Another dumb idea (or two, or three ? )

Other topics
1

Random thought from a relative newbie (started printing last November), so feel free to tell me why this is dumb…

Toolheads are heavy. A decent chunk of that weight is cooling: hotend cooling, part cooling, ducts, fans, etc. Tool changers (Snapmaker, Flashforge, and others) and nozzle-swap systems like Vortek, INDX, and KLITec also have to deal with the size and mass of whatever is moving around the gantry. You need clearance for swapping, and you need to accelerate all that mass every time you move, dock, or pick something up.

What if the printer had two separate XY gantries?

One gantry carries the extruder/hotend. The second carries the cooling system. They’d be synchronized, but not necessarily perfectly. In theory, the motion control could intentionally keep them slightly out of phase to reduce chassis vibration, almost like a built-in dynamic damper.

The cooling airflow doesn’t necessarily have to be perfectly centered on the nozzle. If the airflow column is wide enough, a few millimeters of offset probably wouldn’t matter much. You could even use the second gantry to carry a CPAP-style air supply so it’s delivering cooler outside air. Cooler, denser air might reduce the airflow volume needed for the same cooling effect.

Taking it a step further, what if cooling became dynamic? Instead of blasting the whole part, the system could direct airflow only where needed, adjusting cooling based on geometry, layer times, overhangs, bridging, local feature size, or even slicer-predicted thermal stress.

That changes the tradeoff space quite a bit. If cooling becomes more effective, without as much cost of weight, you might be able to:

  • Improve overhangs and bridging

  • Increase print speed on small layers

  • Reduce toolhead mass

  • Push acceleration higher because the cooling hardware is no longer riding on the printhead

  • Reduce vfa’s caused by resonance from acceleration wobble by calculating offset mass movements between the two gantries.

And while we’re throwing crazy ideas around…

What if there was a third XY system carrying a laser?

It wouldn’t need the full build volume. Just enough motion to maintain line-of-sight slightly ahead of the nozzle path. the laser itself might be aimed without having to move the whole gantry position.

The laser’s job would be to preheat the surface immediately before deposition. In theory, that could improve layer adhesion by bringing the previous layer closer to its glass transition temperature right before the new filament is laid down. It might also help relieve some internal stress by allowing the previously cooled plastic to relax a bit before being bonded to the next layer, potentially reducing warping. Could also be used for adding visual features not only to the top surface, but sliced into the walls or base layer as it is being printed. So you could have laser created visible features to all sides of a model.

Anyway, that’s my naive idiot idea of the day. Now the people who actually know what they’re talking about can explain why I am so dumb.

2

Actually I don’t think that the cooling fans are a major weight factor. The heaviest parts in the toolhead is by far the extruder motor. I think then come metal parts like linear bearings and then in no particular order electronics, housing, hotend, cooling fans, wires, etc.

But there is a different point in externalising the cooling: max airflow.
Instead of an extra gantry, there actually are already systems that have a huge stationary blower fan and air is guided through a hose to the print head. That allows far higher wind speeds / pressure.
For optimal cooling, you want the air to hit the part directly at the nozzle so that it solidifies immediately.

3

The CPAP system used in printers like the Prusa HT90 takes the cooling fan off the toolhead and introduces an air hose, whose mass is sort of desynchronised from the toolhead movement (being flexible). It also has near instant on-off using a valve. This is simpler and does a lot of what your first and second points aim to do.

Preheating the previous layer can be done simpler with a higher chamber temp, many printers do this already, with additional benefits.

Food for thought :slight_smile:

4

I just had another idea… the Z-axis screws for bed movement could be for just gross movement.
Have solenoids on the bed mountpoints that snapped the bed up and down with spring dampers to adjust jerk. The solenoids would be faster and have less slop, also wouldn’t have to move the entire z-axis mechanism, only the bed itself, so less mass, less vibration.

5

Why would solenoid have less backlash?

How heavy would these solenoids be?

At this stage, why not just use a flying gantry?

1 Like
6

The solenoid coils don’t have to move for z-hop. They simply deliver an electric charge to the pins that are attached to the rest of the build plate.

So, when you want to do a z-hop. all (3?) of the solenoids would fire together and yank the build plate down. But the belt, the stepper motor, the 3 cogs, the roller tension pulley, the 3 screws, and the screw rider nuts and the braces that hold the build plate would all remain stationary when a z-hop occurs.

In my envisioned approache the z-screws would only turn when a layer change occurs.
During the printing the 3 solenoids would do all the z-hops. and would only hop the build/hot plate.

7

Flying gantry is what Voron did right?
But with tool changers or nozzle swappers, I think that would be a lot of movement.

8

Regarding those solenoids: that would add a lot extra components: not only the solenoids, but also electronics to activate them, wires going to the moving platform, maybe some chains to guide the wires.
It also complicates control, because suddenly, movements in Z aren’t hard mapped to the z motor anymore but they have to be divided between z motor and solenoids.

And the most important point: I think z-hop is working perfectly fine at the moment, so there isn’t actually a problem that is in search for a solution. :zany_face:

9

All these questions are answered by machine design principles, particularly CNC machine design principles.

I prefer self guided learning for these kinds of questions - ask yourself “what am I trying to fix with this idea?”, “what do others do to fix this?” and “is this really a problem” - THEN, evaluate the specific concept, can it actually fix the problem, what complexities does it introduce, and how much will it actually fix the problem.

I can say that most of what you’re suggesting isn’t really a problem, and the concepts either don’t really fix the problem, or introduce significant cost and complexity.

10

Voron 2.4 is a flying gantry, as is the Phrozen Arco. Bed stays fixed at the bottom, and the print head (with its gantry) moves X, Y and Z.

Voron Trident is like a Bambu. 3 lead screws and bed moves on Z.

11

Yeah, it is complex. I think I explained what I was envisioning?
Lower jerked mass, out of phase masses could be manipulated to reduce wobble VFAs by using the different X-Y gantrie’s masses against each other. Allow more consistent+higher speeds that reduce pressure advance inaccuracies by having a more narrow range of noz speeds.
z-hop speed reduces the amount of nozzle x-y slow/speed up waiting for Z-hops.
Increase z-layer adhesion by having a hot spot under the approaching fresh nozzle layer output, And have better cooling if you had less mass budget to compromise for cooling.