Cables

A first for Hi-Fi News – but testing a DC supply is not dissimilar to testing a digital server: the impact of the 'front-end' is most usefully assessed by examining the difference it makes to the performance of the kit downstream. I looked at compact DAC solutions requiring 5V to 12V DC feeds but, in short, differences were negligible to none. By way of illustration of the many that I tested, the maximum (fixed) output, A-wtd S/N, –10dBFs THD and jitter for iFi Audio's 5V/2.4A NEO iDSD [p72] remained 4.285V, 109.7dB, 0.00016% and 550psec, respectively; for Cambridge Audio's 12V/2.0A DacMagic 100 [HFN Jun '12] this was 2.310V, 108.8dB, 0.0005% and 255psec; for Mytek's 12V/1A Brooklyn Bridge [HFN Dec 19] this remained 6.06V, 112.8dB, 0.00007% and 14psec; and for Chord's 12V/0.5A Chordette Qute EX this was 3.05V, 111.4dB, 0.0008% and 8psec.

Chord, like iFi Audio, has a fine grip over its products' PSU filtering and regulation, however primitive the supplied wall-wart. So there's a difference of less than 1psec in the ±100Hz noise manifest in the Qute EX's jitter spectrum, in favour of the wall-wart on this occasion [Graph 1, below]. Real differences only start to emerge with products that have a higher current demand – specifically DAC/headphone amps or products fitted with small triode tubes – and if the wall-wart's connecting lead is thin. Here the benefit of Ferrum's 4T connection is revealed. Pro-Ject's Tube Box DS2 phono stage [HFN Jul '20] has a 18V/1.0A wall-wart, and while the 75mV/1kHz input overload margin was unchanged, the A-wtd S/N ratio did improve from 84.1dB to 85.0dB with the Hypsos in tow (MM input, 40dB gain).

In practice, the most reliable improvements were obtained with high output DC-powered headphone amplifiers. Arcam's 12V/2.0A rHead [HFN Nov '16] is indicative for while there's no difference in the maximum 1260mW/32ohm power output, the A-wtd S/N is stretched from 96.1dB to 96.7dB and distortion squeezed from 0.0015% to 0.0003% (100mW/32ohm) with the Hypsos. Into very low headphone loads there's a slight boost from 1790mW to 1810mW/8ohm and improvement in THD from 0.0013% to 0.0010% (100mW/8ohm). So there you have it – the Ferrum Hypsos gets a win [see Graph 2]! PM