02 September 2015

David Clark Headphones amplifier design 2

The simple common emitter circuit was redesigned with an increased target collector current of 10mA. This allowed enough current to supply the 150 ohm load and ensure that the transistor was operating as a constant current source.   

Varying emitter degeneration and/or collector current(by increasing the rail voltage)  demonstrated the effect on amplifier voltage gain and dynamic range. As can be seen from the measurements, collector current increase results in amplifier dynamic range improvements (seen as reduced gain compression in the measurements). Also increasing emitter degeneration improves dynamic range but at the expense of amplifier voltage gain.

So you cannot 'have your cake and eat it'. Better dynamic range comes at the expense of increased battery consumption or reduced gain.  With a 7V supply I found the gain limit to be less than 6.5. Increasing the gain to the 10 - 13 range required a collector current of 18mA. Clearly either option was problematic for this application. 

After increasing the emitter degeneration resistor to 15 ohms I then tried the amplifier connected to my K2 and using a Vcc of 7.5 volts.  Listening to our club evening 80m net with high QRN I was very interested to note slight distortion in the headphones. The received CW note sounded undistorted to my ear at normal listening levels. However the background noise sounded slightly distorted. The measured 3dB points of the amplifier were 200Hz - 1800Hz. When the K2 audio gain was turned down to a low listening level no distortion was detected. So in spite of measuring an undistorted (as far as I could tell) 1kHz output sinewave approaching 600mV (p-p)  on my bench. In practice the noisy band sounded distorted. Could this be due to noise spikes from the QRN exceeding the dynamic range of the amplifier? Using a single signal input to simulate a real radio channel is simplistic. 

Next it would be interesting to try a 2 transistor amplifier using an emitter follower on the output. This would allow an increase in the load resistance of the common emitter input stage which would allow improved dynamic range and gain at a lower collector current. However at least 10mA would be needed to drive the emitter follower. Is this a zero sum game?

20mV input, 128mV output, Vcc=7V Gain (V) = 6.4 'undistorted'
100mV input, 500mV output, Vcc=7V. Gain (V) = 5 'distorted'
250mV input, 1.3V output, Vcc=7V. Gain(V) = 5.2 'distorted'
250mV input, 1.55 V output, Vcc = 12V. Gain(V) = 6.2 'undistorted'