I just had a crazy idea that I am sure someone thought of before me.
But I haven’t heard anyone say it?
Make a table with a tilt from left to right, and from front to back.
Tilt could be controlled with lead screws (like the z-axis on the print bed)
Then you could
TILT THE WHOLE PRINTER as you were printing.
The layers would still all be stacked on top of each other, but now you could have different sections of the print (at different layer height ranges) pick the average z-axis tilts to minimize the maximum overhangs. I bet the slicer could be made to control the two tilt angles as part of the print job, and auto calculate the ideal printer tilts for each layer.
Now everyone tell me what an old tired idea that is and how I am really lame and “captain obvious”
Yeah, but since the whole idea would be to reduce the average overhang angles of the last few printed layers as z-printing increases, I think that overall sag risk would be reduced. We usually have chamber temps much lower than HDT temps.
That is cool, even more sophisticated than the concept I was envisioning, since the planes of the layers changes with respect to other layers and the print bed. I think that solution is more advanced and would require other thoughts about the slicer / toolhead-print interference etc..
I was recently trying to print a part that has a section with 70+ degree overhangs and the surface quality sucks. And supports don’t help the surface quality much. I was sitting there and thinking I could just stick a stack of books under one side of the printer and let it print tilted 25 degrees to one side. I think that would work FOR THIS PART.
I know my idea is way more simpleminded than the video you show. But it wouldn’t require drastic changes to the slicer. Heck I think I could even add manual G-Code pauses at the beginnings of certain layers and then adjust the tilt of the print myself. So simple… And it would work with just about any printer I can think of (of course the spools might fall off of external hooks, and the lid would need to be taped on. but the rest should “just work” ?
It is an interesting concept. It can solve a lot of problems. The big problem is the user interface, or UI. Some problems would be worse, some better. The complexity of the machine, increases machine cost. And the result is complex. I have been following several that are moving this forward. Can change strength, and surface finish. Sometimes better, sometimes worse.
yes, but like what @EnoTheThracian posted above, that is making the tool-head - build plate change relative orientation. That involves a whole new layer of slicing and G-Code complexity.
What I am suggesting is a little different. Still a 2.5d print process.
Just that the vector for gravity can be adjusted now and then.
The z-layers in my idea are all still parallel planes, and the head only moves in x-y and then incrementally in z, and z is still always perpendicular to the X/Y plane.
Just tilt the whole printer for certain ranges of layers to average the overhangs to the least average.
In my idea you could do the same regular slicing you do now. Then select 2mm slices of layers in the model calculate the average vector to have least overhang and insert a tilt G-code for the printer.
Then print another 2mm and adjust.. etc.. (repeat, maybe smooth)
It wouldn’t need to be tiltable more than 30 degrees.
Did you try dissimilar support interfaces? PETG for PLA and vice versa? Or BVOH/PVA? They don’t do proper miracles, but they are good for the odd wonder or two .
Depending on the part (single sided overhang, no overhangs exceeding ~45°-[tilt angle] on the opposite side), this may just work. Might be worth giving it a try for a simple shape. Just gotta remember to calibrate the printer after tilting it and making sure it sits tight and nothing falls off.
A support free max-angle arch printed on a tilted printer may actually be an interesting Community Contribution.
Mounting a printer on a controlled turn-table on a 45° ramp would probably enable quite a lot to be printed support free.
Key caveats are of course the massive inertia of the printer and the need to slave the turntable to the print head. Not something I’d expect to be achievable with a Bambu since they are quite “locked-in”, but the principle sounds interesting.
Personally, I’d love to see an add-on to the H2: A (smaller) heatbed and plate mounted on a turntable itself mounted on a rotary axis => True 5-axis printing
I very much doubt that this is high on Bambu’s list of priorities though
The User Interface would be complicated. The main advantage (as I understand what you are describing) would be Support would be very different. You would still need to build on something, from some angle. But the requirements would be very different.
If I understand you correctly, you are describing non-planar printing.
This is achieved by rotating the nozzle or build plate at angle in the x/y with z axis changes as well.
Lots of work has been done on this with many working prototypes in the wild.
The main limitation to widespread adoption though is the print head size and the need for the nozzle to be close to it due to the heating, cooling, and other sensors. This makes the head the part that crashes into the models during printing.
Without a means to allow the nozzle to significantly protrude with heating (and its dissipation) and part cooling systems shrunk, the mass market appeal is small.
The hardest part of this I believe is the algorithm to calculate the tilt and steer the nozzle to not crash - co planar printing, to me, seems like an algorithm problem more than a mechanical one.
Given this, is there a benefit to table tilting Vs multi axis motion components? You also need a proper big machine to upend an H2D, and the thing is big enough already.
My idea is dumb. Think stupider.
Everyone is thinking 5 axis extruder head (or xy extruder and 3 axis build plate).
I think people kind of understand what I am saying…
Tilt the whole table, with a regular printer on top of it.
And not quickly tilt back and forth when the printhead gets to a section of a layer where there is an overhead, and then move back… because I agree the printers are heavy and the head moves around the perimeter Fast. It would be dangerous to be near the printer if it was swinging around like a washing machine with a bowling ball in it.
I just mean that right before each layer is printed, the slicer could calculate the average vector (z+ direction) of the walls being printed in the next 2-3 layers and tilt the printer to match that vector.
So the printer would only move a small amount for each layer of the print (very slow)
But, let’s say you had a cube with the left face leaning outward 80 degrees and the back wall leaning out 60. and you put your printer on the table shown below and clamped it on.
Now the back wall might only be leaning outward 40 and the front wall is also leaning out 20 and the left wall is now leaning out only 45 and the right wall is leaning out 35. See I averaged all the overhangs so that the max overhang before was 80 and the max overhang now is only 45. The printer would be tilting toward you 20 and to the right 35.
You think the problem is easier to solve than it is.
I gave both examples, top down (print head tilts, bottom up (build plate tilts), the table in your example.
Either way you have the identical issue, a cuboid with a pointy bit slightly poking out off the bottom must ,arty with the model being printed.
Any angular parts on the model would necessitate a considerable tilt degree, lets call it 45° for the example.
Now mentally place that cuboid with the nozzle slightly poking out the bottom at the tip of the printed portion if the model. The bottom of the cuboid forms an angle to the nozzle of around 75 to 80°, this cannot touch the rest of the model where the cuboid dimensions intersect with it.
The reason the nozzle only just pokes out is due to the still required part cooling fan.
Your idea isn’t dumb, it has been achieved, albeit, it can’t achieve many models other than things that appear like gently slopping landscapes because of the restrictions.
Actually, the software part isn’t that difficult, nor is the hardware, both have been designed and there are ,any examples from many different people.
The issue as I have stated is the location of the cuboid, the nozzle slightly poking out the bottom relative to the model being printed. Anything with geometry that doesn’t smoothly transition in all axises fails because the print head with the nozzle poking out hits the taller parts already printed.
I think you are still missing what the OP is suggesting. Don’t tilt the build plate, dont tilt the head. Tilt the bench that the printer is sitting on. Not saying it will work (gravity’s a ■■■■■).
.
.
& 
