The prime towers use quite a bit of filament, especially when printing multiple colours, and so do supports. Both will be thrown away after. What about using the supports as a prime tower?
you can enable purge into supports
Flush into objectsâ support (enabled by default)
Since the support part of a model will be removed after printing, it can be printed with flush extrusions. For models with large area supports, enabling this option can reduce a lot of material waste and shorten the printing time.
I knew it was a good idea ![]()
Thanks for the replies
Just out of curiosity. When one of the filaments I am using is support interface filament, will it flush that into your support base also?
Because in that case it would weaken your support structure, and itâs better to use a prime tower I guess.
In my opinion, you should always use the prime tower, although I could be wrong. This is because not every layer has enough substance to balance it out. However, you can make the Prime Tower significantly smaller.
Now, to answer your question: It depends on whether you print the entire support in support filament, which very few people do because itâs expensive. Or you print the support in PLA and only the intermediate layer, i.e., the contact surface, in support filament.
How exactly the slicer does this is beyond my knowledge if you completely dispense the Prime Tower. As I said, all the options are more about keeping the Prime Tower as small as possible. Iâve seen printouts that skip the prime tower, and most of them didnât look good anyway.
But to sum it up briefly: Yes, youâre right, you shouldnât mix different filaments.
__________
What makes more sense, the better alternative: a flush object, but only if you use the same material. You place as many of the objects as possible on the plate that need to look good, and one object that doesnât matter, e.g., a functional part. Then all the excess filament is injected into it.
On the left are the objects that need to look nice, on the right are functional parts where the color doesnât matter. From: Ukdavewood
Agreed - what I would like to see is:
- Being able to prime and where necessary flush into defined useful objects rather than a tower or creating flush waste. Even if the prime object needs a bit of cleanup.
- Also if possible being able to allow the prime object layers to get slighty ahead of behind the main print object - to minimise the amount of non flush/prime in the extra objects - and therefore avoid extra time.
Thatâs already possible, my entire comment above yours is about it, including the video. ![]()
Or, starting at minute seven of the video, you can even see how to set it up. As I said, I donât think it makes sense to leave out the Prime Tower completely, but you can do it.
Here is an example from Reddit:
_________
Here is another interesting video on the subject.
Is Zero Waste possible with a BambuLab AMS 3D printer?
When I last tried it - if you turn off the prime tower then flush-into objects doesnât work properly - but it may be that they have improved BambuStudio since then will recheck when I do my next big print.
Infact I went as far as writing a gcode post processor to strip out the prime tower after BS had done its flush-into slicing.
just tried a test
Without flush into - 29g of waste - flush ratio nearly 300%
with single flush into object - flush waste reduced from 29g to 24g
with huge flush into object - big enough to take pretty much all of the waste
Turning off prime tower - flush waste comes back again
Edit: Just tested it myself. Youâre right. Apparently this is no longer possible without restriction in the latest slicer software. Iâll have to see if the Orca slicer can still do it.
Just checked on GitHub. At some point, Bambu added this limitation. I didnât know that.
Iâve been searching GitHub like crazy, trying to find out which version could do what and what it couldnât. Even in the German forum, people complained about this, but no solution was found. At some point, it must have been possible, as the video shows. All I found was a comment from Bambu on the topic from 2025:
- âWe will discuss whether to remove this restrictionâ
Keep in mind, itâs a âPrimeâ tower, not a âPurgeâ tower. The purpose isnât to flush material, itâs to get the pressure up in the nozzle prior to starting the print. Not sure how much the distinction matters in the context of this discussion. But Bambu might have configured the slicer to âreinforceâ that distinction. Dumping large amounts of filament in to a support might have negative consequences, maybe increasing the probability a support will be deformed by heat during the âpurgeâ.
The question that arose during the discussion was whether an flush object could be used to build up pressure in the nozzle if quality was not an important factor. So yes, a distinction was made in the discussion. ![]()
That used to be possible, as the video shows, and as older Reddit posts confirm, but at some point it was restricted? According to Bambu, this was probably because it caused some 3D models to collapse. But I think every user should decide for themselves whether they want to take that risk.
Take a close look at Ukdavewoodâs example. The slicer doesnât really allow it and actually uses much more filament. I was surprised myself when I tested it. Without prim tower, filament consumption is higher. That should never be the case in this scenario, because it suddenly refuses âinfill in objectâ - because the slicer somehow fails to fill the waste into a flush object nor use it to build presure.
Others are welcome to test it too, maybe Ukdavewood and I made a mistake. ![]()
The problem for Bambu is, they have to provide technical support to all the customers who try and canât make it work. Easier by far for them to restrict functionality they know leads to failures.
Remember, Bambuâs marketing theme is 3D-printer-as-appliance, take it out of the box, plug it in and go even if you know nothing about 3D printing. Theyâre not aiming at us âprosumersâ as much and they have separate product segmentation for enterprise users.
If I were them, Iâd assume any user smart enough to understand this particular issue is smart enough to work out a solution for themselvesâŚ







