Fibre Core

From RepRap
Jump to: navigation, search
Crystal Clear action run.png
Fibre Core

Release status: concept


A hot end for co-extruding with embedded fibre reinforcing
CAD Models
External Link


So who has head of Pultrusion a way they make some continuous glass and carbon fibre composite extrusions, tent rods and such.

This is typically made by drawing a bunch of fibre filaments into a bath of epoxy or similar resin and then out through a nozzle/die and curing (usually heat) the wetted fibre bundle into a rod, tube or other profile.

Well imagine doing it with a 3D printer that lets you build a 3D model that has much increased tensile strength in every extruded path. This is achieved by co-extruding the fibre with the melted plastic.


Fibre wrapping system

First public domain disclosure 18 September 2013, so no longer patentable unless already pending.

One can add an aramid (Kevlar) thread feeder between the extruder gear and the hot end on a RepRap printer. Add a small hot point to tack the reinforcing filament to the plastic filament. Add a small (spider leg gadget that will zig-zag the thread or) winder to wrap it helically around the fibre and tack it to the 3mm filament so there is much more fibre length going into the hot end than plastic filament. To start you just tack it on and send it down. After a while it will be dragged to the melt zone and then out of the nozzle (hopefully automatically or else a manual start) when it will start to extrude at a rate multiplied by the ratio of feedstock cross sectional area divided by nozzle area ((3.0^2/0.45^2)=44 or 1.75^2/0.45^2=15) So for each mm of filament you would need 44 mm of fibre wound around it. For a 0.45mm diameter nozzle that would be about 4.5 turns per mm of 3mm diameter filament (2.7 turns/mm for 1.75mm dia). Have the hot point continuously weld the turns down one side so they feed evenly and do not uncoil (retracting not required really as bridging is done at extrusion speed with a reinforced core as usual. However the amount of thread wrapped around could be increased in calculated advance of a jump where you would extrude thread and very little plastic. The core fibre could be cut at places with shears to move to the next part, a conveyor bed would be useful it is not that easy to make hops from part to part. Using filament in a equal size nozzle would mean that the thread can travel together through the melt zone at full speed and extrude faster as quality would not matter much and might be very easy to test the possibilities.

This method of filament addition could theoretically be done without changes to the hot end of the extruder.

The addition of the filament could also be done outside the printer which would then require very little changes to the printer (firmware to prevent jumping and a quick remove nozzle). It could be a helical fibre or a knitted tube around the plastic that could even be added at the point of exit from the filament extruder (3mm) so it embeds tightly into the raw material. If pulled below the surface of the plastic the thread would be reasonably safe from destruction in a gentle extruder or it could receive an overcoat to fully encapsulate it. The proportion of thread added would have to be carefully controlled and selected for the chosen filament to nozzle area ratios. The helix or knit of fibre would then be pulled straight in the melt zone and extruded at the correct speed. This would be a brilliant consumable lock in opportunity for a manufacturer.

There may even be a way to prepare the filament by pull-truding a filament with very little plastic around it. Then in a second process it would be pushed into a significantly larger diameter bore while warm and fill it randomly line honey dribbling in a pile. This core would then be pulled to size through a die when warm and a final layer over moulded. It would result in a continuous fibre that is as long as the final extruded diameter but randomly jumbled inside.

Research has been done in the use of laser to cut fibres. Low power lasers (<10W) can be effective for darker organic fibres and carbon fibre might lend itself to being cut like that when having to to open jumps, bridging would be no problem as the span is supported by a tension member from the first pass.

Coaxial nozzle

Fibre co-extrusion

New disclosure 9 October 2013.

An alternate method of inserting the fibre is with a co-axial nozzle that will draw the reinforcing fibre into the centre of the extruded bead as it exits the head. A metering system would have to feed out fibre at the speed the nozzle is depositing though it would auto feed once the end was set in plastic. Starting might be easier, cutting might be easier, adjusting the rate of the fibre to path deposition ratio could allow for plastic free jumps that would only need trimming of the exposed fibres between parts or regions. A side port that feeds the fibre in a thin tube into the melt zone to open at the nozzle opening would be needed. The molten plastic must be able to flow around the internal fibre feed tube. It might be easier to test this method as the fibre could be drawn out by head movement.

Wire co-extrusion

This technique could theoretically be used to lay down conductive wires embedded in the plastic. The ends could be left exposed and if they had extra layers and turns of plastic around the extruded one they would be fully insulated. They would be suitable for extruding windings if alternated with plain plastic. This is basically how regular insulated wire is made, it is just a lot easier to to do when the wire is straight and moves at a constant speed. If it is accepted that the wire ends would need some further preparation then pre-insulated wire could be used in the extrusion process and it would be automatically insulated even if it were to touch an adjacent wire in a winding/coil formation when randomly positioned in the extruded plastic. The insulation in 'bell/magnet wire' is a varnish that can be removed by abrading or heating to make a connection at the ends, robust insulation displacement connectors might also be enough to make good contact. The extra insulation removal should not be a large burden in coil applications as it is only needed at the few terminal points. Multi-stranded wire is also an option, Litz wire is a very flexible wire used in many high frequency coil winding applications as it has a much larger surface area (each fine strand is individually insulated) than the same conductor cross sectional area of a solid conductor. It fetches a price premium and is somewhat harder to use usually but has a lower high frequency resistance for the same amount of copper. There are a few 3D printed motors where windings are inserted during interruptions in the build progress. With the co-extrusion it should be possible to make self supporting flat coils that are made of the same material as the rest of the part.

The Co-axial systems would require complex nozzle arrangements.

There is a range of insulated magnet wire (usually copper) that has the outer layer of coating made from a heat or solvent activated glue. There is a informative PDF available from [here].

Meltable Core

I was just wondering if one had two polymers that had been co-extruded at the factory like a flux core in solder but with different characteristics. Perhaps the one was stiffer with greater tensile strength and the other was softer with better adhesion. Having them extruded at the hot-end simultaneously might allow them to retain their linear arrangement and basically have a thin thread of hard plastic laid down with the softer stickier plastic.


  • Controlled direction reinforcing fibres in part
  • Selection of fibres that have higher melting temperatures than the extruded plastic
  • Allow for massive bridging distances due to support core
  • Plastic matrix may be selected for bonding ease instead of tensile strength


  • Increased complexity of extrusion head
  • Second consumable that needs maintaining
  • More complexity for starting new material end
  • No easy way to jump without a connecting thread


Try and fit it all into a small space. Might work better with a Bowden setup. New slicer software would be needed to minimise jumps. Firmware would have to synchronise the winder with the filament drive. A look ahead would be needed to insert extra filament in advance if fibre only jumps were wanted.

Relevant Forum Threads

Other links