Despite the fact my headset project used inexpensive
parts, the resulting amplifier (a diagram of which is shown
in Figure 1) is a high quality device with very low
distortion and noise, and a flat frequency response over
the audio frequency range and well beyond.
Distortion is also excellent — less than 0.0012% from
20 Hz to 1 kHz, increasing to 0.0021% at 20 kHz as
measured with my HP 339A with a 1K load. With a 100
ohm load, distortion increases somewhat to 0.0016% at
low frequencies and 0.008% at 20 kHz. At 10 kHz with
this load, it measures about 0.005%. The level at 1K load
is right at the limit of the sensitivity of this instrument.
Distortion is probably considerably lower than this.
Traditional distortion analyzers measure distortion plus
noise. THD+N is the common nomenclature: Total
Harmonic Distortion plus Noise. Essentially, these
instruments tune out the fundamental frequency signal
and measure everything else, which includes noise and
harmonics. The breadboard circuit produces no audible
noise in the headset with no input signal. Measured noise
at the amplifier output with a 1K load and a measurement
bandwidth of 100 kHz is 20 microvolts.
How good is good enough? It is commonly accepted
that 0.1% harmonic distortion is about the limit of
audibility. A worse problem is that harmonic distortion
indicates non-linearity of the circuit which can produce
what is called intermodulation distortion. Low harmonic
distortion guarantees low intermodulation distortion. I've
tested myself by listening to a pure 500 Hz tone and then
gradually injecting a second harmonic until I could just
hear it. The distortion analyzer indicated 0.08% harmonic
at that point. I feel that it would take considerably more
than that to be heard when listening to music.
How can such an inexpensive circuit have such
impressive performance? The short answer is that the
design is much more important than the cost of the
components. I take no credit for the design except for the
output stage with its inherent shortcircuit protection. The
rest of the project is an implementation of an inexpensive
amplifier and the verification of its performance.
Since the 1960s, there has been one dominant
arrangement or architecture used for most power
Last month, I detailed the assembly of a
headset amplifier. This month, I will discuss
the theory of operation behind that amplifier.
However, most of this discussion will apply
to the power amplifier I wrote about in the
August issue, as well.
Get Amp’ed Up Over the
By Ron Anderson
44 October 2013
Headset Amplifier; September 2013 issue
Power Amplifier; August 2013 issue
Go to www.nutsvolts.com/index.php?/
to post comments.