|
May 1968 Radio-Electronics
 [Table of Contents]
Wax nostalgic about and learn from the history of early electronics.
See articles from Radio-Electronics,
published 1930-1988. All copyrights hereby acknowledged.
|
Here is an interesting
article entitled "Vibrating-Wire Audio Filter and Oscillator," that appeared in
a 1968 issue or Radio Electronics magazine. Author John Rankin
describes a very high-Q (i.e., narrow bandwidth) bandpass filter operating in
the audio frequency band that uses a length of taut wire suspended between the
poles of a horseshoe magnet. He was able to obtain bandwidths as narrow as a
couple Hertz. The useful frequency range proved to be from near 0 through about
20 kHz. Of course the experimental setup used was probably not endure in a
normal operational environment, the principle demonstrated is quite interesting.
It might be considered a mechanical analog to something like a
microwave-frequency magnetron whose resonant frequency is determined by
electrons in a cavity surrounded by a magnet. I have mentioned in the past about
how early (1950s) radio-control system receivers for model airplanes employed a
series of resonant vibrating metal strips (aka
reeds) to demodulate signals from the transmitter.
Vibrating-Wire Audio Filter and Oscillator
Novel amplifier circuit has an exceptionally narrow bandpass
of only 1 or 2 Hz.
By John C. Rankin
If a straight piece of wire is placed in a magnetic field and an audio - frequency
voltage is applied to it, the wire will vibrate and act as a very narrow bandpass
filter.
Wire Filter
Figure 1 shows a 2-inch length of No. 38 wire in a field supplied by a 2 oz. magnet.
The wire resonates near 1000 Hz and behaves like a high-Q LC circuit.
Most 1000-Hz circuits consisting of inductance and capacitance have low values
of Q - between 5 and 10 - due to the high resistance of the large amount of wire required.
Resistance and capacitance are often used in parallel-T circuits for better results.
Although the curve of such circuits indicate good selectivity, close examination
will nevertheless show that they rarely have a Q greater than 15. Sometimes regeneration
is applied to both these circuits to improve selectivity, and at maximum selectivity
the Q may reach 90 in a circuit bordering on instability. The selectivity of any
of these circuits may be classed as poor and leaves a lot to be desired.
The vibrating wire in Fig. 1 has a Q of over 600 near 1000 Hz. It can be one
of the most selective audio amplifiers available for such uses as telemetry, communications
(CW) or wave analyzers.
The short vibrating wire is a low-impedance device. I used the bridge circuit
of Fig. 2 as a bandpass amplifier.
The bridge is balanced by feeding in any frequency except the resonant one of
the wire, then adjusting R2 for a null. When a signal of about 1000 Hz is applied,
the wire vibrates, its impedance changes and the bridge is no longer balanced; so
the 1000-Hz signal is passed.
Fig. 1- Position of pole pieces and magnet is not critical, but
the wire must be taut and not touch the pole pieces.
Vibrating wire in a bridge gives this amplifier an exceptionally
narrow bandpass, making it ideal for remote control work.
A prototype vibrating-wire sinewave oscillator. Its circuit
is unique in that no L, C or R elements are used for tuning.
Transformers T1 and T2 can be any type of speaker transformer used in tube or
transistor radios and connected so that T1 steps down and T2 steps up. The transistor
amplifier makes up for losses in the balancing network; gain from input to output
is about 4.
If the input is too high, the wire will vibrate excessively and will possibly
touch the pole pieces. Under high input conditions, the vibrating wire will act
as a frequency doubler and a 500-Hz signal will produce a 1000-Hz signal at the
output terminals. If signal level is kept reasonably low, however, the device remains
linear. Good operating conditions seem to be about 50 mV in and 200 mV out.
Strong input signals can move the wire to the electrical center of the pole pieces,
which may not be the position in which the wire was originally placed. This change
in position alters the strain on the wire, varying the original frequency. You can
avoid this by sealing the wire at both ends so that it cannot be moved at the end
supports.
Wire Oscillator
The vibrating wire can also be used as an oscillator. The circuit of Fig. 3 oscillates
at around 1000 Hz and is a bit unusual; there are no capacitors in the oscillator
circuit, only in the power supply.
Variable resistor R7 is adjusted for lowest output distortion when used with
two types of wire to be described later. Distortion is 8% at 400 mV out and 3% at
100 mV. These figures probably could be improved, however, by using true complementary
transistors and changing resistor values. Both transistors are high-gain types
and should not be changed unless similar gain units are available. I constructed
several of these oscillators; all performed well.
Vibrating-Wire Construction
The assembly consists of a General Company 2-oz. magnet placed on 2" x 1/2" iron
mending plates made by National Manufacturing Company of Sterling, Ill. Both items
are available in hardware stores. The mending plates are mounted with 1/4" bolts
on a 3" x 1/2" piece of 1/16" plastic. I spaced the plates about 1/16" apart. The
wire is stretched taut in the center of the gap and soldered to a 2-56 screw mounted
at each end of the plastic. Run the wire through a hole at each end of the plastic
before soldering it to the screws; then shim the plastic rod (.025" diameter) at
each wire end. You can cut the thin plastic rod from a plastic whisk brush.
As the heart of the device, the wire must be prepared carefully. Originally,
I used No. 38 enameled copper wire; the enamel was removed with fine sandpaper.
Then I noticed the resonant frequency changed in time, because the copper did not retain its original tautness. A 6" length was then
stretched to 7" before installation - with improved results. At this stage I noticed
something unusual: only half the length of wire had stretched - which meant the
wire had two diameters - and the output of the amplifier had two adjacent peaks similar
to two overcoupled coils.
Fig. 2- Resonant frequency of vibrating wire (VW) sets passband
of this amplifier.
Fig. 3- Unique sinewave oscillator uses a simple vibrating wire
as tuning device.
My best results were obtained with a special nickel alloy wire. An assembly using
this wire was subjected to heat and cold and proved very satisfactory. When it was
frozen for 1 hour and immediately checked, the original frequency of 1000 Hz had
changed to 960 Hz - only a 4% change. Since wire contracts with cold, the frequency
should have increased; the lower frequency may have been due to moisture on the
wire causing a change in its mass. Nevertheless, within 10 minutes the resonant
frequency returned to 1000 Hz, although the unit was still very cold.
Nickel alloy wire is easy to handle and solders easily; it is available in 24"
lengths. For those who may wish to try it, it is called Nickel Alloy 180 with 0.004"
diameter. You can order it from the manufacturer, California Fine Wire Co., 390
Manhattan St., Grover City, Calif. 93433. Cost is $1, including postage in the USA.
Adjustment and Operation
The value of resistor R1 in Fig. 2 depends on the type of wire used as the vibrator;
if you use copper, R1 should be a piece of wire identical to that used as the vibrator.
Nickel alloy has a resistance about 15 times greater than copper, so with nickel
wire, you should use a 30-inch length of No. 38 copper wire, or a carbon resistor
of 1 or 2 ohms, for R1. You will find the correct value when R2 nulls at about mid-range.
Resonant frequency of the vibrating wire is varied, by changing the tension on
the wire with a screw and threaded bushing. I made a unit that had a frequency variation
of 750 Hz to 1500 Hz. You can use larger magnets, but oscillator output may be square
waves, due to overdriving. In this case, it would be necessary to increase the value
of R6 to possibly 150 ohms. An increase in magnet size means an increase in impedance
of the vibrating wire, and in the amplifier this would mean an increase in output
voltage.
To test a vibrating-wire amplifier you need a good audio generator. Q is over
600, and I made this Q measurement by the 3-dB method; this means the bandwidth
was 1.6 Hz at 1000 Hz Obviously, most audio generators cannot read this accurately.
To make the measurements I constructed a generator with a 180° dial covering 10
Hz to 1000 Hz. Allowing for pointer width and dial markings, it was debatable whether
the bandwidth was 1 Hz or 1.6 Hz - which actually puts the Q between 600 and 1000.
Q is controlled by damping, and air causes damping of the vibrating wire -just
as it does to a loudspeaker cone. Therefore, it's quite possible that even higher
Q might be obtained by installing the wire in a vacuum.
The frequency of the wire may be changed by changing its length or diameter or
both. There is no low-frequency limit; units may be constructed for 60 or 120
Hz. The practical high-frequency limit seems to be in the vicinity of 20 kHz.
I chose a bridge circuit for the selective amplifier because it is simple and
easy to get operating. A bridge is not necessary, though; a differential amplifier
- integrated circuit or otherwise - might well prove more satisfactory.
Whichever way you try the vibrating wire, you'll find it quite different from
the LC or RC circuits.
|
