Mechanical Engineer Interview Questions

"I was the kid who took apart the lawn mower to see why it wouldn't start, so mechanical was an easy pick. During my degree I found I liked the design side most, the part where you turn a rough idea into a drawing someone can actually make. My final-year project was a small test rig for measuring friction in bearings, and that taught me how much of engineering is getting the details right, like fits, fixtures and how you measure. After that I spent a year in a machine shop as a trainee, which showed me what happens to a drawing after it leaves the office. What I want now is design work that stays close to manufacturing, and from the job description that's exactly what this role is."

"From what I read, you make pumps and valves for water treatment plants, which means long service life, corrosion and sealing are the problems that matter most. That lines up with what I enjoy. In my last role I worked on a gearbox housing that had to survive outdoor use for years, so I spent a lot of time on material choice, coatings and seal grooves. I like products that run for a long time in tough conditions, because the design decisions really show up in the field. What I'd like to learn here is the hydraulic side in more depth, like how impeller design affects efficiency, because that's where I have the least hands-on experience today."

"In my final-year project our team built a pedal-driven water pump for small farms. The goal was to lift water about five metres using one person pedalling at a comfortable pace. My part was the drive: I picked a chain and sprocket ratio so the crank ran at a normal cycling speed while the pump got the stroke rate it needed, and I sized the crank bearings from the loads. I also designed the frame in CAD and checked the weld joints by hand calculation. When we tested it, we hit the flow we'd aimed for, but the frame flexed more than I'd expected, so we added a cross brace. If I did it again I'd run a quick stiffness check early instead of only checking strength."

"At my last company we made a welded steel cover that took a welder almost an hour each. I noticed most of the welds only joined flat sheets at right angles. I redesigned it as one sheet with bent flanges and just a few tack welds at the corners, which the press brake team could make easily. I checked the stiffness with a simple simulation and it came out close to the old design. We made five covers, fitted them on test machines and ran them through the normal vibration test. They passed, so we released the new drawing. Welding time dropped to a few minutes a part and the finished cover was a little lighter too."

"The first law says energy isn't created or destroyed, it just changes form. For a closed system, the heat added minus the work done by the system equals the change in its internal energy. A kettle is a simple example: the electrical energy goes into the water as heat and raises its temperature. The second law is about direction. Heat flows by itself from hot to cold, never the other way, and you can't build an engine that turns all the heat it takes in into work; some always has to be rejected to a colder place. A car engine shows it: a big share of the fuel's energy leaves as hot exhaust and through the radiator. Another way to say it is that the total entropy of an isolated system never goes down."

"A heat engine has to reject some heat to a colder place to keep running in a cycle, so it can never turn all of its heat into work. The best any engine can do between two temperatures is the Carnot efficiency, which is 1 minus the cold temperature over the hot temperature, and both have to be in kelvin. Here that's 1 minus 300 over 800. 300 over 800 is 0.375, so the limit is 0.625, a bit under two-thirds. A real engine between those temperatures would do noticeably worse because of friction, heat losses, pressure drops and processes that aren't reversible. The formula also tells you how to improve things: raise the hot temperature or lower the cold one."

"Conduction is heat moving through a material, like a spoon getting hot in a pan. Convection is heat carried away by a moving fluid, like air blowing over a car radiator. Radiation is heat sent as electromagnetic waves, which needs no medium, like feeling the sun or a fire on your face. For the wall it's pure conduction, so I use Fourier's law: heat flow equals k times area times temperature difference, divided by thickness. That's 0.8 times 10 times 20, which is 160, divided by 0.2, which gives 800 watts. That treats 20 K as the difference between the two wall surfaces. If it were between the air inside and outside, I'd add the air films on each side as extra resistances in series, which would bring the number down."

"First continuity: the same volume per second passes both sections, so area times velocity is constant. Halving the diameter cuts the area to a quarter, so the speed goes up four times, from 2 to 8 metres per second. Then Bernoulli, and since the pipe is level the height terms cancel. The pressure drop equals half the density times the difference of the velocities squared. That's 0.5 times 1,000 times 64 minus 4, so 500 times 60, which is 30,000 pascals, or 30 kPa lower in the narrow part. This is the idea behind a venturi meter. It assumes steady, incompressible flow with no friction, so in a real pipe the drop would be a bit larger because of losses."

"Stress is the internal force per unit area, so it's force divided by cross-section. Strain is how much the part deforms relative to its size, change in length over original length, and it has no units. In the elastic range they're linked by Young's modulus. For the bar, stress is 10,000 newtons over 100 square millimetres, which is 100 newtons per square millimetre, or 100 MPa. Steel's modulus is about 200 GPa, which is 200,000 MPa, so strain is 100 over 200,000, which is 0.0005. Multiply that by the 2,000 millimetre length and it stretches about 1 millimetre. That's fine as long as 100 MPa is well under the steel's yield strength, which it is for ordinary structural steel."

"A factor of safety is how much stronger the part is than it needs to be. You take the failure strength, often yield for a ductile metal, and divide by the actual working stress. So if the steel yields at 250 MPa and the part sees 100 MPa, the factor is 2.5. How high I go depends on what I don't know. If the loads are well measured, the material is certified and the part is tested, a lower factor is fine and saves weight and cost. If loads include shocks, the material varies, or a failure could hurt someone, I go higher. And for things like pressure vessels or lifting gear, the design code sets the minimum, so I follow that rather than pick my own."

"It starts as a straight line, where stress is proportional to strain and the slope is Young's modulus. That holds up to the proportional limit, and just past it is the elastic limit, the last point where the bar would spring fully back. Mild steel then shows a clear yield point, often an upper and a lower yield, and stretches at roughly constant stress for a while. After that it strain hardens, so the stress climbs again until it peaks at the ultimate tensile strength. Past that, the bar starts necking in one spot, and because the engineering curve uses the original area, the stress appears to drop until it fractures. A brittle material like cast iron would break with almost no plastic part at all."

"For a cantilever with an end load, the bending moment is biggest at the wall, and it's load times length, so 1,000 newtons times 1,000 millimetres, which is one million newton millimetres. The second moment of area for a rectangle is b h cubed over 12. That's 20 times 60 cubed over 12, and 60 cubed is 216,000, so I is 360,000 millimetres to the fourth. The distance from the neutral axis to the outer edge is half the depth, 30 millimetres. Bending stress is M y over I, so one million times 30 over 360,000, which is about 83 MPa. That peak sits at the fixed end, tension on the top face and compression on the bottom, for a load pushing down."

"I'd start from what the bracket has to survive. It'll see vibration from the machine, so fatigue matters as much as static strength, and it'll sit outside in rain, mud and maybe fertiliser, so corrosion is a big one. Then I'd look at how it gets made and in what volume, say a few thousand a year, cut, bent and welded from plate. My shortlist would be a structural steel with galvanising or powder coat, and an aluminium alloy. Aluminium is about a third of the weight but also about a third as stiff, it has no true fatigue limit, and welding weakens heat-treated grades near the weld, so it needs a bigger section. For farm gear where weight isn't critical, I'd most likely pick coated steel for strength, toughness, weldability and cost."

"Casting is my pick when the shape is complex or has internal passages, like a pump housing or an engine block. It's cheap per part once the pattern or die is made, but you have to design for porosity and shrinkage, so strength is less predictable. Forging is for parts that take high or repeated loads, like crankshafts or connecting rods, because shaping the metal gives a grain flow that follows the part and good fatigue strength. The shapes are simpler and the dies cost a lot, so it needs volume. Machining from solid suits prototypes, low volumes or anything needing tight tolerances. In practice they're often combined: most cast and forged parts are finish-machined on the surfaces that matter."

"First I'd quarantine the batch and have quality re-measure a sample, so we know how many are off and by how much. Then I'd look at what that dimension actually does. If it's a clearance on a non-critical face and the stack-up still works, I might recommend using them as is, but only through the formal non-conformance or concession process, signed off by quality and the design owner, not a quiet yes. If it's a bearing fit or anything affecting safety, I'd say no, and look at rework, sorting the good ones, or another source. Either way the line manager hears the plan early. Then I'd ask the supplier for a root cause and corrective action, so the next batch is right."

"Plus-minus dimensions leave two gaps. They don't say where to measure from, and for a hole position they create a square tolerance zone, even though a round zone that touches the square's corners gives about one and a half times the area and still works. GD&T fixes both. Datums define how the part sits for inspection: the primary datum takes three points of contact, the secondary two, the tertiary one. A feature control frame reads left to right: the characteristic, like position, the tolerance zone, with a diameter symbol if it's round, any material modifier, then the datums in order. With position at maximum material condition, the smallest allowed hole, any extra size the hole actually has adds to the position tolerance as bonus."

"At my last company I designed a lifting fixture for moving motor housings between two machines. I started by writing the requirement down with the operators: the weight, the pick points, and that an operator had to hook it on with one hand. I sketched three concepts and compared them on weight, cost and how easy they'd be to use, then picked a simple welded frame. I hand-calculated the main members and welds, then checked the frame with a quick simulation, which agreed within reason. The design review caught that one pin could be put in backwards, so I made it asymmetric. We built one, load tested it at the level our lifting procedure required, and only then released the drawing with the rated load marked on it."

"First I'd find out why it breaks. Usually it's features referencing edges or faces that get renamed when geometry changes, or sketches that aren't fully defined, so they wander. With a review next week, I wouldn't rebuild everything at once. I'd fix the features the review depends on, re-referencing them to base planes, axes or a main sketch, so key dimensions drive the model cleanly. I'd also check the model still matches the last released drawing. After the review, I'd plan a proper rebuild with clear design intent: key dimensions up front and a sensible feature order. And I'd note what I found for the team, so the next model doesn't end up the same way."

"In a four-stroke engine, the piston goes down to draw in air or mixture, up to compress it, down again on the power stroke after combustion, and up to push the exhaust out. That's one power stroke for every two turns of the crankshaft. The big difference is ignition. A petrol engine compresses a fuel-air mixture and lights it with a spark plug. A diesel compresses only air, so hard that it gets hot enough to ignite the fuel as it's injected. That's why diesels use much higher compression ratios, and petrol engines can't, because the mixture would knock. Petrol engines also control power by throttling the air, while diesels just change how much fuel they inject."

"In the ideal cycles, the Otto adds all its heat at constant volume, at the top of the stroke, while the Diesel adds heat at constant pressure as the piston is already moving down. For the same compression ratio, the Otto gets more expansion out of its heat, so it comes out ahead. But real engines don't run at the same ratio. A petrol engine is limited by knock, because it's compressing a fuel-air mixture. A diesel compresses only air, so it can use a much higher compression ratio, and efficiency rises with compression ratio. Diesels also don't throttle the intake, so they avoid pumping losses at part load, and they run lean. Together those more than make up for the cycle difference."

"The compressor takes in low-pressure refrigerant vapour and squeezes it to a high pressure, which also makes it hot. In the condenser it gives off heat to the surroundings and turns into a liquid. The expansion valve then drops the pressure suddenly, so the refrigerant comes out as a cold mix of liquid and vapour. In the evaporator it boils at that low pressure, taking heat from the space you're cooling, and goes back to the compressor as vapour. The COP is the cooling you get divided by the work you put in, so 5 over 1.25 is 4. By the first law, the condenser has to reject both, so it gives off 6.25 kW. A COP above one is normal, because you're moving heat, not making it."

"Breakdown maintenance means you run the equipment until it fails and then fix it. That's actually sensible for cheap, non-critical things with a spare on the shelf, like a light fitting or a small bench fan. Preventive maintenance is done on a schedule, by time or running hours, like changing a gearbox oil every so many hours whether it needs it or not. It cuts surprise failures but can replace parts that still had life left. Predictive, or condition-based, maintenance watches the machine's actual condition with things like vibration readings, oil analysis or thermal imaging, and you step in when the trend says a failure is coming. I'd use it on critical, expensive equipment, like a main compressor, where downtime hurts most."

"On a packaging line at my last plant, a conveyor drive bearing was failing every few weeks. The quick fix had been to fit a new bearing each time. I pulled the old bearings and saw wear marks on one side of the race, which pointed to misalignment rather than a bad bearing. I checked the motor and gearbox with a dial gauge and found the gearbox base plate had cracked and let the gearbox shift under load. So the real cause was the base plate, and the bearing was just the victim. We replaced it with a stiffer plate, realigned everything, and added the alignment check to the preventive schedule. After that the bearing ran for months without an issue."

"First I'd protect the fracture surfaces, so nobody fits the pieces back together and damages them. Then I'd gather the history: running hours, loads, any recent changes, and the drawing and material certificate. On the surface, a failure after months usually points to fatigue, and I'd look for beach marks spreading out from an origin. The origin is often at a stress raiser like a keyway end, a sharp shoulder fillet or a press-fit edge. A small final fracture zone suggests low nominal stress with a long crack growth, while a large one suggests high stress. I'd check the fillet radius and surface finish against the drawing, test hardness, and look at alignment and loading. The fix might be a larger radius, better finish or correcting misalignment."

"First I'd make sure nobody's working on it without isolating and locking it out, and I'd get the standby pump running if we have one, to protect output. Then I'd find out what's tripping it. Is it motor overload, high temperature or a pressure switch? I'd ask the operators when it started and whether anything changed, like a valve setting or a new batch of fluid. I'd check the obvious things: motor current, bearing temperature, vibration, suction pressure and strainers. A blocked strainer or a half-shut suction valve starves the pump and makes it cavitate, while a pump pushed far out on its curve, say by a leak downstream, can overload the motor. Once it's running safely, I'd write up what I found and hand it over to the day team, so the root cause gets properly fixed."

"In my first year, I drew a machined bracket with sharp internal corners in a deep pocket. The machinist came to my desk and told me he couldn't cut a square corner with a round cutter, and the long thin tool needed would chatter. I went down to the shop with him and he showed me the tools he actually had. I changed the corners to a radius slightly bigger than his cutter and made the pocket shallower, which the load calculations allowed. It cut the machining time and the part came out cleaner. Since then I check internal radii and tool reach on every machined part, and I try to get a machinist to look at new drawings before they're released, not after."

"On a shaft drawing, I put a tolerance on a bearing seat that gave a loose fit instead of the light press fit the bearing needed. I'd read the wrong row in the fit table. It wasn't caught at review, and it only showed up when the fitters said the bearings were sliding on by hand, which they never should. I told my lead the same day, worked out which parts were affected, and issued a corrected drawing. The shafts already made were saved with a bearing retaining compound, as a one-off concession that quality and my lead signed off. Since then, I check every fit against the bearing maker's recommendation, not just the general table, and I ask a colleague to check fits on any new rotating part."

"First I'd make sure I'm right, by re-checking my loads and assumptions and asking a colleague to look at it quickly. If the number holds, I'd go to my manager straight away with the calculation, the standard it falls short of, and what could happen if the bracket fails. I'd come with options, not just a problem: maybe a thicker plate or an added gusset that could be done today, or releasing the rest of the design with this one bracket on hold. What I wouldn't do is release it quietly and hope. If they still insisted, I'd put my concern in writing and take it to whoever owns design sign-off. A late part is fixable; a failed one may not be."

"To me, a good safety culture is one where people report near misses without worrying about blame, anyone can stop a job if it looks wrong, and lockout is done every time, not just when a supervisor is watching. My part is first to set the example: I wear the right gear on the floor and never skip isolation because it's a quick job. As a design engineer I also have a bigger lever, because safety starts in the drawing. I think about guards, lifting points, pinch points and how a technician will reach something for maintenance before the design is released. If a machine is awkward to service safely, people will find a shortcut, so the design should make the safe way the easy way."

"I like to get everyone talking early, before the design is set, because most problems happen at the interfaces. On my last project, we had a quick weekly meeting where we went through the shared assembly model together. The electrical engineer could see where cable runs and enclosures had to go, and the controls engineer could check sensor positions and what the motors needed. I also try to explain mechanical limits plainly, like why a motor mount can't move without redesigning a frame, and ask about their limits in return, like sensor range or wiring rules. It stops the classic problem of finding out at assembly that a cabinet door can't open because a guard is in the way."