I don't think there ever was a "corporate standard" about whether air pressure within an instrument should be positive or negative. The issue was largely left to R&D teams, as a function of the specific requirements of every product (power dissipation, noise level requirements, mechanical constraints, cost, MTBF, maintainability, to name a few).
There's an example of the cooling system used in the 54600 oscilloscope series in the February 1992 issue of HP Journal, pages 37 and 38. Simple concept, and interesting reading.
Cheers,
Joel
