A 3D-printed rack bracket lives a strange life. It is asked to hold a 1.2 kg switch perfectly square against a rail at 44.45 mm of vertical pitch, to do that for years, and to do it inside a metal cabinet whose internal air can sit ten to fifteen degrees above room temperature for most of the working day. That is not where most filament samples are tested. Spec sheets are written for a 23°C lab on a rainy Tuesday in Germany, not for the top of a closet that hit 41°C last August because the office HVAC cut out at 6 PM Friday. This post is the long answer to a short question: PETG, ASA, or polycarbonate — which one actually belongs in a rack?
The short answer, for readers who only have a minute: PETG is fine for most homelab brackets, ASA is the right default for anything that sees sustained warm-air exposure or sunlight, and polycarbonate is reserved for the handful of cases where the bracket genuinely sees structural load near a hot exhaust. The rest of the post is the reasoning, the failure modes, and the numbers that put each filament in its slot.
The thermal envelope a rack actually presents
Before any filament discussion, it is worth pinning down what "inside a rack" really means thermally. A small wall-mount cabinet in a temperate office holds inlet air at roughly 22–26°C during business hours. The air leaving a stack of switches and a small NVR will be 8–15°C above inlet. That puts the air immediately above the warmest device somewhere in the 30–41°C band on a normal day. A bracket clamped to a switch chassis with internal hotspot temperatures north of 60°C can see localized contact temperatures of 45–55°C on the surface where it touches the device.
Add a closet on an exterior west-facing wall, or a homelab in an unconditioned attic, and the design number creeps up. A reasonable worst-case envelope to design against is 60°C of sustained air temperature for a closet bracket and 70°C of contact temperature for a bracket directly clamped to an enclosed switch chassis. That is the bar each filament has to clear without softening, creeping, or warping in a way that lets the device sag out of square over months.
PETG: the practical default
Polyethylene terephthalate glycol — PETG — is the workhorse of the 3D-printing world for good reasons. It prints reliably on nearly any printer, bonds well between layers, does not warp aggressively on the bed, costs roughly $20 per kg in bulk, and has impact resistance that is genuinely useful. For a bracket that holds a Cloud Key, a Flex Mini, or a fanless mini-PC in a normal-temperature homelab, PETG is the right answer almost every time.
The number to keep in mind for PETG is its glass transition temperature: roughly 80°C, depending on the formulation. Glass transition is not melting; it is the point at which an amorphous polymer goes from rigid to gradually rubbery. Long before reaching that line, PETG starts to creep under sustained load. Heat deflection temperature at 0.45 MPa — the standard low-stress measurement — sits around 70°C for most untreated PETG. At 1.8 MPa it drops into the high 60s.
What this means in practice: a PETG bracket loaded with a 1 kg switch at 40°C air temperature will hold its geometry indefinitely. The same bracket sitting at 60°C in a poorly ventilated cabinet, with sustained mechanical preload from cable bundles pulling on the back of the device, will slowly creep over months. You will not notice until you go to remove the switch and the screw holes have ovalled by half a millimetre.
PETG also yellows under UV. A closet bracket never sees UV, so this is irrelevant indoors. But a bracket installed in a garage rack near a window, or anything outdoors, will visibly discolour within a year and become brittle within two.
Use PETG for: brackets in conditioned homelabs, brackets that hold low-wattage devices, brackets where you can replace the part in five minutes if it does fail. It covers about 70% of homelab use cases honestly.
ASA: the right default for closets
Acrylonitrile styrene acrylate — ASA — is the filament most users underestimate. It prints with more drama than PETG, demanding an enclosure and good extruder temperature control to avoid layer splits and warping. Once it is printed correctly, though, it does almost everything PETG does and adds two properties that matter in a rack: meaningfully higher heat resistance and complete UV stability.
ASA's glass transition is around 100°C, with heat deflection at 0.45 MPa near 95°C and at 1.8 MPa near 85°C. That is a clean 15–20°C buffer over PETG. A bracket that is borderline in PETG at 60°C air temperature is well inside its safe operating envelope in ASA. The polymer is also genuinely UV stable — the "A" at the end of ASA replaces the butadiene in ABS with an acrylate ester, which removes the failure mode that makes ABS yellow and crack in sunlight.
The ASA tradeoffs are real. The filament costs about $25–$30 per kg, a small premium over PETG. It prints best in an enclosure at 240–260°C nozzle and 90–100°C bed; an open-frame printer in a cold garage will give you splits and curled corners. The smell during printing is noticeable — not as harsh as ABS, but present — so the printer needs to be in a ventilated space or, ideally, an enclosure with a small carbon filter. Layer adhesion in ASA is slightly weaker than in PETG at the same wall thickness, so wall counts often go from three to four perimeters for the same target strength.
For a rack-mount bracket destined for an SMB closet that may or may not always be in spec, ASA is the right default. It survives the design envelope of 60°C sustained air with margin, holds up under cable preload, and does not slowly turn the colour of an old laptop palmrest. The bracket designs in the 3D Rack Mounts catalogue that ship in ASA do so because the thermal headroom is worth the extra print difficulty for the buyer who is installing them inside a metal cabinet they cannot easily re-cool.
Polycarbonate: when the math says you need it
Polycarbonate — PC — is the heaviest hammer in the consumer-filament toolbox. Glass transition is around 147°C; heat deflection at 0.45 MPa sits north of 130°C; tensile strength is roughly double that of PETG. It is, on every spec-sheet axis that matters for a rack bracket, the best filament you can buy at the prosumer price point.
And it is, for most rack brackets, overkill.
PC is hard to print well. It demands nozzle temperatures of 260–290°C, which means an all-metal hotend and ideally a hardened nozzle for any blended PC like PC-CF. It needs a bed at 100–120°C and an enclosure capable of holding 50–60°C chamber temperature, or the part will warp and crack along layer lines from differential cooling. It absorbs moisture aggressively — a spool left out overnight in a humid room will print poorly the next morning unless you dry it. And it costs $40–$60 per kg for the well-behaved blends.
The cases where the cost and trouble are worth it are narrow but real. A bracket clamped to a 10G Pro XG 8 PoE switch that runs at 78°C on the chassis top under full PoE load, in a cabinet that sees 50°C internal air on a bad summer day, is a candidate. A bracket holding a piece of gear inside a sealed outdoor enclosure where direct sunlight pushes the inside to 65°C is a candidate. A bracket that has to absorb a real mechanical shock without cracking — say, on a mobile rack in a truck or trailer — is a candidate. For the closet rack in a spare bedroom, PC is, almost always, more filament than the job needs.
One subtle gotcha with PC: while pure PC is incredibly tough, the carbon-fibre-filled blends marketed as PC-CF trade some of that toughness for stiffness. PC-CF prints more easily, holds geometry better, and is what most "engineering" rack-bracket products are actually made of. Read the datasheet, not the marketing.
Side-by-side: the numbers that matter
For comparison at a glance, here are the spec ranges that actually drive bracket design decisions. Values vary by brand; treat these as a representative band, not a single point.
- PETG: glass transition ~80°C, HDT at 0.45 MPa ~70°C, tensile strength ~45 MPa, impact resistance high, UV stability poor, print difficulty low, cost ~$20/kg.
- ASA: glass transition ~100°C, HDT at 0.45 MPa ~95°C, tensile strength ~45 MPa, impact resistance good, UV stability excellent, print difficulty moderate (enclosure required), cost ~$25–$30/kg.
- PC: glass transition ~147°C, HDT at 0.45 MPa ~130°C+, tensile strength ~70 MPa, impact resistance excellent, UV stability moderate, print difficulty high (heated chamber recommended), cost ~$40–$60/kg.
The temperature columns are the headline. A PETG bracket and an ASA bracket can look identical, weigh the same, and behave identically for two years in a cool homelab. The difference shows up on the bad week, when the room hits 40°C and the cabinet pushes 55°C internal, with cable preload steadily pulling on the back of the device. The PETG bracket starts to creep — slowly enough that you do not see it happening, fast enough that next time you pull the switch out, the holes are oval.
Layer orientation matters more than filament choice
A bracket printed in the wrong orientation will fail in any filament. Rack brackets fail in two ways: the screw bosses delaminate from the body, or the mounting flange snaps away from the cradle. Both failure modes track the layer boundaries.
Print the bracket so the load on the mounting flange runs across multiple layers, not along a single layer line. For most 1U brackets that means the bracket sits flat on the bed with the rack-facing flange on its side, so screws pull the flange in shear across a stack of layers rather than peel a single layer off the body. A PETG bracket printed this way will outperform an ASA bracket printed the wrong way. Filament choice is the second variable; orientation is the first. Wall count and infill come third — three perimeters at 30% gyroid does most of the work on a thin bracket.
What about ABS, PLA, and the others
PLA is wrong for a rack bracket. PLA's glass transition is 60°C. A PLA bracket on a desk holds its shape forever; a PLA bracket in a closet that hits 50°C on a hot day will sag, slowly and irreversibly. Every few months a homelabber posts a photo of a drooping PLA bracket with a switch tilting out of the rack. It is the filament being asked to do something it cannot do. PLA is fine for prototypes you intend to reprint in something else.
ABS is the predecessor to ASA and broadly worse for this application. Similar glass transition, but it yellows and cracks under UV, smells more during printing, and warps more on the bed. ASA exists to replace ABS in everything except a few niche industrial applications. Nylon and PETG-CF are real options at the upper end — nylon for toughness if you can dry it, PETG-CF as a stiffer, slightly more heat-tolerant variant for brackets where stiffness matters more than impact resistance.
A decision rule for the next bracket
Boil all of the above down to a working rule for the next bracket you design or buy:
- Bracket in a conditioned room, light load, no UV exposure: PETG. Save the money and the print time.
- Bracket in a closet that might warm up, or near a warm exhaust, or in any installation you cannot easily get back to: ASA. The thermal margin is worth the small premium.
- Bracket clamped to a hot chassis, or in an environment with sustained 50°C+ air, or carrying real structural load: PC, ideally PC-CF.
- Any outdoor or sunlit installation: ASA at minimum, PC if it also sees heat.
- Never PLA.
The decision rule is conservative on the warm side because the failure mode is silent. A bracket that creeps does not announce itself; it just leaves you with an oval screw hole the next time you touch the rack. Spending five dollars more on the filament avoids that conversation.
Wrap-up
PETG, ASA, and PC are all real engineering filaments and they all earn their place in a rack-bracket catalogue. The distinction that matters is thermal margin against the environment the bracket will actually live in, not against the environment the spec sheet was written in. A homelab in a cool basement is honestly PETG country. A homelab in a converted attic, or a wall-mount cabinet in a small office, sits in ASA's sweet spot. The handful of installations that justify PC are real but rare.
If you remember one thing from this article, make it the orientation point: a 3D-printed bracket fails along its layers, so the print orientation matters more than the filament name on the spool. Get that right first, pick the filament second.
