RF PCB partial mask reason ?

[fmkit]

after googling for no avail. I'm asking here what is thr reason for having some areas of PCB unmasked ? I've designed PCB for myself and measured no difference in masked  / no mask  boards but keep seeing partial mask in commercial design.

[TC]

Most likely to have better control over impedance. Solder mask as a dielectric constant between 3.5 and 4.0 (if I remember correctly). Effects will vary based on the solder mask thickness which isn't precisely controlled. Air has a dielectric constant of 1.0 (nominally).

[yl3akb]

At higher frequencies is may reduce the loss a little bit. See here Fig 3. for example: https://www.circuitinsight.com/pdf/effects_pcb_fabrication_high_frequency_electrical_performance_ipc.pdf

In practice its worth doing on some particular edge coupled microstrip structures such as directional couplers and band pass filters where material/medium between the edges is critical.

[RoV]

I think it's mostly done to limit losses. Besides, it allows some fine tuning (move/add capacitors, add a bit of silver paint somewhere), but this approach has almost disappeared, at least on production boards.

[tchiwam]

Then if you do ENIG it kind of negates the benefit as the ENIG is very rough on the surface. so you are often better off with the solder mask.

Hard gold or ENEPIG is better but more expensive. best being bare copper, for a short while.

Just this to say, YMMV

If you play under 2 or even 3GHz ... Meh don't worry too much.

[ConKbot]

Then if you do ENIG it kind of negates the benefit as the ENIG is very rough on the surface. so you are often better off with the solder mask.

Hard gold or ENEPIG is better but more expensive. best being bare copper, for a short while.

Just this to say, YMMV

If you play under 2 or even 3GHz ... Meh don't worry too much.

That, and the gold is thin enough that for some (surprisingly high, but I forget off the top of my head) frequencies, the RF current will end up in the nickel plating.

So from a pure loss perspective, not the best, but for consistency between board runs, no mask and using ENIG will at least be more consistent impedance/vswr wise than SMOBC, due to varying soldermask thickness or dK.

Can you get hard gold mid-board like that? I'm used to card-edge fingers, which require the bus bar that gets used to electroplate them, but the bus bar gets routed away. Not saying it's not possible, just not familiar with it.