The AMS was the unique selling point to me for buying my X1CC.
I am satisfied with its performance but the AMS annoyed me a bit with its slow changes and huge amounts of wasted material. After a while, I got frustrated motivated enough to embark on improving it myself.
Back in July I published my “Bambu Lab AMS better purge Gcode” on Printables alongside a calibration method. Things evolved, MakerWorld came into existence … but nothing from BL addressing flushing volumes. → So, I now doubled down and my latest G-code alongside two methodical calibration prints made it to MakerWorld with refined upload information:
If you’re inclined to using them or willing to participate as volunteer in the testing phase, please check them out.
Regarding potential problems/bugs/improvements: Tell me about it, I am reading here
I am curious, what do you think?
Filament is cheap, 100mm³ of standard filament isn’t even one cent… Why bother?
Finally, I will use just the calibration prints.
Nice. I will use both, the calibration prints and that custom G-code.
Just the G-code for me. - The shorter printing time is what is of value to me here.
That’s all nice and dandy, but I am waiting (maybe an additional year) for updates by Bambu Lab and a slicer implemented calibration.
Pictures before and after calibration. Silver to other is a bit to low, because the silver shimmer is traceable in the other bands, however, it is so little that i think it doesn’t matter. The other numbers are a bit on the safe side.
A1 Mini
As you may have noticed, Bambu Studio has integrated the (experimental) feature called “Long retraction when cut”.
Long story short, to me this implementation is not worth loosing a lot of words over. With a valid retraction range of 10mm to 18mm, no consideration of the actual type of filament retracted, … I do consider it a far cry from what users could have anticipated.
On a positive note:
My calibration v1 (composed of a reign of flat first layer rectangles) is now fully supported in Bambu Studio as well. → Up to a full AMS or one filament against 7 others can now be calibrated in just one go (not with the A1mini though atm).
After a while I decided to add a second AMS to my printer because four colors somehow were not enough. - I am certain eight will do … for a while.
My standard filament is PETG and that is what has been used here.
One AMS filled with white (BL), grey (Prusa), black (BL), orange (BL) and the other AMS with red (BL), yellow (Prusa), green (?), blue (BL).
That simply is my compromise to get most of my preferred color shemes and also have the colors that are preferred by the kids at hand at all times.
Of course I went to town and determined the necessary flushing volumes for all 56 possible color changes (of course with my modified G-code applied) and compared them to the auto-calculated values from Bambu Studio.
Auto-calculated flushing volumes (+45mm³ prime volume not included)
with
negative volume = BS underestimated how much flushing was needed
positive volume = material that would have been wasted
Put to words:
On average around 110mm³ (+45mm³ if you include the prime tower) of filament is saved per change.
9 out of 56 changes were off by more than 200mm³. (In BL’s defence, not all was Bambu filament - but even within those changes they were off in half of the cases…)
4 out of 56 changes were insufficient and needed more flushing
In addition to the expected low flushing volumes required to switch between “neighboring” colors such as orange and red, changes between yellow and green are also very efficient and only require 30 mm³ more material.
As expected changes to white or from black consistantly need the most flushing volume.
Savings in print time have to be tested separately,
If you want to do this calibration as well, simply use the “Two AMS” profile provided within my calibration v1.
The printouts are distributed across 4 plates, with plate 1 and 4 resembling the calibration of each full AMS
