This is followed by a "knee" in the trace representing the unit's residual distortion level, and a rise as the circuit becomes less linear with increasing level. The apparent reduction in THD as the signal level rises toward 1V is due to the noise increasingly becoming a smaller fraction of the THD+N figure. But again, the Premier 15 needs to see a line preamplifier with a high input impedance if that superb linearity is not to be compromised.įig.4 Conrad-Johnson Premier 15, spectrum of 50Hz at 3.3V RMS into 10k ohms.įinally, fig.5 shows the preamp's THD+noise percentage plotted against its output voltage. Considering the complete absence of negative feedback other than what's inherent in the output cathode-follower stage, this is astonishingly linear behavior. Even so, the only significant distortion components present are the second harmonic (100Hz) and third harmonic (150Hz), at 0.3% and 0.03%, respectively. Fig.4, for example, shows the spectrum of the C-J's output while it drove 50Hz at a very high level (3.3V) into 100k ohms. 10mV input at 1kHz (right channel dashed).ĭistortion levels were below 0.1% into 100k ohm loads, but rose drastically into loads of less than 1k ohm or at high levels, and less so at the edges of the audio band. Some AC components can be seen in this graph-180Hz in the left channel, 120Hz in the right channel, 60Hz in both-but these are still sufficiently low not to be a problem.įig.3 Conrad-Johnson Premier 15, spectrum of noise floor ref. There is a rise in noise at low frequencies due to the gain stage following the RIAA equalization, but the noise is otherwise low in level. Fig.3 shows a wideband spectral analysis of the Premier 15's noise floor, again referenced to a 10mV/1kHz input level. These figures were for the left channel, the right channel being around 3dB quieter. Bottom traces are with subsonic filter switched into circuit.įig.2 Conrad-Johnson Premier 15, crosstalk (from top to bottom): L-R, R-L (5dB/vertical div.).ĭespite its use of tubes, the C-J's noise levels were acceptably low, at least when referred to a 10mV input signal at 1kHz: 68dB unweighted (10Hz-500kHz bandwidth), 71dB unweighted (22Hz-22kHz bandwidth), and 90dB A-weighted. Some high-frequency spuriae could be seen riding on top of the crosstalk, but this was well down in level.įig.1 Conrad-Johnson Premier 15, RIAA error into 100k ohms at 5mV/1kHz input (right channel dashed, 0.5dB/vertical div.). Channel separation (fig.2) was only fair, but still better than any phono cartridge. Otherwise, the Premier 15's response is very flat, the preamp easily meeting its RIAA specification. As well as rolling off the bass (-3dB around 16Hz), the filter drops the overall level by 0.5dB. The top two traces are with the unit's subsonic filter switched out, the bottom pair with it switched into circuit.
The Conrad-Johnson's RIAA error is shown in fig.1. The 12.5k ohms offered by Conrad-Johnson's own ART preamp is about as low as you should go, I feel. The line preamplifier with which the Premier 15 is partnered should have as high an input impedance as possible if you don't want a "lean" balance. The output impedance was slightly higher than specified, at 323 ohms at 20kHz and 328 ohms at 1kHz, with a rise to 1602 ohms at 20Hz. As used by MF, the review sample was set to 47k ohms input impedance I actually measured 41.5k ohms. The Premier 15 was noninverting, as specified, and its voltage gain at 1kHz was also to spec at 53.25dB, with a 0.04dB difference between channels-which is superb matching, considering the lack of any loop negative feedback.
0 kommentar(er)
