A high-fidelity audio signal chain gains several advantages when the system supply voltages are higher – the signal-to-noise ratio, dynamic range and slew rates can all improve with increasing supply voltages. But constraints from battery- and USB-powered systems or limitations on the audio analog-to-digital converter or codec may require signal processing with low-voltage supplies.
Picking an amplifier for these applications isn’t always straightforward. An amplifier built with bipolar technology gives the best possible combination of noise and quiescent current, but has a limited input and output signal dynamic range. Consider a basic amplifier configured as a buffer on a 3.3-V rail. Without a rail-to-rail input, the input signal range is typically limited by 1.2 V. The output stage can also limit the dynamic range. Figure 1 below shows a typical AC-coupled op-amp configuration, with the common mode voltage biased at mid-rail.
Figure 1: Input and output voltage swing
Table 1: Comparison of input and output signal voltage ranges
Nominal consumer audio line-level input voltages are -10 dBV; designing for an additional 10-dB headroom gives a peak-to-peak input voltage of 2.828 V. Table 1 shows that in order to buffer or filter a signal of this magnitude with only a 3.3-V supply rail, an audio-grade, rail-to-rail input CMOS amplifier is required. Additionally, professional wireless microphones must deliver high-fidelity audio within small-sized solutions capable of more than 8 hours of battery operation. These systems require small-form-factor solutions with low quiescent current and extremely low noise floors to enable extended portable operation and high-fidelity audio recording.
In addition to merely limiting the voltage, a bipolar output can have a significant increase in quiescent current when the output is in saturation and can show significant delay when recovering from output overload, leading to harsh harmonics when the signal is at or near the point of clipping. In Figure 2, a schematic example of an overdriven amplifier is given.
Figure 2: AC-coupled, unity gain audio amplifier schematic example
Figure 3 shows the output of the amplifier in Figure 2, simulated using circuit models. For the simulation, the amplifier input signal is driven beyond the rails for two different amplifiers: a standard, 3-MHz, rail-to-rail input/bipolar output amplifier and a new 13-MHz, rail-to-rail input/output CMOS amplifier from TI, the OPA1671.
Figure 3: Transient simulation of clipping on a bipolar vs. CMOS amplifier
In Figure 4, the simulation zooms in closely to the peaks where it is easy to see that not only does the CMOS amplifier exhibit a wider output voltage swing before clipping, but the wide bandwidth and CMOS outputs allow for a much faster recovery from clipping.
Figure 4: Transient simulation of clipping on a bipolar amplifier vs. a CMOS amplifier
Many CMOS amplifiers improve on the headroom limitations of bipolar amplifiers thanks to lower threshold and saturation voltages inherent in CMOS processes. However, voltage noise, total harmonic distortion (THD) and output impedance will typically be compromised.
The OPA1671, a rail-to-rail input and output CMOS audio amplifier, helps provide a solution to these design challenges by providing ultra-low noise, THD and wide bandwidth (13 MHz). The rail-to-rail input and outputs, along with the high input impedance, enable wide signal swings in buffer configurations. The low current and voltage noise, coupled with a low 1.25-mV offset voltage, make the OPA1671 an excellent amplifier for high-gain preamplifier circuits that maximize output swing and can provide very low output-impedance drive to an audio ADC or codec.
Nice post, and nice part! Always happy to see people cooking with SPICE.
Interesting post Michael! :) I usually make sure that my signals have enough headroom but saturation can always occur. Perhaps you could explain the difference between a bipolar output and rail-to-rail output please? Thanks!
The difference is in MOS devices vs. bipolar when the V(source to drain), or V(collector to emitter), respectively, becomes too low. MOS devices go into the "triode" region, where they are effectively a resistor, whereas bipolar devices go into "saturation" region, where they are commonly modeled as a low voltage DC source (~0.2 V). In the saturation case for bipolars, the base current of the output device will increase significantly. When this occurs, it draws increased current from the previous stage that is driving it. The output drivers are not typically designed to operate with the normal output response at this current level, so, the response can be slow, depending on the driving stage architecture. In other bipolar devices I have worked on in the past, there was actually an "anti-saturation" circuit necessary around the bipolar device to keep in from entering this state.
MOS devices, on the other hand, will never have an increase in gate current. While the device will still go into "triode" region that has its own negative effects (loss of Gm, Gate to Drain capacitance will go up, etc.), the loading on the driving stage problem does not exist.
Let me know if this helps.
All content and materials on this site are provided "as is". TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with regard to these materials, including but not limited to all implied warranties and conditions of merchantability, fitness for a particular purpose, title and non-infringement of any third party intellectual property right. No license, either express or implied, by estoppel or otherwise, is granted by TI. Use of the information on this site may require a license from a third party, or a license from TI.
TI is a global semiconductor design and manufacturing company. Innovate with 100,000+ analog ICs andembedded processors, along with software, tools and the industry’s largest sales/support staff.