My setup:
left side has 1 x AMS 2 Pro, 3 x AMS HT
right side has 3 x AMS 2 Pro, 1 x AMS HT
I think the prints where the FTS is helpful are more limited than one might think/hope. But there are also some benefits that aren’t obvious, drawbacks, and some unrealized benefits waiting for software. I’ll try to break down my opinion here.
Single color prints
You get the convenience of printing any material with any nozzle, without swapping anything around. This sounds like a nice convenience. But since the machine doesn’t support using different nozzle sizes in the same print, you probably already have the same size nozzle on the left as the majority of your right nozzles (0.4 most likely). So the single color print could have just been printed with the left nozzle to begin with. I keep a single 0.2 and 0.6 nozzle on the right, so being able to print a left material on these is the only remaining benefit in the single-color case that I can see.
2-color prints
For 2-color prints and a H2D/C machine, you’d typically want one material on either side, to get the fast 10 second changes with no filament cutting. And this can’t be done using the FTS with two materials on the same side, since there is only one feed line per side to the FTS. Using the FTS you could feed two materials to two right-side nozzles, and you get the benefit of shorter filament retractions since it only pulls back past the FTS (not all the way to the AMS), but it still cuts the filament so it’s not a big save. Maybe 10 seconds per change? So the end result is that you are still juggling spools to make sure they are on separate sides, and at that point you definitely want to get the faster switching speed from using both left and right nozzles with no cutting, which could be achieved without the FTS.
The edge case here may be if you have two specialty nozzles on the right, and the FTS would be save some filament loading time by keeping both filaments ready closer to the printer. You’d still have to make sure the spools are on different sides to get any benefit at all.
Many-color prints
Here there just isn’t much of a benefit in the general case. To print efficiently with H2C, you’re usually trying to make sure a base material used throughout the print is always loaded on the left, while the right side swaps between materials. But to make this work, you need to move spools so that the base material is on a different side from the others and the FTS does not change this significantly, since it only has one channel per side (it does let you swap the sides).
The edge case, which truly is an edge case but still the one Bambu touts in their marketing, is when you have different combinations of two materials for different heights of the print. The FTS can often make sure you get material changes with no cutting by making sure the two materials for the layer are assigned to each side. BUT, you still have to make sure all the combinations are spread between the two sides. It works in practice if you have a “base” material that you can keep on the left, and the different accents on the right. But that case already worked well without the FTS.
Problems
The resistance added to the filament path is no joke. It can be made reliable enough, but this is clearly eating into the margin of performance and I had some feeding failures before redoing my lines. This is why they added the silly funnels, since we are now much closer to the limit of AMS feeding.
The missing product
The whole setup is clearly, obviously, missing an “AMS 3 Pro” that has a built in 4-2 switch, so that each AMS can feed to either channel. You’d still want an FTS near the machine, so that the AMS can use one channel as the “next” material for the same nozzle. This setup would allow the next material to always be ready to load very close to the machine, speeding up the nozzle change.
Although to make this truly optimal, you’d want 3 channels. You want to allow the left nozzle to have a material loaded with no cutting, while the right side is changing quickly with the next material queued up. The FTS part for this setup could be simpler, just a filament detector and a merger, since the AMS would decide which side to target. And this merger/detector seems like it could be built into the machine, just like it is on the A1.
The missing software
There are some software updates to the current hardware that could make it a little more useful.
First, the slicer should show a much better preview of my options for nozzle assignment and spool placement. One of my 0.4 nozzles is HF, which is yet another variable that I have to play with manually to find the best tradeoff.
If the slicer supported multiple nozzle sizes in the same print, the general benefit from FTS would go up as well.
If we placed our 4-1 mergers close to the FTS, the AMS units could retract only past this merger instead of all the way to the AMS. Actually this would apply even without the FTS, since the AMS units can retract a fixed distance if the software supported this cleverness. But since each AMS only has one output, the benefit here is limited.
Bambu Slicer also severely lacks the filament assignment strategy of “shortest print time”. That’s the one I want. Without the FTS, this was usually the same as “filament saving”, but with the FTS that is less frequently the case and I don’t think it properly communicates this.
Current best target audience
I think it’s the X2D, which makes sense given their launch alignment. Since aux nozzle is different, you’d want to avoid it when you can and this lets you do that for single color prints, and swap the sides for 2-material prints. This is a clearer benefit, which mostly doesn’t exist on the H2D/C since the left nozzle is the same as the right ones (usually). If you do use specialty nozzles on the H2C, you get a similar benefit.
Conclusion
For the H2C and H2D, the FTS is a convenience benefit in several cases. But it saves only a little time in most practical cases, and it adds serious resistance while also making the thinking process (my poor brain) while slicing much harder if you truly are trying to understand the bottleneck and optimize the print time.
