TPU Feeder Puck + Slack Case: sealed dry box that feeds TPU straight to the Bambu X1C (bearing hub, desiccant inside the hub)

What it is

TPU has to be dry, and most TPU does not work in the AMS. This box keeps the spool sealed and dry while the printer pulls filament straight out of it.

It is my Spool Puck (same twist-lock clear lid, same TPU lip seal, same hygrometer window and label card slot) with three changes inside:

• A fixed peg standing up from the floor in the middle.

• A hub sleeve: a 48 mm drum that slips over the peg and spins on two 608ZZ skateboard bearings pressed into its back. The spool hangs on the drum by its bore. The drum is hollow and holds about 70 g of desiccant beads behind a TPU cork, with vent slots, so the desiccant sits inside the spool where the air is. A small magnet in the peg and one in the drum hold the drum on; it lifts off with a pull.

• An exit port on the body wall: a 12 mm funnel on the inside leads to a 3 mm throat that lets filament through but stops the tube, and an 8 mm seat on the outside holds a TPU liner that grips a 4 mm PTFE tube.

The puck stands on its edge on two runners beside the printer, lid facing you, with the card slot at the top so the label reads from above. The tube runs from the port to the printer's external-spool inlet.

Optional add-on: the Slack Case

EXPERIMENTAL. The slack case interior is fiddly. Read "What is proven and what is not" before printing.

When the X1C unloads an external spool, it cuts the filament and backs it out of the extruder only. About a metre stays in the tube and you pull it out by hand. With TPU that metre must not sit in open air, and winding it back into the dry box means opening the box every time. This case gives it a sealed place of its own, right at the printer inlet.

• A 120 mm round cup, 30 mm tall, with a 60 mm drum in the middle. The filament wraps around the drum under its lip, about five turns for a metre; the drum has a notch for the filament end. The drum is hollow: about 15 g of desiccant beads go inside under a TPU cork.

• Two tube ports on opposite sides, on one straight line, so in use the filament runs straight across the open case: the feeder's tube comes into one port (IN), a short tube leaves the other (OUT) to the printer inlet or the Y-splitter's external-spool hole. Each port has a TPU liner that grips the tube and a 3 mm throat that passes filament but stops the tube.

• A clear lift-off lid with a thumb ear, held by four 10 x 2 mm magnet pairs (lid roof and four pillars in the base). It closes onto a TPU lip ring in a groove, the same seal as my Spool Puck. Peel it off by the ear; the magnets let go one by one.

• An L-bracket that screws to the two rear spool-holder holes on the back of the X1C (the stock rear spool holder comes off) and wraps around the printer's left rear corner (left as you face the printer). On its side plate a curved saddle carries the case from below and two hooks hold it from above; four more magnet pairs pull the case home against the plate. The case hangs ports down, ear up, lid facing you when you stand at the left side of the printer. It pushes straight on and pulls straight off.

Size

• Feeder closed: 239 mm across, 95 mm tall; stands 139 mm wide on its runners.

• Spools up to 210 mm across, 70 mm wide, with a bore of 52 mm or more (any normal 1 kg spool). The spool hangs on the drum and turns with almost no drag.

• Slack case: 120 mm across (137 with the port blocks), about 36 mm tall closed. Lid 125 mm, 135 with the ear.

• Bracket: 133 x 137 footprint, 110 tall, 3.3 mm walls, about 96 g. Its side plate reaches 100 mm above the screw holes; the case top ends about 125 mm above them.

Hardware you need (not printed)

For the feeder:

• 608ZZ bearing, 8 x 22 x 7 mm. Qty: 2. Notes: Common skateboard bearing. The pocket is a press fit designed for a 22.0 mm bearing.

• Disc magnet 4 x 2 mm. Qty: 2. Notes: Neodymium. Both are glued.

• PTFE tube, 4 mm outside. Qty: about 0.5 m. Notes: Standard Bambu/Capricorn tube.

• Desiccant beads, 3 to 5 mm. Qty: about 70 g. Notes: Loose silica gel or molecular sieve; colour-indicating is handy behind the clear lid.

• Superglue (CA). Qty: a few drops. Notes: For the two magnets.

• Mini LCD hygrometer, Amazon B0CJM5R5N6. Qty: 1. Notes: Optional. The one unit the lid window was sized for; no other hygrometer is supported. Snaps into the window.

• Y-splitter for the printer inlet. Qty: 1. Notes: Only if you also have an AMS. I use MakerWorld model 43892 "Instant Y-Splitter" by OurAngryBadger (not mine, not included): tube from this box into its external-spool hole.

The Slack Case adds:

• Disc magnet 10 x 2 mm. Qty: 16. Notes: Four in the lid, four in the pillars, four in the case floor, four in the bracket. All glued.

• PTFE tube, 4 mm outside. Qty: a short piece more. Notes: OUT port to the printer inlet or splitter. The feeder's tube goes into IN.

• Desiccant beads. Qty: about 15 g. Notes: Inside the drum.

• Superglue (CA). Notes: For the sixteen magnets.

• M3 screws. Qty: 2. Notes: The two stock rear spool-holder screws reused (the bracket flange is 3.3 mm where the stock part is about 2.5, so they still bite on mine). If not, M3 x 10 or 12, not longer.

Parts you print

TPU Feeder Puck profile:

• 1. Hub cork and tube liner. Material: Bambu TPU 95A. Plate type: Textured PEI, no glue. Notes: 0.2 mm layers (set in the file). Dry the TPU first.

• 2. Body. Material: Bambu PETG Basic, white (any PETG). Plate type: Textured PEI. Notes: 5 mm brim. 233 x 238 x 90 mm.

• 3. Lid. Material: Bambu PETG Translucent (clear). Plate type: Smooth High Temp plate, thin wipe of glue stick. Notes: Prints roof down. Same as the Spool Puck lid but with a taller stop ring, so the two are not interchangeable.

• 4. Hub sleeve. Material: Bambu PETG Basic, white (any PETG). Plate type: Textured PEI. Notes: Stands 83 mm tall on a 60 mm footprint: print it alone, with the zero-gap brim that is set in the file. Cut the brim off with a knife.

• 5. Lip ring. Material: Bambu TPU 95A. Plate type: Textured PEI, no glue. Notes: 0.2 mm layers.

• 6. Two label cards. Material: Any PETG, your colour. Plate type: Smooth High Temp plate, glue stick. Notes: Glossy face down. Dry-erase marker wipes off the glossy side.

Slack Case profile:

• 1. Base. Material: Bambu PETG Basic, white (any PETG). Plate type: Textured PEI. Notes: 5 mm brim.

• 2. Lid. Material: Bambu PETG Translucent. Plate type: Smooth High Temp plate, glue stick. Notes: Prints roof down. 100 % solid (set in the file).

• 3. Lip ring, drum cork, two tube liners. Material: Bambu TPU 95A. Plate type: Textured PEI, no glue. Notes: 0.2 mm layers (set in the file). Dry the TPU.

• 4. Bracket. Material: Bambu PETG Basic, white (any PETG). Plate type: Textured PEI. Notes: Prints standing as an L on its bottom edges, zero-gap brim set in the file. Cut the brim off.

Print settings (what the 3mf files already contain)

Sliced for the Bambu Lab X1C with a 0.8 mm nozzle at 0.4 mm layers, OrcaSlicer profiles "0.40mm Standard @BBL X1C 0.8 nozzle". Walls are whole numbers of 0.82 mm lines.

TPU Feeder Puck profile:

• 1 Cork + liner. Filament slot in the file: 1, TPU 95A. Bed: 35. Nozzle first layer: 230. Nozzle after: 230

• 2 Body. Filament slot in the file: 2, PETG Basic. Bed: 70. Nozzle first layer: 245. Nozzle after: 255

• 3 Lid. Filament slot in the file: 4, PETG Translucent. Bed: 80. Nozzle first layer: 260. Nozzle after: 245

• 4 Hub sleeve. Filament slot in the file: 2, PETG Basic. Bed: 70. Nozzle first layer: 245. Nozzle after: 255

• 5 Lip ring. Filament slot in the file: 1, TPU 95A. Bed: 35. Nozzle first layer: 230. Nozzle after: 230

• 6 Cards. Filament slot in the file: 2, PETG Basic (change to any PETG). Bed: 70. Nozzle first layer: 245. Nozzle after: 255

The lid and cards are set to 100 % infill with 10 top and 10 bottom layers so the clear parts are solid (sparse infill shows through translucent PETG). Check the plate-type tab on plates 3 and 6 (High Temp Plate) before slicing.

Not tested: a 0.4 mm nozzle. Wall thicknesses were chosen for 0.8 mm lines, and the bearing pocket's crush ribs (0.3 mm) and the TPU liner's bore (printed 4.2 mm to take a 4 mm tube) were sized from 0.8 mm prints. At 0.4 the holes will come out closer to the drawn size, so the liner may be loose on the tube and the bearings may need less force. If you try it, please report.

Slack Case profile:

• 1 Base. Filament slot in the file: 2, PETG Basic. Bed: 70. Nozzle first layer: 245. Nozzle after: 255

• 2 Lid. Filament slot in the file: 4, PETG Translucent. Bed: 80. Nozzle first layer: 260. Nozzle after: 245

• 3 TPU parts. Filament slot in the file: 1, TPU 95A. Bed: 35. Nozzle first layer: 230. Nozzle after: 230

• 4 Bracket. Filament slot in the file: 2, PETG Basic. Bed: 70. Nozzle first layer: 245. Nozzle after: 255

The slack case lid is set to 100 % solid like the feeder lid; check that plate 2 reads High Temp Plate. Not tested at 0.4 mm nozzle (the 10.3 mm magnet pockets and the 4.2 mm liner bores were sized from 0.8 mm prints; at 0.4 they will come out closer to drawn size, so the magnets may be looser and the liners too).

Assembly: the feeder

Names: the body is the big white cup with the peg in the middle. The sleeve is the drum that spins on the peg: cork end open, bearing bore at the back. The liner is the small TPU tube with a rim (flange). The lip ring is the big TPU ring. CA is superglue.

Magnets go in before the bearings (the sleeve's magnet pocket is at the bottom of the bearing bore and the bearings are a press fit). Both magnets are glued because they pull toward each other and would climb out.

1. Peg magnet. One drop of CA in the shallow hole in the tip of the peg, press a 4 x 2 magnet in, either way up. Hold 30 s. It stands about 1 mm proud.

2. Sleeve magnet, the right way up. Sleeve on its cork end, bore up. Second magnet on the tip of a steel hex key. Bring it down toward the peg magnet: it must PUSH AWAY. If it wants to jump on, turn it over on the key and check again. Drop of CA into the small hole at the very bottom of the bore (down the funnel), lower the key until the magnet seats, hold a minute, pull the key straight out.

3. Bearings. Bore still up. Look down the bore first: six small ridges run down the wall and the roof narrows like a funnel. First bearing on the mouth, square, firm thumb push until it stops; it shaves the ridges on the way and seats against the sloping roof. Second on top, same push; it ends about 0.5 mm proud of the sleeve's back face, by design. If a bearing needs more than thumb force, stop; see "What could go wrong".

4. Beads. Turn the sleeve over, cork end up. Pull the cork by its tab, fill with beads to about 10 mm below the mouth, press the cork back in.

5. Sleeve onto the peg. Lower it, bore first, over the peg until the bearings stand on the body floor. It pulls down the last millimetre on its own and spins freely with a finger. If it floats or pushes back, the sleeve magnet is upside down.

6. Liner and tube. Push the liner into the exit hole on the outside of the body wall, flange outward (the flange ends flat against the boss). Push the PTFE tube into the liner until it stops at the step inside.

7. Lip ring. Foot down in the groove on the body rim, lip leaning inward.

8. Filament first, spool second. Pull 30 to 40 cm of TPU off the spool. Reach into the body: the funnel mouth is a 12 mm dish on the inside of the wall near the top. Push the filament end into it and keep pushing until it comes out of the far end of the tube. (With the spool in, fingers cannot reach the mouth.)

9. Spool on. Slide the spool over the sleeve so the filament leaves the TOP of the spool heading toward the exit. Turn the spool to take up slack.

10. Lid. Arrow to arrow, drop it on, turn clockwise to the stop.

11. At the printer. Push the filament out of the tube end into the external inlet (or into the Y-splitter's external-spool hole) until the extruder grabs it, then seat the tube.

Fitting: the Slack Case (in this order)

1. Lid magnets. Glue four magnets into the four pillar tops in the base, any way up. Then for each pillar: let a loose magnet snap onto the pillar magnet, put a marker dot on its exposed face, slide it off sideways, and glue it into the lid pocket above that pillar (lid upside down) with the dot DOWN at the bottom of the pocket. The face that touched the pillar then faces the pillar again when the lid turns over.

2. Case floor magnets. Four pockets open inside the drum floor (under where the beads go). Glue a magnet into each, any way up for now; note which face is up.

3. Bracket on the printer. Unscrew the stock rear spool holder. Hold the bracket against the back panel with both screws loose at the far ends of the two slots, slide the whole bracket along the panel until its side plate touches the printer's left side, then tighten. The slots cover a screw spacing of 40 to 52 mm and a corner anywhere from 68 to 94 mm from the far hole; the bracket's inner corner is cut away where the printer's corner curves.

4. Bracket magnets, oriented from the case. Hold the empty case in the cradle (lid side toward you, ports down). Let a loose magnet snap onto the bracket plate over each pocket position (it pulls through the case floor onto the glued floor magnet), dot the face you can see, lift the case off, glue each magnet into its bracket pocket with the dot OUT, toward the case.

5. Lip ring into the groove, foot down. Beads into the drum, cork in. Liners into both ports, flange outward. Feeder tube into IN until it stops; short tube into OUT until it stops, other end into the splitter's external-spool hole or the printer inlet.

6. Case on: hold it level, lid toward you, ports down, line the bare wall up with the saddle and hooks, push straight toward the printer until the magnets click. Off: hold the base and pull straight out.

Loading and unloading on the X1C

The X1C only pulls an external spool in once you have pushed the filament to the extruder, and on unload it only cuts and backs the filament out of the extruder. The rest of the metre inside the printer you pull back by hand. Pull it back until the end sits inside the tube, then either lid off and wind it back onto the spool, or park it in the Slack Case (below).

With the Slack Case:

• Going to TPU: lid off, push the TPU end from IN across the case into the OUT funnel and on through the tube into the printer until the extruder takes it; lid on.

• Done with TPU: unload on the screen, wait for the cut, lid off, pull the filament back through OUT until the end drops into the case, wind it round the drum under the lip, end into the notch, lid on. The dry box stays shut; the filament stays threaded through the feeder.

What is proven and what is not

The feeder:

• Proven on my printer: bearings press in as described, magnets hold the drum on, the drum spins freely with a spool on it, the liner grips a 4 mm tube, the filament feeds to the extruder without catching, several TPU jobs printed through it, and the hygrometer window and label card work as on the Spool Puck.

• Not tested: how long the beads stay dry (weeks, months); I'm still waiting to see. Other bearing brands (the ribs are sized for a 22.0 mm bearing; a 22.1 bearing will be tighter). Spools narrower than about 50 mm or wider than 70. A 0.4 mm nozzle. Other TPU brands for the liner (hole size varies with flow).

• With this box and the slack case together, the TPU never sits in open air.

The Slack Case:

• Proven on my X1C: bracket screwed on and sitting flat, case docks and stays in place well, tubes connected, TPU threaded, a print completed, unload and the pull-back worked, the metre stored in the case.

• Not proven: any other X1C.

• Not good yet: the inside. Winding onto the drum, hooking the end into the notch and working around the cork is fiddly with the case hanging on the wall. A later version would open this up. This is the main reason the design is marked experimental.

• Not tested: long-term seal, the magnets' hold over months, printers other than the X1C (the P1S back panel is similar but I have not checked it).

What could go wrong

The feeder:

• Bearings too tight. Printed holes come out small. The pocket is 22.6 mm with six 0.3 mm ribs for the bearing to shave; if your print is tight, a few strokes of a round file on the ribs is the fix. Never hammer the bearings: the roof above them is a thin cone.

• Drum will not pull down or pushes back: the sleeve magnet is upside down. It is glued, so test the polarity on the hex key first as described.

• Tube loose in the liner: TPU holes vary with the brand and flow. The liner's bore is 4.2 as drawn and takes a 4 mm tube with friction on my printer. If yours is loose, scale the liner 98 % in the slicer's X and Y.

• Hub sleeve print: 83 mm tall and narrow. The zero-gap brim is set in the file; print it alone, not next to another part, and cut the brim off with a knife.

• Clear lid: glue stick on the smooth plate is there to stop it sticking too hard. Let it cool before lifting.

• Wet TPU prints stringy and the cork and liner come out rough. Dry it first (this box exists for that reason).

The Slack Case:

• Bracket does not fit your printer. The slots absorb a lot, but the flange must clear the AMS cable plug below the holes and the inlet fitting above them; on mine the flange runs from 10 mm below the holes to 25 above. Check against your back panel before printing.

• Lid hard to peel. Four 10 x 2 pairs at 0.2 mm apart pull with several kilograms. If the ear is not enough, leave two of the four lid pockets empty (the opposite pair); do not enlarge the ear.

• Peeling the lid pulls the case out of the cradle. The docking magnets pull through 1.1 mm of floor and are weaker than the lid magnets. Hold the base with the other hand while peeling. If it does come out it slides along the saddle and cannot fall.

• Case too tight or too loose in the cradle. Saddle and hooks are 0.4 mm off the case on radius. Tight: sand inside the saddle. Loose: the magnets take up the rattle.

• Bracket print. It stands as an L; the two hooks climb out of the plate at 45 degrees and their undersides are the roughest surfaces on the part. They do not carry the case; the saddle does.

• Magnets in backwards. All sixteen are glued, so follow the dot method above and check each pair before the glue sets.

Design notes

Own design, a parametric script on top of my Spool Puck. Every underside is a 45 degree chamfer or cone, so no part needs supports; the only bridges are the 6 mm vent slots in the drum. The bearing pocket roof is a cone that touches only the outer ring of the top bearing, so the fixed inner rings are never clamped. The Slack Case reuses the feeder's seal, tube ports, liners and cork; the bracket's hooks start at the plate and climb at 45 degrees so it prints standing with no supports.

Companion design by the same account: Spool Puck (the plain sealed box).