Surface roughness on a drawing: where Ra is needed and where it isn’t

Why a part gets more expensive because of an unnecessary Ra

A single number on a drawing can change the whole machining route. If a designer specifies Ra 0.8 instead of Ra 3.2 without a clear reason, the lathe operator no longer does the normal rough pass. They leave a larger allowance, reduce feed, change cutting mode and often add a finishing operation. Sometimes after turning the part is still sent to grinding or finishing, although the assembly doesn’t need it at all.

The cost rises not because of the Ra symbol itself but because of the actions it triggers. In a standard pass the shop makes the surface in one cycle and quickly moves the part on. When a low roughness is required, the operator spends more time setting up, makes a trial cut, checks sizes more frequently and risks going out of tolerance after removing extra material.

In practice one ordinary pass becomes two or three. Machine time increases, tools wear faster, quality control inspects more points, and the batch timetable shifts.

If the drawing is unclear, the shop almost always hedges. When there’s no note explaining why such an Ra is needed, the process engineer chooses a more cautious route. Their logic is clear: it’s better to spend an extra 15–20 minutes on a part than to receive returns for an entire batch. For the customer this caution turns into an additional charge for finishing that doesn’t affect fit, sealing or wear.

This is especially noticeable on series production. On a single part an extra finishing pass looks minor. On a batch of 200 housings it already adds hours of machine time, delays the next order and increases scrap risk. The longer a part stays in process, the more points where a mistake can happen: removing too much allowance, leaving a risky amount, or failing to meet size after a repeat pass.

Therefore specify surface roughness on the drawing by surface function, not by habit. If an area doesn’t operate in a friction pair, doesn’t hold a seal and isn’t part of a precise fit, an unnecessarily low Ra most often only makes the part more expensive and slower to produce.

What Ra means on a drawing in simple terms

Ra indicates how smooth the surface should be after machining. Simply put, it’s the average height of small irregularities left by a tool, milling cutter, grinding or other operation. The smaller the number, the smoother the surface and typically the more work the shop must do.

When reading roughness on a drawing, people look not only at the value but at where it’s placed. Sometimes it’s a general requirement for the whole part, applying to all surfaces without a separate note. Sometimes the symbol is placed on a specific surface. A local mark overrides a general one because it sets the condition for that particular area.

This isn’t a formality. If you set a strict general requirement, production will have to refine almost the entire part, including outer zones that don’t matter. If you only mark bearing fits, mating faces or sealing areas, unnecessary finishing drops significantly.

There’s no point in automatically choosing the smallest number “just in case.” Usually a level that satisfies the part’s function is enough. Surfaces for fits, sliding and sealing more often require lower Ra. Ordinary external surfaces do not. In practice designers and process engineers aim for a range that can be stably achieved by the chosen operation, not the minimal value at any cost.

For example, it makes no sense to require the same Ra for a housing cover as for a shaft journal under a seal. The external surface may look slightly rougher but still perform the same. Meanwhile machine time and price won’t increase without reason.

Another common mistake: using Ra instead of size and form tolerances. A surface can be smooth but a hole still end up oversized. A face may have good Ra and still be warped. Therefore roughness is specified together with tolerances for size, flatness, concentricity or runout.

Where roughness actually affects part performance

You should tighten Ra not everywhere, but only where the surface directly participates in the assembly. Otherwise the shop will add finishing, and the part won’t become functionally better.

Tightening roughness makes sense for surfaces with close contact, sliding or sealing. Typically these are bearing and bushing fits, shaft journals under seals, guides, sliding surfaces and mating faces that need to be leak-tight.

On those areas micro-irregularities affect friction and wear. A too-rough surface wears a mating part faster, makes noise, heats up and can damage a seal. But too-smooth isn’t always better either: lubrication needs to stay on the surface, otherwise the oil film works worse.

A good example is a shaft journal under a lip seal. If the surface is rough, the seal edge wears faster and begins to leak oil. If the surface is finished too carefully without real need, the assembly life may not change while part cost rises.

For coatings and assembly joints the logic differs. For painting you don’t need mirror smoothness but a clean, even surface free of scale, oil and torn tool marks. For welding edge cleanliness and proper joint prep matter more than low Ra. For adhesives, an overly smooth surface can also be counterproductive: many adhesives bond better to a moderately rough surface.

There are zones where geometry matters more than minimizing Ra. For a mounting face of a housing, an assembly flange or a datum surface, flatness, concentricity, perpendicularity and size are often more important. If the geometry is off, a very smooth surface won’t fix the problem.

The rule is simple: if a surface seals, guides, rotates in a pair or holds a fit, roughness directly affects performance. If it’s just an external wall, a paint area or a secondary base, check form and size first, then decide if a low Ra is needed.

Where a low Ra only adds work

Low Ra is often specified “just in case.” For the shop this isn’t trivial: the operator reduces feed, adds a finishing pass, sometimes changes the tool or introduces an extra setup. The part’s function doesn’t change, but the price goes up.

Most often unnecessary requirements target surfaces that don’t participate in the part’s function: hidden cavities, internal pocket walls, rough datums, external faces under a cover. If a surface doesn’t rub against another part, doesn’t seal and doesn’t define a precise position, a strict roughness there is usually unnecessary.

A separate case is surfaces that will be cut away by the next operation anyway. For example, a face after rough turning will later be finish-turned. Or after milling a pad will be bored together with the mating hole. Setting a strict Ra at an intermediate stage is pointless: the metal will be removed in the next step.

Usually an unnecessarily low Ra gives no benefit on first-operation rough datums, at the bottom of pockets that aren’t load-bearing, on external housing walls without contact, in closed internal cavities without flow or wear, and on areas that will be reworked later.

A single requirement for the entire part is almost always a blunt solution. One housing part may have a bearing fit, a gasket face, mounting ears, ribs and ordinary external walls. For the fit and gasket face a low Ra may be needed. For a rib or sidewall — not. If you put one general requirement on the drawing, the shop will pull all surfaces to the same level even though only two or three are actually important.

Therefore assign roughness by function for each zone. Where there’s sliding, sealing, a precise fit or significant load, the requirement should be explicit. Where the surface just remains after machining, leave a more relaxed value or don’t lock it down without reason.

How to assign roughness step by step

If roughness is specified by habit rather than function, the part’s cost quickly rises. An unnecessary low Ra drags in finishing passes, different tools, more machine time and extra inspection.

It’s easier to start from the surface’s function than from the desire to “make it better.”

1. Break the part down by surfaces and ask a simple question for each: what does it do in the assembly? If it centers, seals, slides, controls a dimension or serves as an assembly datum, consider it separately.
2. Then split surfaces into three groups: functional, datum and secondary. Functional surfaces affect part performance. Datums are needed for setup and repeatability. Secondary surfaces hardly affect anything except appearance and removal of machining marks.
3. Specify Ra only where it’s necessary for the function. A bearing bore, a shaft journal, a gasket face or an assembly datum often require a specific value. Side walls, external zones and hidden cavities usually don’t need a strict requirement.
4. Immediately check the requirement against the actual process. The same Ra behaves differently on steel, aluminum and stainless. The operation matters: one Ra may be easy after turning, while another will require a separate finishing pass or even grinding.
5. Mark any disputed areas directly on the drawing. A short note or a small sketch often saves more time than a long exchange between design and the shop.

A simple example: a housing has a bearing hole, a face for a cover and external walls. It makes sense to set a specific roughness for the hole and the face if fit and sealing depend on it. For external walls a general requirement without extra tightening is often sufficient.

In practice the process engineer will ask the designer two short questions: what will fail if Ra is rougher, and on which surfaces will that be noticeable? If there’s no clear answer, the requirement can often be relaxed without harming the part.

How the process engineer quickly checks feasibility

A process engineer doesn’t look at Ra separately from the process. They immediately link the requirement to the material, tool and cutting mode. The same Ra in steel, aluminum and stainless yields different costs and different scrap risks.

If a designer specifies Ra 0.8, the engineer asks directly: which operation will actually produce that surface? After rough turning the answer is usually obvious: a finishing pass is needed, sometimes a different tool, different speeds, lower feed and tighter control.

Next they check the route. Can the required surface be obtained in one setup or will the part need re-clamping and another alignment? Every extra setup increases time and can spoil accuracy and the surface itself.

Usually the engineer quickly verifies four things: the material on the drawing, the tool already planned in the route, whether the machine can deliver the result without an extra pass, and how much an additional finishing pass and inspection will add to the price.

This is visible fast. Suppose a housing drawing lists nearly the same Ra on almost all surfaces. The engineer reviews the route and understands: bearing fits can be produced cleanly in the current operation, while external faces don’t need that finish. If the general requirement remains, the shop will add unnecessary finishing and then inspect surfaces that don’t affect assembly.

After such a check the engineer doesn’t argue in general terms but proposes a concrete change. For example, keep low Ra only for fits and seals, and give a more relaxed value for the rest. Sometimes a local symbol rather than a new number is enough.

If the issue rests on equipment capability, it’s better to solve it before the batch starts. That’s why a good link between designer, process engineer and the person selecting the machine for the task is needed.

Shop example: a housing part with an unnecessary finishing pass

A simple steel flange-housing came in: a mounting face, a central bored hole, an outer diameter and four bolt holes. The part was straightforward, but the designer specified Ra 1.6 almost everywhere. The shop immediately saw the problem: that value drags unnecessary passes where they don’t affect function.

The engineer built the route as if every surface affected fit and assembly. After rough turning they added a finishing pass on the outer diameter, a separate finish cut on the face, touch-up of some planes and stricter inspection after drilling. Time grew, but the benefit was minimal.

When the engineer and designer reviewed the drawing together, it turned out the low Ra wasn’t needed everywhere. The mounting face stayed at Ra 1.6 because it defines the housing fit in assembly. The central hole also kept Ra 1.6 because a mating part runs in it. Bolt holes were left without extra finishing if there’s no seating zone under the bolt head. The outer diameter and side surfaces were relaxed to Ra 6.3 or Ra 12.5 where permissible.

After agreement the route shortened. The shop removed some finishing passes, stopped refining external surfaces “just in case” and reduced inspection volume. The part still assembled correctly but was made noticeably more simply.

The difference was tangible. Before the change the route took about 34 minutes per part, after — about 21 minutes. On batches of 80–100 pieces that’s a serious saving. In price terms such a correction often reduces cost by 10–18%, sometimes more if the original Ra forced an additional setup.

This example shows a simple point: Ra should be placed where the surface actually works. If you require the same low roughness across a housing, the shop will almost always add unnecessary finishing. The customer pays for that.

Mistakes that cause disputes between design and the shop

A dispute usually starts not because of the machine but because of the drawing notation. The designer wants a certain surface, while the shop sees extra passes, extra time and extra cost. If the requirement is too broad or vague, everyone interprets it their own way.

One of the most common mistakes is using one strict Ra for the entire part. It’s easier on the sheet, but inconvenient in production. A housing, shaft or flange almost always has working and non-working surfaces. If you require, say, Ra 0.8 everywhere, the shop must finish even areas where it makes no difference.

No less dispute arises when people confuse roughness with dimensional accuracy. A dimension can be held tightly while the surface remains not very smooth. And vice versa: a low Ra won’t save you if the size is out of tolerance.

Another costly mistake is failing to specify which surface the symbol applies to. The drawing shows Ra and the shop guesses: is it the face, the fit or the whole contour? Extra questions begin there, and sometimes extra machining, because it’s safer to make it “better” than to redo.

Problems also occur when a low Ra is required before an operation that will spoil the surface. A typical case: a surface is almost finished and then nearby welding, heat treatment, heavy assembly or clamping leaves marks. The shop reasonably asks why waste time earlier than necessary.

There’s also a practical question: how to measure it later. If the surface is in a narrow pocket, near a flange or inside a short bore, a standard instrument may not reach. On paper the requirement exists, but checking it quickly and reliably is difficult.

Usually these disputes can be resolved before production by answering four questions:

• Does this surface really work or was the requirement added “just in case”?
• Is exactly this Ra necessary or is a more relaxed value sufficient?
• Is it clear where to measure the surface?
• Can it be measured after all operations?

A good drawing doesn’t make the shop guess. When the designer marks working surfaces separately, ties Ra to the surface function and considers inspection, the conversation is shorter. The part also comes out cheaper.

A short check before release to the shop

Before handing the drawing to the shop it’s useful to run a quick check. It takes a few minutes but saves paying for unnecessary finishing and explaining why you set a too-low Ra where it doesn’t affect the part.

Most issues arise not from complex geometry but from redundant or vague requirements. If the designer sets roughness “just in case,” the foreman usually chooses a more cautious operation. The part takes longer and the price rises.

Before release check five things. Is Ra specified only on working surfaces? Will the operator understand the expected operation: normal turning, additional finishing or grinding? Does the requirement match the material and batch size? Can the inspector measure the surface after all steps? Are general phrases like “finish cleanly” or “make smooth” removed, since shop and QC interpret them differently?

It’s useful to ask the foreman one more question: which operation will give the required result without extra work? If they answer immediately and without guessing, the drawing is likely clear. If a debate starts about what was meant, fix the requirement before release.

Small example. A designer sets Ra 1.6 for the whole housing while only two seating surfaces matter. The shop adds finishing passes in several places, spends more time and inspects areas that don’t affect assembly. If the requirement is left only where needed, the part comes out cheaper without quality loss.

If at least one checklist item lacks a clear answer, reopen the drawing. It’s easier to correct notation on paper than to remove extra metal on the shop floor.

What to do next

The most useful step is to resolve disputed areas before the part goes to costing. If the designer sets a low Ra “just in case” and the process engineer silently accepts it, the price rises fast. The shop adds extra passes, holds the part in the machine longer and sometimes switches tools where it brings no functional benefit.

A simple workflow works well: the team marks surfaces that affect fit, sealing, friction or appearance, then reviews all others separately. This turns roughness from a blanket precaution into a clear requirement for specific zones.

Keep a short practical kit at hand: a list of typical Ra values for your parts, materials and operations; examples where turning already gives the needed surface without extra finishing; a drawing template without automatic strict Ra for every surface; and a simple rule to discuss disputed items before costing.

Such a kit saves time at the first review. The designer no longer relies on memory, the process engineer doesn’t debate every dimension, and procurement gets a more accurate cost.

If you often produce the same housings, bushings, flanges or shafts, compile an internal table. For each group note material, typical operation and usual Ra. In a month that table removes many repeated questions and significantly reduces drawing revisions.

Also review the template. If it defaults to a too-strict Ra, the mistake will repeat. Correct the template once — it’s easier than explaining each time why non-functional surfaces don’t need costly finishing.

If the issue is not a departmental disagreement but equipment capability, pre-check the task against the machine and route. EAST CNC — the official representative of Taizhou Eastern CNC Technology Co., Ltd. in Kazakhstan — supplies CNC lathes and other metalworking equipment, and supports projects from selection and commissioning to service. For those who regularly face these tasks, EAST CNC blog materials on east-cnc.kz can also be a useful reference: they publish equipment reviews, industry news and practical metalworking tips.