Metzler's Laws of Signals
Many fine circuits have been abandoned or ignored because of
'components' that never appeared on
the schematic.
1. Any conductor
that carries alternating current is considered to be a
transmission line. Any energy that fails to appear at the
far end went
elsewhere. Signals escape by way of capacitance,
mutual inductance,
common resistances (ground loops)
or by radiating as RF. It's a bad
idea to just hope the missing stuff
turned into heat! This includes
power supplies, which must be
assumed to be carrying nasty stuff until
proven clean.
2. Reciprocity:
if stuff can leak out, stuff can also leak in!
3. If the
conductor is << 1/8 wavelength (at the highest excitation
frequency), time delays MAY be
unimportant. In digital work,
excitation
frequencies (edge rates) are way
higher than clock frequencies. In
analog work, distortion products
are way higher than signal frequency
excitations. Is the line still short?
4. If there's a
known resistance in range, try to match to it unless
there's a very good reason not
to. Even a simple series terminator
at
the source end can help. If you get lucky and condition 5 is met,
the
line can be ignored... maybe.
5. ALL lines
have return paths associated with them.
If you don't
control them, Murphy will. In which case return will likely be by
way
of another of your signal
lines. Return is by way of the
lowest
impedance, NOT the lowest
resistance path, even at 'audio' frequencies.
The smallest area loop will carry the signal current. DC powered
amplifiers of ALL kinds work by
shunting current between 2 or more
'power rails', which become the
actual return points. Have you tied
them together? Where and with what? Only a perfect transformer can
keep these current off of your
lines. This includes logic gates.
6. Capacitors
have inductance, lots of it. Resistance too.
Know how
much if you can. People who make capacitors don't like
inductance and
resistance and don't readily admit
to having any!
7. Inductors
have capacitance, lots of it. Resistance too.
Know how
much if you can. People who make inductors don't like
capacitance and
resistance and don't readily admit
to having any!
8. Resistors
have capacitance, lots of it. Inductance too.
Know how
much if you can. People who make resistors don't like
capacitance and
inductance and don't readily admit
to having any!
9. Conductors
are usually decent inductors. Their
capacitance may be
due to lousy dialectics. Make sure yours is good enough. This includes
ANY insulator between signal and return.
10. ALL mismatched lines (most lines in general) are
resonant somewhere
in the spectrum. If they're not resonant, they're
matched, PERIOD!
Sometimes one can get away with matching them only at high
frequencies
(snubbing). Find or control Z and the frequency
(length) rather than
blindly trying out a slew of
resistor and capacitor values.
Never
assume that where they're resonant
isn't hurting your signal in some
way.
11. If something isn't working right and the voltages don't
tell you
why, start looking at the currents.