Akai AD-93
A second-generation DAT deck that arrived in America before DAT itself did — and measured like it.
Overview
The Akai AD-93 occupies a strange and slightly tragic corner of digital audio history: a genuinely excellent DAT recorder that, for a while, you simply could not buy from an authorized dealer in the United States. As Audio (June 1988, p.71) noted at the time, DAT machines were available in Japan and much of Europe, but not through official U.S. channels — a state of affairs the magazine attributed to record-company resistance rather than any technical shortcoming. So the AD-93 arrives in the archive as a machine defined by its engineering first and its market fate second.
That engineering was serious. Audio's Leonard Feldman (June 1988, p.68) profiled the deck and came away regarding it as a "second-generation" DAT recorder — a unit whose designers had clearly studied the first wave of machines and set about improving them. The convenience features of the R-DAT standard, drafted by 84 companies more than three years before the review, are all present here; what Akai added was a transport and circuit topology built to exploit the format properly.
The headline numbers are the kind that made analog cassette engineers wince. Akai rated frequency response at 2 Hz to 22 kHz, ±0.5 dB, with a signal-to-noise ratio of 92 dB and THD of 0.005% or less at 1 kHz and 0 dB. Dynamic range was quoted at 90 dB in record and 96 dB in playback, and wow-and-flutter was specified as below measurable limits — which, on a tape format, is the whole point. Feldman's bench work largely backed this up. His measured THD + N came in around 0.008% at 1 kHz, and he was careful to explain the apparent gap: he was measuring distortion plus noise, whereas Akai quoted distortion alone. The 87-dB A-weighted S/N he measured corresponds to roughly 0.005% on its own, so the combined figure was entirely consistent with the spec. Using the EIA method for dynamic range, he obtained 94 dB across the complete record/play cycle.
Where DAT genuinely separates itself from analog is phase behavior, and here the AD-93 turned in a striking result. Feldman measured interchannel phase error of just −0.89° at 20 kHz through the entire record/play process — a figure he explicitly contrasted with what any accurately aligned cassette deck could manage. He also demonstrated that a 0-dB swept recording stayed flat to within a fraction of a dB at 20 kHz, down only 0.23 dB, and made the point that DAT has no high-frequency tape saturation to worry about at all: zeroes and ones don't care whether their patterns alternate twenty times a second or twenty thousand. That said, the format punishes carelessness in a new way. Push recording level past the 0-dB mark and the distortion curve rises almost vertically — at +2.0 dB it had already exceeded 2%. Engineers raised on analog meters that routinely kissed +3 will need to relearn their instincts.
The build reflects the price and the ambition. The capstan and head-drum motors are both high-torque direct-drive types, with three further DC motors handling reel drive, tape loading, and the cassette tray. The head in the spinning drum is a spatter Sendust type, pairing Sendust's magnetic properties with ferrite's wear resistance, and the drum itself uses a silicon-enhanced aluminum alloy intended to reduce tape and head wear. The chassis carries a 1.6-mm upper panel and triple-layer side panels, sitting on large pedestal feet for vibration control. Six power supplies are on board, with the analog rails kept completely separate from the digital ones, and the audio amplifier circuitry is split into four sections — record, playback, left, and right. On the conversion side, the AD-93 uses two-times oversampling digital filtering with independent A/D converters for recording, and four-times oversampling with independent left- and right-channel D/A converters for playback.
Around the front panel, the deck is well provisioned. A smooth-gliding cassette drawer sits beside a large fluorescent display that can show elapsed time, absolute time, or the time of a numbered program, along with a digital peak-reading level meter calibrated from 0 down to −42 dB and a peak-hold function that retains indications above −21 dB. Overrecording is flagged by a bright red indicator at the meter's right end. A "Tape View" button illuminates the cassette area during playback or recording so you can see how much tape remains — handy, though Feldman thought it no substitute for the accurate remaining-time indicators found on some rival DAT machines. Subcode handling is fully automatic: start ID, program number, and absolute time are written as you record, and start IDs can be erased or added afterward without touching the audio. Up to 60 selections can be programmed from a maximum of 39 numbered selections on a tape, and the supplied remote handles direct program access. Cue and review run at 16 times normal speed, search tops out at 200 times normal, and a full 120-minute tape winds end to end in 50 seconds. The "End Search" function jumps to the end of recorded material on a partly blank tape — a genuinely useful trick given the two-hour capacity. An external timer can trigger unattended recording or playback, and the deck will display one of nine error messages for conditions like moisture condensation, broken tape, or transport faults.
The ergonomic complaints are minor but real. Feldman's chief criticism was that program search has no pause option: call up a program number and the transport zips to the start and begins playing immediately, whether you wanted that or not. He'd have preferred the choice to stop and wait for the Play button. His listening notes were more straightforwardly positive — material reproduced with the noise-free, distortion-free character he associated with CD sound, and CD tracks transcribed through the analog inputs came back barely distinguishable from the originals. On one or two occasions he sensed a little dynamic range slipping away, which he attributed to his own caution in setting levels rather than the deck.
Physically it's a substantial piece: 18.1 inches wide, 4.6 inches high, and 15 inches deep (46 × 11.7 × 38.1 cm), weighing 24.3 lbs (10.9 kg). The sample Audio tested was a 100-V, 60-Hz Japanese-market version, while the published spec calls for 120 V, 60 Hz — a detail worth checking before plugging one in. Analog inputs take 316 mV for 0 dB into 47 kilohms; analog outputs deliver 2 V at 0 dB into 100 ohms; the headphone jack gives 30 mW into 32 ohms. Digital I/O runs at 0.25 V peak-to-peak in and 0.5 V peak-to-peak out, both into 75 ohms. Tape speed is 8.15 mm/s. No price was available at the time of review, which fits a machine that wasn't officially on sale in the States.
For collectors, the AD-93 is a curious proposition. It's a well-built, over-engineered DAT deck from the format's awkward adolescence, documented by one of the era's most rigorous testers, and burdened by a media format that never found its footing. The mechanical complexity — five motors, six supplies, a spinning head drum — is exactly the kind of thing that ages poorly and rewards careful servicing. Anyone chasing one should verify the voltage variant, confirm the transport and head drum are healthy, and accept that blank tape and prerecorded software are their own scavenger hunt. What they'll get in return is a machine that, on the bench, did nearly everything its spec sheet promised.
eBay Listings
As an eBay Partner, we earn from qualifying purchases. This helps support our independent vintage technology research.