[Phys-l] grounding and shielding (was: capacitance problem)

[John Denker <jsd@av8n.com>]

On 03/29/2008 12:25 AM, Bernard Cleyet wrote:

> Did you try air dielectric caps.  e.g. broadcast band tuning  
> (variable) ones.? 

If the objective is to measure stray fields, why use a 
capacitor at all?  It can't possibly help.

There are two incompatible lines of thought here;  take
your choice:
  a) Analysis of the circuit described at 
   http://www.phy.mtu.edu/~suits/PH2260/CapProblem.pdf
  b) Measurement of typical stray fields.

Just because somebody claims (a) and (b) are related doesn't
make it so.  As far as I can tell, the only connection 
between (a) and (b) is via some false assertions about the 
total impedance of capacitors in series.  Any introductory
treatment of series and parallel circuits should suffice
to explain why these assertions are wrong.

It's your choice:  Pick one:
  a) If you want to analyze the circuit, analyze the circuit.  
   You should be able to analyze it, quantitatively, in your 
   head.   http://www.av8n.com/physics/capacitive-divider.htm
  b) If you want to measure stray fields, there are reasonable
   ways to do this, as discussed below.  
   -- Good technique does not require capacitor C1;
   -- does not require capacitor C2;
   -- does not require switch S1;
   -- does not require switch S2; and
   -- certainly does not require an elaborate protocol for 
    opening and closing the switches.



>  Being order 300 pF will require an electrometer,  
> e.g. op amp w/ > 30 gigaohm   (time constant ~ ten seconds).     

No, if you're trying to measure typical stray fields you don't 
need anything like that.

It suffices to use an ordinary oscilloscope.  For stray fields
that are primarily *capacitively* coupled, you can pick up the
fields using the lid from a wide-mouth pickle jar.  If you
want something bigger, a "disposable" aluminum pie pan is
convenient.  You can grab it with the scope probe and wave
it around to estimate the position and orientation of the
fields.

Hint: Comparing the signal you get with 1x, 10x, and 100x 
scope probes allows you to estimate the Thévenin open-circuit 
voltage and Thévenin impedance of the stray capacitive injector.

Note bene:  Any field that is too small to be measured this
way is much to small to have any significant effect in
ordinary real-world circuits, with rare exceptions, provided
standard basic electrical engineering principles are used.
That is, you can intercept the stray fields and divider
them to negligible levels by simple, basic shielding.  This
is why people use coax instead of bare wires.  This is why
people build circuits inside a chassis.  It's not rocket
science, it's just introductory-level physics.

  The exceptions include things like hooking wires to neurons
  during brain surgery, where you don't have a good "chassis 
  ground" and where there is little tolerance for stray currents.  
  This is a specialized and advanced case.  This is a very 
  distant departure from where we started, but we can follow 
  this tangent if anybody is interested.  It's just physics.

In typical circuits, you will find that the dominant stray
capacitances are internal to the circuit, e.g. interwinding
capacitance in transformers.  Hence the market for triple-
shielded transformers, much loved by people who are serious
about grounding and shielding.

By the way, it is also possible to measure the stray
*inductive* contributions using an easily-made device
   http://www.av8n.com/physics/img48/stray-field-pickup.png
plus an oscilloscope and a couple pieces of coax.  Can
you see how it works?  For starters, note that the shield
of the loop of coax protects the circuit against stray
capacitive contributions, so it is only sensitive to stray
inductive contributions.  Remember that 
       V = Phi dot 
is one of the Maxwell equations.  Here Phi is magnetic flux.

In many classroom and/or laboratory settings, you will 
find that these inductive contributions are much larger
and much harder to shield than the previously-considered
capacitive contributions.  That's because the inductive
guys have a near-zero Thévenin impedance, whereas the
capacitive guys have a near-infinite Thévenin impedance.
This is the physics behind the dreaded "ground loops".