Every number in this episode, with the source it came from and the date it was checked. Standards are cited in the revisions current at 2026-09-02.
White XLamp LEDs over 40% efficient, roughly 50 to 60% of input power leaving as heat; the conservative 25% light / 75% heat sizing assumption (Pt = 0.75 x Vf x If); TIM design target below 1 degC/W.
LER 323 lm/W for a typical 3000 K, 80 CRI phosphor-converted white and 414 lm/W for an ideal colour-mixed white; the 165 lm/W warm-white 2020 package value at 25 degC; efficacy = LER x power conversion efficiency; Stokes loss releasing heat in the phosphor.
Measured board plus mounting-interface resistance, XP-E at 0.7 A: 3.81 degC/W metal-core against 8.40 degC/W for 1.6 mm FR-4 with 14 thermal vias in 4 oz copper. Heat-sink-to-ambient 13.60 degC/W; with the 5.8 degC/W datasheet package that is the 25 / 16 / 59 percent split. Via technique not recommended above 5 W per LED.
Defines the package-level junction-to-case link only. Cited for the statement that Rth names one link and that vendors publish junction-to-case or junction-to-solder-point, which are not the same number.
Thermal conductivity by TEMPER at 20 degC: 1100-O 222; 6063 O 218, T1 193, T6 201; 6061 O 180, T4 154, T6 167 W/m.K. Kaiser prints no 6063-T5 thermal value, which is why the table uses T6 201. Also %IACS beside every thermal figure.
As-cast typical values for separately cast test bars: A380 family 96.2, A360 113, A413 121, B390 134 W/m.K, no test temperature printed. Table A-3-3 rates die-filling capacity best for the highest-silicon alloys.
Optimum fin spacing about 6 to 10 mm of free space for a 100 x 100 mm sink with 45 mm fins; below that band the natural-convection airflow is obstructed.
The mass-optimal versus maximum-dissipation geometries, and the finding that for each fin thickness and spacing pair there is a height beyond which fin efficiency and a falling coefficient make more fin add nothing.
Vertical fin-array data scaling as about H^0.68 (fitted on 15 and 25 mm fins), so doubling fin height buys about 60% more, and the vertical orientation gives the highest convective rates. Publisher record: doi 10.1016/j.applthermaleng.2012.10.043.
A SEPARATE study from [3], on a GaN inverter: a 0.84 mm FR-4 board estimated at 3.76 degC/W against 4.23 degC/W for that IMS build, whose 0.152 mm dielectric was thicker than needed; heat-sink term 31.8 of 39.1 degC/W per FET, so the heat sink dominates.
A 76 micrometre dielectric at 0.58 degC.cm2/W, against a 1.6 mm aluminium base at 150 to 200 W/m.K (0.08 to 0.11 degC.cm2/W): the dielectric is the larger term through the stack, about five to seven times. Mirror URL; the product line now sits at tclad.com.
Bench test on a 6 W thermal test die under natural convection: a dry joint at 2.9 degC/W case to sink against 0.9 degC/W with grease, the whole junction-to-ambient path 8.1 to 9.9 degC/W, about 22%. Also the 5 W/mK elastomer at twice the grease resistance with under 30% contact.
Joint resistance falling with contact pressure, and the ranking of interface materials by degC.cm2/W: why a TIM is specified at a stated pressure and area, not by bulk W/m.K.
Anodised aluminium measured at 0.83 to 0.86; a fresh untreated extrusion measured 0.02, and the authors recommend 0.05 for new untreated aluminium.
On a small natural-convection heatsink at emissivity 0.85, radiation carried 28.3 to 30.4% of total dissipation across 5 to 25 W; the bare-metal case (handbook 0.09) ran close to 30 degC hotter. Source of the 0.09 / 0.85 bar pair.
Solar absorptance and infrared emittance are different properties in different wavebands. Cited for the daytime solar-load point on dark outdoor finishes, not as luminaire measurements.
0.4% cooling design dry bulb 45.0 degC; 50-year extreme dry bulb 49.8 degC. The 2021 edition gives 44.9 and 48.3 degC. The 44.2 degC figure printed in an earlier build was the 2005 Handbook value for a different station and is retired.
The isothermal optimum-spacing form S_opt = 2.714 L / Ra_L^0.25 that the slide 6 curve is checked against. (The 2003 paper [8] prints 2.66 and 2.677 for its own arrays.)
Edition 10.0, published November 2024, superseding the 2020 ninth edition; the current active revision. ta is marked only when other than 25 degC; tc is a component case rating distinct from ta. NOTE ON VERIFICATION: the thermal-test and attachment provisions summarised in this episode were read in the 2008 and 2020 editions; clause and annex numbering changed between editions, so no clause number is quoted anywhere in this episode.
The approved method for the in-situ temperature measurement a specifier should ask for, at the temperature measurement point of the hottest LED, with an elevated-ambient procedure. It says nothing about access holes, resealing or IP inference: that sentence on slide 12 is NLC practice.
Flux and colour maintenance of LED packages, arrays and modules, indexed by controlled case (TMP) temperature. It reports the temperature the test HELD, not the temperature a package reaches inside your luminaire, and junction temperature is not the reported quantity.
Edition 2.0, published 2023-01-24. The manufacturer declares tq, the highest ambient temperature around the luminaire related to a rated performance, and it must be reported even when it is 25 degC.
Thermal conductivity 226 Btu/sq ft/ft hr/degF at 68 degF, equal to 391 W/m.K at 20 degC. The copper reference row.
The extrusion process: one constant cross-section pushed through a die; open ends, so a sealed shell needs machining, end caps or another process.
Silicon in solid solution scatters the electrons that carry heat in aluminium casting alloys.
Companion source for the silicon-in-solution mechanism and its interaction with heat treatment.
The oxide film, not the dye, sets infrared emissivity: clear anodising radiates nearly as well as black.
Anodised aluminium: black 0.82 and 0.86, clear 0.76 and 0.84. Aluminium, highly polished 0.039 to 0.057.
Deep, closely spaced fins largely exchange radiation with each other; radiative output does not scale with total wetted fin area.
The channel view factor between fin walls, base and surroundings that governs a finned heat sink's radiative output.
Eddy-current %IACS screening of an incoming casting against its declared value; for aluminium heat-treatment evaluation it is used together with a hardness examination. It does not identify an alloy on its own.