Drawing tolerances and machining cost: what you can relax

Why the price rises even when the drawing dimension is the same

The same part often gets different prices because of a few extra lines on the drawing. The nominal size may be identical, but then tight runout, concentricity, surface finish or a requirement to inspect every part appear. The part's function might not change, but the workload jumps.

Manufacturing cost depends not only on the dimension to be achieved. What matters is how the shop will hold that dimension reliably from the first part to the last. The narrower the tolerance band, the slower the machining, the more frequent the tool changes and the more time spent on inspection.

Sometimes the difference seems small. For example, a shaft with a diameter tolerance can be produced with a normal finishing pass. But if the same surface also requires Ra 0.8, minimal radial runout and a strict datum reference, a single turning operation may no longer suffice. The shop then changes cutting parameters, adds a pass, uses a different insert or moves some work to grinding.

Several costs increase at once: tooling and fixtures, machine setup, the machining time itself and inspection. The last item is often underestimated. When tolerances are wide, an operator checks size quickly with a micrometer or gauge. When a tolerance is tight, the part is allowed to cool, measured in several points, and sometimes taken to a more accurate machine. That is labor time too, and it adds to the price.

So the same nominal size on paper does not mean the same cost. If a designer leaves unnecessary precision where the part gains nothing in function, production pays for stability, measurement and slower work. On the drawing the dimension looks the same. In the shop it becomes a different task.

Which dimensions actually control the part's function

When reviewing a drawing, it's useful to separate dimensions that the part truly needs for function from those that merely describe its shape. The most important ones are usually related to fits, assembly and mutual position of surfaces: diameters for shafts or bearings, center‑to‑center distances, mounting datums and the depth of seating surfaces.

If a dimension doesn't affect the joint with another part, it can often be loosened. Overall body length, the width of a free shelf, chamfer radius or the size of an unmachined area are often needed only for contour. They should stay reasonable, but they don't always need high accuracy.

It's helpful to mentally split the drawing into three groups: sizes involved in fits or assembly, external contour dimensions and requirements for form and location. The last group is better checked separately. A linear size alone doesn't tell how the part will behave in the assembly. Sometimes the diameter is easy to hold, yet the assembly doesn't fit because of runout, concentricity or datum flatness. Other times a tight linear tolerance is set when the real risk is a form tolerance.

For each requirement ask one clear question: what exactly will fail if we allow more variation? If the answer is concrete, the tolerance is likely justified. For example, play will increase, concentricity will be lost, the part will tilt, sealing will worsen or a subsequent operation will lose reference. If the answer is just "it's safer" or "just in case", the requirement should be reconsidered.

Surface finish deserves separate attention. For seating surfaces, seal faces and datums it is often as important as the size. On surfaces that neither rub nor mate, an overly smooth finish only raises the price.

A good drawing keeps tight values only where the part actually works. Then production spends time on requirements that affect function, not on precision for its own sake.

What raises the cost the most

Cost doesn't rise because of a single number in a tolerance box. It's usually a combination of demands: how smooth the surface must be, how the part must be fixtured, how many dimensions need to be held simultaneously and how all of that is to be inspected.

Overpayment frequently starts where precision is specified "just in case." The machine may be able to make the size, but the shop will spend more time on setup, remove stock more slowly, add finishing passes and perform extra inspection.

Money most often leaks away in four areas. First, tight linear tolerances on length, width or height where there is no mating surface. If the size doesn't affect assembly, clearance or the position of the working zone, there's no need to treat it as a fit. Second, low surface roughness on hidden or nonfunctional surfaces. Achieving a smooth finish means lower feed, more passes and sometimes a different tool. Third, overly precise chamfers and radii. They look harmless but easily add time. If a chamfer is only to break a burr, a wider tolerance usually suffices. Finally, strict runout requirements where the part does not function in a rotating pair — such demands lead to more rigid fixturing and separate inspection.

On turned parts this becomes obvious quickly. If you specify Ra 0.8 on all outer and end faces of a bushing, while only the bearing diameter actually works, the price will rise with no benefit to the assembly. The same happens when a flange is required to have minimal runout even though it is only bolted and does not center the neighboring part.

The simplest way to lower cost is to separate working surfaces from the rest. Where there is no fit, friction, sealing or precise mutual position, a looser tolerance usually yields the same result in the product and significantly simplifies shop work.

What a designer can often relax without harm

The link between drawing requirements and machining cost is usually simple: the more unnecessary strictness on the drawing, the longer the part takes. Yet not every dimension affects function. If a dimension does not define a fit, concentricity, sealing or assembly position, it can often be released.

Overall dimensions are often tightened out of habit. For housings, covers or brackets the outer length, width and height rarely need tight tolerances unless another part bears on those surfaces. The shop doesn't need to chase hundredths where the part just needs to fit into a space.

The same applies to step lengths. If a step doesn't position a bearing, gear, bushing or another part on the axis, a tight requirement doesn't improve the product. It only adds an extra pass, inspection and risk of scrap.

Surface finish is also frequently over‑specified. A low Ra value is not required everywhere. If a surface doesn't slide, seal or serve as a precision datum, a normal finish pass is usually enough. A too‑smooth surface costs more and often gives no benefit.

Chamfers and fillets are often specified too precisely. If the edge only needs deburring and easier assembly, an exact size and angle are not always necessary. The same for fillets: if they don't relieve stress or affect a mating part, a reasonable range is enough instead of a single tight number.

Nonfunctional fastener holes are another common case. If a bolt merely passes through and centering is provided by another surface or sleeve, there's no point in holding the hole diameter and position too tightly. A bit of clearance usually does not harm assembly.

A simple filter works: ask whether this surface contacts another part in operation, whether this size sets assembly position, whether the hole centers a fastener, or whether the surface roughness is necessary for friction, sealing or a datum. If the answers are all "no," the requirement can usually be relaxed. The part's function doesn't change, while price and lead time typically improve.

How to review a drawing step by step

Looking at a drawing as a collection of numbers is not helpful. It's much better to read it as a map of the part's functions. Then it becomes clear that it's not the numbers themselves that cost more, but the demands that force the shop to set up longer, measure more and risk scrap.

Start with a printout or a PDF copy and highlight the working surfaces. Mark everything that actually contacts another part: fits, datum faces, stops, fastener holes and sliding surfaces. Mark everything not directly involved in the assembly separately.

Next, write the role of each dimension in one word. Four labels are usually enough: fit, datum, clearance, contour. Already at this stage you see which requirements hold the function and which migrated from an old template.

After this markup, go through the drawing in order:

1. Check sizes that only relate to the external contour. If the part is not centered by that surface and does not stop against it, a very tight tolerance is often unnecessary.
2. Look at dimensions with small deviations where there's a clear assembly gap. Often they can be relaxed without harming assembly.
3. Compare linear tolerances and geometry. Sometimes the same area has a precise size, concentricity and runout. The shop ends up controlling nearly the same thing twice.
4. Check surface finish. If the surface doesn't work in friction, sealing or precise fixturing, a very low value only adds extra passes.
5. Mark disputed items and discuss them with the technologist before launching production.

In practice, a short conversation with the technologist often removes more unnecessary precision than a long lone review. The technologist quickly sees where a tolerance conflicts with real fixturing, where it complicates clamping and where it doesn't affect function or machining stability.

A good outcome is simple: the drawing keeps tight tolerances only where the part actually works. The rest stops inflating cost before the batch starts.

A simple example with a bushing

Take a simple bushing: it has a bearing diameter, an overall length and two chamfers. On the drawing all three requirements may look equally important. In reality that is almost never so.

If the bushing mates with a shaft, do not touch the bearing diameter. This dimension defines the fit: whether the part will press in, have a light interference or the required clearance. Changing 20.00 +/-0.01 to 20.00 +/-0.03 may make assembly too tight or introduce play.

Overall length is often easier. Suppose the drawing shows 40.0 +/-0.02, but the bushing does not abut a neighboring part and does not locate the assembly axially. Such precision does not help the part work better. You can relax it to +/-0.10 or even +/-0.20 if calculations and assembly allow.

Chamfers are usually even less sensitive. If they are only to break sharp edges and ease assembly, you don't need to hold them tightly. Instead of an exact note like 1.0 +/-0.05 x 45°, a looser requirement usually suffices.

For this part the logic is simple: keep the bearing diameter strict, check overall length by function and often loosen it, and specify chamfers without unnecessary precision. In production this is immediately visible. The lathe operator spends less time approaching the tool and on extra finishing passes. Inspection becomes easier: you don't hunt hundredths where they don't matter. Defects on secondary dimensions also fall.

This is why cost depends not on the number of tight dimensions but on their meaning. If a bushing keeps tight only what actually affects the shaft and assembly, the part functions as intended and the price typically drops without unpleasant surprises in the shop.

Common drawing mistakes

Overpayment often starts not at the machine but earlier — on the drawing itself. The designer adds strict requirements "just in case," and the technologist then spends more time on machining, tooling and inspection.

One frequent mistake is applying the same tight tolerance to almost every dimension. This is habit: if a part is "critical," then let everything be precise. But function rarely depends on every dimension equally. Usually two or three sizes matter: fit, concentricity, distance to a datum. The rest can be made looser and the part will still work.

The same goes for surface finish. If every surface has a low Ra requirement, production faces extra operations. Sometimes a different insert is needed, sometimes a slower feed, and sometimes grinding instead of a usual finishing pass. For an external bearing surface this may be justified. For an internal closed face inside an assembly often not.

Another costly habit is carrying over requirements from an old drawing without checking. An old part may have been made for a different assembly, material or process. Sometimes tolerances live on the drawing for years and no one can explain them. If a size doesn't affect fit, clearance, runout or function, it should be reconsidered.

Duplicated dimensions cause trouble too. For example, a size is given from the datum and then repeated via a chain of elements. On paper this seems convenient, but on the shop floor questions arise: which dimension is primary, which to inspect first, what to do if they disagree. In the end the price rises not only because of machining but also because of increased risk.

Inspection is a separate cost item. A complex tolerance may be achievable but expensive to measure. If a special gauge, more time on QA or a separate inspection protocol is needed, the cost grows even when the machining itself hardly changes.

A good drawing is not one where everything is squeezed to the limit. A good drawing clearly shows what really matters and what can be left less strict.

A short checklist before releasing the drawing

Before sending a drawing to production it helps to run through a few questions. This takes minutes but can prevent paying extra for needless operations, inspection and rework.

• Check mating sizes. Tight tolerances belong where the part fits onto a shaft, into a housing or works with another component.
• Look at nonfunctional surfaces. They rarely need low roughness unless they seal, guide or contact another part.
• Evaluate chamfers and fillets. If they only remove burrs and sharpness, they usually don't need the same precision as the fit.
• Compare geometric tolerances with the part's function. Concentricity, runout, flatness and perpendicularity are necessary where they truly affect assembly and operation.
• Review inspection requirements separately. If the drawing forces measurement of things that don't affect function, the price will rise without benefit.

What to do next

The most useful step is to show the drawing to the technologist before placing the order. They will quickly see where a tolerance is needed and where the price rises only from habit. This check takes less time than reworking an order after an expensive quote.

Next, review cost not in general terms but by changing one requirement at a time. Remove one unnecessary demand and request the quote again. For example, keep the fit diameter strict but relax surface finish on an external nonfunctional zone or remove a form tolerance that doesn't affect assembly. This shows more clearly which item drives cost.

For serial parts the discussion should go beyond the numbers on the drawing. Cost depends heavily on which machine will do the part, how it will be measured and how long each cycle takes. If you discuss this early, it's easier to choose a calm cutting mode and avoid paying for precision the part doesn't need.

It's useful to set a simple rule for the team: before sending to procurement, have the technologist give a short review; change requirements one at a time and compare prices after each change; then incorporate successful solutions into templates and internal rules. One careful template reduces costly mistakes more than long debates over each new part.

If the issue is not only the drawing but also machine capability, discuss that in advance. At EAST CNC we usually look at the whole task: machine selection, commissioning and service directly affect what accuracy the shop can reliably maintain in a series without extra cost.