CAPXL AMS Mod - Print TPU/Flex ≥85A on your AMS by chemape

chemape

3D Printer>3D Printer Parts

Print Profile(1)

All

P1S

P1P

X1 Carbon

X1E

A1 mini

0.2mm layer, 2 walls, 15% infill

41 min

1 plate

5.0(2)

Open in Bambu Studio

Boost

492

904

457

104

Released

Description

Summary

What is it?

This mod for the Bambu Lab AMS enables the use of TPU (tested down to 85A) while maintaining compatibility with standard AMS-supported filaments. It opens up the possibility of printing TPU parts with complex geometries using support materials, as well as multi-color TPU prints.

Dipped in IPA for 15 seconds for perfect delamination at each TPU/PLA interface

Who is it for?

For people who want to print flexible materials and are willing to tinker with it and aid in further development. This is a work in progress. I have tested it extensively on my printer, but I can't guarantee that it will work for you. Your feedback will help me improve the design.

100 back to back filament changes. PETG (white), TPU (black)

Is it reversible?

Yes, this mod is 100% reversible. There are no changes to your printer or AMS that can't be reverted. If, for some reason, it doesn't work or you're unhappy with it, you can go back to your old setup in a matter of minutes.

Limitations

This mod currently only works for nozzles ≥ 0.6 mm, and the AMS needs to be placed next to the printer in an elevated position. I guesstimate that the reliability is around 98% for me, which is not quite as good as the standard AMS.

__________________________________________________________________________________________________________________

Warning

While this mod is fully reversible, Bambu Lab may still consider your warranty void, as stated on their website/wiki. Proceed at your own risk.

How does it work?

There are three main reasons why the standard AMS can't handle soft filaments:

The filament slot spring is too stiff to allow soft filaments to retract
The filament hub spring is so stiff that there is not enough resistance created by the flexible filament reaching the extruder for the gears to start slipping. If the gears don't slip the hall sensor doesn't send a signal to stop, so they just keep turning indefinitely and start shredding the filament to bits.
The standard tubing is too long and has too many curves, which causes the filament to either not reach the extruder at all, not have enough force remaining to actuate the filament buffer, or not have enough force to overcome the filament sensor inside the extruder.

These issues can be solved by using custom springs and adjusting the filament path.

!!!ATTENTION!!!

I received reports from a few users that they experienced retraction issues. The cause is typically a misfit of the travel stop. After I did some testing I learned that depending on how long the PTFE tubes are and how the cables are layed, some tension can be applied to the filament slot mechanism. That results in the travel stop sometimes being too large or too small. I have designed an adjustable mechanism as a replacement (image below), which I will hopefully release soon, when I get more TPU delivered. In the mean time, I recommend that you wait with attempting this mod.
Sorry for the inconvenience.

Instructions

1. Print the travel stop and the hub spring both in TPU 85A (I used Extrudr Semisoft TPU).

Edit: I've been informed by two users that 95A TPU works as well. I will test that when I get around to it.

2. Replace the metal springs with your new printed springs/travel stops as shown below.

3. Get two pieces of standard Bambu Lab PTFE tubing (2.5 mm inner diameter, 4 mm outer diameter) with a length of 40-45 and 55-60 cm.

The longer piece goes into the extruder and the shorter one into the AMS. Remove the buffer from the printer and connect the two pieces of tubing through the buffer. Put the AMS in an appropriately elevated position, so that the tubing has enough slack for the buffer to actuate. Avoid using the Bambu tube connector pieces, as they narrow down to a straight edge on which the filament can catch and cause malfunctions.

4. Done!

5. Try to run the 100x filament change test. You should get less than 5 errors per 100 filament changes and they should only occur on TPU. If you have any failures with filaments other than TPU, something is wrong!

Things to watch out for:

If you aren't already, use a filament guide. I use this one, but I'm sure others work too. Without a filament guide you might get filament caught on the edge of the filament sensor.
Make sure that the tubing that enters the buffer can travel freely. If it rubs against the guide hole the resistance can lead to failures to feed TPU. If you can't solve the issue by adjusting the tubing, you can print the optional buffer housing. It has a large opening on the spring side and no guide hole.
If the AMS filament slot has trouble feeding the filament towards the hub, it means that the travel stop is undersized. Scale it up in the Z-direction by 0.5 mm and try again (happened to me on one slot)
Remember, TPU gets soft at elevated temperatures. Run your printer with the door and/or lid open.
TPU has a hard time purging other filaments - adjust your purge ratios accordingly!

Roadmap

AMS position

Currently the AMS has to be placed next to, or behind the printer to keep the filament paths short, and relatively straight. I did try to put the AMS on top of the printer, but then the tubing was either too long for the filament to have enough force to feed, or it was curved to a point where the buffer couldn't properly actuate. If the 4 pin bus cable that connects the printer with the buffer was a bit longer, you could adjust the angles appropriately. Unfortunately, as of me writing this, I haven't been able to find a longer version of this cable yet.

0.4 mm nozzles

This mod was initially developed with the standard 0.4 mm nozzle, but the backpressure during purging caused the TPU to be deformed by the extruder gears, which later led to failures to feed. Additionally, purging didn't work properly for the same reason, which caused frequent nozzle clogs. I tried to remedy this problem with custom G-code, but since I'm not very experienced I couldn't come up with an appropriate solution. If you have an idea for custom G-code and some time to spare, please contribute!

Purge routine

The current purge routine works, but it isn't pretty. Because it is necessary to purge quite a large amount of filament when changing to TPU, the purge piles up in the shute and it forms a blob around the nozzle.

Maybe someone could write some G-code that wipes off the purge half way trough the routine, instead of just at the end.

Another possible solution would be to use custom G-code that reduces filament waste, as shown here, but I haven't been able to test that yet.

This project was inspired by a video by Noozle3D on Youtube. I would also like to thank the members of the X1Plus discord for providing valuable feedback.

Changelog:

19.06.24 Re-uploaded the hub spring model. Due to an error I made in CAD it was modeled as three parts that could be split by the user.

12.06.24 Initial release

Comment & Rating (73)

@user_434820605

Boosted

This is definitely real - I printed this with 95A TPU and it allowed me to print TPU in the AMS. This is an incredibly important feature for me so thank you very much!

The designer has replied

@chemape

I'm glad you like it. Thanks for the boost!

@chemape

Hey! Can you tell me which filament exactly you used so I can add it to the description?

@user_434820605

Boosted

Replying to @chemape :

Hi, sorry for late response I used the Black Geeetech TPU 95A.

@peschka

Boosted

Same situation as described by @alaric.hyland I am also using Hydra AMS Pro. I printed this mod with Bambu 95A TPU. No problem with grabbing TPU and delivering it to the extruder. But there is a problem with the retraction. AMS hub is the reason for sure, because if I try to pull it back by hand - no problems all the way to the ams hub and hard stuck after that. Any suggestion how I need to scale the hub spring replacement - up or down, or any idea how I can troubleshoot this further ?

@JackZwykle

Same problem here, started experimenting, any luck on your side with retraction?

@peschka

Boosted

Replying to @JackZwykle :

Sadly, but no. I will try with softer/smaller or harder/bigger spring and report back. More than 500 layer changes without problems with the feeding but 5/100 are with problematic retraction and it's a hard one.

@JackZwykle

Replying to @peschka :

Thank you for the reply. Seems we have different issues. My printer was bought two months ago and has the "V2" AMS. In my case the spring plays no role on retraction. It is the filament sensors in the AMS hub that literally wedge the filament against the square tunnel. They wedge it so hard that it's impossible to pull the filament in any direction through the hub, up to breaking it, it needs to be disassembled and inside you can see the filament pinched so hard it is about 0.5mm wide and very flat. This happens 100% of time, not a single success. Thinking about redesigning better AMS hub body, but hell... that's a lot of work for this.

(Edited)

@user_434820605

Boosted

Updates: I have been using this for some time and it is working very well. Bambu P1S combo AMS in stock position on top of the printer ShortStack Low Interference Version in ABS 95A main spring and end stop replacement in 3rd slot from right (only did one slot so far) 0.4mm nozzle as stock Replaced the Filament Cutter with a new one Tape the PTFE going to the filament buffer with the AMS cable underneath so there is a section of totally straight pipe (at least 50mm) in all directions. Still working all slots with normal filament like PLA PETG ABS, no failure at all. I have only tried printing 95A so far. When using 95A, the printer successfully goes through tens of filament changes with no problems. When using spools of TPU in the AMS, it may be necessary to cut a thin piece of cardboard to cover the spool so that when the TPU is unloaded, the extra TPU filament doesn't pop out of the spool and tangle. With this modification, it works very well. ShortStack is necessary, as is using a new filament cutter blade (an old one will not be able to cut TPU). Sometimes the loading will be retried once but so far it has been doing it much less than before. Hope this helps someone!

(Edited)

@DigitalAbstract

Is this a V1 or V2 AMS unit that you are using?

@Brestling

works flawlessly, even in stock Position. I tried to remove as much friction as possible to even include a y-splitter. Have not encountered any issues so far! thanks a lot!🙏🏻

@fr3tles5

Which splitter did you end up using? I don't quite have room for my AMS to be to the side, so I am interested in your set-up!

@mhoff67

The AMS stand is superb, I wonder if my old Gerthsen will work on it too ;-)

(Edited)

Show original

@Makos

The print profile is no longer public

printed out perfectly, will try to install soon after current print is finished

@alaric.hyland

Has this been tested with a Hydra/Hydra Pro AMS mod? I think the angles for the filament path are slightly different and I'd like to know if anyone else has had luck before doing this myself. Or I'll test it if I find the time.

The designer has replied

@chemape

Actually I'm fairly confident that using the hydra mod will have a positive effect on reliability because the angles are more favorable. The regular AMS slot causes friction and stretching at the entrance to the filament slot. I will do testing with hydra at a later point. I have not done so because I wanted it to have as few pre-requisites as possible.

@alaric.hyland

Replying to @chemape :

So I gave it a shot alongside my hydra pro mod. Travel stops work nice and can grab/release the TPU well, and the TPU made it to the buffer. However, on retraction, the filament jammed in the hub motor. I printed all pieces out of MatterHackers Pro Series TPU (https://www.matterhackers.com/store/l/pro-series-tpu-thermoplastic-polyurethane-1lb/sk/M396ANY7), which is supposedly 95A but feels more flexible and stretchy than that in person. This is the same material I tried running through the AMS as well. I have some ninjaflex I can print with instead, think I should redo the hubspring?

@chemape

Replying to @alaric.hyland :

Remove the travel stop and load/unload a couple of times (you might have to assist in loading for the first couple of cm until it reaches the hub). If unloading works reliably with the travel stop removed, that means the travel stop was oversized. Print a new one scaled down by 0.5 mm and try it again. I will add a troubleshooting guide to the description asap. Thanks for your feedback.

@user_2698062189

For 0.4mm nozzles you could probably loosen the tension screw on the extruder to make it stop grinding up the TPU. It's the only way I've gotten the X1C to print 85a TPU properly.

The designer has replied