PE Mechanical Machine Design common mistakes
PE Mechanical Machine Design · Understand the traps and how to avoid them.
Mixing mass and force units - carrying lbm into a stress or energy equation without dividing by gc (32.174 lbm·ft/(lbf·s²)), or converting horsepower with the wrong constant.
The exam mixes USCS and SI, and the handbook tables give density in lbm/in³ while loads are in lbf. Candidates plug numbers straight in and get answers that are off by a factor of 386 or 32.2, then pick the closest distractor.
What to do instead
Pick one unit system at the start of the problem and convert everything into it before computing. Write the hp conversions on your scratch sheet: 1 hp = 550 ft·lbf/s = 33,000 ft·lbf/min, and P(hp) = T(in·lbf)·n(rpm)/63,025. If a stress comes out in the thousands of psi or the ten-thousands range for a normal steel part, suspect gc before suspecting the method.
Combining tolerances statistically (root-sum-square) when the problem describes a worst-case or guaranteed-interference condition, or adding symmetric ± values to limit dimensions without first converting them to a nominal-and-tolerance form.
Statistical stack-up is faster to compute and candidates remember it from design-for-six-sigma material, so it gets reached for by default. Limit dimensions in the prompt also look additive, and the sign of an asymmetric tolerance is easy to lose.
What to do instead
Default to worst-case (arithmetic sum of absolute tolerances) unless the problem explicitly says statistical, states a Cpk or defect rate, or asks for 'probable' variation. Convert every limit dimension to nominal ± first, then sum. Check the answer against the worst-case physical assembly: the extremes must be achievable with every part at its limit.
Misreading a fits callout - treating H7/g6 as an interference fit, or assuming a fit is clearance based on the numeric grade rather than the letter case.
Candidates memorize a few common fits and pattern-match, and the upper/lower-case hole/shaft convention gets rusty. The IT grade number says nothing about clearance; only the fundamental deviation letter does.
What to do instead
Uppercase letter = hole, lowercase = shaft. Hole deviations A through H lie at or above basic size, so an 'H' hole paired with any shaft letter a through h gives clearance; j through n gives transition; p through zc gives interference. When in doubt, compute the minimum and maximum clearance from the actual tabulated limits rather than reasoning from the callout alone.
Computing bearing rating life as L = (C/P)^3 and reporting it directly in hours, or using the ball-bearing exponent for a roller bearing.
The catalog rating C is defined at one million revolutions, and candidates forget the implicit 'millions of revolutions' unit. The exponent difference (3 versus 10/3) is a single character in the handbook and is easy to skim past.
What to do instead
Keep the sequence explicit: L10 = (C/P)^p in millions of revolutions, p = 3 for ball, 10/3 for roller; then L10h = (10^6 / (60·n))·(C/P)^p with n in rev/min. Confirm C and P are in the same force units and that P is the equivalent dynamic load (including the X and Y factors if there is thrust), not just the radial load.
Multiplying both the alternating and mean stress components by the fatigue stress concentration factor Kf, or applying Kt directly without the notch sensitivity q.
Stress concentration is learned first as a single multiplier, and applying it 'everywhere' feels conservative. The distinction between Kt (geometric) and Kf (fatigue-effective) blurs under time pressure.
What to do instead
Compute Kf = 1 + q(Kt − 1) using the notch radius and material strength to read q. Apply Kf to the alternating component when entering Goodman/Gerber. For the mean component, apply Kf only if you are accounting for local yielding at the notch (Neuber's rule); otherwise the nominal mean stress is used. Also confirm whether the handbook chart is for bending, axial, or torsion - the values differ.
Using Se = 0.5·Sut as the endurance limit directly, skipping the Marin surface, size, load, temperature, and reliability factors.
The 0.5·Sut shortcut is the most memorable line in the fatigue chapter, and the modifying factors feel like refinements rather than required inputs. On short-duration problems candidates take the shortcut to save time.
What to do instead
Build Se = ka·kb·kc·kd·ke·kf·Se' every time. Use the surface factor coefficients keyed to the specified finish (ground, machined, hot-rolled, forged), take kb from the actual diameter (and note the no-size-effect range for axial loading), use kc = 1 for bending, 0.85 for axial, 0.59 for torsion, and apply ke for the stated reliability. If the answer choices differ by 10–30 percent, the missing Marin factor is almost certainly the cause.
Adding the entire external load to the bolt preload (Fb = Fi + P) instead of the joint-stiffness fraction C·P, or using the nominal bolt diameter in place of the tensile stress area At.
The bolt-joint stiffness ratio is the least intuitive part of the topic, and 'the bolt sees the whole load' is the naive physical picture. Nominal diameter is also the number printed on the callout.
What to do instead
Write the joint model before calculating: C = kb/(kb + km), Fb = Fi + C·P, Fm = Fi − (1 − C)·P, and separation occurs when P > Fi/(1 − C). Compute Fi from At·Sp with At from the thread table and a preload fraction tied to the connection type (roughly 0.75 of proof load for reused joints, higher only when stated). Check that At (not the nominal area) is used in the preload and in the bolt stress.
Taking a material property off the table without reading the condition or temper column - using 4140 annealed values when the drawing says quenched and tempered, or 6061-O when the spec says T6 - and using typical rather than minimum specified properties for a design check.
Property tables are dense and the condition column is narrow; candidates scan to the familiar alloy number. Handbook 'typical' values are also frequently the only ones in view, and they are 5–15 percent above the specified minimums.
What to do instead
Before using any number, state aloud the alloy, form, and temper demanded by the problem. Use minimum specified values for design calculations and typical values only for deflection or stiffness where the question asks for a nominal behavior. If the problem gives a hardness or a temper designation instead of strength, convert with the appropriate table and note the conversion is approximate.
Applying a distortion-energy (von Mises) or maximum-shear criterion to a brittle material such as gray cast iron, a ceramic, or a hardened tool steel with negligible ductility.
The ductile failure theories dominate coursework and practice, so they become the default. The material's ductility is not always stated explicitly - only the tensile and compressive strengths are given.
What to do instead
First classify the material using elongation at break (roughly under 5 percent is brittle) and check whether compressive strength differs markedly from tensile strength - a large difference is the tell for cast iron. For brittle materials use maximum normal stress (Rankine), or modified Mohr when the compressive strength is much larger. Also flag that stress concentration must be applied to brittle parts in both static and fatigue loading.
Converting hardness to tensile strength with the 500·HB approximation outside its valid range or across material classes (stainless, aluminum, cast iron, cold-drawn sections), or ignoring elevated-temperature strength reduction.
The hardness-to-strength correlation is quick and appears in many worked examples for carbon and low-alloy steels, so it gets generalized. Temperature effects are also easy to overlook when the problem states an operating temperature casually in the middle of a sentence.
What to do instead
Restrict the ~500·HB estimate to carbon and low-alloy steels in the mid-hardness range and prefer tabulated Sut when it is available. Do not apply Brinell-to-strength to aluminum, austenitic stainless, or cast irons. If the problem states an operating temperature above roughly 400 °F for steel or above the aging temperature for an aluminum alloy, go back to the temperature-derated property table rather than the room-temperature value.