Jim you asked:

(stuff deleted)

>In the past, I have always considered this to be a 6.5 mil circuit.
>Is this correct thinking?
>
>A review of IPC-T-50 terms and definitions says that the conductor
>width is "The observable width of a conductor at any point chosen at
>random on the printed wiring board normally viewed from directly
>above unless otherwise specified."
>
>If this circuit is viewed from above with backlighting, the
>observable width is 6.5 mils.  When viewed from above with top
>lighting, the appearance is 5.5 mils.

Sounds like one of those circular arguments we so frequently waste time
with in this business. I thought that Bob Vanech had it right:

> I would
> consider this to be a 6.0 mil finished line in my calculations for end
> product requirements and let the pundits debate how to measure the trace.
>
>                                 Bob Vanech

Been there myself. Must be some QA types involved who sometimes seem to
revel in pointless debates. I'd probably measure with an eyepiece, looking
at reflowed tin-lead only, and call the trace width 0.006, but there is
some chance it might really make a difference functionally (dubious) if so,
the original drawing should have had notes to make this very clear.
Stripline boards are a different breed. The process tolerance in
conventional fab processes (etch factors, etc.) would normally be too wide
to make the discussion you describe reasonable.

All too often, you have design engineers or inspectors who somehow believe
that the board should be _exactly_ like their film, and they think in two
dimensions only.

This kind of stuff is what makes me have premature grey hair, I think.

regards,

Jerry Cupples