• Greater than 120 dB of dynamic range
• SNR of 120 dB
• Output voltage noise of 5 μVrms at gain = 2V/V
• Power supply range: ±5V to ±15V
• 1300 V/μs slew Rate
• Independent power supplies for low crosstalk
• Short circuit and thermal protection
The total harmonic distortion plus noise (THD+N),
dynamic range, signal-to-noise ratio (SNR), and slew rate
(the maximum rate of change of a signal at any point in a
circuit) are excellent. For comparison, the THD+N figure
for a National Semiconductor LM386 is 0.2%.
Although great specifications don’t necessarily
translate to great sound, they do set a baseline for what is
possible. See the Texas Instruments website ( www.ti.com)
for the official datasheet, application notes, and user
guide for the TPA6120A2. More importantly, download
the documentation on the TPA6120A2 evaluation module,
which provides drawings of suggested component layouts
and ground plane configuration. I used the ground plane
configuration from the evaluation module as a model for
the design presented here.
The Analog Devices AD8610 is a surface-mount JFET
input op-amp with low offset voltage and drift, low current
noise, and low input bias current. Two of these wide
bandwidth op-amps are used in the project as precision
signal-level buffers. See Analog Devices website (www.
analog.com) for the datasheet with detailed specifications.
In brief, the noise figure and slew rate of the AD8610
complement those of the TPA6120A2. Even so, feel free
to substitute your favorite low-noise op-amp for the AD8610.
Many op-amps are pin-compatible with the AD8610 and
you should be able to use the existing component values.
Why use a different op-amp? Some audiophiles claim to
hear a difference in the audio produced by different
op-amps — but I confess, I can’t detect a difference.
Resistors are not created equal. Ordinary thin film
surface-mount resistors — while inexpensive — are noisier
and less stable than metal film resistors. I suggest you use
metal film resistors throughout the project. You might save
a dollar or two by using thin film surface-mount resistors
instead, but at least consider metal film resistors for the
input circuit to the AD8610s. Resistor-generated noise is
more noticeable when inserted early in the amplifier
chain. Most of the capacitors used for the signal path are
low-noise PPS film. You may be tempted to substitute less
expensive ceramic capacitors for PPS film capacitors, but
you’ll have better results with the film variety.
The connectors used in this project are RCA jacks for
audio input and 1/4 inch audio jacks for instrument input
and audio output. The instrumentation input audio jack is
a mono phone jack with transfer circuit that connects the
inputs of the left and right channels together when a 1/4
plug is inserted into the jack. The only splurge item in the
connector category is the set of gold and Teflon RCA
jacks, shown in Figure 3. The connectors, available from DIYCable.com ( www.diycable.com), are about double the
price of what you can pick up from Digi-Key, but the
higher quality is obvious. In addition, the Teflon insulators
allow you to float the ground until the signal reaches the
input of the amplifier or input attenuator.
The input circuit shown in Figure 1 is designed for a
fixed, tube-based preamp input. If you want to work with
a variety of input sources, then consider adding a stereo
potentiometer to the circuit. For each channel, feed the input
signal across the full resistance of the potentiometer and
take the signal from the wiper arm as the input to the
An inexpensive audio taper pot from Alpha or
RadioShack will cost about $3. For about $40, you can
use an audiophile-quality pot by ALPS. The third option is
to use a variable attenuator, which is a switched series of
discrete resistors. Popular brands for switched attenuators
are DACT and GoldPoint, at about $170. I’ve also seen
stepped attenuator kits from China on eBay, starting at
about $30. I chose a 50K stepped attenuator from
GoldPoint for this project (see Figure 4).
What do you get from a stepped attenuator for such a
huge outlay? One is the ‘feel’ of the stepped level control,
especially when coupled with a large, heavy knob. The
second is precision relative tracking, as shown in Figure 5.
As shown in the figure, as the inexpensive 50K Alpha pot
is moved through its full range of resistance, the difference
in resistance between left and right elements ranges from
0 to about 5K, with a marked peak around 18K. The
tracking of the more expensive ALPHA potentiometer is
much better, with a maximum difference of about 1.5K.
The GoldPoint stepped attenuator, in comparison, showed
no significant difference throughout the 50K ohm range.
We could debate whether the variation in relative
tracking for a given pot is detectable — most humans can
■ FIGURE 3. Gold
and Teflon RCA jacks.
June 2008 51