Wednesday, Jan. 28, 2015

Electrostatic potential
We'll start with something we weren't able to cover in class on Monday - electrostatic potential. 




I'm not enough of a mathematician to be able to explain why this is true (other than demonstrating that the curl of a gradient is zero).  We'll just have to accept that on faith.  And actually figuring out what the scalar function needs to be is another problem.

The curl of the r (a vector) over r3 (magnitude of r cubed) term in the expression for electric field is zero.


The scalar function in this case is - (1/r)


We'll insert this into the expression for electric field.



It is often much simpler to determine the electrostatic potential because it is a scalar quantity.  The electric field can then be determined by taking the gradient of the potential.


We didn't cover any of this short digression in class.

When the curl of a vector field is zero, the vector field is irrotational.  There's a pretty good  online animation  that shows both irrotational and rotational vortices (spinning fluids).


The pattern of electric field vectors around a positive charge would look something like the figure below.  We can imagine that the arrows represent fluid motion and can picture what would happen if a small object were placed in the moving fluid (as was done in the animation above).  It would move outward without any rotation.


An irrotational field has another important property.  A line integral from from one point to another will be independent of the integration path. 




Because the curl is zero we can use Stokes' Theorem to say the line integral of the vector around a closed loop is also zero.  The rest of the argument follows fairly simply from that.


Sometimes rather than starting with Φ and then determining the E field, we might know the E field and want to determine Φ




We can determine the potential as shown above (there's no reason rref  needs to be the upper limit, it could just as easily be the lower limit)

Let's assume a point charge and substitute in an expression for E into the left integral above.

We'll set this equal to the earlier expression


We get our earlier result (provided we assume that Φ(r =) is zero)


We can write the electric field as the gradient of the electrostatic potential and then substitute that into Gauss' Law.



We obtain Poisson's Equation.  Laplace's equation applies in situations where the volume space charge density is zero.  We'll be using Laplace's equation in our next lecture.  Here is a handout with vector differential operators (Laplacian, curl, gradient and divergence) in cartesian, cylindrical, and spherical coordinate systems.


"Fast" and "slow" E field antenna systems


Another example of a field mill record, one hour of actual thunderstorm and lightning fields recorded at the Kennedy Space Center, is shown below (from:  Livingston and Krider (1978)).  The abrupt transitions are caused by lightning and are superimposed on a static field of about 3 kV/m (negative potential gradient corresponds to a positive E field pointing upward toward negative charge in the bottom of a thunderstorm).


A field mill can be used to determine when a thunderstorm becomes electrified and monitor electrical activity in a thunderstorm.  Note that a fairly large dynamic range is needed (-12 kV/m to +12 kV/m) is needed to insure that the E field remains on scale.  Later in the semester we will see that measurements of the lightning field changes at multiple locations can be used to determine the magnitude and location of the charge neutralized by the lightning discharge. 

A lightning discharge only lasts 1 second or so and appears as just an abrupt field change on the record above.  We're going to find that a lightning discharge consists of a series of different processes that occur on millisecond, microsecond and even sub microsecond time scales.  The figure below illustrates this (don't worry about all the details and names at this point, we'll come back to this later in the semester).  An electric field mill can measure static and slowly varying fields but wouldn't faithfully resolve all the field changes and variations that occur on these faster time scales.  We need a different kind of measuring system.




One way of measuring these faster time varying electric fields is to use a flat plate antenna (aka flush plate dipole antenna).  It basically consists of a large flat grounded plate that would be positioned on the ground (preferably flush with the surrounding ground).  A smaller circular insulated sensor plate is found inside a center hole as shown in the photograph below (the antenna is on the classroom floor in this photograph).  The center plate "senses" the electric field and is insulated from ground.




We look under the top plate of the antenna in the next picture.



The center sensor plate is supported by insulating nylon or Teflon spacers.  The top end of the supports are covered with "rain hats" to try to keep the insulators dry during rainy weather.  A wire connection to the center plate connects to a coaxial cable to carry the signal to processing and recording equipment.

In some ways the operation of this antenna is similar to the field mill.  In this case a time varying E field causes current to flow to and from the center sensor plate (you don't need to repeatedly cover and uncover the sensor plate).


This current is proportional to the time derivative of the electric field (σ in the equation below is the surface charge density on the sensor plate). 

Integrating the current gives an output signal that is proportional to E.


In the circuit above the antenna is connected to a capacitor, this is a passive integrator.  Some kind of measuring device would then be connected across the capacitor.




The value of the RC decay time constant determines whether the antenna works as a "slow" or a "fast" antenna system. 

A Slow E field system with a long, 1 to 10 second decay time would be appropriate if you wanted to study an entire lightning discharge on a faster time scale.  A couple of actual Slow E field records are shown below (also from the Livingston and Krider (1978) paper cited earlier).

Note first the much faster time scale, 0.4 seconds full scale in this case.  The step changes in the E field are lightning strokes to the ground.  The top example shows a 3 stroke cloud-to-ground discharge.  The second discharge has 4 strikes to ground.

Because the Slow E antenna system does not have DC response the static E field (which can be several kV/m) is effectively filtered out (like switching from DC to AC coupling on an oscilloscope).  Lightning field changes can be examined with more gain.  Because the signals of interest last from 0.5 to 1 second, a decay time constant of 10 seconds would be appropriate here.
  The slow E field would decay back to zero during the interval between lightning discharges.

To give you some appreciation for how recording methods have changed, the signals above were (I believe) displayed on a storage oscilloscope and photographed with a Polaroid film camera!

Because of the long decay time constant, charged precipitation falling on a Slow E field antenna can drive the signal off scale.  Inverted antennas are sometimes used to avoid that problem.







This antenna is mounted on the roof of the Penthouse atop the PAS Building.  Because of its exposed position and the chance that it could be struck by lightning, signals are brought into the Penthouse on a fiber optic cable.

The E field variations that occur during a individual return stroke could be examined by increasing the vertical gain and displaying the Slow E field signals on a faster time scale.  However, as sketched below, the long time decay would mean the signal might not decay back to zero in the interval between return strokes.


A slow E field record (a few seconds long decay time constant) displayed on a much faster time scale.

The solution is to shorten the RC decay time constant.  This turns the slow antenna into a fast antenna system.




Same antenna but with a much shorter (milliseconds long) decay time constant.


A decay time of about 1 to 10 milliseconds would be long enough to accurately record the Fast E field variations but would allow the signal to decay back to zero in the interval between strokes. A short decay time constant would also mean charged precipitation would be less likely to drive the Fast E signal off scale.


The operation of the flat plate antenna was demonstrated in the last few minutes of class.  The setup is shown below


We first need to create a time varying electric field of known amplitude.  We do this by placing a flat metal screen on insulators a known distance (0.125 m) above the top of the antenna.  A square wave signal from a function generator is connected to the test plate.

The electric field created by the test screen is just the voltage on the screen (10 volts) divided by the distance between screen and the antenna (0.125 m); 80 V/m.
 




The signal coming from the antenna is a current that is proportional to dE/dt, the time derivative of the electric field.  The antenna signal is connected to a 1000 pF integrating capacitor.  The output voltage across the capacitor will be proportional to E, the electric field.  This voltage signal is displayed on an oscilloscope.




The test signal and the antenna output signal are shown above.  The test signal was a 0 to 10 volt square wave and is displayed on a 5 volt/div scale above.  The antenna signal is shown below.  Note the exponential decay of the signal that occurs because the 1 M Ω input impedance of the oscilloscope is connected in parallel with the 1000 pF integrating capacitor.  The signals are being displayed on a 1 ms/div time scale.  A closer view of the oscilloscope display is shown at right.

We'll do a quick calculation of the voltage that we would expect from this setup and see how well it agrees with the actual measurements above.



We would expect a peak signal of about 80 mV, we seem to be getting about 60 mV.  That's not too bad.  I suspect the reason for the difference is that there may be some stray capacitance between the antenna and ground that we would need to add to the 1000 pF in the integrating capacitor.

As a final test, the antenna was connected to a small (50 pF) integrating capacitor.  That should increase the amplitude of the antenna signal and shorten the decay time constant.




The oscilloscope display above confirms this.  Though again the signal isn't as large and the time decay as short as calculations would predict.  Stray capacitance is probably again the reason for the discrepancy.  The stray capacitance may well be larger than the 50 pF integrating capacitor that we are using in this test. 




References:
J.M. Livingston and E.P. Krider, "Electric Fields Produced by Florida Thunderstorms," J. Geophys. Res., 83, 385-401, 1978.