chamber.
Problem: In the winter my office where the printer lives is cooler than in the summer, but regardless for temperamental filament like PC or Nylons, I would like the chamber to be above 25C, 30C+ if possible.
Solution: Before I would just set it to 100C and leave it for 45 min or so and fired up the print job later. But I wanted a more orchestrated way to do it, and also have fans blowing over the build surface to help circulate the air and transfer heat more effectively to the upper parts of the volume. I was going to add more fans, but then realized I could just use the part fans. This works because I also made the part fan duct/nozzle out of very high temp CarbonX so sitting over a 100C or 110C build surface wont deform it.
I came up with a Macro in RepRapFirmware 3.2 on my Duet 2 that will be called in certain job scripts from the slicer and not only run the preheat, but then also drop the build table back down to a reasonable temperature before the print starts.
The Macro below does the following key steps.
- Moves the heads to the back sides of the build plate and within 5mm of the surface, so the part fans blast air across the build surface.
- Spins up the part fans on both extruders to 100%
- Raises the bed temp to 100C
- Then waits to the existing DHT22 sensor which is a few inches above the level of the build surface to hit 27C . The latent heat in the build plate will carry the chamber to 28 or 29C.
- Power down the build plate and let the temp get back to 80C
- Shutdown the parts fans and start the print.
;Set Chamber Heat to 28+C ;Find where we are G28 ; home all axes ;Unload any tools T-1 P0 ;Move Heads into position ;Turn Part fans to 100 M106 P0 S255 M106 P1 S255 ;Turn the Heated Bed to 100C M140 S100 ;Disable Steppers M18 M291 P"Heating Started" T2 S1 ;Wait for chamber temp to come up to 28C while sensors.analog.lastReading < 27 G4 S1 ;Echo Report its done M291 P"Heating Complete Cooling the bed to 80C" T2 S1 ;Post Heating Process: ;Powerdown the bed M140 S0 ;l while sensors.analog.lastReading > 80 G4 S1 ;reset the fans M106 P0 S0 M106 P1 S0
This took what would be a 45 – 60 minute warm up is now a 15 – 20 minute and uses all the hardware already in the printer. The macro leverages objects in the logic feature of RRF 3.1.1+ to handle the loops. The native M commands in RRF for waiting on heat up are not functional under 40C for some random reason, so objects and logic needs to be used for loops.
These updates fixed a lot of the minor issues I had with the printer, and now it prints better than ever!
Next steps are to design a E3D Hemera mount for E1. Hemera is not an easy extruder to mount from the top but now that I have accurate designs from scratch for the mount on the E1 position, I should be able to design a functional mount. The goal is to get the Hemera mounted in a way that it is very stable but also very compact, which is a great challenge in the E1 position because the wires leave vertically right about where the back belt runs.
I have the Hemera in hand and am going to start working on mount designs coming up. The E0 position will stay a Titan Aero as that extruder offers some better flexibility to use 2.85mm filament, and Hemera cant be installed in a mirror mode currently.