Surface Grinding Machine Selection for Precision Parts in ASEAN

Buy a surface grinder to an accuracy standard, not to a spec sheet. For a horizontal-spindle machine with a reciprocating table that standard is ISO 1986-1:2014; for a vertical-spindle machine it is ISO 1985:2015, which runs its geometric tests through ISO 230-1 and ISO 230-7. Name flatness, parallelism and achievable Ra as numbers in the purchase order, and require a witnessed test-piece grind before the machine leaves China. This matters because those numbers are usually absent from a grinder datasheet altogether: capacity gets documented in detail, capability does not.

What buyers compare, and why it does not predict accuracy

Almost every field on a grinder datasheet describes capacity, not capability. Take the [Mking 307NC2](/en/products/mking-precision-machinery-hangzhou-mking-307nc2-cnc-precision-surface-grinding-machine), the most completely documented grinder we carry: 24 spec fields, 23 of them numeric — table 300 x 700 mm, wheel 355 x 40 x 127 mm, 7.5 HP spindle at 1,480/1,780 rpm, 200 kg maximum table load, 3,800 kg net. All useful for deciding whether your part fits. All silent on whether the part comes off flat.

The field most often mistaken for accuracy is handwheel graduation. A "0.001-0.01 mm vertical handfeed" describes the dial. It is the smallest increment you can command, not the flatness the machine holds across a 400 mm pass after two hours of running.

The asymmetry nobody flags

There is a strange asymmetry in how machine tools are documented. Machining and turning centres routinely cite a positioning-accuracy standard — VDI 3441 or ISO 230-2 — right on the datasheet. Surface grinders, the finishing machines those centres depend on to rescue a part, rarely cite anything at all. Buyers are handed more accuracy documentation for the roughing machine than for the machine that has to hold the final tolerance.

The one complete datasheet belongs to a used machine

The exception on our own site is second-hand. The [KENT KGS-4080WM1](/en/used) states flatness <= 0.003 mm / 400 mm, parallelism <= 0.004 mm across the full table, spindle face runout < 0.002 mm and surface finish Ra 0.16-0.4 um, alongside its 800 x 400 mm chuck, 0.001 mm servo downfeed, stress-relieved Meehanite casting and hand-scraped Turcite-B rails.

This is not an argument that used beats new. It is proof that the disclosure exists and can be demanded. Send that list to a new-machine supplier and ask them to fill in their own figures.

The four numbers, and the conditions attached

An accuracy figure without conditions is marketing. Ask for each number, then ask what it was measured under:

| Figure | The question that makes it real | |---|---| | Flatness | Over what length? On what test piece? | | Parallelism | Full table, or partial travel? | | Surface finish Ra | Which wheel, dressed how, at what feed? | | Spindle face runout | Cold, or after warm-up? Measured with what? |

Then add the two conditions every acceptance test needs and few quotations mention: ambient temperature during the test, and warm-up time before it.

Heat is a purchased tolerance in a tropical shop

ISO 1:2016 fixes exactly one standard reference temperature — 20 C — for size, location, orientation, form and surface texture. Every tolerance on your drawing and every accuracy figure in your supplier's quotation implicitly assumes it.

A grinding bay in Rayong, Bac Ninh or Johor is not at 20 C. Steel expands roughly 11-13 um per metre per C (CTE Magazine gives 7.3 x 10-6 per F for steel; grey cast iron runs lower, near 10-11 x 10-6 per C). Taking 11.5 um/m/C and a 32 C shop, a 400 mm steel part sits about 0.055 mm larger than its 20 C drawing length. That is arithmetic from stated assumptions, not a measurement — but it is roughly eighteen times the KENT's declared flatness, and your micrometer is at 32 C too.

Two effects are often confused. A uniform temperature offset moves size; temperature gradients — spindle heat, hydraulic oil, sun on a steel roof, one hot wall — bend form. Only the second attacks flatness directly. The practical response is cheap: soak parts before measuring, measure at the temperature you ground at, and air-condition a metrology corner rather than the whole building.

The accuracy chain is not in the quotation

The machine does not deliver your Ra alone. The chuck, wheel, dresser, balancer and coolant filtration do — and they are rarely specified on the grinder datasheet at all.

• **The chuck must be ground in place** on the machine after installation. That grind-in is commissioning work, not an accessory. Schedule it, and agree who performs it.
• **Wheels carry safety marks** — ANSI B7.1 in the United States, EN 12413 in Europe — covering maximum operating speed and handling. Confirm the mark before anything is mounted.
• **Unbalance appears as chatter** long before it becomes vibration you can feel. Our [ISO 1940-1 balancing grade calculator](/en/tools/balancing-grade) gives permissible residual unbalance for a given rotor mass and speed.

Buying used, and the pre-shipment test

Machinists on trade forums — practitioner reports, not controlled studies — describe one failure repeatedly: worn table ways make the table travel in a curve, the chuck is reground to compensate, and the chuck is then no longer flat. One reported 0.030 inches of way wear. Their checks are cheap: blue and a straightedge across the chuck, spin the spindle by hand for rough spots, run a precision level along the table travel.

The decisive test applies to new and used alike. **A supplier who will not grind a test piece to your drawing, in front of your inspector or on video, before payment, has told you something.** Static inspection misses what appears only under load. Several ASEAN countries also restrict used-machinery imports by age and require an assessment certificate, so confirm the current rule for your HS code before committing.

Conclusion

A surface grinder is the one machine bought purely for accuracy, and the one whose accuracy is least often written down. Fix that in the purchase order: four figures, their test conditions, ambient temperature, warm-up time, and a witnessed test-piece grind. See how we screen suppliers in our [factory audit checklist](/en/news/verify-chinese-equipment-supplier-audit) and [certification guide](/en/news/quality-certifications-chinese-equipment), read how we handle [China sourcing](/en/china-sourcing), or [talk to us](/en/contact) about writing the acceptance clause.