[sdiy] Wanted: non mathematical description of the functionof RC-filters

[David G Dixon]

> There is a lot between black and white. Some people will 
> appreciate that you don't handle them as stupid ones, though 
> they wil not have the same level of insight as you. The word 
> for this social ability is "respect".
> 
> Every teacher should have this ability.

The problem here is that every analogy put forward only makes the filter
harder to understand, because there really isn't another system which
behaves quite like it.

Thinking about the voltage divider, there is basically a pipe T, where one
branch of the pipe is supplied with water (current) at a given pressure
(voltage), and the other branch empties out into ambient pressure (ground).
At the T there is a pressure gauge.  The roughness of the pipe (which
creates friction and causes the pressure to drop) corresponds to resistance.
Hence, if the pipe tee separates two long pipe sections of identical length
and roughness, the pressure measured at the junction will be exactly half
the (gauge) pressure of the upstream source.  This is analogous to an
unloaded voltage divider with two equal resistors.  If a third pipe is
connected to the T, then the divider is loaded, and the pressure measured at
the T will be less than half the source pressure because there is additional
flow through the third pipe.  This is the analogy of a loaded voltage
divider.

Of course, with an RC filter, one of the resistors becomes a
frequency-dependent resistance.  Assuming a LP filter, the capacitor is
between the T and ground.  If the frequency is zero (DC), then the capacitor
is a perfect insulator, and no flow goes to ground.  In this case, the
pressure at the T is exactly the same as the source, since no flow (current)
implies no friction (resistance) through the resistor.  The higher the
frequency, the less resistive the capacitor becomes, and hence, the more
flow will leak out from it.  At a very high frequency, the capacitor has
little or no resistance to current flow, and the pressure at the T will
start to look a lot like ground.  Of course, this means that the signal
there will be inaudible.

The point I'm trying to make here is that any meaningful explanation of an
RC filter is going to have to involve current flow and impedance.

I have a few DIY friends for whom the whole concept of "impedance" is very
hazy.  In my opinion, the true understanding of electronics begins with an
understanding of impedance -- an intuitive understanding of impedance is
probably the single most important weapon in any designer's arsenal, and yet
it can be a fairly difficult concept (I am thankful that I have a solid
grounding in fluid mechanics, because I think it makes electronics much more
intuitive).