Validation.

How close MeshWeave’s heat path comes to the numbers on vendor datasheets and in published measurements. Thermal pastes, gap pads, phase-change films, solders, sintered silver, braze, bought heat sinks, heat pipes and chip packages on a test board: each one typed into the app the way you would type it, and the answer set beside the published one.

15 of 60 scored values within tolerance; 19 shown but not scored, each with the reason. Run on v3.1003, 2026-11-20. No case is left out for missing, and no tolerance was widened after the run: they were set before it.

Thermal pastes, pads and phase-change filmsJoint resistance per area

Each sheet’s conductivity and bond line (for a gap pad, its thickness once squeezed) typed in as a stated layer. The sheets measure the whole joint, the two surfaces it touches included, so a layer worked out from conductivity alone should read low. How low is the point of the table.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
Shin-Etsu MicroSi G751R'' (laser flash)20 psi, bond line 73 µm (datasheet), k 4.7 W/mK17 K·mm²/W15.5 K·mm²/W−8.6 %±15 %within
Shin-Etsu MicroSi X23-7921-5R'' (laser flash)20 psi, bond line 25 µm (datasheet), k 6.0 W/mK5 K·mm²/W4.17 K·mm²/W−16.7 %±15 %outside
BERGQUIST GAP PAD TGP 1000HDThermal impedance (D5470)10 % deflection: 1.016 mm pad squeezed to 0.914 mm, k 1 W/mK1,097 K·mm²/W914 K·mm²/W−16.6 %±15 %outside
Thermal impedance (D5470)20 % deflection: 1.016 mm pad squeezed to 0.813 mm, k 1 W/mK1,026 K·mm²/W813 K·mm²/W−20.8 %±15 %outside
Thermal impedance (D5470)30 % deflection: 1.016 mm pad squeezed to 0.711 mm, k 1 W/mK948 K·mm²/W711 K·mm²/W−25.0 %±15 %outside
BERGQUIST GAP PAD TGP 1300Thermal impedance (D5470)10 % deflection: 1.016 mm pad squeezed to 0.914 mm, k 1.3 W/mK1,090 K·mm²/W703 K·mm²/W−35.5 %±15 %outside
Thermal impedance (D5470)20 % deflection: 1.016 mm pad squeezed to 0.813 mm, k 1.3 W/mK91 K·mm²/W625 K·mm²/W−31.3 %±15 %outside
Thermal impedance (D5470)30 % deflection: 1.016 mm pad squeezed to 0.711 mm, k 1.3 W/mK813 K·mm²/W547 K·mm²/W−32.7 %±15 %outside
BERGQUIST GAP PAD TGP 1500RThermal impedance (D5470)10 % deflection: 0.508 mm pad squeezed to 0.457 mm, k 1.5 W/mK69 K·mm²/W305 K·mm²/W−55.8 %±15 %outside
Thermal impedance (D5470)20 % deflection: 0.508 mm pad squeezed to 0.406 mm, k 1.5 W/mK568 K·mm²/W271 K·mm²/W−52.3 %±15 %outside
Thermal impedance (D5470)30 % deflection: 0.508 mm pad squeezed to 0.356 mm, k 1.5 W/mK529 K·mm²/W237 K·mm²/W−55.2 %±15 %outside
BERGQUIST GAP PAD TGP 2000Thermal impedance (D5470)10 % deflection: 1.016 mm pad squeezed to 0.914 mm, k 2 W/mK626 K·mm²/W457 K·mm²/W−26.9 %±15 %outside
Thermal impedance (D5470)20 % deflection: 1.016 mm pad squeezed to 0.813 mm, k 2 W/mK574 K·mm²/W406 K·mm²/W−29.2 %±15 %outside
Thermal impedance (D5470)30 % deflection: 1.016 mm pad squeezed to 0.711 mm, k 2 W/mK516 K·mm²/W356 K·mm²/W−31.1 %±15 %outside
BERGQUIST GAP PAD TGP 5000 APThermal impedance (D5470)10 % deflection: 1.016 mm pad squeezed to 0.914 mm, k 5 W/mK265 K·mm²/W183 K·mm²/W−30.9 %±15 %outside
Thermal impedance (D5470)20 % deflection: 1.016 mm pad squeezed to 0.813 mm, k 5 W/mK219 K·mm²/W163 K·mm²/W−25.9 %±15 %outside
Thermal impedance (D5470)30 % deflection: 1.016 mm pad squeezed to 0.711 mm, k 5 W/mK194 K·mm²/W142 K·mm²/W−26.5 %±15 %outside
BERGQUIST GAP PAD TGP A2000Thermal impedance (D5470)10 % deflection: 1.016 mm pad squeezed to 0.914 mm, k 2 W/mK671 K·mm²/W457 K·mm²/W−31.9 %±15 %outside
Thermal impedance (D5470)20 % deflection: 1.016 mm pad squeezed to 0.813 mm, k 2 W/mK645 K·mm²/W406 K·mm²/W−37.0 %±15 %outside
Thermal impedance (D5470)30 % deflection: 1.016 mm pad squeezed to 0.711 mm, k 2 W/mK613 K·mm²/W356 K·mm²/W−42.0 %±15 %outside
BERGQUIST HI FLOW THF 1400PThermal impedance (D5470)bond line 0.001 in (0.0254 mm), k 1.4 W/mK; published at 50 psi (10-200 psi spans ±10 %)11 K·mm²/W18.1 K·mm²/W−83.5 %±15 %outside
Thermal impedance (D5470)bond line 0.0015 in (0.0381 mm), k 1.4 W/mK; published at 50 psi (10-200 psi spans ±10 %)135 K·mm²/W27.2 K·mm²/W−79.9 %±15 %outside
Thermal impedance (D5470)bond line 0.002 in (0.0508 mm), k 1.4 W/mK; published at 50 psi (10-200 psi spans ±10 %)155 K·mm²/W36.3 K·mm²/W−76.6 %±15 %outside

What it says: for greases at a thin bond line, conductivity alone comes within 9 to 17 %. For gap pads conductivity alone reads 17 to 56 % low, and for a phase-change film almost all of its resistance is at the surfaces, so conductivity alone reads about a fifth of it. For a pad or a film, type the datasheet’s own impedance as a stated R′′, not its conductivity.

Die attach: solder against sintered silverEffective conductivity

The joint library’s solder and sintered-silver entries, picked in the card, against conductivities measured on real bond lines.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
Understand thermal characterization of high-power LEDs for reliable SSLSAC305 effective kslope of R against bond line, 20 LEDs61.6 W/mK57.3 W/mK−7.0 %±15 %within
Sintered silver effective kprocess not stated; compared with the pressureless entry (chosen before the run: LED die attach is normally pressureless)207 W/mK125 W/mK−39.6 %±15 %outside
Thermal characterization methodology for thin bond-line interfaces with high conductive materialsPressureless sintered silver effective ktransient method, thin bond line115 W/mK125 W/mK+8.7 %±15 %within
SAC305 effective ktransient method, thin bond line55 W/mK57.3 W/mK+4.1 %±15 %within
LOCTITE ABLESTIK SSP 2020Sintered silver k, stated as a lower bound (>100 W/mK)pressureless profile, 250 °Cover 1 W/mK125 W/mK——within

What it says: the solder entry is within 7 % of both measurements. The pressureless silver entry matches the pressureless measurement within 9 % and is 40 % under the LED study, whose silver process is not stated; a denser silver would read higher.

Braze jointsFew published numbers exist

Measured braze-joint resistances are scarce in anything public. Two rows can be compared; the rest are listed below with the reason they cannot.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
Direct Metal-to-Metal Contact Thermal Resistance in Cold Plate DesignBrazed Al joint R'', stated as a range 0.1-1 K·mm²/Wlibrary default bond line; low-confidence source (range only)0.10 to 1 K·mm²/W0.38 K·mm²/W——within
Wesgo/Morgan Cusil (BAg-8) data sheet, 06/2018Ag-Cu eutectic braze k (bulk)this datasheet is where the library number comes from: shown, not a check371 W/mK371 W/mK+0.0 %—not scored the library entry is taken from this datasheet

What it says: the aluminium braze sits inside the only published range found, and that range is wide and from a low-confidence source. The silver-copper braze row is where the library’s number comes from, so it is shown and not scored.

Heat sinks against airflowSink resistance, base to air

Seven bought BGA heat sinks rebuilt from their drawings as a fin design in the card: base size and thickness, fin height, thickness, gap and cut length, at the datasheet’s air speeds, with the heat spread over the whole base. The drawings leave some fin sizes open, so each row gives the weight of the fins as read against the weight on the datasheet.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
Same Sky (ex-CUI Devices) HSB02-101007 BGA heat sinkRth, still airnatural convection, 1 W. Fins as read weigh 0.90 g with the base; datasheet 1.78 g.35.3 K/W86.8 K/W+145.9 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 0.90 g with the base; datasheet 1.78 g.13.6 K/W11 K/W+708.6 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 0.90 g with the base; datasheet 1.78 g.13.6 K/W17 K/W+25.2 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 0.90 g with the base; datasheet 1.78 g.9.6 K/W37.9 K/W+294.8 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 0.90 g with the base; datasheet 1.78 g.9.6 K/W10.2 K/W+6.6 %±25 %within
Same Sky HSB28-606022 BGA heat sink with push pins,Rth, still airnatural convection, 1 W. Fins as read weigh 119 g with the base; datasheet 108 g.4.5 K/W10.6 K/W+134.8 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 119 g with the base; datasheet 108 g.1.7 K/W16.2 K/W+854.3 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 119 g with the base; datasheet 108 g.1.7 K/W0.79 K/W−53.5 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 119 g with the base; datasheet 108 g.1.1 K/W5.41 K/W+391.9 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 119 g with the base; datasheet 108 g.1.1 K/W0.43 K/W−61.2 %±25 %outside
Same Sky HSB33-272710 low-profile BGA heat sink,Rth, still airnatural convection, 1 W. Fins as read weigh 6.3 g with the base; datasheet 6 g.21.9 K/W33.6 K/W+53.3 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 6.3 g with the base; datasheet 6 g.6.4 K/W306 K/W+4679.5 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 6.3 g with the base; datasheet 6 g.6.4 K/W3.61 K/W−43.6 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 6.3 g with the base; datasheet 6 g.4 K/W78.1 K/W+1852.1 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 6.3 g with the base; datasheet 6 g.4 K/W1.82 K/W−54.5 %±25 %outside
Same Sky HSB38-707025P BGA heat sink with push pins,Rth, still airnatural convection, 1 W. Fins as read weigh 69.8 g with the base; datasheet 119 g.4.4 K/W7.15 K/W+62.6 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 69.8 g with the base; datasheet 119 g.1.4 K/W36.5 K/W+2505.7 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 69.8 g with the base; datasheet 119 g.1.4 K/W0.58 K/W−58.3 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 69.8 g with the base; datasheet 119 g.0.90 K/W9.58 K/W+964.8 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 69.8 g with the base; datasheet 119 g.0.90 K/W0.30 K/W−66.8 %±25 %outside
Same Sky HSB39-252509P BGA heat sink with push pins,Rth, still airnatural convection, 1 W. Fins as read weigh 6.1 g with the base; datasheet 8 g.22.3 K/W30.4 K/W+36.4 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 6.1 g with the base; datasheet 8 g.7.2 K/W24.5 K/W+239.8 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 6.1 g with the base; datasheet 8 g.7.2 K/W9.28 K/W+28.9 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 6.1 g with the base; datasheet 8 g.5 K/W11.8 K/W+136.7 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 6.1 g with the base; datasheet 8 g.5 K/W6.09 K/W+21.9 %±25 %within
Same Sky HSB42-353519P BGA heat sink with push pins,Rth, still airnatural convection, 1 W. Fins as read weigh 47.7 g with the base; datasheet 34 g.13.5 K/W20.3 K/W+50.7 %±25 %outside
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 47.7 g with the base; datasheet 34 g.3.1 K/W66.3 K/W+2039.8 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 47.7 g with the base; datasheet 34 g.3.1 K/W1.54 K/W−50.5 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 47.7 g with the base; datasheet 34 g.2.2 K/W22.1 K/W+902.6 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 47.7 g with the base; datasheet 34 g.2.2 K/W0.86 K/W−60.8 %±25 %outside
Same Sky HSB44-606010P BGA heat sink with 4 push pins,Rth, still airnatural convection, 1 W. Fins as read weigh 49.9 g with the base; datasheet 54.6 g.10.9 K/W13 K/W+19.0 %±25 %within
Rth at 200 ft/min (open)1.016 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 49.9 g with the base; datasheet 54.6 g.3 K/W24.6 K/W+721.1 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 200 ft/min1.016 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 49.9 g with the base; datasheet 54.6 g.3 K/W2.06 K/W−31.2 %±25 %outside
Rth at 400 ft/min (open)2.032 m/s approach, 1 W, heat over the whole base, open: air can go round the fins. Fins as read weigh 49.9 g with the base; datasheet 54.6 g.1.9 K/W8.27 K/W+335.2 %—not scored the open choice, shown beside the card's default (ducted)
Rth at 400 ft/min2.032 m/s approach, 1 W, heat over the whole base, ducted: all the air through the fins. Fins as read weigh 49.9 g with the base; datasheet 54.6 g.1.9 K/W1.12 K/W−40.8 %±25 %outside

What it says: on the five larger sinks MeshWeave reads 31 to 67 % low (too good). Those scored rows take every bit of the approaching air through the fins (Ducted); for the datasheet numbers only about a third to two thirds of it would have to, the size of the air a sink in open air loses round its fins. The two smallest read within or close to the sheet now: their cut fins are as short as they are thick, so they are worked out as a bank of short pins, pressure drop and heat transfer from one tube-bank correlation (Gaddis & Gnielinski with Martin’s Lévêque equation). The smallest one’s own sheet also says it was measured in open air (its conductance grows with the square root of the speed, as an open laminar sink’s does). The card’s Open choice works out how much air goes round from the fins’ own pressure drop, and here it lets far too little through (shown beside each row, not scored); the card says so in red. Still air now counts the outer faces as well as the channels between the fins: 1.2 to 2.5 times the datasheets at 1 W, and within -29 to +37 % at their own higher powers apart from the 10 mm sink. For a sink with a datasheet, type its number as a stated K/W.

Heat pipesCapacity and resistance

One copper/water pipe typed into the card’s heat-pipe fields, between a small copper block and a near-perfect sink so the vapour sits at the temperature the source states. Where the source gives no pipe, a 6 mm sintered pipe 150 mm long is used, chosen before the run. The tilt, bend and flattening rows are ratios against the same pipe level, straight and round.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
Quickcool QG-SHP-D6-200MN heat pipe 6 x 200 mmQmax at 70 °Cmesh wick, Ø6 x 200 mm, horizontal; ends not stated, 50/50 mm assumed45 W54.2 W+20.4 %±20 %outside
Thermal resistancesame pipe, 1 W0.050 to 0.10 K/W1.18 K/W——outside
CCI Heat Pipe Design Guide Rev A06 (2013): thermal resistance and test methodThermal resistance, straight sintered pipeØ6 x 150 mm (length chosen: the guide spans 50-350), 15 mm heater, 70 mm condenser, 10 W, ~60 °C0.30 to 0.40 K/W0.16 K/W——outside
Wakefield Thermal heat pipe overview tutorial (DigiKey PTM) slide 8: orientation effectCapacity against gravity / levelsintered Ø6 x 150 mm, 50/50 mm ends, 60 °C (pipe chosen: the source gives none); evaporator 90° above0.50 ×0.45 ×−10.5 %±20 %within
Capacity with gravity / levelsintered Ø6 x 150 mm, 50/50 mm ends, 60 °C (pipe chosen: the source gives none); evaporator 90° below. "Can nearly double"2 ×1.59 ×−20.7 %±20 %outside
Heat Pipe RFQ Checklist and Selection Quick ReferenceCapacity after one 90° bend / straightsintered Ø6 x 150 mm, 50/50 mm ends, 60 °C (pipe chosen: the source gives none); bend radius 18 mm (3 x OD). Rule of thumb: 5-10 % per 90°0.90 to 0.95 ×0.95 ×——within
Capacity flattened to 4 mm / roundsintered Ø6 x 150 mm, 50/50 mm ends, 60 °C (pipe chosen: the source gives none); flattened to 2/3 of its diameter. Rule of thumb: 10-30 % lost0.70 to 0.90 ×0.89 ×——within
Combined effect of bending and flattening on heat transfer performance of cryogenic sintered-wick heat pipeResistance ratios: bent, flatteneda nitrogen pipe at cryogenic temperature1.22 ×not run: the card models water pipes——not scored a nitrogen pipe; the card models copper/water pipes

What it says: tilt, bending and flattening move the capacity the way the vendors say they do: against gravity it keeps 45 % (they say half), one 90° bend costs 5 % and flattening to two thirds costs 11 %. With gravity it gains 59 %, short of the doubling one vendor quotes. The capacity of the one mesh pipe listed reads 20 % high, right at the edge. The resistance is the weak side: a mesh pipe reads more than ten times the listing, and a sintered pipe reads half of a fixture figure that also counts the heater block’s contact. For a pipe with a datasheet capacity, use that number.

Packages on a JEDEC test boardJunction to air, still air

A package’s θJA is measured on a standard 4-layer test board (76.2 x 114.3 mm, two buried copper planes) in still air. Here it is the sheet’s own θJB plus the board’s resistance to the air from the Board card’s layered solve: 4 layers, 1.6 mm, 2 oz outer copper at 10 % (traces), 1 oz planes, no vias, a square of the same area, and 10 W/m²K on both faces. That air coefficient is our choice for still air with radiation; the sheets do not give one.

datasheetvalueconditionpublishedMeshWeaveerrortoleranceverdict
TI DRV8825 (SLVSA73F) 7.4 Thermal Information, PWP HTSSOP-28 with PowerPADθJA, still air, JEDEC 2s2p boardHTSSOP-28, 9.7 x 4.4 mm (footprint 6.53 mm square); the sheet's θJB 5.6 K/W plus the board to the air31.6 K/W35.1 K/W+11.2 %±25 %within
TI LM61460-Q1 (SNVSB70F) 8.4 Thermal Information, RJR QFN-14, JESD51-7 versus EVMθJA, still air, JEDEC 2s2p boardQFN-14, 4.0 x 3.5 mm (footprint 3.74 mm square); the sheet's θJB 19.2 K/W plus the board to the air58.7 K/W64.5 K/W+9.9 %±25 %within
TI TPS2492/TPS2493 (SLUSA65C) Thermal Information, PW TSSOP-14θJA, still air, JEDEC 2s2p boardTSSOP-14, 5.0 x 4.4 mm (footprint 4.69 mm square); the sheet's θJB 53.8 K/W plus the board to the air116 K/W92 K/W−21.0 %±25 %within
TI DLPA2000 (DLPS043B, rev Feb 2018) section 6.5 Thermal Information, YFF DSBGA-56θJA or ψJT——not run——not scored the sheet gives θJA alone, no θJB to build it from
TI AN-1520θJA or ψJT——not run——not scored a 9 sq in board that is not the JEDEC one, and no θJB
TI SPRA953D Semiconductor and IC Package Thermal MetricsθJA or ψJT——not run——not scored ψJT (junction to the top of the case) needs the package solved inside, which the card does not do

What it says: all three packages land within 25 %, from 21 % low to 11 % high. The board is the hard part of a θJA, and it needs every copper layer resolved. The junction temperature on a part now uses the same layer-by-layer board; the averaged-board estimate it used before would have read these three boards two to four times high.

Read, but not scoredWith the reason
How to read itSaid plainly

The error is MeshWeave’s value against the published one. Joints are held to ±15 %, heat pipes to ±20 %, heat sinks to ±25 % because their fin sizes are read from drawings that leave some of them open, and packages to ±25 % because the board’s air coefficient is ours. A range or a lower bound on the datasheet is scored as inside or outside it. Every source was read twice, by two readers working apart; the 4 cases where they disagreed are not used: Indium Corp 'High-Performance Thermal Interface Materials' handout, Form 99541 R1, A Brief Overview of Heat Pipe Design Guidelines, ATS Round Heat Pipes catalog (R11_0219): Qmax vs effective length, sintered 3-7 mm, Design considerations when using heat pipes (Electronics Cooling, 2016): flattening, bend and tilt derating.