Vintage

While Philips' very early players used the 14-bit TDA1540 DAC – the CD100 [HFN Oct '11], CD104 [HFN Apr '14] and CD303 [HFN Nov '13] – and latterly the 16-bit TDA1541 in the CD450 [HFN Aug '14] and CD960 [HFN Jul '20], all with an external 4x oversampling filter, the CD380 was the first to employ the 'new' TDA1543. This 'economy' monolithic DAC and op-amp analogue filter stage still delivers the standard 2V maximum output but the source impedance is a fairly high 920ohm/1kHz increasing to 1.55kohm at low (20Hz) frequencies. The A-wtd S/N ratio is a respectable better-than-16-bit 99dB while the DAC's low-level linearity is an unexpectedly fine –0.1dB/90dBFs with an error of just –0.4dB at –100dBFs. The deviation is greater at high frequencies (+4.6dB/–100dBFs/20kHz) but this is not atypical. The CD380's response is also very flat to within ±0.1dB (20Hz-20kHz) and includes ±0.03dB ripples from the linear phase digital filter, the latter offering a 40.7dB stopband rejection.

Where the TDA1543 differs most substantially from the TDA1541 is in both its distortion profile (versus level and frequency) and digital jitter. The latter is some 10x higher in the CD380 at ~1550psec [see Graph 2, below], comprising mainly odd-order data-induced sidebands. Distortion, at peak level, increases from 0.0027%/20Hz to 0.0055%/1kHz and 0.012%/20kHz but there's a sharp transition – a step reduction and reset in THD – at ~60dBFs where the TDA1543's 11-bit passive divider hands over to the 'least significant' 5-bit passive divider [see Graph 1, below, and DAC schematic, p121]. Back in 1988, and along with the jitter measurement, this was not a diagnostic test available to either reviewers or designers. PM