Vintage

Entirely its own design, right down to the 'J91-0341-05' optical mechanism, Kenwood's flagship DP-1100SG employed a pair of Burr-Brown's PCM56P DACs, seen a year earlier in Sony's CDP-C5M HFN Feb '21], among other players. However, while the latter employed 2x oversampling, the DP-1100SG ups the ante with 4x oversampling courtesy of NPC's SM5804D. The impulse response [inset, Graph 2] shows the pre- and post-ringing associated with this cascaded 80-tap/15-tap linear phase digital filter while the frequency response, with its mild +0.04dB presence boost and –0.35dB/20kHz dip, also exhibits a markedly reduced ±0.02dB HF ripple. Combined with Kenwood's analogue filter, the stopband rejection amounts to 74dB while distortion at the peak (0dBFs) output of 2V is very low at ~0.0007% right across the 20Hz-20kHz audio range.

That 2V output and 105.7dB A-wtd S/N ratio might suggest a ~17-bit resolution but, in practice, its low-level linearity deviates quite significantly below –80dBFs where signals appear at a level some 3-4dB lower than expected. At 16-bits (–96dBFs) they are nearly 2-bits adrift at –109dB. The PCM56P DACs employed 'ultra-stable nichrome (NiCr) thin-film resistors to provide monotonicity, low distortion, and low differential linearity error' but our results suggest some drift has occurred over time. This is also clear in the distortion vs. digital level [see Graph 1] although the fact that THD is generally lower at 20kHz than 1kHz (0.007% vs. 0.027% at –30dBFs) is as much a function of the analogue filtering pinching off digital harmonics above ~22kHz. Jitter, meanwhile, is moderate at 360psec and while free of PSU sidebands is almost entirely data-induced in nature. Back in 1988, this was not something any of us could measure... PM