How to Reduce Manufacturing Cost Through Mechanical Design

Every rupee below is hypothetical teaching maths. Not a BrahmWorks fee, not a supplier quote, not a yield and not a time from a factory. The method is the part you can reuse. The remainder at the end is smaller than the uncertainty in the inputs, which is the point.

A cheaper housing quote can raise the product cost. Ask which invoice moves.

Leave tax, freight and duty as their own lines. A design review that folds them into "unit cost" cannot tell you which decision to undo. Cosmetic polish belongs in the same table. It is a process, not a property of the CAD model.

Every custom part carries a drawing, a supplier, an incoming check and a way to be wrong. A bracket that exists only because a boss was placed badly is not a commodity. Move the boss, delete the bracket, and you delete a screw, a mask step and a rattle. Merging can also thicken a tool or force a slide. The test is the whole set of invoices, not a smaller BOM as a vanity number.

Look for left and right parts that are mirrors. Two mirror mouldings are two tools, or one confused cavity. A common part used twice is one tool, but only when the product can stand the symmetry. Do not force a common part that makes the assembly ambiguous. A lid that fits both ways round, and only works one way, is not a saving. It is a field failure with a shorter BOM.

Standard fasteners beat drawings. One thread-forming screw, one length if you can, bought standoffs instead of a turned custom spacer. Mix machine screws and thread-forming screws only when the joint demands it, and then make the difference obvious in the hand, not only in the work instruction.

A tight tolerance on every dimension is a machining strategy and a moulding argument you will lose. Hold tight where a fit, a seal or a safety gap needs it. Leave the rest at the process capability you and the supplier agree. A texture can hide a mild sink and remove a paint line. It can also demand more draft and a dearer tool finish. Choose it because the product needs the appearance, then put the draft in the model. Painting a moulded part to hide flow is often a second factory. If the product does not sell on that surface, do not build the factory into the cost.

Secondary machining on a moulded part should be a red mark in review. A drilled side hole, a milled flat, a tapped metal insert added because the boss was an afterthought: each one is a queue. Redesign into the draw, or accept the queue in the costed design and staff it. An uncosted queue reappears as "the moulder is late".

Pins that locate before screws clamp, and a shape that will not accept a reversed lid, remove time and remove a defect. Watch someone who did not design the unit build it. Count hesitations and extra hands, not only the stopwatch. A fixture you can design away is a cost removed. A fixture you still need should be simple enough that the pilot can build it, and the work instruction should match the fixture, not an ideal sequence from CAD.

Cable routes that depend on a technician "tucking it clear" are assembly cost and a pinched-wire return. Give the wire a channel or a clip that is in the tool. If the channel needs a slide, put that slide in the comparison below rather than pretending the tuck is free.

Invented stack, not a BrahmWorks product. An indoor logger: moulded housing, a small sheet-metal bracket, a board on two screws, a third screw through the bracket because the boss missed the board's mounting hole by a few millimetres. The bracket is masked before the housing is coated. None of those steps is doing product work. They are repairing the boss.

The redesign in the exercise moves the boss under the existing board hole, deletes the bracket and its screw, and drops the mask. It may add nothing to the tool if the boss was already in the draw. It may add a tool change if the cavity is already cut. Both branches are costed next, so the slogan "fewer parts" cannot outrun the steel.

Assume assembly labour at INR 6 a minute, which is INR 360 an hour, invented. Design A, with the bracket, takes 3.0 minutes: INR 18. Design B takes 1.5 minutes: INR 9. Labour saving is INR 9 a unit. The deleted bracket and screw are assumed at INR 16 a unit. Combined saving INR 25 a unit. Quantity assumption: 4,000.

Gross = 25 × 4,000 = INR 1,00,000.

Branch 1. The boss moves in CAD before any tool. Tool change INR 0. Engineering time assumed at INR 30,000. Remainder = 1,00,000 − 30,000 = INR 70,000 on these lines only.

Branch 2. The cavity exists. A tool change is assumed at INR 55,000, engineering still INR 30,000. Remainder = 1,00,000 − 85,000 = INR 15,000.

Branch 3. The moved boss also forces a slide, assumed at INR 85,000 extra, on top of branch 2. Remainder = 15,000 − 85,000 = INR −70,000. Fewer parts lose money.

At 2,000 units, branch 2 gross is only INR 50,000, which does not cover INR 85,000 of tool change and engineering. The same CAD edit is rational before the tool, marginal after a simple change, and a loss if it buys a slide or if the quantity is the smaller one. Tax, freight, scrap and the coating line's minimum charge are not in the sum. The INR 15,000 remainder is smaller than a wrong labour minute. Do not treat it as a decision. Treat the three branches as the decision.

The same arithmetic belongs in the production handoff, where a pilot will pay these minutes every day: How to Take a Hardware Prototype to Production.

Score invoices, not adjectives.

• Piece, tool, assembly, secondary, fasteners and service are separate lines.
• Each custom part has a reason that is not "it was in the prototype".
• Left and right parts are common, or the reason they cannot be is written.
• One fastener family, unless a joint truly needs another.
• Tight tolerances are listed against a function. The rest are the agreed process.
• Paint, print and post-machining are either deleted or present in the cost.
• Texture and draft are one decision.
• A reversed or skipped part is hard to assemble, not merely forbidden in text.
• Cables have a path that does not rely on a tuck.
• Any merge that needs a tool change or a slide is costed at the real quantity.
• The quantity in the sum is the quantity you will buy, not the pitch volume.
• A cheaper resin is checked against chemical, heat and snap duty before it is scored as a saving.

Not by itself. A resin that cracks on the disinfectant you specified, creeps in a latch, or moves with moisture until a fit opens, spends the saving on returns. Compare grades on the duty first. Only then compare price at the quantity, with the same tool assumptions. A family change can also scrap the tool's shrink allowance.

No. Screws are sometimes the service path and the only joint you can torque and inspect. Snap fits can be cheaper in the piece and expensive when a user or a technician must open the product without destroying it. Remove screws that exist because two parts were never designed to meet. Keep screws that are the designed joint.

No. Less material can mean a harder fill, a weaker boss or a warp you then fixture. Lightweighting pays when the material is a large share of the piece and the duty still holds. In a small enclosure the tool, the assembly and the secondary work usually dwarf a gram of resin. Weigh the invoice, not the part.

On the feature that is the product, and on anything that is a safety or a compliance argument. A thinner wall across a creepage path, a deleted strain relief, or a seal lip moved to avoid a slide are not cost reductions until the relevant requirement is re-checked. Cosmetic savings on a surface the buyer never sees are fair game. Savings that erase the evidence you still need are not.

Bring the BOM, the tool status, the assembly sequence and the quantity you will actually place. BrahmWorks can separate a CAD edit that deletes work from a change that only moves the invoice into the tool.

Review the next stage with BrahmWorks