Polycarbonate on a Desktop Printer: Toughness, Heat and the Moisture Problem Nobody Warns You About

Ask ten makers which filament is tough and most will say PETG. Ask which one survives a dropped enclosure or a hammer blow and the honest answer is different: polycarbonate. PC is the material that clips, guards, brackets and tool housings are actually made of in injection moulding, and it is available for desktop printers. It is also the material most likely to be abandoned after two spools, because it asks for conditions most printers do not have by default — a genuinely hot hotend, a chamber, an aggressive first layer and bone-dry resin.

This is not a review of one brand's PC. It is a working explanation of what the polymer is doing while it prints, why the failures look the way they do, and how to design and run parts so the toughness you paid for is still there in the finished object.

What Polycarbonate Actually Is, And Why That Matters At The Nozzle

PC is a polyester of bisphenol A and phosgene. The backbone is stiff and regular, the chains pack tightly, and that is the source of both its virtues and its problems. Stiff aromatic chains with strong intermolecular attraction mean:

• a high glass transition (Tg around 145–150 °C), far above ABS at roughly 105 °C;
• high notched impact strength, and impact that stays high at low temperature rather than going brittle;
• resistance to creep under sustained load that PETG simply does not have;
• strong absorption of water into the amorphous bulk, and a sharp drop in melt viscosity when that water is present.

That last point is the whole game. PLA and PETG also absorb moisture, but they print "merely worse" when wet. PC prints wrong: the water flashes to steam inside the melt, the melt loses viscosity and the extrusion becomes a foam. You get a rough, matte, silvery surface, spitting and popping, and — most importantly — a part with voids in the bead. A part with voids has no impact strength, because toughness depends on uninterrupted material to carry the crack tip. You paid for polycarbonate and shipped something closer to a cracker.

Practical consequence: dry PC first, always. 80 °C for four to six hours, or 70 °C overnight. Not 50 °C, which barely moves the water. If you want the full reasoning on why some polymers forgive moisture and others do not, the material-by-material moisture guide covers the mechanism; the storage and rotation side is in storing and reviving filament stock. For a workshop the simplest reliable habit is a dry box that the spool lives in rather than visits, like the FilaBox drying box or, for several spools at once, a drying and storage cabinet.

Why PC Needs A Hot End That Is Actually Hot

PC extrudes around 260–290 °C, and the top of that range is not decoration. At the low end the melt is still viscous enough that layer adhesion is weak — the bead lands, cools before it has wet the layer beneath, and you get a part that looks perfect and delaminates along the Z axis the first time it is loaded. The range that works depends on the specific grade and on print speed, but the principle is fixed: PC wants to be printed hot and fast enough that the interface is still molten when the next bead arrives.

Three hardware limits bite here:

1. Hotend temperature ceiling. An all-metal hotend rated to 300 °C is a requirement, not a suggestion. PTFE-lined hotends degrade above roughly 250 °C and are unsafe as well as unreliable.
2. Heater power and flow. At 280 °C with a 0.6 mm nozzle you are pushing a lot of heat into the melt zone. A weak heater shows up as temperature droop mid-print, which shows up as inconsistent layer bonding. Larger nozzles and higher flow demand more power, not just more temperature.
3. Nozzle wear and geometry. PC is abrasive enough that hardened or tungsten nozzles are common, and a worn nozzle changes the extrusion width quietly. The trade-offs are covered in hotend upgrades worth the money and, for diameter and flow selection, nozzles and hotends decoded.

A starting point that behaves reasonably across most desktop machines:

Parameter	Starting value	What to change if it fails
Nozzle temperature	270–285 °C	Raise 5 °C if layers separate; lower if stringing is severe
Bed temperature	100–110 °C	Hold at 100 °C minimum; below that, corners lift
Chamber temperature	45–55 °C	Passive enclosure is usually enough; avoid hot air blowing on the part
Part cooling fan	0–20 %	Effectively off. Fan on PC is a delamination generator
Print speed	30–50 mm/s (0.4 mm nozzle)	Faster needs more temperature, not less
First layer speed	15–20 mm/s	Slower helps adhesion; too slow invites elephant's foot
Nozzle	0.6 mm preferred	Larger nozzle improves layer bonding and impact strength

If your machine cannot hold 280 °C or cannot keep a chamber warm, the honest answer is that PC is the wrong material for that machine. Choosing a material your printer can execute well beats insisting on a datasheet number you cannot reach — the same logic that runs through matching material to project.

The Chamber Is Not Optional, And Not Only For Warping

Everyone knows PC warps. Fewer people know why it warps differently from ABS. PC has a high mould shrinkage and — critically — a high coefficient of thermal expansion combined with a stiff glassy state below Tg. When a bead cools from 280 °C to room temperature inside a cold room, it tries to contract; the layer below it is already stiff and resists; the stress has to go somewhere. On a long part it goes into curl. On a short part it goes into internal stress that only reveals itself when you cut, drill or drop the part.

A chamber at 45–55 °C does two things: it slows the cooling rate so the stress has time to relax while the polymer is still above its softening range, and it reduces the temperature gradient across the part. Both raise the effective strength of the finished object without changing a single slicer setting.

What a chamber does not fix is airflow. A part cooling fan or an auxiliary fan blowing room-temperature air across a PC part will delaminate it regardless of chamber temperature. Turn part cooling essentially off and let the chamber do the work. If you are new to enclosures, the reasoning in to enclose or not applies directly, including the fume management that PC also requires.

Bed Adhesion: PC Bonds To The Wrong Things Aggressively

PC adheres hard. On a smooth PEI sheet at 100 °C it can bond so completely that removing the part tears the sheet. This is the single most common expensive mistake with the material.

• Use a release layer. Glue stick on smooth PEI, or a dedicated PC-compatible surface. The glue is not there to make the part stick, it is there to let it come off.
• Let it cool. PC releases dramatically as it drops below 60 °C. Pulling a part off hot is how sheets die.
• Brim rather than raft. A 5–8 mm brim gives the corner the extra anchorage it needs without the surface damage a raft leaves.
• Watch first-layer squish. Over-squashed PC generates a wide foot that welds to the plate. A slightly higher first layer than you would use for PLA is often correct here.

Plate choice matters, and the comparison in build plate science is worth reading before you commit a sheet to PC. Both a Cold PEI plate and a smooth PEI plate work with PC — with a release agent, and with patience during cooldown.

The Failure That Looks Like A Settings Problem And Is Not

Here is where most people lose a week. The part prints, looks fine, and then:

• Delaminates along layer lines under a light load. Almost always: insufficient nozzle temperature for the speed, or fan on, or wet filament. The three produce nearly identical results, which is why diagnosis requires isolating them one at a time rather than turning three knobs at once.
• Cracks at a corner or hole weeks later, with no load applied. Classic residual stress plus a stress riser. The fix is geometric, not thermal: radii, and fewer sharp internal corners. The mechanics of where cracks start are in creep, fatigue and preload.
• Shrinks off dimension by 0.5–1 %. PC is not a dimensionally forgiving material. If the part is a fit, calibrate for it rather than trusting nominal CAD. Getting dimensionally accurate parts describes the calibration flow.
• Surface is matte and rough with popping sounds. Wet filament. Stop printing and dry it. No amount of temperature tuning fixes water.

The general principle deserves stating plainly: a PC part that looks good can still be structurally worthless. With PLA, surface quality is a decent proxy for strength. With PC it is not, because the failure modes — voids from moisture, weak interlayer welds from cold extrusion, residual stress from fast cooling — are all internal. Inspect by breaking a test coupon, not by admiring the surface. The inspection mindset in correct parts, wrong assembly is the right one: measure and test, do not assume.

Designing For PC: Where The Material Earns Its Price

PC is expensive and difficult, so use it where the properties are actually doing work:

1. Impact and drop. Enclosure shells, tool housings, guards, brackets on equipment that gets knocked. This is PC's home ground and it is a large improvement over PETG, which is tough in tension but notch-sensitive in impact.
2. Heat near the machine. Parts near a heated bed, a motor, an extruder or an engine bay. A Tg near 145 °C means a PC bracket will not sag where a PETG one will slowly yield. Compare with sustained-load behaviour before assuming any polymer is safe near heat.
3. Transparent and translucent parts. PC prints translucent, not optically clear from FDM, but it is far clearer than PLA and can be sanded and polished. If clarity is the goal, resin and post-processing are usually a better route — see plates, nozzles and finishing consumables.
4. Parts that must not creep. A clamp, a load-bearing spacer, anything held under compression for months.

And where it does not:

• Decorative parts, jigs, prototypes — PLA+ is cheaper, easier and stiff enough.
• Large flat panels, where warping will win.
• Outdoor UV exposure. PC yellows and embrittles under UV unless it is a UV-stabilised grade. ASA is the better outdoor choice and the reasoning is in ASA versus ABS.
• Anything requiring chemical resistance to solvents, fuels or strong alkalis. PC is attacked by many of them.

When PC is not quite right, the usual next step is a blend or a fibre-filled grade. ABS-CF and PA6-CF20 trade some impact strength for stiffness and much easier printing; PPS-GF20 goes further on heat and chemicals at a much higher price and difficulty. A comparison of where the money goes is in which engineering filament earns the spool cost.

Geometry Rules That Matter More On PC Than On Anything Else

Because PC combines high shrinkage with a stiff glassy state, geometry decisions that are merely good practice on PLA become decisive here.

• Round every internal corner. Sharp internal corners concentrate stress and are where residual-stress cracks initiate. A radius of at least half the wall thickness, more where possible.
• Keep wall thickness even. Thick-to-thin transitions cool at different rates and pull against each other. This is the same mechanism behind warping, localised.
• Avoid large flat unsupported areas. If the part must have them, orient them vertically or add ribs. Rib geometry is covered in shell count, wall thickness and infill, and orientation strategy in strong parts are designed before they are sliced.
• Design for the anisotropic strength you will get. Interlayer bonds are the weak direction on every FDM part, PC included. Orient so that the main load runs in the XY plane. The full reasoning is in load paths, not shapes.
• Use fillets and chamfers, not sharp edges. A chamfer costs one line of CAD and removes the single most common crack initiation site.

Annealing: Sometimes Worth It, Often A Trap

Annealing PC can relieve residual stress and raise heat resistance further, but it also shrinks the part — typically a fraction of a percent, more on large parts — and it can warp a part that was previously flat. If dimensional accuracy matters, anneal a test piece first and measure, or do not anneal at all. The mechanism, the temperature windows and the cases where annealing makes parts worse are covered in detail in annealing 3D prints.

Hardware That Makes PC Manageable

PC is not a material to fight with a stock machine. In rough order of impact:

1. A dry box or drying cabinet. Non-negotiable. This alone eliminates the majority of "PC is impossible" reports.
2. An enclosure. Passive is fine for most parts; active heating is better for large ones.
3. A hotend rated for 300 °C. Look for all-metal construction and adequate heater power. A high-flow nozzle helps push the melt faster without drooping temperature.
4. A hardened or tungsten nozzle. Wear resistance and stable extrusion width over long prints. See the tungsten nozzle option.
5. A machine with real bed temperature headroom. 110 °C is not a peak rating to be visited briefly; it is a working temperature. A machine like the Bambu Lab H2D or a large industrial unit such as the ANTBULL656 is built for sustained high-temperature work; a machine rated to 100 °C bed temperature will struggle with PC on anything sizeable.

What PC Does Not Solve

Two honest limits. First, PC is not a metal replacement. A well-designed printed PC bracket is a good bracket; it is not a steel bracket. Second, PC's toughness is direction-dependent and void-dependent. A wet, cold-extruded, fan-cooled PC part can be weaker than a well-printed PETG part, and look better while being so. The material sets the ceiling; the process decides where under that ceiling you land.

If you need to go beyond what any desktop FDM machine can reach on temperature and crystalline engineering polymers, that is a different class of equipment entirely — see the PEEK-capable printer for what that step costs and requires.

A Short Checklist Before You Load The Spool

1. Dry the filament at 80 °C for 4–6 hours. No exceptions.
2. Confirm the hotend is all-metal and rated to at least 300 °C.
3. Confirm the bed reaches and holds 100–110 °C.
4. Close the chamber and let it stabilise at 45–55 °C before starting.
5. Set part cooling to zero or near zero.
6. Apply a release layer to the plate. Do not print PC bare on smooth PEI.
7. Add a 5–8 mm brim; skip the raft.
8. Print a small test coupon first: a 30 × 10 × 3 mm bar snapped by hand tells you more than a surface inspection.
9. Let the part cool fully before removing it.
10. Check the critical dimension against CAD before you assemble anything.

Do those ten things and PC stops being the material that defeated you and starts being the material you reach for when a part genuinely has to take a hit. The rest of the engineering-filament landscape — where PC sits relative to ABS, ASA, PETG and the filled grades — is mapped in inside the filament lineup. And if you are weighing a spool against a roll of something easier, the arithmetic in filament economics is the honest way to decide.