Someone else on the forums had the brilliant idea of using the TPU feed assist module to help with feeding PPA-CF, I tested this today and can confirm it works perfectly as a solution to feeding problems. The way I am feeding it is AMS HT bypass > TPU feed assist module > TPU bypass inlet (since the buffer has an issue with PPA-CF).
Very interesting, I’ll have to give that a try!
Do we know if the TPU assist uses a hardened feeder gear?
The module looks like a P1/X1 extruder adapted to its role. It would be pretty straight forward to upgrade the gears to hardened steel if they are not already. I will take it apart when I have time to confirm.
TPU feed assist module:
X1 extruder:
Just took it apart and it does have standard gears, X1 hardened gears will drop right in.
Dang it! I knew I should have kept a set of those when selling my x1 lol.
LOL… im in the same boat. Had like 3 sets and gave them all away after I sold the X1 and P1. No biggie they are pretty cheap, just had to add some other stuff to make the shipping free.
mine seem to work just fine when using the tpu inlet to feed from left side of printer from a roller. but using assist motor is always good ![]()
Been printing a lot of PPA-CF last few days and the only downside with using the TPU assist module is it causes more oozing from the nozzle. Not surprising since it is basically a second extruder shoving filament into the nozzle, but easy to cleaned up post print.
I suspect that could potentially be cured through retract settings; the default Bambu filament profile would have been made assuming the filament was being fed by only the extruder, it’s just having a much easier time now so it could benefit from more aggressive retract.
Seems like the oozing issue was unrelated to the TPU feed assist. I fed it directly and had the same problem with this particular object with fine features. I switched from HF to a standard hot end, lowered the temp to standard (I usually print at 320c for best layer adhesion), increased the retraction length a bit and its printing cleanly now.
Could be just flowing too well ![]()


