> As it happens, by the way, computers happen to pretty well avoid the bad
> frequencies, since the harmonics of 33Mhz (66, 99, 132, 165) all miss the
> navigation band as do the harmonics of 25Mhz (50, 75, 100, 125).  While
> computers do also generate lower frequencies, the likelihood of computers
> spending much time in identical loops is pretty remote.

The actual clock frequency plays a fairly small role in RF emissions. The
killer is the switching edges (di/dt), which is where the energy transfer
takes place, and loop aeas.

If you want a graphic demonstration of this, find a board with an external
data bus and something relatively slow (say an 8051 @ 10MHz). Now replace
the HC logic with the equivalent AC family parts and watch the emissions
hit the roof.

Rule of thumb - Use the slowest speed parts you can.

A good example is the difference between an inbuilt crystal oscillator and
using an external can type. With something like a PIC, you have a nice sine
wave on the outside with no harmonic content and low energy. It doesn't
become digital until it's inside the chip where the loop areas are small.
On the other hand. If you buy an oscillator unit, it will be one of a range
that may go as high as say, 50MHz. The output is therefore designed to give
suitable switching edges for a 50MHz clock, even though yours is say, 4MHz.
Now you have all the harmonics of a high frequency impulse superimposed on
the harmonics of the 4MHz square wave. Given all those harmonics floating
around the PCB (and Mr. Fourier says there are quite a few), what are the
chances of one of them finding a track that's just the right length to
resonate ?

I'm quite impressed that so many PC's are able to pass FCC limits. When you
have things like PCI that rely on reflections over a few inches in order to
work and VRAM being clocked at God knows what. Add to that cables all over
the place and big holes in the front, it's suprising the cordless phone
ever had a chance.

Steve.

======================================================
  Very funny Scotty.  Now beam down my clothes.
======================================================
Steve Baldwin                Electronic Product Design
TLA Microsystems Ltd         Microcontroller Specialists
PO Box 15-680                email: steveb@kcbbs.gen.nz
New Lynn, Auckland           ph  +64 9 820-2221
New Zealand                  fax +64 9 820-1929
======================================================