Class A MOSFET headphone amplifier

Class A MOSFET headphone amplifier

I wasn't entirely satisfied with how my computer's sound card drove my 32-ohm headphones, so I decided to build my own Class A MOSFET headphone amplifier. As with most of my projects, the goal was to keep it simple, reduce costs, and try using some recycled parts. This is a simple DIY headphone amplifier project, primarily based on Greg Szekeres' Class A MOSFET headphone driver project, with some reference to Mark's DIY Class A 2SK1058 MOSFET amplifier project. The amplifier concept is simple, following a typical single-ended Class A circuit, using an active constant current source (CCS) instead of a passive resistor. The CCS doubles the circuit's efficiency compared to using a passive load resistor, reaching up to 25%.

Several points need to be noted. FET follower circuits can provide high current, but the voltage gain will be less than one. This amplifier is only suitable for applications where the input signal does not require voltage amplification (e.g., MP3 players or computer outputs). Furthermore, simple single-ended circuits like this lack power supply ripple suppression, so any noise from the power supply will pass directly through the amplifier. Therefore, you will need a regulated power supply. Suitable, inexpensive regulated (wall-mounted) power supplies can be purchased at Radio Shack. 10-20VDC and 750mA should suffice.

DIY Headphone Amplifier Construction

This headphone amplifier will primarily reside on my work desk, so it needed to fit the office environment. Fortunately, I had a spare, broken Plextor external CD-ROM that served as the perfect enclosure and matched my desk perfectly. Even better, it already had a power switch, power adapter socket, and RCA input on the back, and a headphone jack on the front. Perfect! The opening visible on the back is where the USB port is located, but I previously recycled that for another project.

Plextor external CD-ROM enclosure

The amplifier is built on a roughly 1.75-inch square prototype board (276-148) from Radio Shack, but any board will work. I only used parts I had on hand, and you can see I didn't use any high-end components. There are standard (but matched) metal film resistors, a 1uF polyester film input capacitor, and a 0.47uF polypropylene bypass capacitor at the output. The 0.1uF decoupling capacitor is also polypropylene. Some people might prefer to use higher quality input and bypass capacitors, which should improve the sound. You can use carbon resistors, but I recommend metal film resistors, especially for CCS, as they have better temperature stability than carbon resistors.

DIY headphone amplifier on prototype board

The heatsinks were salvaged from various broken components. The smaller heatsinks, about 1.75 inches square, only became slightly warm, but remember that they are attached to a metal chassis, which also helps dissipate some heat. Make sure to isolate the MOSFETs and regulators from the heatsinks.

Construction of headphone amplifier

The headphone amplifier was first tested using a regulated power supply at very low voltages (smoke test). The bias was set by adjusting a 100k variable resistor until the output side (source) of the MOSFET was at half the supply voltage (drain). You'll need to check and reset the bias several times in the first few hours, as it will drift once everything stabilizes. The amplifier works well between 10 and 20VDC, but seems to work best at 13V and above. There is no audible hum when using a regulated power supply. This is not the case when using an unregulated power supply.

Set headphone amplifier bias

Next, I had the opportunity to try out my new USB oscilloscope. It's a DSO-2150, a dual-trace oscilloscope with a bandwidth of 60MHz and a maximum sampling rate of 150MS/s. For those interested in this type of oscilloscope, here's some more information about my experience using the DSO-2150 USB PC Based Oscilloscope. I checked the sine wave response, and as expected, the results were good in the 20Hz to 20kHz range (the limitation of my function generator). Below are two screenshots of the square wave response at 100Hz and 4800Hz.

The top track (green) is the input waveform, and the bottom track (yellow) is the output. My signal generator isn't very good, which is reflected in the quality of the input waveform. If you compare the input and output voltages, you'll see that the circuit's gain is around 0.8. As you can see in the 100Hz track, the square wave response is slightly skewed but stable. The skew gradually decreases as the frequency increases, and above about 300Hz, the square wave response is excellent up to 20kHz, which is the limitation of my signal generator. Since music is primarily composed of sine waves, this isn't a problem because the sine wave response is good within the audible range.

The final touches involved gluing the CD-ROM front panel to the aluminum plate with epoxy resin and reassembling the casing. Since the volume will be controlled by an MP3 player or computer, there is no potentiometer on the amplifier. The original volume control knob of the CD-ROM was shortened and glued in place.

Completed Class A MOSFET headphone amplifier

For a simple single-ended amplifier design, the sound is quite good to my ears. This amplifier easily drives my Grado SR80 headphones, while my portable MP3 player cannot. I even prefer this sound compared to the integrated headphone amplifier built into my NAD C162 preamplifier.