The Bore That Wore Out Early

A buyer sent us a sintered iron bushing last year, returned from the field with a worn bore. The drawing said Ra 0.8 micrometers. Our mill certificate said Ra 1.6 micrometers as-sintered. Both were right. The drawing was written for a machined part, not a sintered one, and the spec was never reconciled to the process.

It was not the first time we had seen that. On a porous surface, the same physical part can read Ra 0.8 micrometers or Ra 1.6 micrometers depending on which stylus, which cutoff, and which pore happens to fall under the trace. The buyer thought they were getting a defective part. The shop thought the callout was unrealistic. Nobody was wrong about their own number; the callout was the problem.

The bushing was a hydraulic seal face. It should never have been specified as Ra 0.8 micrometers as-sintered, because as-sintered cannot hold that on a seal surface. It needed either sizing, plating, or a different callout (the bearing surface that mates with a steel shaft, not the seal face that mates with an elastomer). It wore out in service because the open pores on the seal face were micro-leaking pressure, and the surface finish was accelerating that leakage with every cycle.

That is the lesson: on sintered parts, surface finish is decided by what the surface has to do, not by what a generic datasheet says is possible. The rest of this article is the working out of that lesson, with numbers and callouts you can put straight on your drawing.

What Surface Finish You Get at Each Process Stage

The numbers below come off the same iron-base PM grade on our press shop floor, measured on the same part geometry at three process stages. They are not marketing figures and they are not textbook figures. They are the numbers our QC team records on inspection reports, and they are what you should expect from a controlled press-and-sinter process.

Two things to read off that table. First, the as-sintered Ra band is roughly five times the wrought-machined reference band. That gap is the whole reason PM surface finish is its own conversation. Second, sizing roughly halves Ra on die-formed features for almost no extra cost, because sizing is a re-pressing operation in a second die, not a machining step. Most drawing callouts under-use sizing, and the part pays for it in either an unnecessary grinding step or a worn-out seal face.

Why Sintered Surface Finish Is Different from Wrought Metal

Sintered metal is not a solid block of steel with a defined grain. It is a network of metal powder particles that have been pressed into a shape and heated until the particles diffusion-bond at their contact points. That process leaves a structure with two features a wrought surface does not have: residual porosity and a powder-particle texture on the surface.

Residual porosity is the dominant effect. Conventional press-and-sinter parts land at 85 to 92 percent of theoretical density, which means 8 to 15 percent of the volume is open or closed pores. Most of those pores are sub-surface, but enough of them intersect the surface to create the rough, slightly pock-marked texture you see on an as-sintered part. You can see it under a 10x loupe, you can feel it with a fingernail, and a profilometer will measure it as higher Ra than the same part in wrought steel.

The powder-particle texture is the secondary effect. The surface of a sintered part is essentially a frozen cross-section of the powder compact, with each particle leaving a peak or a facet roughly 50 to 150 micrometers across. Sizing compresses those peaks and roughly halves the surface roughness, but it does not change the underlying pore structure. To get a genuinely pore-free surface, you need a separate operation: plating, infiltration, or a higher-density process like MIM or hot isostatic pressing.

What Each Secondary Operation Does to Surface Finish

Most drawing callouts name only one finish. In practice there are five common secondary operations, and they do different things to the surface. Picking the right one starts with what the surface has to do.

Sizing (re-pressing in a second die)

Compresses the as-sintered peaks and trues the dimensions. Ra drops from 0.8 to 3.2 micrometers down to 0.4 to 1.6 micrometers. Cost is low because the part is already formed; sizing is just a second press cycle. Best for bores, ODs, gear tooth flanks, and any die-formed feature that needs to be tighter in dimension and finish at the same time.

Grinding and honing

Material removal with a wheel or stone. Ra drops to 0.1 to 0.4 micrometers, the same band as a machined wrought surface. Cost is high because every ground feature pays for a separate operation, and grinding on a sintered part generates wheel-loading from the open pores. Best for high-cycle fatigue surfaces (gear tooth roots, bearing journals) and seal faces that have to mate with elastomers.

Steam treatment (black oxide)

Heats the part in steam to grow a magnetite (Fe3O4) layer that fills near-surface pores. Ra drops slightly, from roughly 1.6 to 3.2 micrometers down to 0.8 to 1.6 micrometers. The main benefit is sealing the pores and adding corrosion resistance, not finishing. Best for parts that need to hold oil and resist rust, like automotive shock absorber components.

Electroless nickel plating (ENP)

Deposits a nickel-phosphorus layer, 5 to 25 micrometers thick, that bridges the surface pores and gives a genuinely pore-free finish. Ra on the as-plated surface is 0.2 to 0.8 micrometers. Best for parts that need a seal-quality finish, corrosion resistance, or a cosmetic appearance, and where you are willing to pay for plating.

Tumbling and vibratory finishing

Mechanical surface smoothing with abrasive media. Light effect on Ra (maybe 30 to 50 percent reduction), but it breaks edges and removes burrs. Best as a cosmetic or pre-plate step, not as a primary surface finish callout.

Measuring Surface Finish on Porous Sintered Metal: Why Ra Lies (ISO 23519)

Now the part of the conversation most engineers have not seen. The standard surface finish parameter, Ra, is defined in ISO 4287 and ASME Y14.36 for solid, homogeneous surfaces. It works by taking a stylus of a defined radius (usually 2 micrometers) and dragging it across the surface along a defined evaluation length (4 mm typical), then averaging the absolute deviation of the profile from the mean line. The result is a single number that summarises roughness.

On a porous sintered surface, Ra misleads. Here is why. The stylus has a physical tip radius. When it crosses an open surface pore, the tip falls into the pore and records a deep valley that is not really a roughness feature, it is a hole. Ra counts it as roughness, and the more pores the trace crosses, the higher Ra reads, even when the actual solid surface between pores is smooth.

ISO 23519, published in 2021 and applicable to all powder metallurgical parts with or without machining, prescribes a different parameter set for exactly this case. The two key parameters are:

• Rk (core roughness depth): the roughness of the core surface, after the deep valleys (pores) are filtered out. This is the number that reflects how the actual solid surface between pores looks.
• Rpk (reduced peak height): the height of the peaks above the core. This tells you how much the surface sticks up above the core profile, which matters for wear and contact mechanics.

On a porous surface, Rk and Rpk together describe the real surface behaviour far better than Ra does. Rk on a sized PM bore typically sits in the 0.2 to 0.8 micrometer range, even when Ra reads 0.8 to 1.6 micrometers on the same surface. The difference between the two numbers is the pore signature.

How to use this on a drawing. If you are buying a sintered part and the inspection report shows Ra, ask for Rk and Rpk as well. The Rk number will tell you whether the actual contact surface is what you need it to be; the Ra number alone may not. We now include Rk and Rpk on every PM inspection report by default, not on request, because the question of which parameter to read on a porous surface comes up often enough that it should not be a special ask.

Apply the Right Finish to the Right Surface

Most drawing errors on sintered parts come from putting the same finish callout on every surface. In practice, the finish a surface needs is a function of what that surface does in service. This is the question to ask before you write the callout.

How to Spec Surface Finish on a PM Drawing Without Overpaying

1. Leave the callout off by default. As-sintered is the default state and the cheapest state. Put a finish callout only on the surface that has a documented functional reason to need it.
2. Match the callout to the function, not to a datasheet. A bearing bore is sized (Ra 0.4 to 0.8 micrometers). A seal face is plated or ground (Ra 0.2 to 0.8 micrometers, with Rk and Rpk on the report). A structural wall carries no callout.
3. Use Rk and Rpk on porous contact surfaces. On a seal face or a bearing bore, the standard Ra callout misleads. ISO 23519 (2021) prescribes Rk (core roughness) and Rpk (reduced peak height); together they describe the real contact surface. We include Rk and Rpk on every PM inspection report by default.
4. State the cutoff and evaluation length in the callout. Ra is meaningless without a cutoff and evaluation length. Specify 0.8 mm cutoff and 4 mm evaluation length (or per ISO 23519, 0.8 mm cutoff on the contact surface). If you do not, the report value is ambiguous.
5. Avoid calling out Ra on every surface. That is the most expensive sintered-part drawing mistake. It forces grinding or plating on surfaces that did not need either, and it does not improve the surface that actually mattered.
6. Confirm finish routing in a DFM review before tooling. A 30-minute review of the finish callouts against the function of each surface catches more cost than any other step in the print sign-off. We do this as a free 48-hour DFM pass on every drawing.