The Cost Drivers Hidden in Your Design: A Machinist's Guide to Features That Drain Your Budget

Sharp Internal Corners: The Single Most Expensive Feature

We've written about fillets before, but the sharp internal corner deserves its place at the top of any cost-driver list because it's the most frequent and the most expensive. A rotating cutting tool is round. Every internal corner it cuts will have a radius equal to at least the tool's radius. If the drawing demands a sharp 90-degree corner, the machinist must either use a non-rotating process (broaching, slotting, EDM) or leave a radius and ask for a deviation. Both options are expensive. Broaching requires a custom tool and a secondary operation. EDM requires a shaped electrode, slow burn times, and a specialized machine. Both can add hundreds or thousands of dollars to a part that could have been milled in minutes if the corner had been modeled with a reasonable fillet.

The cost multiplier is nonlinear. A pocket with corners that allow a standard 12mm endmill can be roughed and finished in one efficient operation. The same pocket with corners that demand a 3mm endmill takes longer because the smaller tool must take lighter cuts, and the cycle time balloons. If the corner must be smaller than 1mm, the part may need to be EDM'd, multiplying the cost by five to ten times.

Old Li's rule: "Every time you halve the corner radius below the natural tool size, you double the machining time. Every time you demand a zero radius, you quadruple the cost. The sharp corner is a luxury; make sure the function justifies the price."

Deep, Narrow Cavities and Pockets

Deep features — pockets, slots, holes — where the depth is more than about 4 times the diameter or width, force the use of long-reach tooling. Long tools deflect. Deflection causes taper, chatter, poor surface finish, and broken tools. To compensate, the machinist must slow down dramatically, take lighter cuts, and often use multiple tools of increasing length to "step down" into the cavity. The cycle time for a deep pocket can be ten times that of a shallow pocket of the same area.

Deep cavities also trap chips. In aluminum, recut chips can weld to the tool. In steel, they can pack and snap the cutter. The machinist must add pecking cycles, dwells, and coolant-flushing pauses, all of which add time. If the cavity is blind (closed at the bottom), chip evacuation is even worse; a through-hole at the bottom, if the design allows, can slash the cycle time by providing a chip escape path and a coolant entry point.

We've quoted parts where a single deep, blind pocket accounted for 70% of the total machining cost. The designer could have split the part into two pieces bolted together, eliminating the deep pocket entirely, and reduced the cost by more than half. Old Li calls deep blind pockets "money pits." He once drew a little cartoon of a dollar sign spiraling into a dark hole and taped it to the programmer's monitor.

Thin Walls and Fragile Sections

Thin walls — typically less than about 0.5mm for metals, or less than 1mm for plastics — are a triple threat. They distort under clamping pressure, they vibrate during cutting (causing chatter and poor finish), and they can warp after machining as residual stresses relax. To machine a thin wall accurately, the machinist must take very light cuts, use special fixturing to support the wall, and often resort to multiple stress-relief steps.

Thin walls also raise the scrap rate. A part that's 99% complete can be ruined by a single vibration spike that crumples the thin section. The risk of losing the part late in the process means the shop must build a higher scrap allowance into the quote.

A better design either thickens the wall to a more robust dimension, adds ribs or gussets for support during machining (which can sometimes be machined off later), or uses a material and process better suited to thin sections, like sheet metal fabrication or injection molding. Old Li's thin-wall test: he taps the section with a fingernail. If it rings like a tuning fork, he says, "That's not a wall. That's a drum head. It's going to sing, and the song is going to cost you."

Undercuts and Back-Side Features

An undercut — a recess that can't be reached by a tool coming from the primary spindle axis — requires either a special tool (a T-slot cutter, a lollipop cutter, a dovetail cutter) or a secondary setup that flips the part to access the feature from another direction. Special tools cost money, have limited reach, and often require custom grinding if a standard tool doesn't match the geometry. A secondary setup adds time, requires another fixture, and introduces a potential source of misalignment.

Internal undercuts — inside a bore, behind a shoulder — are the worst. The tool must reach through the bore, expand to cut the undercut, and then retract without hitting the part. The geometry of the tool shank, the bore, and the undercut profile must all be compatible. We've seen undercuts that were mathematically impossible to machine with any known standard tool, and the designer had to redesign the part.

Old Li's undercut test: he draws a stick figure of a tool trying to reach the feature. If the tool has to bend around a corner or magically expand after passing through a small opening, he puts a big red X through the feature. "Tools are not made of rubber," he says. "They don't turn corners."

Multiple Tight Tolerances on a Single Part

A part with one tight-tolerance feature — a bearing bore, a dowel pin hole — is manageable. A part with tight tolerances on multiple features in different locations and orientations is exponentially more expensive. Each tight tolerance requires careful setup, often on a different machine or in a different fixturing, and each must be inspected independently. The interactions between tight tolerances compound the difficulty: a bore that must be ±0.005mm and also positioned within 0.01mm of a datum that's on the other side of the part demands a level of process control that can double or triple the total machining time.

The cost curve for tolerances is steep and nonlinear. A ±0.1mm tolerance is usually achievable with standard practices. ±0.05mm requires more attention but is still reasonable. ±0.01mm demands a climate-controlled environment, multiple inspection loops, and possibly grinding or honing. ±0.005mm or tighter enters the realm of specialized equipment and extremely high scrap rates. Every zero after the decimal point adds a zero to the cost.

We ask designers to identify the true functional tolerances — the ones that affect fit, function, or safety — and relax everything else to a standard commercial tolerance. Old Li keeps a chart on his wall that plots tolerance against cost, with a curve that looks like a hockey stick. His caption: "The steep part of the curve is where your budget goes to die."

Cosmetic Finishes on Non-Cosmetic Surfaces

A surface finish callout of Ra 0.8 µm or better on a surface that no user will ever see is a common and costly mistake. Achieving a fine finish requires additional finishing passes, possibly grinding or polishing, and careful handling to avoid scratches. If the surface is inside a deep pocket or on a complex contour, the difficulty multiplies.

We've had parts where every surface, including the inside of a cable routing channel, was called out to a fine finish because the designer applied a global surface finish note. The channel functioned perfectly with a coarse, as-machined finish, but the print forced us into a slow, expensive polishing operation. We asked, the designer agreed to relax the callout on non-cosmetic surfaces, and the cost dropped.

Old Li's surface finish rule: "If the surface touches a human eye, make it pretty. If it touches another part, make it functional. If it touches nothing but air, leave it alone."

Threads in Difficult Locations and Non-Standard Threads

A standard metric coarse thread in a through-hole is essentially free — it's a quick drill-and-tap cycle. A blind thread near the bottom of a deep pocket, a thread in a hard material, a left-hand thread, an obscure thread standard, or a thread that must be milled because the material won't tolerate tapping — these all add cost. Thread milling is slower than tapping and requires a more expensive tool. Obscure threads may require a custom tool or a specialized thread grinding operation.

We encourage designers to use standard thread sizes and pitches, to avoid blind threads whenever possible, and to leave adequate clearance for the tap or thread mill. If a blind thread is necessary, the drawing should specify the minimum full thread depth and allow generous extra hole depth for chip clearance and tap lead.

Old Li's thread test: he looks at the thread callout and asks, "Can I buy this tap from a catalog? Can I reach the hole with a standard tap wrench? If both answers are no, you're paying for a custom tool and a special setup."

Unnecessary Material Removal

A part that starts as a solid block and ends as a thin, skeletal frame wastes material and time. A high buy-to-fly ratio — the weight of the starting billet divided by the weight of the finished part — means you're paying for a lot of metal that becomes chips. Features like large, flat-bottomed pockets that remove 90% of the material from a solid block are extremely costly. Where possible, design the part to be closer to net shape: use extrusions, castings, or weldments as the starting point, or break the part into simpler components that can be machined more efficiently.

We've seen parts where simply flipping the part and machining from both sides, rather than hogging a deep pocket from one side, halved the material removal and the cycle time. Old Li's material efficiency rule: "Every chip you make is a gram of material you paid for and threw away. The best chip is the one you never cut."

The Horror Shelf Gets a Feature Gallery

We've created a dedicated section on our horror shelf called "Feature Cost Drivers." It includes:

• A small aluminum block with a sharp-cornered pocket, next to an identical block with a filleted pocket, with their respective cycle times and costs written below.

• A deep, blind cavity that trapped chips and snapped a long-reach endmill (the broken tool is still wedged in the corner).

• A thin-walled titanium ring that warped into a potato chip shape during machining.

• An impossible undercut that required a $600 custom form tool.

• A part with a mirror finish on the inside of a non-visible wire channel.

• A threaded hole so deep and narrow that three taps were sacrificed to it.

Old Li curated this shelf with museum-style lighting. A sign at the top reads: "Every feature tells a story. Make sure it's the story you want to pay for." He gives tours to visiting engineers. The tours are effective; we've had several designers go back to their offices and revise parts immediately after seeing the shelf.

A Checklist for Cost-Conscious Design

If you're designing a machined part and want to keep the cost down, here's the checklist we wish every engineer would use:

1. Fillet all internal corners. Use the largest radius the design allows. At minimum, match the radius to the tool size expected for the pocket depth.

2. Avoid deep, narrow pockets. If a pocket depth exceeds 4x its width, look for ways to split the part, open the bottom, or change the geometry.

3. Keep walls thick enough to be rigid. Avoid walls under 0.5mm for metals, 1mm for plastics. Add ribs if thin sections are unavoidable.

4. Limit undercuts and back-side features. If an undercut is essential, design it to be accessible with a standard T-slot or dovetail cutter, and provide clearance for the tool shank.

5. Reserve tight tolerances for functional features. Use a general tolerance block for everything else. Every tight tolerance should have a documented functional reason.

6. Specify surface finishes only where needed. A global finish note is a global cost increase.

7. Use standard threads. Metric coarse, through-hole if possible, with clearance for the tap.

8. Design for the starting stock shape. Minimize material removal. Consider near-net-shape starting forms.

9. Talk to your machine shop early. A 15-minute call during the design phase can save thousands of dollars and weeks of lead time.

Old Li summarized it best, as he often does: "The cheapest feature is the one you don't put on the print. The second cheapest is the one you design with the machine in mind. Every other feature has a price tag. Read the price tag before you sign the drawing."

Have you ever redesigned a part and seen the cost drop dramatically? What feature surprised you with its cost impact? I'd love to hear your design-for-cost war stories — share them in the comments, and let's build a library of features worth their price and features that aren't.