DP1 is an in-progress high performance headphone amplifier. It is centered around the LM4562 operational amplifier. It is to be powered by any standard 5V USB power source, and will feature a micro-USB port.
While the circuit is powered by a 5V USB port, the circuit will utilize an isolated DC-DC converter. The V+ and V- lines from the USB port enter the R05P209D, and the 2-ported output will have +9V across its terminals. An isolated DC-DC converter is used prevent any current loops, differences in ground between the power lines and audio lines, or any other strange circumstance. The new +9V supply will then be put through a voltage divider with 220k resistors, resulting in 4.5V, and fed into Texas Instruments' BUF634 IC, which creates a virtual ground. The new 'ground' (the negative side of the new 9V supply) is -4.5V, the positive side is +4.5V, and the 'split' is now considered virtual ground. My original intent was to use the TLE2426 rail splitter IC, and not rely on a voltage divider, but the chip's output current capability was far too low. The negative terminal of the audio input will connect to the virtual ground, and the +-4.5V will be connected to the rail supplies of the op amps. As seen in the schematic, many decoupling capacitors will be utilized across the power supply path to ensure a smooth supply, as the USB supply from a computer is known to be rather noisy.
The left and right channels enter the circuit and immediately pass through noise-suppressing ferrite beads. The signals then go through a dual-gang 10 kΩ potentiometer, used to adjust the input signal amplitude to adjust volume. After passing through DC-blocking capacitors, the signal enters the amplifier. The circuit uses a non-inverting amplifier configuration with a gain of 11.
Additionally, the circuit utilizes a pair of TVS diodes. One connects the audio output nodes to the virtual ground, and one connects the audio input nodes to virtual ground. This is for ESD protection.
The LM4562 has an output current capability of +- 26 mA. While this may be enough for certain applications, my goal is to have this amplifier be able to drive high impedance headphones, where that current may not be enough. The amplified output of the op amps will be fed, finally, into TI's LME49600 chip, a high performance, high fidelity, and most importantly, high current headphone buffer. It maintains a similar THD figure to the LM4562, but with an output current capability of 250 mA, while maintaining unity gain.
The current layout is shown below. The back of the board has a copper pour (not shown, for clarity), connected to the virtual ground. Many components simply have a via right next to them, connected to it.
DIGITAL TO ANALOG CONVERSION
An enticing future direction for this project is to incorporate an onboard DAC. While some may find it cumbersome to have a board that requires a power input and a signal input, utilizing an onboard DAC reduces the external connections to a single USB connection, which would handle power, as discussed above, and data.
A chip being considered for this is TI's PCM1789, a 24-bit delta-sigma DAC. This chip comes in an easily hand-solderable package (24 pin TSSOP), and, as stated above, can handle a 24 bit, high quality digital signal. However, using this chip would require a separate USB driver chip, as this DAC uses a serial communication protocol. For this reason, another chip being considered is TI's PCM2704, a 16 bit DAC that features on onboard USB interface. The D+ and D- USB lines would simply need to be connected in order to output a stereo analog signal to use as the headphone amplifier input. This chip also comes in a TSSOP package (28 pin). The downside to this chip is that it can only handle 16 bits instead of 24. While this difference is most likely negligible to the ears of most humans, the goal of this project is high performance. For this reason, more research needs to be done into other possibilities, such as utilizing a USB interfacing IC with the PCM2704, or searching more deeply to find a USB compatible 24 bit DAC IC. I have not worked very much with digital and analog interfaces simultaneously.