Introduction by KK:
With the arrival of eBay and feverish activity in the second-hand market, a lot of audiophiles find old reviews useful. Simply Stax asked if I minded them running my old reviews - flattered, I said, 'No problem!' The only provisos I would add is that you read these in the context that they're old reviews, they cover out-of-production models, the prices are no longer applicable, and my ears and taste may have changed slightly in the intervening years. Otherwise, I hope they help you find some answers.
Ken Kessler
-------------------
A few observations by editor Neil McCauley
What you are reading here is the original review as submitted by Ken Kessler some 27 years back. You probably know that the ‘editor’s word is final’ and this means that this text was possibly edited when it eventually appeared in the magazine. Those of you who are sharp-eyed and patient might feel inclined to see, if you have a copy of Hi-Fi News November 1990 the changes – if any.
You’ll notice the absence of photos/images. This is because the editor and/or a colleague decide which images to include. For me to have done this for this posting would have been presumptuous. Hence no images.
Finally, you’ll notice the absence of headings and sub-headings. It’s traditional for the editorial team to invent and insert these rather than the author. Once again, for me to have done this for this posting would have been presumptuous. Hence no sub heads. Yes, I know it makes it harder to read, but as this is a very rare opportunity for ‘civilians’ to view a review before it arrives at the destination magazine, I don’t want to tamper with it in any way. Anyway – enjoy!
Neil McCauley
-------------------
HFN NOVEMBER 1990 STAX REVIEW
D/A = DIGITAL-TO-AWESOME?
STAX has released a tube/MOSFET hybrid D/A converter tipping the
scales at £8900. Say what?
by Ken Kessler
You really ought to meet Tony Faulkner. My fellow HFN/RR contributor and tube fanatic is known around the world as a masterful knob twiddler, and his recordings are among the few commercial releases produced with an audiophile at the helm. He's a matter-of-fact type of person, in my experience not prone to exaggeration. When he speaks, I tend to listen.
Some months ago, Tony said he was looking for a new D/A converter for his studio and that he wanted something that sounded wonderful. He'd been through all of the pro-sector offerings, designed by number rather than by ear; he wanted something better. What was the best I'd heard from the tweak sector? I suggested he take a look at the STAX DAC-X1t. Although my exposure to it had been limited, I hadn't heard anything to compare with the sounds it reproduced.
The next thing I know, Tony has dipped into his own pocket for a STAX DAC-X1t. He was so convinced of its greatness that he even wrote a paean to it in the late, lamented Which CD. After wiping the sweat off my brow – I'd have hated to learn that Tony thought I was deaf – I rejoined the queue for a review sample.
As STAX only makes 10 a month for the entire planet, they're hard to nail down; Tony bought his, so I can understand why he had preference over a mere reviewer after a loan.
Now that I've fed a STAX with thirty or so 'difficult' discs in my own system, I can understand why he was prepared to invest a fat chunk of his company's capital in it. Still, I'd love to have been there when he explained to the bank manager that it really is a lot better than any one of a dozen D/As selling for under £500.
One look at the DAC-X1t and you know you're in the presence of something out of the ordinary.
It looks as chunky as it is; to the best of my knowledge, there are no other stand-alone D/A converters on the market weighing 19kg. It occupies a space measuring 480x147x425mm (WHD), not counting the gap you must leave above it to allow the pair of valves to breathe.
You'd swear it was hewn from solid. The gunmetal case, mounted on height-adjustable feet, is made from 'extremely rigid, non-ferro-magnetic aluminium' to reduce mechanical vibration. The bottom plate and fascia panel are 10mm thick. Internally, there's a thick copper chassis while the digital section is encased in a pure copper, shielded box within the unit to reduce RF interference. Bolts are visible everywhere, but not in the half-assed British fashion which suggests an inability to perform decent design engineering. With this monster, their appearance seems deliberate as per the Cello components. It's as if they're to have some psychological effect on the owner: they scream 'bomb-proof'.
Recessed into the front panel is a bank of switches and indicators, more than you'd think a D/A converter could justify. The top row consists of four LEDs, one showing de-emphasis (selected automatically according to a disc's requirements) and three for the sampling frequencies of 32kHz, 44.1kHz and 48kHz. Again, the STAX chooses the frequencies automatically.
Across the middle of the recess are eight toggles, each with an LED above to indicate operation; as these are press-press types, the use of LEDs is mandatory as there is no physical up/down positioning. You press down on the toggle, release, and it springs back to its original position. On the extreme left is the most useful switch of them all, for polarity inversion; the phase swapping occurs in the digital domain and the effect is not subtle.
The first cluster of three toggles provides a choice of three co-axial inputs; next are three more toggles for choosing among the trio of optical inputs. When you switch on, the STAX defaults to Optical Input 1. Last in the row is a mute selector which comes on 25 seconds after the power switch is pressed. The STAX automatically mutes when there is no signal present, so don't waste time fiddling with it when your CD is in, say, Optical Input 2 and you've selected the unused Co-axial Input 3. Below the row of toggles is the main power switch.
Across the middle of the back panel are three TOSlink sockets, all of which were surprisingly 'oversized' for all of the four entirely different optical cables I had to hand. Next to the optical inputs are three top-quality gold-plated phono sockets for co-axial connection. At the extreme ends of the back panel are the left and right outputs, reminding you of the near-dual-mono construction. You have a choice of unbalanced phono connection or three-pin, cannon-type balanced output.
What about the mains? you might be wondering. I'd suggest you run down to your nearest electrical supply shop if you brought this home and had only one spare mains outlet awaiting its arrival. The STAX requires three, one for the mains to the digital section's power supply and one each for the left and right analogue stages. The three mains leads connect to the underside, not the back, with STAX thankfully supplying angled IEC mains connectors which just fit the gap created by the chunky feet.
The DAC-X1t is a well-filled box, partly because it sports three power supplies and, to a lesser degree, because its actual DACs – two per channel – are larger than your basic IC. But first the power supplies.
The P/S for the digital processor consists of a small transformer, the AC line fuse, rectifiers and three 4700 microfarad capacitors, plus three 3-pin voltage regulators. Small by-pass caps are used for most of the chips in this section. The two main supplies for the separate left and right D/A converters and analogue stages are mirror-imaged and located at the front of the STAX; I've seen a lot smaller in whole amplifiers. Each consists of a large transformer with its own, aforementioned mains input, eleven filter caps and seven regulation stages; in total, the DAC-X1t contains 17 fully regulated power supply stages.
The heart of the digital section is Ultra Analog's D20400 DAC; STAX claims the DAC-X1t to be the first converter in the world to use it. Just dropping a model number like that is bound to induce yawns. Better I should say that Ultra Analog manufactures the ADC20048 128-times oversampling A/D converters which make Chesky CDs sound so wonderful. The D20400 is a dual, 20-bit DAC operating at 8-times oversampling; no Bitstream here, my friends.
The device itself consists of over 100 components including discrete devices and a custom LSI chip, all mounted on a double-sided PCB. This lot is then packaged in an encapsulated module measuring 50x75mm. The module also incorporates the universal serial interface, a precision low-noise reference and output deglitcher circuits. The DAC-X1t employs a double PLL synchronizer and high-stability jitter-free clock generation system, with a separate power supply for the second PLL clock generator; there's a patent pending on this circuit, said to ensure for the analogue signal greater immunity to jitter from the digital signal.
Apparently, this DAC costs ten times the price of yer basic OEM job, which helps to account for the elevated pricing. The use of dual DACs to provide balanced differential operation tells you that STAX's accountants probably weren't consulted until the design was completed.
The analogue circuit and output stage of the DAC-X1t is a hybrid, based around a Gold Aero-sourced 'Platinum' 12AX7 twin triode, one per channel. The valve is driven by a FET, with each half of the triode handling the opposite polarity signal for balanced operation. STAX describes the rest of the analogue circuit as 'very simple, with just a few resistors and capacitors associated with each tube'. The MOSFET-plus-triode super cathode follower output was chosen for the 'sake of isolation' from the next item in the chain, e.g. your pre-amp.
STAX believes that a vacuum tube cathode follower circuit has better 'uni-directional' characteristics compared with solid-state buffers under VHF or UHF external interference.
Other details include the use of PC-OCC copper for all wiring, right through to the transformer windings. All of the capacitors are German-made polypropylene types with non-ferrous leads, while the metal-film resistors are lead-less and hand-soldered to the PCB tracks. You get the impression that STAX actually pays attention to the tweaks living outside of Japan.
The best transport in my arsenal is that of the Marantz CD-12, which enabled me to save much time by revealing very quickly that fibre optics still have some way to go before they can beat the vintage stuff. Coincidentally, Path Premier, who import STAX into the UK, recommended MASter Link cable for the co-axial connection, which is what I'd planned to do anyway. Fibre optic cables included three types of Furukawa and the Marantz cable. References against which the STAX faced off included the Meridian 203, the MAS DCC-1 DAC/pre-amp and my cherished CAL Tempest II SE CD player.
The ink was still wet on the shipping carton, so I gave this unit a nice, long burn-in period. In normal use, it needs about a half-hour to settle down, but it still sounds dandy after the minute or so it takes to get the auto-muting to relax. More critical than the warm-up, though, are the placement relative to the rest of one's system and the choice of cables, whether you use optical or co-axial connections.
The positioning, despite construction which is supposed to make the DAC-X1t immune to external influences, is determined by ear. Move the thing about while an uncluttered disc is playing and you'll hear a slight improvement in background silences and texture. The minimum distance from your transport and pre-amp is around six inches. Sensitivity to cables is such that the less adventurous among you may just wish to take my word for it by going straight to the MASterlink.
First, I found that even the best optical cables -- and some are very good -- made the music sound somewhat dry and slightly compressed. As the dynamic capabilities of this unit are among its most cherished traits, you don't want to compromise them; to do so would be to eliminate much of the justification for the eye-popping price tag. But just as noticeable was a graininess, a texturing of the sound which was not present when listening in the co-axial mode. The latter is especially evident on mellow instruments and female vocals of the C&W warbler variety. I'd even suggest that STAX supplies a copy of The Judds' Greatest Hits with each DAC-X1t as a cable selection test disc.
Regardless of the cable selection, what couldn't be masked was a warmth and smoothness matched only by the CAL. Using the Joemy Wilson 'hammered dulcimer' CDs, the acoustic feast of Mastercraftsmen on Marco Polo or the a cappella of the Persuasions, there was no doubt that the lesser D/A converters stripped the music of whatever it is that explains the difference between 'hi-fi' and 'lifelike'. What seemed to play the biggest role was the nature of the decay of either the echo of the recording venue, or the decay of the actual notes.
With the STAX and the CAL, the notes trailed off gradually and smoothly, like Sinatra throwing a trenchcoat over his shoulder and drifting offstage. Depending on the nature of the decay, the other converters either truncated it prematurely or they failed to maintain a constant rate of decay from the peak to the eventual silence.
The contribution this makes shouldn't be underestimated because it quite emphatically affects the way the listener perceives the space of the original event, which should be reproduced with accurate dimensions and scale. There's no other way you can achieve the miracle of 'reproducing a musical event in the home'. I'm more willing to accept a slight reduction in scale provided that the spatial relationships are maintained rather than a soundstage which seems life-size but which lacks any semblance of '3-D'. The Stax matched the CAL for vivid three-dimensionality and the ratios of stage width to stage depth were spot-on, but the CAL did approach 'lifesized' with greater accuracy.
Where the STAX left the CAL is in its balancing of tube virtues and computer era accuracy. Why the CAL is no slouch for detail and the sweet top end is not achieved by chopping it off, the STAX manages to mirror these virtues but with far greater transparency, clarity, control, transient attack and slam. Amusingly, though, the CAL seemed to edge out the STAX in both extension and 'realism'. The STAX was tighter, better damped, but the CAL had the swing. But that's it. In every other area, the STAX is, to me, the best CD playback machine I've ever experienced. I'm almost tempted to say that it's closed the gap between CD and LP.
An early reaction was 'Nine grand plus a transport is a hard way to simulate a CAL Tempest', and if it's just a tube-y sound you're after, you can do all that by sticking £249's worth of Croft pre-amp in the chain. But the STAX isn't what I would call a 'valve-sounding' device because it's too precise in too many areas. The warmth and musicality mustn't be confused with the deception of 'romance', and no solid-state supporter will have any reason to suspect the presence of glass. It's simply a perfect blend of a transistor's virtues with the dynamic sweeps and composed behaviour-under-duress of valves.
In no way is the STAX a bargain, even after you factor in the build quality, the rarity, whatever. But nine big ones is what it now costs for what may be the best D/A in the world. I only hope that most of the 120 purchasers of each annual batch are members of the recording industry, like Tony, or -– better still – STAX's competitors.
They'll learn just how far they have to go.
And I'd like to think that the sound of the STAX DAC-X1t will be available to all of us for around three figures within a couple of years. Until then, all you can do is dig deep into your pockets for this quite amazing proof of CD's potential.
Ken Kessler.
