A Study of E-meter Frequency Response

An Electrical Review by Perry Scott, BSEE


"By itself, this meter does nothing. It is solely
for the guide of Ministers of the Church in
Confessionals and pastoral counselling. The
Electrometer is not medically or scientifically
useful for the diagnosis, treatment or preven-
tion of any disease. It is not medically or
scientifically capable of improving the health
or bodily function of anyone and is for
religious use by students and Ministers of
the Church of Scientology only."

[ Introduction | Method | Results: Mark 7 E-meter | Results: Mark 6 E-meter | Discussion | Other Observations | Conclusion ]

Introduction

Scientology relies on the e-meter for auditing, which they have described as a form of "spiritual counselling". Hubbard claims that by asking a preclear questions and reading their response on the meter, an auditor is able to find areas of upset in the preclear's life. Hubbard also claims that the e-meter measures "mental mass", or energy buildups caused by previous upsets. In yet another claim, Hubbard states the e-meter reads the "carrier wave" of the thetan (spirit), and goes on to define its frequency as some ridiculously high number, on the order of 1069 Hz. (By comparison, microwave ovens run at 109 Hz.)

In actuality, the e-meter is simply a high-precision differential ohmmeter. Changes in the person's resistance are indicated by the meter movement. The auditor adjusts the "Tone Arm" (don't ask...) to center the needle on the dial. The auditor can also adjust the gain of the E-meter amplifier to register only gross changes or every little flinch of the preclear.

It should be noted that a physical meter movement has mass, which according to Newton, resists a change in motion. Because of this, meter movements (and especially large meter movements like the one used in Hubbard's E-meter) are incapable of measuring frequencies beyond a few hertz. This experiment seeks to determine the frequency response of the E-meter.

It should be noted that the low frequency response of the e-meter places an upper limit on a thetan's "carrier wave". Using Planck's constant, one can relate frequency to energy. When a "carrier wave" in the 1 Hz region is related to energy, it can be seen that Hubbard's thetans possess very little energy indeed; certainly not enough to be "at cause over Matter, Energy, Space, and Time."

Method

To measure the frequency response of the e-meter, a Wein Bridge oscillator, oscilloscope, and single-ended power supply were used. (Ironically, a bridge oscillator is being used to measure a bridge detector.)

The oscillator uses a TL074 quad op-amp, although only two of the op-amps are actually used. Op-amps typically require two power supplies, one positive and one negative. To avoid this expensive requirement, a "virtual ground" was created for the oscillator at precisely the midpoint between true ground and the supply voltage. The first op-amp (U1A) supplies the virtual ground while the second op-amp (U1B) implements the oscillator.

U1A is configured as a "voltage follower", that is to say its output tracks the voltage on U1A+. The 10-turn pot at U1A+ centers virtual ground at one-half of the input voltage. The U1B- 10-turn pot adjusts the gain of the oscillator. When enough gain exists, a sine wave appears at Output. The two diodes stabilize the oscillator. Oscillation occurs when the reactance of one RC arm is equal to the reactance of the other RC arm, and occurs at a frequency of 1/(2*pi*RC). Due to capacitor and resistor tolerances, frequency will deviate somewhat from the equation. Therefore, frequency is measured directly from the oscilloscope.

The theory behind the Wein Bridge oscillator is beyond the scope of this discussion. For more information, the reader is directed to a college textbook on oscillator theory.

A Dual Trace Storage Oscilloscope was used for this experiment. By storing the waveform, 10 or more cycles can be displayed and a more precise frequency measurement can be obtained. Before each E-meter measurement is taken, the oscillator voltage is precisely controlled to yield an output of 10.0 Volts peak-to-peak (Vpp), +/- .05 volts. The E-meter has 40-50 divisions on its face, so the reading is simply the number of divisions the needle "swings" in response to the standard 10 Vpp signal. The Tone Arm is used to center swinging needle on the meter and to further adjust the lower side of the swing exactly on a division mark.

Oscillator Startup

For each frequency measurement:

The E-meter alligator clips hold a 5.6K resistor. Hubbard uses a 4.99K precision resistor to calibrate the E-meter for females, and 7.50K for males. Thus, 5.6K is within the specified operating range of the meter. The oscillator is always set for the same amplitude for each frequency, in this case 10.0 Vpp Output, which is coupled to the E-meter with a 150K resistor. This sets the injected current into the E-meter at 10/150K = 66 ua pp. Oscilloscope ground, power ground, and one side of the E-meter are tied to a common point.

Results: Mark VII and VI Frequency Response

The Mark 7 frequency response curve was obtained by varying the value of R in the oscillator. The frequency was adjusted by using various resistance values. Three measurements per frequency decade were taken, with intermediate measurements taken when something "interesting" was happening. The meter face is divided into 40 units, and the width of the meter oscillation was recorded for each frequency. Frequency response graphs are plotted log10(f) vs 20*log10(V/V0).

A Mark 6 became available, and it was felt the experimental results could be improved by obtaining more data points for the frequency response curve. This was accomplished by substituting a dual variable resistor for R in the oscillator, as well as replacing the 2.2uf capacitor with six capacitors on a dual rotary switch. This allows a continuously variable frequency range from .1 Hz to 100,000 Hz, and is useful for finding the precise resonant frequency of the meter movement.

Mark 7
 Freq          R
 (Hz) Units (ohms)
 .071  21    1.0M
 .15   22    470K 
 .32   24    220K
 .71   33    100K
 1.5   34     47K
 2.1   9.5    33K
 3.3   4.0    22K
 4.9   1.5    15K
 7.1    .5    10K

Mark 6
 Freq
 (Hz) Units
 .108  18
 .136  18.6
 .157  20.3
 .177  21.0
 .215  22.1
 .223  22.0
 .267  23.2
 .325  24.0
 .384  24.0
 .430  24.6
 .521  26.6
 .665  31.6
 .728  36
 .751  36.5
 .873  42
 1.02  35.6
 1.05  34.5
 1.14  26.5
 1.25  19.5
 1.43  10.5
 1.70  4.3
 1.99  2.0
 2.55  1.0
 3.42  0.5
  

Various Mark 6 control and electrical measurements were also taken:
TA
[Tone Arm]
Gain Calibration
Resistor (Ohms)
Vpc (max) Ipc
1 - - .9165
2 32+2.3 4999 .9870 197 microamps
3 32+1.0 12500 1.927 154 microamps
6 - - 2.503

Other Observations

The Mark VI has roughly the same meter movement as the Mark VII. This particular Mark VI was slightly more sensitive in the low-amplification setting than the Mark VII, so a 180K-ohm resistor was used to inject the 10 v-pp sine wave.

The Mark VI comes with two calibration resistors - 4999 and 12500 ohms for TA=2.0 and TA=3.0, respectively. This differs from the Mark 7 values of 4999 and 7500 ohms. The GAIN and TRIM controls are adjusted until the meter needle is at the SET position. It was also observed that adjusting the TONE ARM or GAIN setting required a corresponding adjustment of TRIM, which is an annoying design flaw in the human/machine interface.

When the e-meter probes are placed across a 5.6K resistor, the oscilloscope shows a pure DC sampling current. DC is not preferred, because it causes electrolysis at the skin/metal junction. In effect, this creates a mini "battery" and can invalidate the measurement. When AC is used, this "battery" is reversed at every cycle, minimizing the effect. Martin Kratz wrote an article detailing this effect, as well as other perils of the E-meter that resulted from Hubbard's psuedoscience.

The E-meter has three ferrite toroids to protect it from induced RF pickup. There are no other frequency-sensitive components in the detector circuit that would account for the low corner frequency.

Discussion

This curve is immediately recognized by an electrical engineer to be a classical low-pass filter. The frequency response below 0.2 Hz is relatively flat.

The "bump" at 1 Hz indicates that the system is underdamped. This is confirmed by Arnaldo Lerma's contention that Hubbard required an underdamped meter movement for "proper" performance. (It should be noted that electrical systems design engineers attempt to achieve what is called critical damping. When a system is critically damped, it will "settle" as quickly as possible. An underdamped system will tend to oscillate after a large step in input, as seen during this experiment.)

Finally, the response rolloff at 2Hz is 9dB in a frequency range of 7.1/2.1 = 3.38. This corresponds to a rolloff of 17.3 dB/decade. The theoretical single-pole rolloff is 20 dB/decade, so this matches well to a graphical determination based on data close to the corner frequency.

Determining the actual corner frequency is done by graphical inspection. By examining the response curves of other underdamped low-pass filters with known corner frequencies, professional experience would place the corner frequency somewhere under the peak at approximately 1.0 Hz.

Conclusions

Given a lack of any electronic filtering in the 1 Hz range, it is apparent that the mass of the meter movement is the biggest contributor to the limited frequency response of the E-meter. Higher frequency response could be gained by reducing the rotational inertia of the needle - essentially reducing its mass at the tip, as well as reducing its overall length. Of course, a modern system would dispense with the mechanical movement entirely and use an analog-to-digital converter inside a PC (a.k.a. SoundBlaster), displaying the result on the computer screen. It is also noted that a fully-loaded PC is also less expensive than an E-meter and you can even use it to surf the Web when you're done auditing your buddies.

The underdamped nature of the E-meter will cause certain measurements to look the same. For example, a 1Hz square wave will cause the same motion on the needle as a 1Hz sine wave. In biofeedback terms, a rapid change (square wave) will be smoothed out (sine wave). These two conditions indicate radically different underlying physiological phenomena, so a device which confuses the two is decidedly non-optimal.

The E-meter exhibits the classical frequency response of an underdamped single-pole low pass filter with a corner frequency of approximately 1.0 Hz. Other medical research (EEGs, EKGs, etc) show that bodily processes occur at frequencies several orders of magnitude higher than this. As such, the E-meter is an extremely poor diagnostic tool.

Hubbard's notion of "carrier waves" is pure unadulterated hoakum. The E-meter uses a DC sampling current and the device cannot measure anything beyond 10 Hz. Taken along with its other deficiencies, it is clear that Hubbard's primary use of the E-meter was to lend an air of scientific respectability to an area fraught with psuedoscience.