I wonder if I have misunderstood something - or whether it’s a setting I have wrong.
A two-colour file printing on my X1C, with the two colours in each of channels 1 & 2 of the AMS - 23 hours to print.
Exact same file, in two colours on the H2D, with one colour in the AMS2 and into the RH print head, the second colour fed from an external spool to the LH head. 40-50g less poop created. But the H2D takes 29 hours to print.
So the H2D is six hours LONGER in printing the same file, same settings, with far fewer colour changes!
That’s not exactly what I signed up for; I was thinking that having two nozzles in the toolhead would allow for much faster swaps between two colours - instead of AMS colour swaps, with all that spooling and rewinding and purging. I expected the H2D to have significantly REDUCED print times, perhaps by a third or even a half. That seemed to be what Stefan got at CNC Kitchen.
Which makes me think that either I have some setting wrong, the setup is wrong (spool holder instead of second AMS), or there is a bug somewhere.
Thank you, that is good to know for when the ordering window opens again. I was just grasping at something I saw on your screen shot and the fact that the ad for the H2D says it comes with a high flow hotend. I did not realize that was an additional purchase, so now my shopping list just got larger.
The toolbead is much heavier so they limit the max acceleration to half the speed. For small objects you need fast acceleration and I’m guessing the default profile has the acceleration set way down.
Yes but when you have an entire plate of very small objects then purging doesn’t take that much time out of the total since it’s only probably 100 layers max and 1 color change per layer. At that point the slow acceleration is worse then the purging.
It looks like you set the video as “safe for kids.” If you did this intentionally as you know what it does, ignore what comes next.
Setting a video as ‘safe for kids’ means it will work on YouTube for kids, as such it can’t be saved to a playlist to be watched later by adults.
“Safe for Kids” doesn’t mean it is safe for everyone or there is nothing that could be harmful to kids, it is targeting kids and is afforded all the limitations that categorisation brings.
I like to save things to playlists to watch later. It makes it easier for me to share with others.
I think most people use 3rd party clients for youtube, such as NewPipe, and these allow all that stuff, plus downloading of videos. (Assuming you still use US product and have not joined the boycott).
I did some test slicing with a simple model and came up with the following rough estimates for each colour change type:
H2D nozzle to nozzle - 11.7secs
H2D AMS same nozzle colour change 76 secs
X1C AMS colour change 95 seconds.
So each H2D 2 colour change should be about 83 seconds quicker than an X1C - which should add up to about 134 minutes saving for the 97 tool changes in your print.
Whats strange is that if you add up the projected filament usage plus amount saved your h2d print it comes to about 207g, vs 194g for the X1C.
The layout of the parts on the bed looks slightly different between the two prints
The X1C print times are pretty accurate IME; this is something that I print regularly, and if it says it takes 23 hours on screen then it’s taking around 23 hour IRL.
Done a bit more digging this morning instead of going to work
Tried setting up a single object - which I’d never print, as it’s too wasteful in both time and material. Same object, same place on the build plate, same filaments; everything at their defaults. Excuse the numbers - I’m rounding everything here just to keep things simple.
Total time - H2D - 2.5 hours. X1C - 4.5 hours.
Makes sense; the X1C is spending 2.5 hours swapping filaments, purging etc, whereas the H2D isn’t.
But the actual times spent printing are less impressive - H2D 2.5 hours, X1C 2 hours. So the H2D is quite a lot slower than the X1C.
Scale it up to my 50+ parts and that purging time stays the same (2.5 hours for the X1C), but as the X1C significantly outpaces the H2D, the total time for the H2D is a lot longer - hence the +6 hours total time I am seeing when I slice the usual batch of parts.
Everything is running off default profiles; I want to print things, not bugger about tweaking a print profile or tuning the printer (this isn’t 2015, people).
Maybe there are inefficiencies in this early set of profiles and drivers. But right now I’m very irritated by this big downgrade.
Most people ??..lol I have a YouTube channel with 23k subscribers and have never heard anything of what you are speaking of . Lol. I think most people still use regular YouTube anyways , back to the topic . A. It disappointed. It understanding of the acceleration limitations . Was not mentioned in any of the reviews I watched . Regardless still an amazing new printer
My bet is that its a bug in the slicer. There was already a post timilar to this on youtube shorts i commented on where the user sliced both prints in .20 layer heights but the calculation that bambu studio was doing was based from a .12 layer height fir the H2D.
Looks closely at the sliced files and see if you can identify any oddities.
Does the H2D load/unload in one nozzle while the other is still printing?
I have heard the acceleration is lower, which could affect things, but if the machine does load one nozzle while the other is printing, something is definitely off…
Yes, there is only one extruder motor, but with the arrangement of the extruder gears and the switching mechanism using the cam at the top, it would theoretically be possible to retract the filament from one nozzle while simultaneously feeding the other.
Once the filament has been retracted above the extruder gear, the AMS could take over and handle the full retraction, then feed the new filament back into the extruder.
To make this work, however, the filament on the switching side would need to be cut shortly before the changeover.
This function isn’t implemented yet, but in theory, it could save quite a bit of time
I actually shared some sketches in another post showing how this could work with the current setup!