does anyone got any new concerning this function ? because that is a huge + when printing, having a tiny nozzle for detail, and a big one for speed, lot of talk since march, but now we are in august, can we get an ETA about it ?
4 Likes
Pretty sure they are working on it, but how to actually apply it is very challenging and dynamic. E.g., in which case it should use the second nozzle?
are you asking me ?
i can think of a lot of scenario, like i said in the previous post, the most obvious reason that anyone want to have is to print faster, you get to print the whole structure of your print that doesnât require lot of detail with a 0.6 or 0.8, like the inside of the print too, then you go to 0.4 or 0.2 for finest detail like text or stuff where a big nozzle isnât working.
another scenario is for strengh, why do multiple pass for a infill density where you can do fewer but with a bigger nozzle ? you donât care about the look of it anyway since its inside.
same goes for support, quicker to have one big pass instead of losing time going up slowly
They should try and figure out how prusa does it
2 Likes
I know, but itâs hard to specify exactly which type of lines shall be printed with the second nozzle
One interesting use case for the H2D having different nozzle sizes on the left and right toolhead would be combining high-resolution and high-throughput printing in a single job. For example:
- Use a 0.4 mm nozzle on the left head to print outer walls and fine detail where surface quality matters.
- Use a 0.6 mm or 0.8 mm nozzle on the right head to handle infill, support structures, or internal geometryâareas where speed matters more than resolution.
This hybrid setup could significantly reduce print time while still maintaining aesthetic quality on visible surfaces. Itâs particularly useful in large, semi-functional prints where structural strength and print time are more important than perfect finish throughout.
However, there are software-level challenges with this approach:
- Toolpath planning becomes complex, especially when slicer logic is built on the assumption that both nozzles are identical in diameter and performance characteristics.
- Seam placement, wall ordering, and infill overlap must be recalculated per tool, or else you risk under- or over-extrusion where the nozzles switch.
- Cooling behavior and retraction settings vary between nozzle sizes, and a mismatch here can cause stringing, blobs, or poor layer bonding.
- AMS and material assignments would need to be mapped correctly per toolhead, especially if one nozzle is fed with high-flow filament while the other isnât.
Currently, Bambu Studio doesnât offer robust support for mixed nozzle diameters, and most users whoâve tried it end up running into unexpected issuesâespecially in multi-color or dual-material prints.
Until native slicer support matures, itâs a cool idea with real potential, but not yet practical in most real-world scenarios.
2 Likes
Yeah. I want to add that, the only ways that a larger nozzle helps are about:
- To print a wider line so that you can print less internal walls, when you need more walls than 2 (but more uneven outer wall results)
- To print a wider line for 100% solid infills, so that your solid infills would print faster
- To print a higher layer height useing infill combination and print walls normally.
The print speed difference would be pretty marginal for most prints, and for the prints that actually benefits a lot on speed from it, you might face even more warping because now youâre pumping out heat much faster than before. unless youâre using fibre filaments.
A feature Iâve long wanted, however, 4 months in, my opinion on the potential general usefulness of it has shifted quite a bit. No, this is not an âacceptance copeâ, but based on more active observations with the feature in mind.
Set aside software implementation challenge because itâs doable, the biggest factor I see limiting its impact is just âhow muchâ you can save.
Weâre looking at models big enough to take too long for a .2 or .4 nozzles, but you want the details and sharpness of those. But how much time is spent doing infills anyway? This will be different for every model, so in my case, most of my self designed models tend to feature complex walls&surfaces, with a baseline 10%~15% sparse infill and enough walls to ensure aesthetic quality, usually 18%~25% of total print time on Sparse & Internal Solid Infills, while up to 50% was spent doing walls. You still want to do at least 3 walls with the same extruder to ensure surface consistency and quality, so in such cases, speeding up infills will only have marginal benefits.
On more âboxyâ models, infills can indeed rise up to near 50% while walls less than 30%.
I think this feature is thus more limited to a few scenarios (to yield significant time saving)
- Big, structural / boxy part, add fine details like texts and logos as outer add-ons - this would necessitate self designed models that split object into parts (or implement AI algos to smart split STLs - questionable)
- Resin-like prints where supports nearly outweighs the actual object
- You want .2 quality details but the print has very large solid layers that would clog up .2 nozzles from pressure creep (not something Iâve tested much on H2D, a significant problem for A1)
- Some others I probably missed involving strength / going beyond nozzle layer height limits
2 Likes
.2 print with .4 for PVA support or other support filaments might be a good case
1 Like
Comes in handy when you have a print with fine details like tiny letters. Printing the whole thing in.2mm takes forever
4 Likes
This would be great if they can do it. I make Keychains and detail in some leaves holes in the work. I think the limitation is in the software, not the actual machine since functionally it doesnât care what its shooting through, its all G-Code. The Slicer would need to be programmed to know which nozzle to shoot the goo through.,
Embossed details that sit on top surfaces is definitely the easiest to implement, and will achieve pristine quality with just a prime tower without needing further tweaks.
It just gets much complicated doing it on outer walls / organic surfaces like a sculpted arm.
Layer change naturally makes for a clean cut without introducing additional un-retractions, seams, where for many filaments that realistically mean small blobs of over-extrusion (the âtipâ is more a teardrop than one end of chopsticks, if you will). This can be mitigated by hiding it on the bonding surfaces.
Probably inevitable tho is that those will better be embossed, not engraved (you can, by printing an entire section separately, more ugly seams), and with a somewhat coarser crispness (just like un-retraction blobs, filaments will take a short moment to reach ideal flow so with such a tight space between un-retraction and retraction, the quality will degrade somewhat)
BBS already handles staggered layer heights between different objects and object supports, and all sorts of line width crack, I think the slicer side is logically straightforward, but the amount of new options and UI change needed seems massive - you now need to define a whole lot more stuff previously left untouched, as there was no need to define with only one option.
btw, on more simpler top surface embossing, you can laser cut PETGs and print PETG on to it. Iâve done it to decent repeatable success, but itâs definitely laborious, needs decent CAD, and not for slicer beginnersđ
This is one thing necessary for make this printer better.
I believe native support for non-planar 3D printing would be a valuable addition.
While Iâm aware that the current toolhead design may not allow for full implementation, partial support could still be introducedâleveraging what the printer is mechanically capable of.
1 Like
Native as in built into Bambu Slicer? That would be a monumental project, it would be easier to add H2D support to an existing non-planer slicer. Iâm assuming since vase mode works that extruding to a XYZ coordinate works as well as the usual XY, but not having tried it I donât know for certain.
Non-planar printing isnât just a feature you casually enable like vase modeâitâs an entirely different class of complexity.
To begin with, implementing true non-planar printing would introduce a massive number of new challenges. One of the most immediate issues is collision risk: the nozzle would be moving in all three axes simultaneously, including dipping into previously printed geometry. This makes it extremely easy to crash into parts of the model unless the toolpath is carefully adapted in real time. Traditional slicers and gantry-based print heads simply arenât designed for this.
Thereâs also the mechanical side. Most current print heads and motion systems are optimized for planar layer deposition. For non-planar printing, youâd ideally need a redesigned print head with better clearance, potentially even a tilting or articulated nozzle, to avoid interference with already printed surfaces.
And then thereâs the software. Youâre not just slicing in flat layers anymoreâeach toolpath needs to be curved and height-aware, something only a few experimental slicers or custom G-code generators currently support. Weâre talking about fundamentally rethinking path planning, extrusion rates, and surface mapping.
Yes, they could release a new toolhead engineered specifically for non-planar applications, but it would require new hardware, firmware, and slicer logic. Research groups like MIT and TU Delft have shown this is possible using 5-axis robotic arms or custom gantries, but itâs still far from consumer-ready tech.
So againâthis is nothing like vase mode. Itâs not even close.
I was trying to clarify exactly what you meant by ânativeâ. As I mentioned, the printer already seems to be able to print in vase mode which is non-planar, so what is it you are asking for? Non planar support on the existing slicer would be, as I said, a Herculean effort, is that what you want? Or do you just want the H2D to support existing non planar slicers? That is principally an issue for the maintainers of those slicers, but it would be magnitudes less work than adding that support to Bambu slicer.
Just to clarify â vase mode is not non-planar printing.
Vase mode (spiralized outer contour) still prints in perfectly flat horizontal layers. The only thing that changes is the Z-height gradually increasing during extrusion, creating a spiral â but the nozzle never deviates from the XY plane in terms of extrusion surface. Itâs still planar by definition.
True non-planar printing means the nozzle actively moves in Z during each layer to follow a curved or angled surface. This allows for stronger part orientation, better surface finish on sloped areas, and reduced support material, but it also introduces a huge set of challenges â including collision risks, complex toolpath planning, and hardware/firmware limitations.
So no, vase mode is not an example of non-planar printing â they are fundamentally different both in theory and implementation.
As for slicer support: yes, enabling compatibility with existing experimental non-planar slicers would be easier than building native support into Bambu Studio. But even then, youâd need:
- Firmware that allows free Z movement within a G1 move,
- A toolhead with proper clearance to avoid collisions,
- A path planning algorithm that accounts for extrusion angle and nozzle orientation.
Simply sending G-code from a non-planar slicer wonât work unless the printerâs firmware and motion system are ready for it â and right now, consumer printers (including the H2D) arenât designed for that.
I havenât worked on the vase mode code in studio/orcaslicer, and I certainly donât have any insight into the innards of the motion planner, but from what I can remember it supports XYZ G1 moves and that is how vase mode is implemented. Are you saying that the motion planner doesnât actually process them as a straigh-line move between two points in 3 space and simply does the Z all at once? That is entirely unsurprising, and as I said initially I havenât tested it. But it certainly seems to support âfreeâ Z movement in a G1 now, thatâs how vase mode works.
Youâre correct that G1 supports movement in X, Y, and Z simultaneously, and vase mode does use that to create a spiral contour. However, that doesnât make it non-planar printing.
In vase mode, Z increases gradually but the toolpath still lies on a flat 2D surface. True non-planar printing involves varying Z dynamically across the entire layer to follow curved geometry â not just interpolating linearly between two points in space.
While G1 X Y Z is accepted, most firmware (including Bambuâs) processes it as a straight move between two points, not as a continuous surface-following motion. Thatâs the key difference â and why vase mode is not non-planar printing.