Design Book › Fasteners and manufacturing

Fasteners and manufacturing

Pick the process before you draw the part, and draw to that vendor's limits. For bolted joints, a designer must do these.

  1. Set preload and carry its uncertainty [1].
  2. Size strength from maximum preload, and slip or separation from minimum preload.
  3. Apply factors to the load, never to preload [2].
  4. Give each joint a lock that does not depend on preload [3].
  5. Check the geometry: edge distance, spacing, thread engagement, blind holes [4], [5], [6], [7].

The fastener guidelines come mainly from NASA spaceflight documents. They are sound for any joint where a loose bolt causes harm. Scope is on each guideline. Categories: fasteners and screws. Fastener price rows cover 202 of 389 parts, and some rows are for packs. At quantity 1, machine screws run $0.01 to $88.96, standoffs $0.40 to $6.02 and shoulder screws $1.99 to $4.99.

Preload

Torque control is coarse. Friction can vary from 0.04 to 1.10 with material and lubricant [8]. Use T = K * F * d [9], and pick K for your actual friction: 0.2 is the common assumption [10], and 0.15 is more realistic for steel on steel [11].

Preload
How preload is setPreload uncertaintySource
Hand torque wrench, lubricatedplus or minus 25 percent[12], [13]
Torque control, dry or as receivedat least 35 percent[14]
Turn of nut or turn angleat least 25 percent[15]
Bolt stretchat least 10 percent[16]
Instrumented load-sensing boltplus or minus 5 percent[12]

The 25 and 35 percent torque-control values are for non-separation-critical joints [13], [14]. The "at least" rows apply without a statistical basis, and a larger tested variation takes over.

Subtract relaxation: 5 percent of minimum initial preload for all-metal joints, and a tested value with any non-metal part or coating. With multiple faying surfaces 5 percent may be non-conservative, so test or run a creep-relaxation analysis [17]. Metal joints typically embed 2 to 10 percent [18].

The two NASA sources differ on how near yield to go. One prestresses to 65 to 90 percent of yield [19]. The other says not to torque close to yield, because exact yield torque needs extensive instrumentation, and says proof load is usually 75 percent of theoretical yield [20]. Without a bolt load measurement, use the lower.

Joint failure modes

Joint failure modes
LevelStatementCheckSource
shouldA preloaded joint must not separate under loading; use a separation factor of safety of 1.2 for structural joints and 1.4 for pressure systems.calc: P_min / (SF_sep * P_ext_load_share) > 1, SF_sep = 1.2 structural, 1.4 pressure[21]
mustUse the separation factor from NASA-STD-5020 Figure 1; a factor greater than 1.0 gives margin between service load and preload. A joint classed non-separation-critical at factor 1.0 could separate above limit load, including during qualification test.review: each joint is classed separation-critical or not, and the factor is recorded[22]
mustDo not count friction as a shear load path at ultimate load; friction may be used only at limit or yield load.review: ultimate shear margin uses bolt shear and bearing only[23]
mustUnless substantiated by test, use joint-slip friction coefficient at most 0.20 only for uncoated, non-lubricated metal cleaned by a qualified process and visibly clean at and after assembly.calc: mu_slip <= 0.20 clean bare metal[24]
mustUnless test shows otherwise, use a joint-slip friction coefficient of no more than 0.10 for all coated, lubricated, painted or non-metal surfaces.calc: mu_slip <= 0.10 coated, lubricated or non-metal surfaces[25]
shouldDo not load threads in shear or bending; use a shoulder bolt where the fastener must carry shear on a body.review: shear plane passes through a full body or shoulder, not through threads[26]
shouldDo not share a load between bolts and rivets or between bolts and dowel pins with interference fit, because the fitted part takes all load until it yields.review: shear pattern uses one load path type[27]
shouldDo not apply high torque to a bolt whose dominant load is shear; keep the shear-tension interaction sum below 1 for a positive margin.calc: R_s^a + R_t^b < 1, exponents from the chosen RP-1228 interaction curve[28]
shouldDesign a joint not to slip, or use close-tolerance bodies, cured thread fillers, or dowel pins, because cyclic slip loses preload.calc: friction capacity at min preload > shear load, or review of dedicated shear device[29]
shouldFor an eccentric load on a bolt pattern, find the centroid, divide the load across the bolts, then add the moment shear vectorially for each bolt.calc: P_c = R / n; P_e = M r / sum(r_n^2); vector sum on the outermost bolt[30]
mustDesign threaded fastening to survive the full service life with a life scatter factor, since fatigue is a real failure mode alongside rupture.calc: fastener fatigue damage over cycles * scatter factor < 1[31]
shouldCheck dissimilar bolt and joint materials at operating temperature: contraction can unload or separate the joint, and expansion can overload the fasteners.calc: preload change = (alpha_j - alpha_b) * dT * L * joint stiffness terms, for hot and cold[32]
shouldUse a fitting factor of at least 1.15 multiplied into the required ultimate factor of safety when load paths are uncertain.calc: design load = limit load * FS_u * 1.15[33]

Keep friction out of ultimate shear [23]. NASA-STD-5020 allows threads in the shear plane with its interaction check, while MIT 2.72 says not to load threads in shear [26]. The second is the simpler guideline for a robot joint.

Locking and geometry

Locking and geometry
LevelStatementCheckSource
mustInclude at least one locking feature that does not depend on preload to work; jam nuts may not qualify.review: every bolted joint lists a locking feature that works without preload[3]
mustUse a mechanical lock such as cotter pin or safety wire on a bolt that serves as an axis of rotation; friction locks alone are not suitable.review: pivot bolts list a mechanical locking feature[34]
guideEstimate self-loosening moment as preload times thread pitch divided by 2 pi. The capture stores the pi as a private-use symbol-font glyph right after the 2; reading it as a plain 2 would be wrong.calc: M_loosen = P * p / (2 * pi)[35]
mustDo not use set screws to transmit torque between a shaft and a component on the shaft.review: shaft-hub joints have a key, spline, clamp or pin[36]
shouldUse multiple-turn spiral-wound retaining rings where a retaining ring is used.review: retaining ring type in the bill of materials[37]
shouldKeep a fastener edge distance of at least 1.5 fastener diameters from the hole centreline.model-lint: edge distance / d >= 1.5[38]
shouldKeep the edge ratio e/D at 2.0 or more to avoid shear tear-out; never go below 1.5.model-lint: e/D >= 2.0, hard limit 1.5[4]
guideUse a nominal fastener spacing of 4 diameters; thin material can buckle between fasteners.model-lint: pitch / d near 4; check buckling in thin plates[5]
mustChoose fastener length so the bolt does not reach the bottom of a blind hole or the incomplete internal threads there.model-lint: bolt end clears incomplete threads in the tapped hole[7]
mustChoose grip and washers so the mating threads do not reach the incomplete runout threads; these can extend up to twice the pitch.model-lint: internal thread stops short of the runout (up to 2 * pitch) on the bolt[39]
shouldProvide clearance for the head-to-shank fillet with a chamfered hole or a countersunk washer, countersink toward the head.review: bolt head seat has a chamfer or countersunk washer[40]
shouldHave at least one bolt thread protrude past a self-locking or plain nut.measure: bolt end protrudes one thread or more[41]
mustWhen a bolt threads into a weaker material and needs full bolt strength, determine tapped-hole depth for the weaker material from its shear strength.calc: pullout load from thread shell area and F_s of the tapped part > bolt strength[6]
shouldThread length beyond 1.5x diameter adds cost but no strength; keep threads no longer than 2x the hole diameter.model-lint: thread length / diameter[42]
shouldAt the bottom of a blind threaded hole, leave a plain, unthreaded section at least half the diameter long.model-lint: blind hole depth[43]
mustProvide enough thread engagement that the bolt shank yields in tension before the threads strip in shear; compute the length from stress area, major diameter and pitch, and lengthen it when the female thread material is weaker than the screw.calc: Le from tensile stress area, D and pitch; add length for weaker nut material[44]
shouldThe FAA has determined that, in general cases, at least one thread must protrude from any bolted connection.review: count protruding threads[45]

Guidelines overlap here. Edge distance has a floor of 1.5 D [38] and a target of 2.0 [4]. Spacing is a nominal 4 D [5] or 3 to 5 D [46]. One diameter of thread engagement gives thread shear strength equal to bolt tensile strength [47]. For a weaker tapped part that needs full bolt strength, determine the depth from its shear strength [6].

Threads in plastic

Do not rely on a bolt threading directly into FDM plastic for repeated assembly [48]. Assess an insert by its specified strength, not a thread-stripping calculation [49].

Threads in plastic
StepValueSource
Blind hole depthInsert length plus about 1 mm[50]
Hole diameterFind by test specimen, per print orientation; printed holes usually come out undersize[51]
Iron temperature, PLANear 225 C[52]
Iron temperature, PETGNear 245 C[53]
SeatingMelt about 90 percent of depth, press the rest, hold still[54]

Stainless-steel or phosphor-bronze helical wire inserts repair stripped threads, and give stronger threads in soft base materials [55]. Dimensions for M3 and other heat-set inserts are on their part pages; the source file states no units.

Choosing a process

Choosing a process
ProcessUse it forAvoid it forGuidelines start at
FDMBrackets, covers, jigs, gearbox housings; same-day iterationLoads across layers; heat near motors in PLA[56]
SLAFine, accurate small parts, fit-checksLoad-bearing structure: resins are often brittle, and UV degrades them[57]
SLS and MJFTough housings, cable guides, hinges, printed-in-place assembliesLarge flat parts, which warp[58]
CNCBearing seats, shafts, flanges, press fitsThin tall walls; deep holes; cost rises with removed volume[59]
Sheet metalChassis plates, brackets, motor mounts, battery traysAnything beyond flat cut and folded shapes[60], [61]
DMLSPrinted metal: titanium, stainless, AlSi10MgNo class note or price in this edition[62]
PCBMotor drivers, sensor boards, power distributionLimits differ by vendor and copper weight[63]

FDM guidelines by vendor

FDM guidelines by vendor
ValueProtolabs NetworkPrusaSource
Minimum wall0.8 mm, supported or notOne perimeter is about 0.45 mm with a 0.4 mm nozzle[56], [64], [65]
Default shell3 perimeters or 1.2 mmInteger multiples of 0.45 mm[66], [65]
Minimum feature and hole2.0 mm eachnot listed[67], [68]
AccuracyPrototype 0.5 percent, floor 0.5 mm; industrial 0.3 percent, floor 0.3 mmAt least 0.2 mm, depending on print and material[69], [70], [71]
Moving clearancenot listedAt least 0.3 mm; tune with test prints[72]
OverhangAvoid where possible; not over 45 deg45 to 60 deg, by nozzle and settings[73], [74]
Overhang, newer machinesnot listedUp to 75 deg on MK4S, CORE One/+ and CORE One L with Nextruder and 360 deg cooling[75]

Thin walls: round internal corners to at least 0.5 mm [76], make ribs 50 to 75 percent of the wall [77], and avoid holes under roughly 1 to 2 mm where possible [78].

Bambu Lab filaments

Orient parts so that the main tensile load runs along the layers. Design Z-loaded features with the Z value. Tensile values are typical X-Y / Z, dry state for PAHT-CF and PET-CF, heat deflection is at 1.8 MPa, from each maker sheet. See the filament category.

Bambu Lab filaments
FilamentTensile X-Y / Z, MPaHeat deflection, COverhang limit, degBridge limit, mmSource
PLA Basic35 / 31not shown5530[79], [80], [81]
PLA-CF38 / 26545530[82], [83]
PETG Basicnot shownnot shown7030[84], [85]
PETG HF34 / 23not shown7030[86], [87], [88]
ABS33 / 28847040[89], [90]
ASA37 / 31927040[91], [92]
PC55 / 341177040[93], [94]
PA6-CF102 / 481647040[95], [96]
PAHT-CF92 / 47170not in guidelinesnot in guidelines[97]
PET-CF74 / 35182not in guidelinesnot in guidelines[98]
TPU 95A HFnot shownnot listed5520[99], [100]

Fibre does not cure layer weakness. PA6-CF drops from 102 to 48 MPa across layers, a larger loss than PLA Basic from 35 to 31 MPa. PLA and PETG soften at lower temperatures than ASA, PC or fibre-filled nylon (compare the heat deflection column), so do not use them next to a hot motor. Use ASA, PC or fibre-filled nylon there, and see the filament category.

SLA guidelines by vendor

SLA guidelines by vendor
ValueProtolabs NetworkFormlabs Form 2, Clear resin, 100 micronFormlabs Form 4, Grey resin, 50 micronSource
Supported wall0.5 mmover 0.4 mmover 0.2 mm[101], [102], [103]
Unsupported wall1.0 mmover 0.4 mmnot listed[57], [104]
Minimum hole0.5 mm0.8 mm0.5 mm[105], [106], [107]
Minimum feature0.2 mmEngraved 0.4 mm, embossed 0.1 mmnot listed[108], [109], [110]
Drain or escape hole4.0 mm3.5 mm0.75 mm[111], [112], [113]
Body clearancenot listed0.5 mm0.4 mm[114], [115]
Horizontal overhangnot listed3 mm5 mm[116], [117]
Weak overhang angle, from level, on a 35 x 10 x 3 mm overhangnot listed19 deg or less10 deg or less[118], [119]
Beam span, 5 x 3 mmnot listed21 mm29 mm[120], [121]
Tolerancenot listednot listedTypically 0.15 percent on 1 to 30 mm features, floor 0.02 mm, in a test at 100 micron with five-minute room-temperature post-cure[122]

Protolabs SLA ABS-like resin allows 0.305 mm walls at normal resolution [123]. On Form 2 with Clear resin at 100 micron, Formlabs recommends 0.4 mm diameter for a 7 mm wire and 1.5 mm for a 30 mm wire [124].

SLS and MJF guidelines by vendor

SLS and MJF guidelines by vendor
ValueProtolabs NetworkFormlabs Fuse, Nylon 12 Powder baselineProtolabs nylonSource
WallUnsupported 1.0 mm, supported 0.8 mmVertical over 0.6 mm, horizontal over 0.3 mmSLS 0.762 mm, MJF 0.508 mm[58], [125], [126], [127], [128], [129]
Other materialsnot listednot listedSLS PP 1.016 mm[130]
Hole1.0 mm1.0 mmnot listed[131], [132]
Fine detail0.8 mm detail0.8 mm pin0.762 mm channel gap[133], [134], [135]
Assembly clearance0.4 mm0.2 mm under 20 mm2, 0.4 mm abovenot listed[136], [137]
Moving parts, printed in place0.5 mmnot listednot listed[138]
Powder removalnot listedTwo drain holes of 3.5 mm; lattice gaps of 8 mmnot listed[139], [140]
Thick sectionsMaximum wall 20 mmnot listedAbout 3.175 mm uniform; hollow to a 2.54 to 3.175 mm shell[141], [142], [143]

DMLS guidelines

Protolabs, normal resolution. The database has no DMLS class note or price.

DMLS guidelines
MaterialMinimum wallSource
Ti-6Al-4V0.381 mm[62]
316L0.381 mm[144]
17-4PH0.381 mm[145]
AlSi10Mg0.762 mm[146]

CNC guidelines by vendor

CNC guidelines by vendor
ValueProtolabsXometrySource
Default tolerance, metalTypically plus or minus 0.005 in (0.13 mm)plus or minus 0.005 in[59], [147]
Default tolerance, plasticnot listedplus or minus 0.010 in[148]
Minimum wall, metal0.020 in (0.5 mm) feature, which does not mean a ribbed heatsink can use it; 0.040 in (1.02 mm) part1/32 in (0.794 mm) as a rule of thumb; thinner may be cost-prohibitive[149], [150], [151]
Minimum wall, plasticnot listed1.5 mm[152]
Wall height to thicknessSlot depth under 6x width10:1 or less[153], [154]
Hole depthMuch deeper than 6x diameter may need two-sided4x diameter; up to 10x at higher cost[155], [156]
Maximum pocket depth2 in (50.8 mm) from either sidenot listed[157]
Internal cornerCutters down to 0.040 in (1 mm); small-cutter pockets to 0.375 inRadius about 130 percent of cutter radius[158], [159]
Threads#2 to 0.5 in, M2 to M12; do not model them in CADUp to 2x diameter; add 0.5x relief in blind holes[160], [42], [43]
Small toolsnot listedFeatures under 2.5 mm need special tools[161]
TextHard metal 0.033 in strokes; soft 0.018 in; 0.0118 in deepnot listed[162], [163], [164]

Turning at Protolabs: tolerance typically plus or minus 0.005 in [165]; walls at least 0.020 in [166]; any diameter at least 0.030 in [167]; axial holes at least 1 mm and up to 6x diameter [168]; radial holes at least 2 mm [169]; external grooves between 0.047 in wide and 0.95 in deep [170]. SendCutSend lists plus or minus 0.005 in for 0.270 in CNC-routed Delrin and 0.009 in for 0.125 in laser-cut Delrin [171].

Sheet metal and cutting guidelines

Minimum flange at Protolabs is a multiple of thickness t plus the bend radius r. Keep holes and features away from bends by the same multiple [172].

Sheet metal and cutting guidelines
Thickness, inAluminium, copperCold-rolled, galvanized, stainlessSource
0.024 to 0.0503.0 t + r3.5 t + r[60], [173]
0.060 to 0.0902.5 t + r2.75 t + r[174], [175]
0.100 to 0.1342.5 t + r2.5 t + r[176]
0.160 and upnot listedStainless 4.0 t + r[177]
Sheet metal and cutting guidelines
ValueProtolabsXometrySendCutSendSource
Bend radiusTooling 0.01 in to 1 in; larger radii by bump formingnot listedVendor sets radius and K factor[178], [179]
Bend limitsnot listednot listedUp to 130 deg; U- and C-channel base at least 2x flange, 3x for polycarbonate[180], [181]
Smallest bent partnot listednot listed0.375 in by 1.5 in flat[182]
Hole toleranceplus or minus 0.005 in; hardware +0.003/-0.000 innot listedHardware +0.003/-0.000 in[183], [184], [185]
Bend to hole, one surfaceplus or minus 0.015 in; non-cumulative 0.030 in across two bendsnot listednot listed[186]
Bend to bendplus or minus 0.015 in, two bends; non-cumulative 0.030 in, three or morenot listednot listed[187]
Edge to holeplus or minus 0.005 innot listednot listed[188]
Countersinknot listedDepth at most 0.6x t; 8x t apart, 4x from an edge, 3x from a bendnot listed[189], [190]
Curls and hemsnot listedCurl radius 2x t; hem return 4x tnot listed[191], [192]
Tapped hole near bendnot listednot listed0.315 in in 0.125 in 5052; look up your thickness[193]

Cut limits at SendCutSend: internal geometry at least 0.015 in [61]. Holes and bridges at least 50 percent of thickness [194], [195], so 0.050 in in 0.100 in aluminium [196]. Laser-cut 6061-T6 holds plus or minus 0.005 in [197], and waterjet carbon fibre plus or minus 0.009 in [198]. To tap a 1/4-20 thread in 0.125 in 5052, keep the hole centre at least 0.167 in from the nearest cut feature [199]. SendCutSend says no hole-to-edge rule of thumb exists, because spacing varies with hole radius and part geometry [200].

PCB guidelines by vendor

PCB guidelines by vendor
ValueJLCPCBPCBWaySource
Track and space, standard0.10 mm on 1 oz, 1 and 2 layers; 0.09 mm multilayer0.1 mm[63], [201], [202]
Track and space, 2 oz0.16 mmnot listed[203]
Track and space, fineHDI 3 mil, 2.7 mil extremeAdvanced roadmap: 3/3 mil for 2016, 2 mil inner and 3 mil outer for 2017, 2/2 mil on part of a board for 2018[204], [205]
Annular ringMultilayer 1 oz: 0.20 mm recommended, 0.15 mm absolute minimum0.15 mm on pads with vias in the middle; advanced 3 mil[206], [207], [208]
DrillStandard 0.15 to 6.3 mm; 0.1 mm microvias on boards up to 1 mm thick with ENIG or OSP0.15 to 6.0 mm[209], [210]
Via0.15 mm hole, 0.25 mm via diameter; 0.1/0.2 mm only on boards up to 1 mm thick with ENIG or OSP; 0.2 mm hole preferrednot listed[211]
Non-plated hole0.50 mmUnder 0.45 mm flagged[212], [213]
Hole tolerancePlated +0.13/-0.08 mmPlated plus or minus 0.08 mm, non-plated 0.05 mm[214], [215]
Outline tolerance0.2 mm regular; 0.1 mm high precision, on boards of at least 50 x 50 mm with three tooling holes; 0.4 mm V-cut, at 0.6 mm thickness or more0.2 mm routed, 0.5 mm V-score[216], [217]

A finished track varies by plus or minus 20 percent, so run current and impedance at the low end [218]. Check the board against the vendor's own manufacturing limits.

Sources

Each guideline in this chapter with its source, the date it was read and the exact words of the source. Level: must means breaking it causes failure or harm, should is good practice, guide is a rule of thumb.

  1. Size strength from the maximum preload and slip or separation from the minimum preload, including method variation, relaxation, creep and maximum and minimum temperature effects. Level: must. standards.nasa.gov 2026-11-08“effects of maximum and minimum expected temperatures.”
  2. Do not apply factors of safety to preload or to thermal preload change; apply them to limit load only. Level: must. standards.nasa.gov 2026-11-08“safety do not apply to fastener preload or changes in preload as a”
  3. Include at least one locking feature that does not depend on preload to work; jam nuts may not qualify. Level: must. standards.nasa.gov 2026-11-08“hardware shall incorporate a minimum of one locking feature that does not depend upon preload”
  4. Keep the edge ratio e/D at 2.0 or more to avoid shear tear-out; never go below 1.5. 1.5 Level: should. ntrs.nasa.gov 2026-11-08“maintain minimum e/D ratios of 2.0 or more. Occasionally the hardware design does not permit maintaining the 2.0 factor and the ratio must be reduced. In this situation, the e/D ratio may be reduced to as low as 1.5,”
  5. Use a nominal fastener spacing of 4 diameters; thin material can buckle between fasteners. 4 D Level: guide. ntrs.nasa.gov 2026-11-08“1.5D. The nominal distance between fasteners is 4D, but the”
  6. When a bolt threads into a weaker material and needs full bolt strength, determine tapped-hole depth for the weaker material from its shear strength. Level: must. ntrs.nasa.gov 2026-11-08“of the tapped hole must be determined for the weaker material”
  7. Choose fastener length so the bolt does not reach the bottom of a blind hole or the incomplete internal threads there. Level: must. standards.nasa.gov 2026-11-08“the bottom of the hole or interfering with incomplete internal threads.”
  8. Thread and underhead friction can vary from 0.04 to 1.10 with material and lubricant, so adjust any torque table to the real friction condition before using it. 1.1 Level: should. ntrs.nasa.gov 2026-11-08“The coefficient of friction can vary from 0.04 to 1.10,”
  9. Estimate tightening torque with T = K F d, where T is torque, F axial load and d bolt diameter. Level: should. ntrs.nasa.gov 2026-11-08“T = KFd, where T denotes torque, F denotes axial load, d”
  10. K = 0.2 is the commonly assumed torque coefficient, but do not use it without checking the actual friction condition. 0.2 Level: should. ntrs.nasa.gov 2026-11-08“The commonly assumed value for K is 0.2, but this value”
  11. Use K = 0.15 as a more realistic typical torque coefficient for steel on steel than the common assumption in FST-TORQUE-002. 0.15 Level: guide. ntrs.nasa.gov 2026-11-08“value for K would be 0.15 for steel on steel. Note”
  12. Carry a preload uncertainty of plus or minus 25 percent for a hand torque wrench on a lubricated fastener; instrumented load-sensing bolts reduce this to plus or minus 5 percent. 25 percent Level: should. ntrs.nasa.gov 2026-11-08“operated torque wrench used on a lubricated fastener is ±25 percent (ref. 1). For comparison, if load sensing (instru-”
  13. For non-separation-critical torque control without a statistical basis, use the greater of 25 percent for lubricated fasteners and tested variation. 25 percent Level: must. standards.nasa.gov 2026-11-08“(1) 25 percent (if”
  14. For non-separation-critical torque control without a statistical basis, use the greater of 35 percent for non-lubricated or as-received fasteners and tested variation. 35 percent Level: must. standards.nasa.gov 2026-11-08“(2) 35 percent (if non-”
  15. Without a statistical basis, use the greater of 25 percent and tested preload variation for turn-of-nut or turn-angle control. 25 percent Level: should. standards.nasa.gov 2026-11-08“(1) 25 percent (2) Variation from tests”
  16. Without a statistical basis, use the greater of 10 percent and tested preload variation for bolt-stretch control. 10 percent Level: should. standards.nasa.gov 2026-11-08“(1) 10 percent”
  17. Subtract short-term relaxation from minimum preload: 5 percent for all-metal clamped parts, a test-derived value for any non-metal parts or coatings. With multiple faying surfaces 5 percent may be non-conservative; testing or creep-relaxation analysis is recommended. 5 percent Level: must. standards.nasa.gov 2026-11-08“Short-term relaxation of Zero 5% of minimum initial”
  18. Metal-to-metal joints in a non-corrosive environment typically embed 2 to 10 percent of preload; TM-106943 treats about 5 percent as a safe design assumption. 5 percent Level: guide. ntrs.nasa.gov 2026-11-08“2 and 10 percent (ref. 4). For design and analysis purposes it is safe to assume the preload loss to be about 5 percent”
  19. A preloaded bolted joint should carry working load mainly through release of flange compression, with the bolt prestressed to 65 to 90 percent of its yield strength. Level: guide. ntrs.nasa.gov 2026-11-08“the bolt is usually prestressed very near its working limits (usually 65 to 90 percent of its yield strength)”
  20. Do not torque close to bolt yield; exact yield torque needs extensive instrumentation, and proof load is usually 75 percent of theoretical yield. 75 percent of theoretical yield Level: should. ntrs.nasa.gov 2026-11-08“This value is usually 75 percent of theoretical yield, to prevent”
  21. A preloaded joint must not separate under loading; use a separation factor of safety of 1.2 for structural joints and 1.4 for pressure systems. 1.2 Level: should. ntrs.nasa.gov 2026-11-08“factor of safety for joint separation, SFse p, is equal to 1.2 for structural applications and 1.4 for pressure system applications.”Review note: Source calls 1.2 and 1.4 the recommended separation factors, so severity must is too strong.
  22. Use the separation factor from NASA-STD-5020 Figure 1; a factor greater than 1.0 gives margin between service load and preload. A joint classed non-separation-critical at factor 1.0 could separate above limit load, including during qualification test. Level: must. standards.nasa.gov 2026-11-08“Primary structure with joints classified as "non-separation critical" and FSsep = 1.0 could”
  23. Do not count friction as a shear load path at ultimate load; friction may be used only at limit or yield load. Level: must. standards.nasa.gov 2026-11-08“Beneficial effects of friction are not permissible in analysis for ultimate load conditions.”
  24. Unless substantiated by test, use joint-slip friction coefficient at most 0.20 only for uncoated, non-lubricated metal cleaned by a qualified process and visibly clean at and after assembly. 0.2 Level: must. standards.nasa.gov 2026-11-08“(1) 0.20 for uncoated, non-lubricated metal surfaces that are cleaned by a qualified”
  25. Unless test shows otherwise, use a joint-slip friction coefficient of no more than 0.10 for all coated, lubricated, painted or non-metal surfaces. 0.1 Level: must. standards.nasa.gov 2026-11-08“(2) 0.10 for all other surfaces, including nonmetallic (coated or uncoated) surfaces”
  26. Do not load threads in shear or bending; use a shoulder bolt where the fastener must carry shear on a body. Level: should. ocw.mit.edu 2026-11-08
  27. Do not share a load between bolts and rivets or between bolts and dowel pins with interference fit, because the fitted part takes all load until it yields. Level: should. ntrs.nasa.gov 2026-11-08“Bolts and rivets should not be used together to carry a load,”
  28. Do not apply high torque to a bolt whose dominant load is shear; keep the shear-tension interaction sum below 1 for a positive margin. Level: should. ntrs.nasa.gov 2026-11-08“Rj + R;< 1 is a requirement for a positive margin of safety.”
  29. Design a joint not to slip, or use close-tolerance bodies, cured thread fillers, or dowel pins, because cyclic slip loses preload. Level: should. standards.nasa.gov 2026-11-08“With shear pins (dowel pins) or other dedicated shear-transfer devices.”
  30. For an eccentric load on a bolt pattern, find the centroid, divide the load across the bolts, then add the moment shear vectorially for each bolt. Level: should. ntrs.nasa.gov 2026-11-08“When the load on a fastener group is eccentric, the first task”
  31. Design threaded fastening to survive the full service life with a life scatter factor, since fatigue is a real failure mode alongside rupture. Level: must. standards.nasa.gov 2026-11-08“without fatigue failure.”
  32. Check dissimilar bolt and joint materials at operating temperature: contraction can unload or separate the joint, and expansion can overload the fasteners. Level: should. ntrs.nasa.gov 2026-11-08“to unload (or separate); differential expansion can cause”
  33. Use a fitting factor of at least 1.15 multiplied into the required ultimate factor of safety when load paths are uncertain. 1.15 Level: should. standards.nasa.gov 2026-11-08“1.15 as a multiplier of the required ultimate factor of safety. The selection of a fitting factor for”
  34. Use a mechanical lock such as cotter pin or safety wire on a bolt that serves as an axis of rotation; friction locks alone are not suitable. Level: must. standards.nasa.gov 2026-11-08“[TFSR 17] A mechanical locking feature shall be used on any bolt subject to rotation in”
  35. Estimate self-loosening moment as preload times thread pitch divided by 2 pi. The capture stores the pi as a private-use symbol-font glyph right after the 2; reading it as a plain 2 would be wrong. Level: guide. standards.nasa.gov 2026-11-08“the thread pitch divided by 2. For reference, consult”
  36. Do not use set screws to transmit torque between a shaft and a component on the shaft. Level: must. standards.nasa.gov 2026-11-08“Set screws shall not be used to transmit torque between a shaft and a component mounted”
  37. Use multiple-turn spiral-wound retaining rings where a retaining ring is used. Level: should. standards.nasa.gov 2026-11-08“[DDMR 63] All retaining rings used shall be multiple-turn spiral-wound retaining rings.”
  38. Keep a fastener edge distance of at least 1.5 fastener diameters from the hole centreline. 1.5 D Level: should. ntrs.nasa.gov 2026-11-08“The minimum edge distance should not be less than 1.5D.”
  39. Choose grip and washers so the mating threads do not reach the incomplete runout threads; these can extend up to twice the pitch. Level: must. standards.nasa.gov 2026-11-08“exist over a distance up to twice the pitch on most commercial aerospace fasteners.”
  40. Provide clearance for the head-to-shank fillet with a chamfered hole or a countersunk washer, countersink toward the head. Level: should. standards.nasa.gov 2026-11-08“Clearance shall be provided for the head-to-shank fillet radius as needed to avoid”
  41. Have at least one bolt thread protrude past a self-locking or plain nut. Level: should. raanz.org.nz 2026-11-08“have at least one thread of the bolt protruding”
  42. Thread length beyond 1.5x diameter adds cost but no strength; keep threads no longer than 2x the hole diameter. 2 x diameter Level: should. xometry.com 2026-11-08
  43. In blind threaded holes, add an unthreaded length of at least half the diameter at the bottom. 0.5 x diameter Level: should. xometry.com 2026-11-08
  44. Provide enough thread engagement that the bolt shank yields in tension before the threads strip in shear; compute the length from stress area, major diameter and pitch, and lengthen it when the female thread material is weaker than the screw. Level: must. en.wikipedia.org 2026-11-08“Bolted joints are designed so that the bolt shank fails in tension before the threads fail in shear, but for this to hold true, a minimum thread engagement must be achieved.”
  45. The FAA has determined that, in general cases, at least one thread must protrude from any bolted connection. 1 thread Level: should. en.wikipedia.org 2026-11-08“the FAA has determined that in general cases, at least one thread must be protruding from any bolted connection.”
  46. Keep material under a bolt head one to three bolt diameters thick, and bolt centres three to five diameters apart. Level: guide. pergatory.mit.edu 2026-11-08
  47. Thread engagement equal to one bolt diameter gives thread shear strength equal to bolt tensile strength; more adds little. Level: should. pergatory.mit.edu 2026-11-08
  48. Do not rely on a bolt threading directly into FDM plastic for repeated assembly; use a brass heat-set insert. Level: should. cnckitchen.com 2026-11-08
  49. Assess a nut or threaded insert by its specified strength, not a thread-stripping calculation, since it can expand under load. Level: should. standards.nasa.gov 2026-11-08“be based on the strength specified for that item rather than on thread-stripping analysis.”
  50. For heat-set inserts, make a blind hole slightly deeper than the insert, approximately 1 mm. 1 mm Level: should. cnckitchen.com 2026-11-08
  51. Printed holes usually come out smaller than CAD; use a test specimen to determine insert bore diameters in different printing orientations. Level: should. cnckitchen.com 2026-11-08
  52. Set the iron near 225 C for PLA, which is 10 to 20 C above print temperature. 225 °C Level: guide. cnckitchen.com 2026-11-08
  53. Set the iron near 245 C for PETG parts. 245 °C Level: guide. cnckitchen.com 2026-11-08
  54. Melt a heat-set insert only about 90 percent of the depth, then press the last part with a tool, and hold it still because it tends to creep back out. 90 % Level: guide. cnckitchen.com 2026-11-08
  55. Stainless-steel or phosphor-bronze helical wire inserts repair stripped internal threads and provide stronger threads in soft base materials. Level: guide. en.wikipedia.org 2026-11-08“These inserts provide a convenient means of repairing stripped internal threads.”
  56. On Protolabs Network FDM service, a wall that nothing supports on its sides must be at least 0.8 mm thick. 0.8 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  57. SLA walls with no support on two sides must be at least 1.0 mm. 1.0 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  58. SLS walls with no support must be at least 1.0 mm. 1.0 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  59. Protolabs says it can typically hold ±0.005 in. (0.13 mm) machining tolerance at its fully automated facility. 0.005 in Level: guide. protolabs.com 2026-11-08
  60. Aluminum or copper sheet 0.024 to 0.050 in thick needs a flange of at least 3.0x thickness plus bend radius. 3.0 x t + r Level: must. protolabs.com 2026-11-08
  61. No laser-cut internal geometry should be smaller than 0.015 in; check the material chart for the real minimum. 0.015 in Level: must. sendcutsend.com 2026-11-08
  62. Printed Ti-6Al-4V (DMLS) walls at normal resolution must be at least 0.381 mm. 0.381 mm Level: should. protolabs.com 2026-11-08
  63. JLCPCB lists 1 oz track/space at 0.10 mm for 1–2-layer boards and 0.09 mm for multilayer boards, with 3 mil accepted in BGA fan-outs. 0.1 mm Level: should. jlcpcb.com 2026-11-08
  64. On Protolabs Network FDM service, a supported wall must be at least 0.8 mm thick. 0.8 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  65. With a 0.4 mm nozzle and PrusaSlicer, one perimeter makes a wall of about 0.45 mm, so walls should be integer multiples of 0.45 mm; walls under one perimeter do not print. 0.45 mm Level: should. help.prusa3d.com 2026-11-08
  66. Protolabs Network states that FDM parts are printed with three perimeter shells or a 1.2 mm wall. 1.2 mm Level: guide. hubs.com 2026-11-08
  67. Keep any FDM feature at or above 2.0 mm on Protolabs Network; finer detail is out of reach for the process. 2.0 mm Level: should. hubs.com 2026-11-08
  68. Holes drawn for Protolabs Network FDM should be at least 2.0 mm in diameter. 2.0 mm Level: should. hubs.com 2026-11-08
  69. Protolabs Network lists prototyping (desktop) FDM accuracy as ±0.5%, with a ±0.5 mm lower limit. 0.5 mm Level: guide. hubs.com 2026-11-08
  70. Industrial FDM on Protolabs Network holds 0.3 percent with a floor of 0.3 mm. 0.3 mm Level: should. hubs.com 2026-11-08
  71. Prusa states that an Original Prusa is accurate to at least 0.2 mm, with tolerances dependent on print and material factors. 0.2 mm Level: guide. help.prusa3d.com 2026-11-08
  72. Start moving or hinged FDM parts with at least 0.3 mm clearance, then tune with test prints. 0.3 mm Level: guide. help.prusa3d.com 2026-11-08
  73. Protolabs Network recommends avoiding overhangs when possible and not designing angles over 45 degrees. 45 deg Level: should. hubs.com 2026-11-08
  74. Prusa gives a clean overhang range of 45–60°, depending on nozzle diameter and settings. 45 deg Level: guide. help.prusa3d.com 2026-11-08
  75. Prusa MK4S, CORE One/+, and CORE One L models with Nextruder and 360° cooling can print overhangs up to 75° without supports. 75 deg Level: guide. help.prusa3d.com 2026-11-08
  76. Round internal corners of thin-walled 3D prints with at least 0.5 mm radius to reduce stress cracking. 0.5 mm Level: guide. hubs.com 2026-11-08
  77. Make ribs on thin walls 50 to 75 percent of the wall thickness. 50 % Level: guide. hubs.com 2026-11-08
  78. Across 3D-printing processes, avoid holes under roughly 1–2 mm where possible; FDM holes may close or need hard-to-remove supports. 2 mm Level: should. hubs.com 2026-11-08
  79. Bambu Lab lists typical PLA Basic tensile strength as 35 ± 4 MPa in X-Y and 31 ± 3 MPa in Z. 31 MPa Level: guide. cdn.shopify.com 2026-11-08
  80. Bambu Lab lists a maximum overhang angle of 55° for PLA Basic. 55 deg Level: should. cdn.shopify.com 2026-11-08
  81. For Bambu Lab PLA Basic, keep unsupported bridges to 30 mm or less. 30 mm Level: should. cdn.shopify.com 2026-11-08
  82. Bambu Lab lists a maximum overhang angle of 55° for PLA-CF. 55 deg Level: should. cdn.shopify.com 2026-11-08
  83. For Bambu Lab PLA-CF, keep unsupported bridges to 30 mm or less. 30 mm Level: should. cdn.shopify.com 2026-11-08
  84. Bambu Lab lists a maximum overhang angle of about 70° for PETG Basic. 70 deg Level: should. store.bblcdn.com 2026-11-08
  85. Bambu Lab lists a maximum bridging length of about 30 mm for PETG Basic. 30 mm Level: should. store.bblcdn.com 2026-11-08
  86. Bambu Lab lists typical PETG HF tensile strength as 34 ± 4 MPa in X-Y and 23 ± 4 MPa in Z. 23 MPa Level: guide. store.bblcdn.com 2026-11-08
  87. Bambu Lab lists a maximum overhang angle of about 70° for PETG HF. 70 deg Level: should. store.bblcdn.com 2026-11-08
  88. Bambu Lab lists a maximum bridging length of about 30 mm for PETG HF. 30 mm Level: should. store.bblcdn.com 2026-11-08
  89. Bambu Lab lists a maximum overhang angle of about 70° for ABS. 70 deg Level: should. cdn.shopify.com 2026-11-08
  90. Bambu Lab lists a maximum bridging length of about 40 mm for ABS. 40 mm Level: should. cdn.shopify.com 2026-11-08
  91. Bambu Lab lists a maximum overhang angle of about 70° for ASA. 70 deg Level: should. cdn.shopify.com 2026-11-08
  92. Bambu Lab lists a maximum bridging length of about 40 mm for ASA. 40 mm Level: should. cdn.shopify.com 2026-11-08
  93. Bambu Lab lists a maximum overhang angle of about 70° for PC. 70 deg Level: should. cdn.shopify.com 2026-11-08
  94. Bambu Lab lists a maximum bridging length of about 40 mm for PC. 40 mm Level: should. cdn.shopify.com 2026-11-08
  95. Bambu Lab lists a maximum overhang angle of about 70° for PA6-CF. 70 deg Level: should. cdn.shopify.com 2026-11-08
  96. Bambu Lab lists a maximum bridging length of about 40 mm for PA6-CF. 40 mm Level: should. cdn.shopify.com 2026-11-08
  97. For dry-state PAHT-CF, Bambu Lab lists typical tensile strength as 92 ± 7 MPa in X-Y and 47 ± 5 MPa in Z. 47 MPa Level: guide. cdn.shopify.com 2026-11-08
  98. For dry-state PET-CF, Bambu Lab lists typical tensile strength as 74 ± 6 MPa in X-Y and 35 ± 5 MPa in Z. 35 MPa Level: guide. cdn.shopify.com 2026-11-08
  99. Bambu Lab lists a maximum overhang angle of about 55° for TPU 95A HF. 55 deg Level: should. cdn.shopify.com 2026-11-08
  100. For Bambu Lab TPU 95A HF, keep unsupported bridges to 20 mm or less. 20 mm Level: should. cdn.shopify.com 2026-11-08
  101. Supported SLA walls must be at least 0.5 mm. 0.5 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  102. On Form 2 with Clear resin at 100 microns, supported walls must exceed 0.4 mm. 0.4 mm Level: must. formlabs.com 2026-11-08
  103. On Form 4 with Grey resin at 50 microns, supported walls should exceed 0.2 mm. 0.2 mm Level: must. formlabs.com 2026-11-08
  104. For Form 2 with Clear resin at 100 μm, unsupported walls of 0.4 mm or less may warp or detach. 0.4 mm Level: must. formlabs.com 2026-11-08
  105. SLA holes should be at least 0.5 mm in diameter. 0.5 mm Level: should. hubs.com 2026-11-08
  106. For Form 2 with Clear resin at 100 μm, holes under 0.8 mm may close during printing. 0.8 mm Level: should. formlabs.com 2026-11-08
  107. For Form 4 with Grey resin at 50 μm, holes under 0.5 mm may close during printing. 0.5 mm Level: should. formlabs.com 2026-11-08
  108. Smallest SLA feature on Protolabs Network is 0.2 mm. 0.2 mm Level: should. hubs.com 2026-11-08
  109. For Form 2 with Clear resin at 100 μm, engraved details should be at least 0.4 mm deep and wide. 0.4 mm Level: should. formlabs.com 2026-11-08
  110. For Form 2 with Clear resin at 100 μm, embossed details should be at least 0.1 mm thick and tall. 0.1 mm Level: should. formlabs.com 2026-11-08
  111. Protolabs Network recommends a minimum escape-hole diameter of 4.0 mm for SLA parts. 4.0 mm Level: should. hubs.com 2026-11-08
  112. For Form 2 with Clear resin at 100 μm, enclosed cavities need drain holes of at least 3.5 mm to avoid cupping blowout failure. 3.5 mm Level: must. formlabs.com 2026-11-08
  113. For Form 4 with Grey resin at 50 μm, enclosed cavities need drain holes of at least 0.75 mm to avoid cupping blowout failure. 0.75 mm Level: must. formlabs.com 2026-11-08
  114. For Form 2 with Clear resin at 100 μm, parts may fuse when clearance is under 0.5 mm. 0.5 mm Level: must. formlabs.com 2026-11-08
  115. For Form 4 with Grey resin at 50 μm, parts may fuse when clearance is under 0.4 mm. 0.4 mm Level: must. formlabs.com 2026-11-08
  116. For Form 2 with Clear resin at 100 μm, horizontal overhangs beyond 3 mm do not completely form and increasingly deform. 3.0 mm Level: should. formlabs.com 2026-11-08
  117. For Form 4 with Grey resin at 50 μm, horizontal overhangs beyond 5 mm do not completely form and increasingly deform. 5.0 mm Level: should. formlabs.com 2026-11-08
  118. For Form 2 with Clear resin at 100 μm, an overhang at 19° or less from level can break off during peel. The 19° value is listed for a 35 × 10 × 3 mm overhang. 19 deg Level: should. formlabs.com 2026-11-08
  119. For Form 4 with Grey resin at 50 μm, an overhang at 10° or less from level can break off during peel. The 10° value is listed for a 35 × 10 × 3 mm overhang. 10 deg Level: should. formlabs.com 2026-11-08
  120. For Form 2 with Clear resin at 100 μm, a 5 mm wide, 3 mm thick beam is more likely to fail beyond a 21 mm span. 21 mm Level: guide. formlabs.com 2026-11-08
  121. For Form 4 with Grey resin at 50 μm, a 5 mm wide, 3 mm thick beam is more likely to fail beyond 29 mm. 29 mm Level: guide. formlabs.com 2026-11-08
  122. In Formlabs' Form 4 test using Grey resin at 100 μm and five-minute room-temperature post-cure, 1–30 mm features had typical tolerances of ±0.15% with a ±0.02 mm lower limit. 0.15 % Level: guide. formlabs.com 2026-11-08
  123. Protolabs SLA ABS-like resins need walls of at least 0.305 mm at normal resolution. 0.305 mm Level: should. protolabs.com 2026-11-08
  124. For Form 2 with Clear resin at 100 μm, Formlabs recommends 0.4 mm diameter for a 7 mm wire and 1.5 mm for a 30 mm wire. 0.4 mm Level: should. formlabs.com 2026-11-08
  125. Supported SLS walls must be at least 0.8 mm. 0.8 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  126. Fuse series unsupported vertical walls must exceed 0.6 mm. 0.6 mm Level: must. formlabs.com 2026-11-08
  127. Fuse series horizontal walls must exceed 0.3 mm. 0.3 mm Level: must. formlabs.com 2026-11-08
  128. Nylon SLS minimum wall is 0.762 mm (0.030 in). 0.762 mm Level: must. protolabs.com 2026-11-08
  129. Nylon MJF minimum wall is 0.508 mm (0.020 in). 0.508 mm Level: must. protolabs.com 2026-11-08
  130. Protolabs lists a minimum normal-resolution SLS PP Natural wall thickness of 1.016 mm. 1.016 mm Level: should. protolabs.com 2026-11-08
  131. SLS holes should be at least 1.0 mm. 1.0 mm Level: should. hubs.com 2026-11-08
  132. On Fuse Series printers, with the Nylon 12 Powder baseline, holes under 1.0 mm may close; use a reamed pilot hole for precisely concentric holes. 1.0 mm Level: should. formlabs.com 2026-11-08
  133. Smallest SLS detail is 0.8 mm. 0.8 mm Level: should. hubs.com 2026-11-08
  134. On Fuse Series printers, with the Nylon 12 Powder baseline, pins and wires under 0.8 mm diameter may deform or break during printing. 0.8 mm Level: should. formlabs.com 2026-11-08
  135. Leave channel gaps of at least 0.762 mm between features so powder does not sinter them together. 0.762 mm Level: should. protolabs.com 2026-11-08Review note: Source says 'We recommend' minimum channel gaps; must is over-strong.
  136. SLS assembly clearance is 0.4 mm. 0.4 mm Level: should. hubs.com 2026-11-08
  137. On Fuse Series printers, with the Nylon 12 Powder baseline, leave 0.2 mm clearance between separate mating parts for features under 20 mm² or 0.4 mm for larger features. 0.2 mm Level: should. formlabs.com 2026-11-08
  138. Printed-in-place SLS moving parts need 0.5 mm clearance. 0.5 mm Level: should. hubs.com 2026-11-08Review note: Source table column is 'Recommended size', not a requirement; severity must is over-strong.
  139. On Fuse Series printers, with the Nylon 12 Powder baseline, Formlabs recommends at least two drain holes of 3.5 mm or more in enclosed cavities. 3.5 mm Level: should. formlabs.com 2026-11-08
  140. On Fuse Series printers, with the Nylon 12 Powder baseline, use lattice gaps of at least 8 mm for easy powder removal in the Fuse Sift. 8 mm Level: should. formlabs.com 2026-11-08
  141. Protolabs Network lists a maximum SLS wall thickness of 20 mm. 20 mm Level: guide. hubs.com 2026-11-08
  142. For parts that warp, aim for about 3.175 mm (0.125 in) uniform thickness. 3.175 mm Level: guide. protolabs.com 2026-11-08
  143. For thick nylon features, Protolabs recommends hollowing to an approximately 2.54–3.175 mm shell. 2.54 mm Level: guide. protolabs.com 2026-11-08
  144. DMLS 316L stainless walls at normal resolution must be at least 0.381 mm. 0.381 mm Level: should. protolabs.com 2026-11-08
  145. DMLS 17-4PH walls at normal resolution must be at least 0.381 mm. 0.381 mm Level: should. protolabs.com 2026-11-08
  146. DMLS AlSi10Mg walls at normal resolution must be at least 0.762 mm. 0.762 mm Level: should. protolabs.com 2026-11-08
  147. Xometry's default CNC tolerance for metal parts is plus or minus 0.005 in unless a tighter dimension is called out. 0.005 in Level: guide. xometry.com 2026-11-08
  148. Xometry's default CNC tolerance for plastic parts is plus or minus 0.010 in. 0.01 in Level: guide. xometry.com 2026-11-08
  149. Keep milled features thicker than 0.020 in (0.51 mm) and nominal part thickness above 0.040 in (1.02 mm). 0.02 in Level: must. protolabs.com 2026-11-08
  150. Protolabs lists 0.020 in. (0.5 mm) minimum feature thickness and 2 in. (51 mm) maximum feature depth, but that does not mean a ribbed heatsink can use those dimensions. 0.5 mm Level: should. protolabs.com 2026-11-08
  151. As a rule of thumb, Xometry gives 1/32 in. (0.794 mm) minimum CNC wall width; thinner walls may be possible but cost-prohibitive. 0.794 mm Level: guide. xometry.com 2026-11-08
  152. Xometry's standard minimum wall is 0.794 mm for metals and 1.5 mm for plastics. 1.5 mm Level: should. xometry.com 2026-11-08
  153. Keep slot-like milled features shallower than 6x the feature width and leave at least 0.5 mm wall beside them. 6 x width Level: should. protolabs.com 2026-11-08
  154. To keep machining costs low, Xometry recommends a thin-wall height-to-thickness ratio no larger than 10:1; larger ratios may cause chatter marks. 10 ratio Level: should. xometry.com 2026-11-08
  155. Protolabs says holes much deeper than six diameters become challenging because of endmill length and may require machining from both sides. 6 x diameter Level: should. protolabs.com 2026-11-08
  156. Limit machined hole depth to 4x diameter; up to 10x is possible at higher cost. 4 x diameter Level: should. xometry.com 2026-11-08
  157. Maximum depth milled from either side is 2 in (50.8 mm). 2 in Level: should. protolabs.com 2026-11-08
  158. Milling cutters go down to 0.040 in (1 mm), and small-cutter pockets are limited to 0.375 in (9.52 mm) deep; use generous internal radii. 1 mm Level: should. protolabs.com 2026-11-08
  159. Add a radius of about 130 percent of the milling cutter radius to internal vertical edges. 130 % Level: guide. xometry.com 2026-11-08
  160. Protolabs automated milling threads UNC/UNF #2 to 0.5 in and metric M2 to M12; do not model internal threads in CAD. 12 mm Level: should. protolabs.com 2026-11-08
  161. Features smaller than 2.5 mm need special tools; avoid them to keep cost down. 2.5 mm Level: guide. xometry.com 2026-11-08
  162. Engraved text in hard metals needs strokes at least 0.033 in (0.838 mm) wide and 0.0118 in deep. 0.838 mm Level: guide. protolabs.com 2026-11-08
  163. Engraved text in plastics and soft metals needs strokes at least 0.018 in (0.457 mm) wide and 0.0118 in deep. 0.457 mm Level: guide. protolabs.com 2026-11-08
  164. Milled text strokes and spacing must be at least 0.020 in (0.5 mm); use recessed text, 20 pt or larger sans-serif. 0.5 mm Level: guide. protolabs.com 2026-11-08
  165. Protolabs says it can typically maintain ±0.005 in. (0.13 mm) machining tolerance for turning. 0.005 in Level: guide. protolabs.com 2026-11-08
  166. Turned walls under 0.020 in (0.5 mm) typically do not survive machining. 0.5 mm Level: must. protolabs.com 2026-11-08
  167. Diameter in any region of a turned part should be at least 0.030 in (0.76 mm). 0.76 mm Level: should. protolabs.com 2026-11-08Review note: Source says diameter 'should not be' less than 0.030 in and 'small features may be allowed'; must is over-strong.
  168. Axial holes on Protolabs lathes: minimum 1 mm diameter, depth up to 6x diameter. 1 mm Level: should. protolabs.com 2026-11-08
  169. Radial holes on Protolabs lathes should be at least 2 mm in diameter. 2 mm Level: should. protolabs.com 2026-11-08Review note: Source says radial holes 'should be at least' 2 mm (recommendation); must is over-strong.
  170. External grooves on turned parts cannot be narrower than 0.047 in (1.2 mm) or deeper than 0.95 in (24.1 mm). 1.2 mm Level: must. protolabs.com 2026-11-08
  171. SendCutSend lists ±0.009 in. tolerance for 0.125 in. laser-cut Delrin and ±0.005 in. for 0.270 in. CNC-routed Delrin. 0.005 in Level: guide. sendcutsend.com 2026-11-08
  172. Keep holes and features away from bends by the same multiples as the minimum flange table (for example 3.0x thickness plus bend radius for thin aluminum). Level: must. protolabs.com 2026-11-08
  173. Cold-rolled, galvanized steel and stainless sheet 0.024 to 0.050 in thick need a flange of at least 3.5x thickness plus bend radius. 3.5 x t + r Level: must. protolabs.com 2026-11-08
  174. Aluminum or copper sheet 0.060 to 0.090 in thick needs a flange of at least 2.5x thickness plus bend radius. 2.5 x t + r Level: must. protolabs.com 2026-11-08
  175. Steel and stainless sheet 0.060 to 0.090 in thick need a flange of at least 2.75x thickness plus bend radius. 2.75 x t + r Level: must. protolabs.com 2026-11-08
  176. All sheet metals 0.100 to 0.134 in thick need a flange of at least 2.5x thickness plus bend radius. 2.5 x t + r Level: must. protolabs.com 2026-11-08
  177. Stainless sheet 0.160 in and thicker needs a flange of at least 4.0x thickness plus bend radius. 4.0 x t + r Level: must. protolabs.com 2026-11-08
  178. Protolabs offers internal bend-radius tooling from 0.01 to 1 in.; larger radii use bump forming. 0.01 in Level: guide. protolabs.com 2026-11-08
  179. SendCutSend sets bend radius and K factor per material thickness; model with their values because custom radii are not offered. Level: must. sendcutsend.com 2026-11-08
  180. SendCutSend bends up to 130 degrees; angles beyond that are out of range. 130 deg Level: should. sendcutsend.com 2026-11-08
  181. For SendCutSend U- and C-channel bends, the sheet-metal base must be at least 2× flange length and the polycarbonate base at least 3×. 2 ratio Level: must. sendcutsend.com 2026-11-08
  182. A flat part sent to SendCutSend for bending must be at least 0.375 in by 1.5 in before bending. 0.375 in Level: should. sendcutsend.com 2026-11-08
  183. Sheet-metal hole diameter tolerance is plus or minus 0.005 in (0.127 mm). 0.005 in Level: should. protolabs.com 2026-11-08
  184. Protolabs lists hardware hole diameter tolerance as +0.003/-0.000 in. 0.003 in Level: guide. protolabs.com 2026-11-08
  185. SendCutSend resizes hardware holes to the catalog size with +0.003/-0.000 in. 0.003 in Level: should. sendcutsend.com 2026-11-08
  186. Protolabs lists bend-to-hole tolerance as ±0.015 in. on one surface and non-cumulative ±0.030 in. across two bends. 0.015 in Level: guide. protolabs.com 2026-11-08
  187. Bend-to-bend tolerance is ±0.015 in. for two bends and non-cumulative ±0.030 in. for three or more. 0.015 in Level: guide. protolabs.com 2026-11-08
  188. Edge to hole position on one surface is plus or minus 0.005 in. 0.005 in Level: should. protolabs.com 2026-11-08
  189. Countersink depth in sheet metal is at most 0.6x material thickness. 0.6 x t Level: should. xometry.com 2026-11-08
  190. Countersinks must be at least 8x thickness from each other, 4x from an edge and 3x from a bend. 8 x t Level: must. xometry.com 2026-11-08
  191. Outside radius of curls must be at least 2x material thickness. 2 x t Level: must. xometry.com 2026-11-08
  192. Xometry gives a return length of at least 4× material thickness for open and tear-dropped hems. 4 x t Level: should. xometry.com 2026-11-08
  193. In 0.125 in 5052, SendCutSend puts tapped and countersunk holes at least 0.315 in from the bend line to avoid distortion; look up your own thickness. 0.315 in Level: must. sendcutsend.com 2026-11-08
  194. When laser cutting, make holes and cuts at least 50% of material thickness. 50 % of t Level: should. sendcutsend.com 2026-11-08
  195. Minimum bridge width between two cut lines is 50 percent of material thickness. 50 % of t Level: guide. sendcutsend.com 2026-11-08
  196. For 0.100 in aluminum, minimum hole and bridge are both 0.050 in or more. 0.05 in Level: guide. sendcutsend.com 2026-11-08
  197. Laser-cut 6061-T6 holds plus or minus 0.005 in. 0.005 in Level: should. sendcutsend.com 2026-11-08
  198. Waterjet-cut carbon fibre sheet holds plus or minus 0.009 in. 0.009 in Level: should. sendcutsend.com 2026-11-08
  199. For a 1/4-20 tap in 0.125 in 5052, keep hole centre at least 0.167 in from the nearest cut feature; add radius plus the vendor table value. 0.167 in Level: must. sendcutsend.com 2026-11-08
  200. SendCutSend says there is no hole-to-edge rule of thumb because spacing varies with hole radius and part geometry. Level: guide. sendcutsend.com 2026-11-08
  201. PCBWay standard min trace is 0.1 mm (4 mil). 0.1 mm Level: should. pcbway.com 2026-11-08“Min Trace 0.1mm/4mil”
  202. PCBWay standard min spacing is 0.1 mm (4 mil). 0.1 mm Level: should. pcbway.com 2026-11-08“Min manufacturable spacing is 4mil(0.1mm)”
  203. JLCPCB 2 oz copper: minimum track width and spacing 0.16 mm (6.5 mil) on 2-layer boards. 0.16 mm Level: should. jlcpcb.com 2026-11-08
  204. HDI boards allow 3 mil track and space (2.7 mil extreme). 3 mil Level: should. jlcpcb.com 2026-11-08
  205. PCBWay's advanced roadmap lists 3/3 mil track and space for 2016, 2 mil inner and 3 mil outer for 2017, and partial-board 2/2 mil for 2018. 3 mil Level: guide. pcbway.com 2026-11-08“Minimum track line 3mil-inner 3mil-outer”
  206. For multilayer 1 oz JLCPCB boards, the recommended PTH annular ring is 0.20 mm or more and the absolute minimum is 0.15 mm. 0.2 mm Level: should. jlcpcb.com 2026-11-08
  207. For PCBWay pads with vias in the middle, minimum annular-ring width is 0.15 mm (6 mil). 0.15 mm Level: should. pcbway.com 2026-11-08“Min Width of Annular Ring 0.15mm(6mil)”
  208. On the advanced PCBWay service the annular ring can be kept at 3 mil. 3 mil Level: should. pcbway.com 2026-11-08“Min. annular ring can be kept 3mil 3mil 3mil”
  209. JLCPCB lists standard 2-layer and multilayer drill diameters of 0.15–6.3 mm, with 0.1 mm microvias for boards ≤1 mm thick using ENIG/OSP; holes ≥6.3 mm are routed. 0.15 mm Level: should. jlcpcb.com 2026-11-08
  210. PCBWay drills 0.15 to 6.0 mm; sizes under 0.2 mm or over 6 mm cost extra. 0.15 mm Level: should. pcbway.com 2026-11-08“Min drill size is 0.15mm, max drill is 6.0mm.”
  211. For 2-layer and multilayer boards, JLCPCB lists a 0.15 mm via hole with 0.25 mm via diameter, or 0.1/0.2 mm only for boards ≤1 mm thick with ENIG/OSP; the preferred minimum hole is 0.2 mm. 0.15 mm Level: should. jlcpcb.com 2026-11-08
  212. Smallest non-plated hole is 0.50 mm. 0.5 mm Level: should. jlcpcb.com 2026-11-08
  213. PCBWay flags non-plated holes under 0.45 mm for review. 0.45 mm Level: should. pcbway.com 2026-11-08“for non-plated holes, the min size should be 0.45 mm”
  214. JLCPCB lists through-hole tolerance of +0.13/-0.08 mm; ±0.05 mm press-fit tolerance applies to 0.55–2.0 mm finished circular holes on multilayer ENIG boards. 0.13 mm Level: guide. jlcpcb.com 2026-11-08
  215. PCBWay hole tolerance is plus or minus 0.08 mm plated, 0.05 mm non-plated. 0.08 mm Level: should. pcbway.com 2026-11-08“PTH: +/-0.08mm, NPTH: +/-0.05mm”
  216. JLCPCB lists CNC-routed tolerance of ±0.2 mm regular and ±0.1 mm high precision; high precision requires 50 × 50 mm minimum boards and three tooling holes, while ±0.4 mm V-cut tolerance requires ≥0.6 mm thickness. 0.4 mm Level: guide. jlcpcb.com 2026-11-08
  217. PCBWay outline tolerance is plus or minus 0.2 mm routed, 0.5 mm V-scored. 0.2 mm Level: should. pcbway.com 2026-11-08“±0.2mm for CNC routing, and ±0.5mm for V-scoring”
  218. Finished track width varies by plus or minus 20 percent; a 0.1 mm track ends between 0.08 and 0.12 mm. 20 % Level: should. jlcpcb.com 2026-11-08