GATE PI common mistakes
GATE PI · Understand the traps and how to avoid them.
Spending 60+ minutes on the 1-mark questions first because they 'look easy', then running out of time on the 2-mark numericals that carry the same marks per minute of work.
The paper mixes 1-mark and 2-mark questions with no section split, and 1-mark items often hide multi-step calculations (e.g. a quick Taylor tool-life calculation) that feel cheap but are not. Candidates anchor on 'do the small ones first' without checking the actual time cost per question.
What to do instead
Before the exam, do 3–4 full-length timed mocks and record minutes-per-mark for each question you attempt. Build a rule: if a 1-mark question is not solved in ~2 minutes, mark it and move on. In the first pass, only take questions you can finish in under 90 seconds; return to the rest in the second pass. Track cumulative time against a per-section budget (e.g. 25 minutes for Aptitude, 55 for Production, 55 for Industrial, 15 for review).
Attempting every question with no negative-marking discipline, guessing blindly on 2-mark numericals where four options are plausible calculated values.
GATE applies negative marking (typically 1/3 for 1-mark and 2/3 for 2-mark MCQs; no negative for MSQ and NAT). Candidates who 'never leave a bubble empty' lose marks because random guessing on a 2-mark MCQ has expected value of roughly (0.25×2) − (0.75×0.667) ≈ −0.0, i.e. break-even at best, and worse when one distractor is a common-error answer. Many candidates do not know which question types are exempt from negative marking.
What to do instead
Memorise the marking scheme for the current year's paper and apply a two-option rule: guess only when you can eliminate at least two options, and never guess on 2-mark MCQs without elimination. Always attempt NAT and MSQ questions you can partially reason about, since there is no penalty. In mocks, log every blind guess and compute your net guess score; most candidates find it is negative and should stop.
Losing marks on machining-parameter numericals by mixing units - using feed in mm/rev where mm/min is needed, or cutting speed in m/min with diameter in mm without the 1000 factor.
Production Engineering formulas (turning time = L/(f·N), N = 1000V/(πD), MRR = V·f·d) are simple, but the paper deliberately gives diameter in mm, speed in m/min, and feed in mm/rev in the same question. Under time pressure candidates substitute directly and get answers off by 1000 or π, then pick the matching distractor.
What to do instead
Adopt a fixed unit protocol: write the formula, then write every given quantity with its unit converted to a consistent set (mm, min) before substituting. For turning, always compute N = 1000V/(πD) as a separate line and sanity-check it (e.g. 100 m/min on a 50 mm bar gives ~637 rpm, not 6.37 or 6370). Practise 20 mixed-unit numericals until the conversion is automatic.
Confusing metal-cutting tool-life and force relations - applying Taylor's equation with the wrong exponent sign, or swapping the specific cutting energy and the cutting-force coefficient in Merchant's circle.
Taylor's VT^n = C requires n between roughly 0.1 and 0.5 and is often rearranged as V = C/T^n; sign and exponent errors are easy under pressure. Similarly, Merchant's shear-angle relation and the force diagram (Fc, Ft, Fs, N) are frequently memorised as a picture rather than derived, so candidates mis-map which component is which when the question rotates the geometry.
What to do instead
Re-derive Taylor from the log-linear form once by hand and note that doubling T reduces V by a factor of 2^n (a small change, ~1.1–1.4×), which is a fast sanity check. For Merchant, redraw the circle for every problem, label the rake angle, friction angle and shear angle explicitly, and check that the shear angle plus friction angle minus rake equals 45° in the minimum-energy condition. Do 10 problems where the geometry is rotated or the rake is negative.
Treating casting and welding calculations as memorised facts rather than applying the governing relations - e.g. using Chvorinov's rule with the wrong characteristic dimension, or forgetting the arc-efficiency and heat-input factors in welding.
Casting solidification time (t = k(V/A)^2) needs the correct surface area of the actual mould cavity, and candidates plug in the riser or the pattern dimension. In welding, heat input is Q = ηVI/v, and the melting-efficiency and arc-efficiency terms are often dropped, giving answers that differ by 20–80%. These topics appear in a small number of questions, so they get under-practised relative to machining.
What to do instead
For Chvorinov, always write V and A for the specific cavity shape and confirm the units of V/A are length. For welding, write the heat-input formula with η explicitly and keep a short table of typical efficiencies (e.g. SAW high, GMAW moderate, TIG lower) to sanity-check. Solve one numerical each for a cylindrical riser, a plate weld, and a resistance-spot weld before the exam.
Misapplying the metal-forming yield and load formulas - using the wrong flow-stress model in forging, or forgetting the friction/back-tension term in wire drawing and rolling.
Forming questions give a stress-strain relation (e.g. σ = Kε^n) and expect the average flow stress, but candidates integrate incorrectly or use the yield stress at the final strain. In drawing, the ideal work term is often used while the question includes die angle and friction, and in rolling the condition for bite (tan α ≤ μ) is confused with the neutral point.
What to do instead
Memorise that average flow stress for σ = Kε^n over a strain range is K·ε^n/(n+1) evaluated appropriately, and always state which stress model you are using. For drawing, write the total work as ideal deformation work plus friction and redundant work, and check whether the question supplies a friction coefficient or die angle before choosing the simple formula. For rolling, separately check the bite condition and the neutral-point location.
Skipping the non-traditional machining and metrology/measurement topics because they 'rarely appear', then meeting 2–3 questions on ECM material removal rate or on limits, fits and gauge design.
Candidates allocate revision by perceived frequency and drop ECM/EDM/USM and the fits-and-tolerances portion of metrology. But ECM MRR (Faraday-based, I/(F·z) with atomic weight) and tolerance-stack questions are formulaic and quick marks if prepared, and they appear often enough to matter in a competitive paper.
What to do instead
Build a one-page formula sheet for ECM (MRR from current and valency), EDM (MRR vs discharge energy), USM (MRR vs amplitude and frequency), and for fits: hole/shaft basis, clearance/interference from maximum and minimum metal conditions, and gauge tolerance (Taylor's principle: GO gauge checks maximum material condition and should be full-form). Solve 5 questions per topic; these are high-yield per hour of study.
Getting sequencing and line-balancing numericals wrong by computing cycle time from total work content instead of the bottleneck station, and by confusing cycle time with throughput time.
Assembly-line balancing questions give task times and a required output rate; candidates divide total task time by the number of stations and call it cycle time, or compute the theoretical minimum number of stations and forget to round up. The terminology (cycle time, lead time, throughput time, takt time) is used loosely in textbooks, so the wrong quantity gets substituted.
What to do instead
Fix definitions: takt time = available time / demand; cycle time of the line = largest station time; minimum stations = ceil(sum of task times / cycle time). Always compute the bottleneck station time explicitly and check line efficiency = sum of task times / (number of stations × cycle time). Practise precedence-diagram balancing with a stated demand rate so the takt time must be derived, not assumed.
Mixing up inventory and forecasting formulas - using EOQ with the wrong holding-cost basis (per unit vs percentage of unit cost), or applying exponential smoothing with α and (1−α) swapped.
EOQ = sqrt(2DS/H) requires H as an annual holding cost per unit; questions often give holding cost as a percentage of item cost, and candidates forget to multiply by the unit price. In exponential smoothing, F(t+1) = αD(t) + (1−α)F(t), and the roles of α and 1−α are easy to reverse under time pressure, especially when the question asks for the forecast two periods ahead.
What to do instead
Write the EOQ derivation once and note the units of H (currency per unit per year); convert percentage holding costs to per-unit before substituting. For smoothing, always write the recursion with the newest demand multiplied by α, then compute period by period in a small table rather than in your head. Check that a high α tracks recent demand closely - if your result does not, you swapped the terms.
Losing easy marks in PERT/CPM and queuing by computing only the critical path length and not the floats, or by using the wrong service-rate convention in queue models.
PERT/CPM questions frequently ask for total float, free float, or the probability of completion by a due date, which requires the variance of the critical path (sum of activity variances) and a normal-table lookup. Candidates stop at the critical path. In queuing, λ and μ are often given as rates but sometimes as mean times, and the utilisation ρ = λ/μ must be less than 1 for the steady-state formulas to apply.
What to do instead
For networks, always tabulate earliest start/finish and latest start/finish for every activity, then compute total float and free float from the table. For the probability question, sum variances only along the critical path and convert to z = (T − mean)/σ. For queues, convert any mean-time data to rates first, check ρ < 1, and write the model name (M/M/1, M/M/c) before substituting.