Mechanical Engineer Interview Questions for Freshers

“I'd say strength of materials. I liked that you can take a real part, draw a free body diagram and get a number that tells you whether it survives. I used it properly in my final-year project, where our team built a small manual press. I did the hand calculations for the frame and the screw, picked the section sizes, and then checked the frame in a simulation tool to see if my numbers were close. They were within a sensible margin, and where they weren't, I found I'd missed a load on one member. That project made the subject feel practical rather than just exam work. So if you'd like to test me, that's the area I'd most like to be tested on.”

“In my final-year project, four of us built a small solar dryer for crops. We split on the frame: two of us wanted welded mild steel, and two wanted bolted aluminium sections so it would be light and easy to carry. It started going round in circles, so I suggested we list what actually mattered: weight, cost, strength and how easily a farmer could repair it. I did quick hand calculations for both frames, and another teammate got prices from a local supplier. Steel came out stronger and cheaper and easy for any local welder to fix, but aluminium was much lighter. We went with steel but used thinner sections, which kept the weight reasonable. It held up fine in testing. What I took away is that putting numbers on the table ends arguments faster than opinions do.”

“I did six weeks of training at a small company that machines parts for pumps. In class, a tolerance was just a number on a drawing. On the floor, I watched an operator spend a long time holding a very tight tolerance on a face that, when I asked, didn't mate with anything. The supervisor explained that the drawing had been copied from an older part, and nobody had questioned it. That stayed with me. Every tight tolerance costs machine time and scrap, and the person paying that cost isn't the one who drew it. I also learned to ask operators first when something looks odd, because they usually know the reason. When I make drawings, I want to be able to say why each tolerance is there.”

“The first thing I'd remember is that the modulus of steel hardly changes between grades, so if my number is far off, the problem is almost certainly my measurement, not the steel. The most common cause is how strain was measured. If we used the crosshead movement of the machine instead of an extensometer on the specimen, that movement includes the grips slipping and the machine frame stretching, so the strain looks bigger and the modulus comes out lower. I'd also recheck the cross-section area with a micrometer, the gauge length, and my unit conversions, because a millimetre and metre mix-up is easy. Then I'd look at the curve and make sure I took the slope only from the straight elastic part, not the early bedding-in region. If possible, I'd repeat the test with an extensometer and compare.”

“Strength is how much stress a material can take before it yields or breaks, and we get it from a tensile test. Hardness is resistance to indentation or scratching, measured with a Brinell, Rockwell or Vickers test. Toughness is the energy a material can absorb before it fractures, which is roughly the area under the stress-strain curve, and we check it with an impact test like Charpy or Izod. Ductility is how much it can deform plastically before breaking, shown as elongation in the tensile test. They don't always go together. Glass, or a hardened steel file, is very hard, but it's not tough: hit it sharply and it shatters. Mild steel is less hard but tough and ductile, which is why it bends and warns you before it fails.”

“In welding, the base metals themselves melt at the joint and fuse, usually with a filler metal added, so a good weld can be as strong as the parent metal. In brazing, the base metals don't melt. Only a filler metal melts, at a temperature above about 450 °C but below the base metal's melting point, and it's drawn into a tight joint by capillary action. Soldering works the same way but with a filler that melts below that temperature, so the joint is weaker. It's used for electronics and small plumbing joints. I'd choose brazing over welding when I'm joining two different metals, like copper to steel, when the parts are thin and would burn through or distort with welding, or when I need a neat, leak-tight joint in something like refrigeration tubing. The trade-off is lower strength at high temperature.”

“On a lathe the work turns and a single-point tool removes material. The common operations are plain turning to reduce diameter, facing the end square, taper turning, cutting threads, drilling and boring along the axis, knurling, and parting off. For the speed, cutting speed is the surface speed of the work past the tool, V equals pi D N over 1000 when D is in millimetres and V is in metres per minute. So N is 1000 V over pi D. That's 1000 times 100, which is 100,000, divided by pi times 50, which is about 157. That gives roughly 637 rpm. On a real lathe with fixed speeds, I'd pick the nearest step below that, not above, so I don't overheat the tool. The right cutting speed itself depends on the tool material and the workpiece material.”

“All four heat the steel and then control how it cools. In annealing, I heat it above the critical temperature, hold it, and let it cool very slowly in the furnace. That makes it soft and ductile and relieves stresses, so it's easier to machine or form. Normalising uses a similar temperature but lets the part cool in still air. That gives a finer, more uniform grain, so it's a bit stronger and harder than annealed steel. Hardening means heating and then quenching quickly in water or oil, which forms a very hard structure called martensite, but the part becomes brittle and full of internal stress. That's why we temper it afterwards. Tempering reheats it to a lower temperature, below the critical point, which gives up some hardness but gains a lot of toughness, so a tool or gear won't crack in use.”

“Temperature tells you how hot something is. It's linked to the average kinetic energy of the molecules and it's measured in degrees or kelvin. Heat is energy that moves from a hotter body to a colder one because of a temperature difference, and it's measured in joules. A bathtub of warm water holds far more thermal energy than a cup of boiling water, even though the cup is at a higher temperature. For the calculation I'd use Q equals m c delta T. The mass is 2 kg, the specific heat of water is about 4.18 kJ per kg per kelvin, and the rise is 60 kelvin. So 2 times 4.18 times 60 gives about 502 kJ, ignoring any heat lost to the surroundings or the container.”

“A system is just the part I choose to study, with a boundary around it. In a closed system, energy like heat or work can cross the boundary but mass can't. Gas trapped in a cylinder with the valves shut is a good example. In an open system, both mass and energy cross, so a turbine, a pump, a nozzle or a car radiator are open systems, and we usually analyse them as steady flow. An isolated system exchanges neither mass nor energy. A perfectly insulated flask is the nearest real example. As for properties, intensive ones don't depend on how much stuff there is, like temperature, pressure and density. Extensive ones scale with size, like mass, volume and total energy. If I cut a tank of gas in half, the temperature stays the same but the volume halves.”

“Adiabatic compression needs more work. In isothermal compression, heat is removed as I compress, so the temperature stays constant and all the work I put in leaves as heat. In adiabatic compression no heat leaves, so the work I do raises the gas's internal energy and its temperature goes up. A hotter gas pushes back harder, so the pressure rises faster for the same drop in volume. On a p-V diagram the adiabatic curve is steeper, and the area under it, which is the work for a compression process, is larger. You can feel this with a bicycle pump, which gets warm at the end. It's also why industrial air compressors often use two or more stages with an intercooler in between. Cooling the air between stages brings the process closer to isothermal and saves power.”

“A four-stroke engine completes intake, compression, power and exhaust over two crankshaft revolutions, so it fires once every two turns. A two-stroke does the same job in one revolution, so it fires every turn. Small petrol two-strokes usually use ports in the cylinder wall instead of valves and a camshaft, which makes them light, simple and powerful for their size. That's perfect for a chainsaw, a leaf blower or a small outboard. The downside is that intake and exhaust overlap, so some fresh fuel-air mixture escapes straight out of the exhaust. Many small ones also mix oil with the fuel for lubrication, which burns. Both make fuel economy and emissions worse, which is why cars moved to four-strokes. Interestingly, very large marine diesel engines are two-strokes and very efficient, because they scavenge with air, not fuel mixture.”

“Power equals torque times angular speed. With speed in rpm, that's P equals 2 pi N T over 60. Rearranging, torque is 60 P over 2 pi N. So I'd take 60 times 10,000 watts, which is 600,000, and divide by 2 pi times 1440, which is about 9,048. That gives roughly 66 newton metres on the motor shaft. If a gearbox reduces the speed, the power stays nearly the same apart from friction losses, so the torque goes up by roughly the same ratio. A four to one reduction would give about a quarter of the speed and nearly four times the torque, a bit under because the gearbox isn't perfectly efficient. That's why the output shaft of a gearbox is thicker than the input shaft.”

“For a solid circular shaft in torsion, the maximum shear stress is 16 T over pi d cubed, so d equals the cube root of 16 T over pi times tau. I'd work in newtons and millimetres so the units match, because one megapascal is one newton per square millimetre. The torque is 500,000 newton millimetres. So 16 times 500,000 is 8 million, and pi times 50 is about 157. Eight million divided by 157 is about 50,900, and the cube root of that is roughly 37 mm. I wouldn't pick 37 mm, though. I'd go up to the next standard size, say 40 mm. And I'd mention that this only covers pure torsion. A keyway weakens the shaft, and if there's bending from pulleys or gears, or the load fluctuates, I'd need a combined stress and fatigue check, which could push it bigger.”

“Spur gears have straight teeth and connect parallel shafts. They're simple and cheap but noisy at high speed. Helical gears also connect parallel shafts, but the teeth are angled, so they engage gradually and run quieter. The catch is they also push along the shaft axis, so the bearings need to take that axial thrust. Bevel gears connect shafts that meet, usually at a right angle, like in a car's differential. A worm and worm wheel give a big reduction in one step between shafts at right angles that don't meet, and some worm sets can't be driven backwards, which is handy for hoists. For the calculation, speed ratio is the inverse of the tooth ratio. Sixty over twenty is three, so the output runs at 1200 divided by 3, which is 400 rpm, in the opposite direction, with roughly three times the torque minus losses.”

“Ball and roller bearings are rolling bearings, and a journal bearing is a sliding one. In a ball bearing the balls touch the races at a point, so friction is low and it suits high speeds and moderate loads, like an electric motor or a fan. A roller bearing has line contact, so it carries heavier radial loads for its size, which is why you see cylindrical rollers in gearboxes. Tapered roller bearings take radial and axial load together, which is why they're used in wheel hubs. A plain journal bearing is just the shaft running in a bush, separated by a film of oil once it's turning. It handles very heavy loads and shock quietly, so you see it on engine crankshafts and large turbines, but it needs good lubrication and suffers most at start-up.”

“A fit describes how a shaft and hole go together once you allow for their tolerances. In a clearance fit the largest shaft is still smaller than the smallest hole, so there's always a gap and the parts can slide or rotate, like a pin in a hinge. In an interference fit the smallest shaft is still bigger than the largest hole, so you need a press or heat to assemble it, like a bearing pressed onto a shaft or a gear hub. A transition fit sits in between: depending on the actual sizes, you might get a tiny gap or a slight interference, and it's used where location matters, like a dowel. In the hole-basis system, the hole tolerance is fixed, usually the H grade, and we vary the shaft tolerance to get the fit we want. That's common because holes are made with standard drills, reamers and plug gauges, while a shaft is easy to turn or grind to any size.”

“Both are ways of laying out orthographic views, but the views end up in opposite places. In first-angle projection, the object is imagined between the viewer and the projection plane, so the top view is drawn below the front view and the left side view appears on the right. In third-angle projection, the plane is between the viewer and the object, so the top view sits above the front view and the right side view appears on the right, which many people find more natural. Which one you use depends on the standard your company or customer follows. To tell which a drawing uses, I look for the projection symbol in the title block. It's a small truncated cone shown in two views, and the position of the circle view relative to the side view tells you first or third angle. I'd never guess from the layout alone.”

“The least count is the smallest length an instrument can read. For a vernier caliper, it's one main scale division minus one vernier scale division. On a common caliper, 50 vernier divisions span 49 millimetres, so the least count is 0.02 millimetres. For a micrometer, it's the screw pitch divided by the number of divisions on the thimble. A typical one has a 0.5 millimetre pitch and 50 divisions, so it reads to 0.01 millimetres. To take a reading, I first close it gently using the ratchet and check the zero. Then I measure the part, read the last visible mark on the sleeve, say 7.5 millimetres, and add the thimble line that lines up with the datum line, say 23 divisions, times 0.01. That gives 7.73 millimetres, then I correct for any zero error I found.”

“The Reynolds number compares inertia forces to viscous forces in a flow. When it's low, viscosity wins, the fluid moves in smooth layers, and the flow is laminar. When it's high, inertia wins and the flow becomes chaotic and turbulent. The formula is density times velocity times diameter over dynamic viscosity, or V D over the kinematic viscosity. For water at room temperature, the kinematic viscosity is about one times ten to the minus six square metres per second. So 1 metre per second times 0.05 metres, divided by that, gives about 50,000. For pipe flow, it's usually laminar below about 2000 and turbulent above about 4000, so this is well into turbulent flow. That matters because turbulent flow has higher friction losses in the pipe, but it also mixes and transfers heat much better.”

“Cavitation happens when the pressure at the pump's suction, usually near the eye of the impeller, drops below the vapour pressure of the liquid. The liquid starts to boil locally even though it's not hot, and vapour bubbles form. As those bubbles move into the higher pressure region of the impeller, they collapse violently, and over time that pits and erodes the metal. You'd notice a noise like gravel going through the pump, extra vibration, and a drop in flow or head. To prevent it, the net positive suction head available has to stay above what the pump needs, with some margin. In practice that means a short, wide suction pipe with few bends, not throttling the suction valve, keeping the liquid cooler, or placing the pump lower relative to the tank so there's more pressure at the inlet.”

“If the rail is free, it simply grows by alpha times length times the temperature rise. That's 12 times ten to the minus six, times 10,000 millimetres, times 40, which is 4.8 millimetres. If both ends are fully held, it can't grow, so that thermal strain, alpha times delta T, is squeezed back to zero and turns into stress. By Hooke's law, stress equals E times strain, so it's 200,000 megapascals times 12 times ten to the minus six times 40, which works out to 96 megapascals. It's compressive, because the rail wants to expand and the supports push back. The length doesn't appear in the stress at all, which surprises people. In a long slender member, that compression can also cause buckling, which is why railways and pipelines leave gaps or use expansion loops.”

“First I'd draw the free body diagram. Because the load is at the centre and the beam is symmetric, each support takes half, so the reactions are 5 kN each. I can confirm by taking moments about one support: the 10 kN load at 2 m gives 20 kN·m, and the other reaction at 4 m must balance it, so it's 5 kN. The shear force is plus 5 kN from the left support to the centre, then it drops by 10 to minus 5 kN up to the right support. The bending moment rises linearly from zero at the supports to a peak at mid-span, where the shear force changes sign. That peak is the reaction times half the span, 5 times 2, which is 10 kN·m. The general formula for a central point load is W L over 4, which gives the same answer.”

“In bending, stress is zero at the neutral axis and grows towards the top and bottom edges, following sigma equals M y over I. So the material near the middle does very little work, while the material far from the centre does most of it. An I-section puts most of its metal in the flanges, far from the neutral axis, which gives a much bigger second moment of area, and a bigger section modulus, than a solid rectangle of the same weight. The thin web in the middle carries the shear force and keeps the flanges apart. The weakness is that it's only strong in one direction. Bent about its weak axis, or twisted, it's far less stiff, and a long, unbraced I-beam can buckle sideways under load. That's why beams are braced and why hollow sections are used when there's torsion.”