Match the gold thickness to the number of times the connector is mated. Under 10 cycles, gold flash at about 3 µin is enough and it is what most catalog parts already have. Between 10 and 50 cycles, specify 10 µin. Above 50 cycles, or in humid and sulphur-bearing air, specify 30 µin. Everything above 30 µin is a different conversation, and usually a different connector.

The reason is that gold is not there to look expensive. It is there so the contact surface does not grow an oxide layer between mating cycles. Tin does exactly that, which is why a tin header that worked in the factory fails two years later in a coastal cabinet.

What each thickness actually buys

Plating Thickness Typical mating cycles Where it fits
Gold flash ~3 µin (0.08 µm) 10–20 Soldered once, never unplugged
Gold 10 µin (0.25 µm) 50 Assembly and service access, indoor
Gold 30 µin (0.76 µm) 100+ Test fixtures, field service, humid sites
Matte tin 100–200 µin 20–50 Wave solder, dry indoor, cost-driven

Those cycle counts are not a guarantee, they are a planning number. What shortens them in practice is contamination, not wear: dust, flux residue, and hand oils on the pins. A 30 µin part wiped with a dirty glove is worse than a 10 µin part that was never touched.

The pore problem nobody specifies

Gold is porous at thin thicknesses. A 3 µin flash has pinholes through to the nickel underneath. In dry air that is harmless. In sulphur-bearing air, the base metal corrodes through the pores and the contact resistance climbs from milliohms to ohms without anything visible on the surface.

This is why the specification has to name the nickel underplate as well. We use 50–100 µin of nickel under 10 µin or more of gold. Nickel alone does not stop the corrosion, but it slows it enough that the part survives its service life. Gold over bare copper, no nickel, is a cost reduction that shows up as a field return in year two.

Where the cost actually goes

Gold is priced by weight, and the weight scales with surface area times thickness. Going from flash to 30 µin multiplies the gold on the part by roughly ten. On a 2×20 header that is a few cents; on a 100-position board-to-board pair it is real money.

  • If the part is unplugged fewer than ten times in its life, flash is not a compromise, it is correct.
  • If a technician unplugs it during every service visit, 10 µin is the minimum that survives a warranty period.
  • If the same connector is mated daily in a test jig, use 30 µin and expect to replace the jig harness, not the board header.

How to write it on a drawing

Say the contact area, not just the part. "Gold 10 µin min over 50 µin nickel min on contact area, tin on solder tails" gives the shop something it can inspect. "Gold plated" does not, and it is how you end up with flash on a part you specified as 10 µin.

Is gold flash the same as gold plated?

No. Flash is a thin cosmetic layer, usually under 5 µin, often with no nickel under it. Gold plated usually means 10 µin or more over nickel. The two behave very differently after fifty mating cycles.

Can I specify selective gold plating?

Yes, and it is the normal way to control cost. Gold on the contact area only, tin on the solder tails, so the part solders easily and still survives the mating cycles. It needs a masking step in plating, which adds a small setup charge.

Does thicker gold reduce contact resistance?

Only slightly. The resistance is dominated by the contact area and the normal force, not the gold thickness. Thicker gold buys wear life and corrosion resistance, not lower milliohms.

Tell us the mating cycles and the environment and we will quote the thickness that fits, rather than the one that sounds safest. Samples at any of the three thicknesses ship in 2 working days from our pin header range.