mark@mips.COM (Mark G. Johnson) (01/22/91)
I confess I was skeptical when reading J. Peter Moncrieff's recent
diatribe [in IAR Hotline] on the differences in sound of different
chemical compositions of Printed Circuit boards.
However, the analog instrument folks seem to concur. In the
January 03, 1991 issue of _EDN_ (a trade magazine aimed at electronics
design engineers) there is a relevant article on page 47:
Brian Kerridge, "VXIBUS Measurement Modules: Instruments prove
tough to design"
It contains these passages on page 48:
"Additionally, [benchtop DMM] designers mount sensitive components,
such as the ac input attenuator, well above the PC board to limit
the effects of PC board dielectric. The stray capacitance from
components to the PC board surface depende somewhat on the dielectric
constant of the board material. Because this constant varies with
frequency, it wreaks havoc with the ac frequency flatness of the
input circuitry."
Later in the article it talks about how instrument manufacturer Datron
chooses to deal with the problem:
"Predicting the dielectric characteristics [of PC board] material
from production batch to production batch and from supplier to
supplier is impossible. Instead, Datron selectively cuts out
sections from the PC board adjacent to sensitive components because
the dielectric constant of air is at least predictable and stable.
Datron's design yields an ac accuracy of 0.05% for frequencies
of 40Hz to 20kHz and an accuracy of 4% at 1MHz."
So it appears that the instrument folks are working at audio frequencies
and they feel PC board dielectrics are a problem at those frequencies.
Well I'll be dipped in, uh, warm mud.
--
-- Mark Johnson
MIPS Computer Systems, 930 E. Arques M/S 2-02, Sunnyvale, CA 94086
(408) 524-8308 mark@mips.com {or ...!decwrl!mips!mark}