Kovar Machining: Why It Cuts Nothing Like Stainless

By CHX Metal Engineering Team · Published 2026-11-16

Kovar is an iron-nickel-cobalt alloy, roughly 54% iron and 29% nickel, with thermal expansion near 5 ppm/°C so it can be sealed into borosilicate glass. It cuts gummy and smears, so it wants sharp carbide, flood coolant, a stress-relief anneal between roughing and finishing, and inspection at 20 °C.

A drawing line that is not a specification

A drawing that says “Kovar, per spec” names an alloy family and nothing else. It does not say which condition the bar arrives in, what stress state the part is in when it leaves, or which dimension is the one that actually seals. Last winter a batch of sealed collars came back from the plater 8 µm over the top limit on the seal bore, and the argument that followed was not really about the plater. It was about a bore dimension that had been signed off before the deposit existed.

The argument that did not happen would have been worse. Had those collars been turned in one pass with no intermediate anneal, they would have measured fine on the bench, moved during the glass-sealing cycle, and cracked at the customer's furnace rather than at ours. Kovar is not hard to cut. It is hard to hand over in the state the next process needs.

The alloy, and the one property it exists for

Kovar is roughly 54% iron, 29% nickel and 17% cobalt, the Chinese designation 4J29, controlled-expansion alloy in ASTM F15 language. Its whole commercial reason is thermal expansion: published figures sit near 5 ppm/°C through the range where borosilicate sealing glasses soften, against roughly 17 ppm/°C for 304 stainless. Seal a stainless pin into hard glass and the glass goes into tension on cool-down, because the metal shrinks three times faster than the glass wants to. Match the expansion and the joint survives.

BehaviourKovar (4J29)304 stainless
Expansion, 20–400 °Cnear 5 ppm/°Cnear 17 ppm/°C
At the cutting edgeGummy, tends to weld and smearWork-hardens, but breaks a chip
Left bareRusts in humid air; needs plating or a sealed environmentPassivates and stays put
Joining routeGlass or ceramic seal, brazingWelding, mechanical fixings
Bar priceA multiple of stainlessCommodity

That last row shapes quoting more than buyers expect. Kovar bar costs several times what 316 does and arrives in small mill lots, so a shop that treats it as just another nickel alloy loses the job on material yield rather than on cycle time.

Where the tooling loses

Nickel-rich alloys pressure-weld to a cutting edge rather than shearing cleanly. The symptom is a built-up edge that grows, breaks off, and takes surface finish with it, so a part comes off looking torn in patches and the operator blames the insert grade. Thermal conductivity is the second problem: Kovar conducts heat roughly an order of magnitude worse than aluminium, so the heat stays in the cut instead of leaving with the chip, and an edge that is already smearing gets hotter still.

What works, in our experience, is unglamorous: a genuinely sharp edge rather than a honed one, positive rake, flood coolant aimed at the flank, moderate depth of cut and no dwelling in the corner. A dull tool rubs instead of cutting, and rubbing is what produces the smear. On small bores and thin walls the same physics shows up as push-off, so we take those features in lighter radial steps and accept a longer cycle rather than fight deflection with feed.

Heat is the part that bites

The anneal decides whether the seal survives

Turning leaves residual stress in the surface, and Kovar's reason for existing is that it will be taken to a temperature where that stress lets go. A part roughed and finished in one setup measures round on the bench and comes out of the sealing cycle oval, or with a flat face that no longer sits. The fix is structural, not clever: rough, stress-relieve, then finish, so the material moves before the last pass instead of after it.

Mill datasheets for controlled-expansion alloys generally call for an anneal in dry hydrogen or vacuum in the 800 °C band, and the atmosphere matters because it is also doing cleanliness work. Where a drawing calls for a hydrogen anneal we write it into the route and coordinate it rather than improvise a furnace cycle - a badly run anneal on Kovar is harder to detect than one on steel.

Twenty degrees is part of the tolerance

At 5 ppm/°C, a 40 mm bore grows about 0.2 µm for every degree it sits above 20 °C. For a ±0.005 mm diameter that is noise. For a seal fit toleranced at a couple of microns it is a real part of the budget, and it explains the emails where two shops measure the same parts and disagree. Our inspection runs on a Leyden CMM and the report says the temperature it was taken at, because a number without that line is not a measurement.

The same expansion logic is why we ask which dimension is functional before quoting. On a Kovar housing the seal bore and the mounting face carry the function and the rest is clearance; tightening everything to ±0.005 mm buys nothing and adds cost. There is a longer version of that argument in tolerance stack-up on turned parts.

Plating moves the fit, so plan it in

Electroless nickel lands on every surface it can reach, which means a bore closes by twice the deposit thickness. An 8 µm coating takes 16 µm off a bore diameter, and the h6 fit drawn for a press pin becomes a rattle. Gold over nickel does the same with less thickness. The answer is not tighter turning; it is pre-plate dimensions on the drawing and masking on the fits that must stay bare, settled at RFQ rather than at goods-in.

What we take on, and what we hand to a specialist

Kovar runs on the same floor as 316L and TC4 in Shajing, Bao'an, Shenzhen: 220 machines, of which 100 are machining centres with 30 of them 3+2, 50 CNC lathes and 70 Tsugami Swiss-type lathes, at ±0.005 mm where the fit asks for it. A minimum order of one piece makes the trial cheap, prototypes come back in 3–7 days and production in 7–15 once the route is agreed. The floor turns out up to 5 million pieces a month with 100+ staff across 5,000 m², thirteen years into this. Terms are EXW Shenzhen only - we are happy to recommend reliable freight forwarders.

The boundary is worth saying plainly. We machine and inspect the alloy; the glass-to-metal seal, the plating chemistry qualification and the sealing furnace belong to the package houses that do them for a living. A shop that tells you it owns the whole chain on a sealed Kovar package is usually describing a supplier list. For the wider material picture, the CNC materials guide covers the alloys we run day to day.

If the drawing already says Kovar, the useful next lines are how it is heat treated, which feature actually seals, and whether the deposit goes inside the bore. Everything else follows from those.

Machining capacity behind this article

Address: Building B, 3rd Floor, Haosan Maozhoushan Industrial Zone, Buyong Community, Shajing Street, Bao'an District, Shenzhen  |  Tel: +86-755-27427365  |  Email: mc@chxmetal.com

Frequently asked questions

Why is Kovar used instead of stainless for glass seals?

Because its thermal expansion is close to that of borosilicate sealing glass, near 5 ppm/°C against roughly 17 ppm/°C for 304. A stainless pin sealed into hard glass puts the glass in tension as it cools and the joint cracks. Matching expansion keeps the seal under controlled stress.

Can Kovar be machined on a standard CNC lathe?

Yes, on ordinary CNC turning and milling equipment with sharp carbide, positive rake and flood coolant. The difficulty is not the machine but the route: rough, stress-relieve, finish, then plate. Skip the anneal and the part moves during the sealing cycle instead of on the bench.

Does Kovar need heat treatment after machining?

Usually yes. Turning leaves residual stress that releases at sealing temperature, so a stress-relief anneal between roughing and finishing is standard. Controlled-expansion alloy datasheets typically specify an anneal in dry hydrogen or vacuum around the 800 °C band; the atmosphere is part of the specification, not an optional extra.

How much does plating change a Kovar bore dimension?

A coating closes a bore by twice its thickness. Electroless nickel at 8 µm removes 16 µm from the diameter, which destroys an h6 press fit. Pre-plate dimensions and masking have to be settled at RFQ, not found at incoming inspection.

What tolerance can you hold on Kovar parts?

±0.005 mm on the features that carry the function, with unmarked dimensions defaulting to ISO 2768-m. Parts are measured on a Leyden CMM at 20 °C and the report states that temperature, since 5 ppm/°C on a 40 mm bore is about 0.2 µm per degree.

Send your drawings for a DFM review and a quotation

STEP, IGES, DXF, DWG or PDF — note the material, quantity and finish. Quotes usually go back within one working day. Email mc@chxmetal.com or use the quote form.