distortion due to Miller non linearity
I wonder has anyone done any research or are familiar with this concept. I recently designed and built a low THD headphone amplifier but the only way I could get THD to 0.001% across the audio band was to engage in time consuming analysis of data sheets and mass testing of devices for lowest and most linear junction capacitance. I eventually settled on Motorola bc549 / 559 for IS, VAS and current sources. The motorola devices achieved far better subjective and measured results than the other branded equivalents. The amplifier is basically a D. Self generic class A design with a few variations. I discovered that the IS current mirror was largely to blame for the Miller contributions mainly due to the low voltage across CB. I've ended up using Motorola MPSA18s in the current mirror because they offer the lowest Miller and also have the advantage of high beta. Not surprisingly they also achieved the best subjective result.
BTW, the output stage is complementary CFP using Toshiba 2SC5171 / 2SA1930 at 100mA qiescent. These really are fantastic devices.
Actually, I'm not really sure how low the THD is now because It's at or below the limits of my 24 bit analyser. By comparison my Musical Fidelity A32cr preamp looks very ordinary but in fairness the distortion is mainly 2nd order. Out of curiosity I compared the headphone amp (as a preamp) with the MF into the hi-fi. My CD source is a Marantz SA-11s1 which also plays SACD. After a week of listening I was finding myself more and more tending to listening to the home brew design due to slightly less grain and more hf detail. The difference was even more apparent on SACD.
The results are encouraging. I wasn't expecting to build something that outperformed a reasonable commercial product.
|All times are GMT. The time now is 10:52 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright ©1999-2016 diyAudio