If you’re a genuine cassette deck enthusiast, you’ll grow to love Harman/Kardon—if not at first sight, then at first or second listen. A warm recommendation for this deck, provided it is fully functional.
HARMAN/KARDON TD392: Modesty is a virtue?
Submitted on: 30 Dec 25
Category: Analog recorders/players
Website Rating:
Author's Description:
A few days ago, my wonderful Hi-Fi friend Boža sent me a message and asked whether I would like to receive two cassette decks as a gift: an Onkyo TA-2066 and a partially working Harman/Kardon TD392.
THANK YOU, BOŽA!!!
You can imagine what my enthusiastic reply was, although I knew there was a risk that I might spend that night—or perhaps the following ones—as a homeless man in a park once my wife saw what I had dragged into the house, for the hundredth time. But the best wife in the world tolerated my Hi-Fi nonsense once again.
The TD392 would actually pair very nicely with its bigger brother—the silver CD491—which has been sitting packed away in my garage for years.
Boža told me that the deck played tapes fine, but that recording on one channel wasn’t good and that it went into clipping. Otherwise, it was well preserved.
And indeed it is very well preserved—almost perfect—and shows very little wear. While making test recordings, I noticed exactly what Boža had mentioned: the left channel worked fine, but the right one didn’t and could only reach a certain low level. Increasing the recording level didn’t help. The process of elimination began: I inserted a tape recorded on another deck and everything sounded fine except for poor azimuth—meaning the problem was in the recording circuit.
But let’s start from the beginning.
About the model
The TD392 was produced from 1985 to 1990 and cost around 1,500 DEM (€750), which was the price of top-tier Sony, Aiwa, or Akai decks at the time. For 200 DEM more, you could buy a Nakamichi BX-300.
The TD392 was one of only two three-head models Harman/Kardon produced; the CD491 sat at the very top of the lineup. In any case, the 392 was far from cheap. This series was later replaced by the TD4200–4800 range, which I’ve already written about and which relied more on plastic than metal—both in the chassis and in the relative simplicity of the transport.
Personally, I find the TD392 to be a beautiful deck (even though it isn’t silver). It has a kind of classic styling and a slightly ascetic, aristocratic elegance that’s rarely seen. It was made in Japan, although I don’t know who the OEM manufacturer was.
What’s interesting is the design shift when you look at the model numbering. The predecessor to the TD392 was the TD391—but the difference between them is significant.
Here are two pictures: Harman/Kardon TD391 i the latter – TD392.
- Harman TD391 is a two-head deck, while the TD392 is a three-head deck. Both sit below the flagship CD491. This means the lineup changed: two three-head decks instead of one. By the way, don’t let the “CD” and “TD” designations confuse you—“CD” originally stood for Cassette Deck, before Compact Disc players existed. Once CD players appeared, cassette decks became Tape Decks, hence the change to “TD.” The Harman CD491 is simply an older model that kept its original designation.
- Pricing was also interesting: the TD391 cost 1,600 DEM, while the TD392 cost 1,500 DEM. So the newer model was three-head—and cheaper.
- The TD391 offers both bias and sensitivity calibration, while the TD392 only has bias calibration—clearly a cost-cutting move.
- The TD391, despite being two-head, features an excellent Sankyo three-motor transport with a closed-loop dual capstan system, just like the CD491, and it uses a direct-drive capstan motor. The TD392, on the other hand, uses a single motor for the entire transport, driving a single flywheel via a rubber belt, from which power is then transferred to the rest of the mechanism.
Effectively, the TD391 is a two-head version of the CD491 with some reduced features (such as a digital counter), while the TD392 is more modest in terms of transport, although internally it resembles the CD491 quite a bit. My guess is that Harman/Kardon wanted to target another market segment with a cheaper three-head deck, while still keeping enough separation from the flagship so the TD392 wouldn’t compete directly with the CD491.
Let’s get back to TD392 description…
The construction is classic: cassette on the left, everything else on the right, with a door hiding some of the controls.
Functions are nicely grouped, although secondary ones use small buttons with not-so-readable labels. It’s good to see that the main transport controls are metal rather than plastic, which wears out more easily.
Tape type (normal/chrome/metal) is selected manually, and convenience features such as track search are almost nonexistent. However, there are features aimed at assisting recording, including a tone generator for optimal bias adjustment, hidden—along with the bias potentiometer—behind the sliding front panel.
There’s also peak weighting for the peak meters, making peak values easier to read during recording, which is quite useful. The peak meters are LED-based, 12 LEDs per channel, and unfortunately lack peak-hold functionality.
The deck uses an early implementation of the HX Pro system, which was still relatively new at the time.
Interestingly, the service manual states that the TD392 uses a direct-drive capstan motor, although in reality it uses a standard DC motor driving the flywheel via a belt. This is likely a documentation error carried over from the CD491 manual. I noticed a few other mistakes, but nothing critical.
Despite its overall ascetic design, the TD392 has no illumination at all except for the power LED above the power button. There’s no lamp lighting the cassette well, making it hard to see anything in low light. The same applies to the status LEDs and peak meter section—it’s so dark that I actually checked the service manual to see if there was supposed to be a bulb that had failed. At the very least, Harman could have taken inspiration from the Dual C844: it cost less, had a far more robust transport, and a beautifully lit VU meter section. This kind of solution costs little but looks and works great.
INSIDE
With the cover off, the Harman looks very tidy for its production year—no “snake nests” of wires, which is a pleasure to see. Removing the bottom cover gives access to nearly all components, which is wonderful compared to newer decks where this option was eliminated for cost reasons, making servicing far more difficult. Access to the PCBs from below helped me a lot during the repair.
Basically, the deck is divided into two sections: the transformer on the left, and everything else—starting roughly from the transport midpoint—on the right, including the rectifier, voltage regulation, and transport control logic. Everything is neat, with only a few necessary wires connecting the boards, transport, and erase/record/playback heads.
It’s a pity the TD392 didn’t inherit its predecessor’s transport—that would have made it a real winner.
That’s not to say the TD392 transport is bad. The flywheel is reasonably sized, transferring power via a belt to a plate that drives the system. It’s a somewhat dated transport design and reminded me of the Bang & Olufsen Beocord 9000—a phenomenal deck with a transport that is clearly good, but not as complex or refined as one might expect. I’ll write about the Beocord another time; B&O had a very clear explanation for choosing that transport design.
The TD392 transport also includes gears for fast winding and several electromagnetic actuators, so it’s not as noisy as one might assume. The actuators drive a cam that lifts the heads and performs other servo functions.
On paper, this transport has nearly twice the wow and flutter of the Sankyo mechanism used in its predecessor—though in practice, Sankyo transports often met specifications only with certain tapes, but were still very good overall.
The tape tension system is also interesting. In most single-capstan three-head decks, tension is provided by felt on a fixed shaft against which the erase head presses, with the tape running between them. In the TD392, things are reversed: the erase head is fixed, while the felt-covered shaft moves and applies pressure to the tape via a spring mechanism, maintaining proper tension. There are adjustment screws for this system, but interestingly, the service manual doesn’t explain how the adjustment should actually be performed.
Harman/Kardon also claimed these decks used a “Uniplane Transport,” which is even printed on the cassette door. What that actually means is anyone’s guess—I’ve never found an explanation in Harman catalogs or online. ChatGPT offered one, but since it hallucinated features the TD392 doesn’t have, I can’t consider it reliable.
REPAIR
This wasn’t a trivial repair, like resoldering an RCA connector.
One issue I noticed was odd behavior of the right-channel peak meter: two green LEDs in the middle would work fine until the next green LED should light up. Instead of lighting, the previous two would go dark and the next LED to illuminate would be the red 0 dB one—leaving a “gap” in the display. Lowering the signal back down would make the problematic LEDs work again.
After some time, the issue disappeared.
The repair here is both possible and impossible: direct repair on the LED/driver module isn’t practical. The LEDs themselves aren’t faulty, but the driver circuit is, and it can’t be replaced except by swapping the entire module. There is a third-party replacement available online, using SMD LEDs and a separate board with Siemens UAA180 LED driver ICs. It may not be as accurate as the original Harman implementation, but it’s functional—at a price of €80 plus shipping. Worth it? Maybe.
The second issue was the previously mentioned recording problem on the right channel.
I noticed that some electrolytic capacitors in the Dolby circuit had been replaced with Matsushita ones—someone had clearly serviced the deck before. Possibly coupling capacitors.
Tracing the signal with a test tone and oscilloscope, everything looked fine up to the heads. Testing was done in Rec/Pause mode, where the entire recording system is active except for the erase and bias oscillators.
Once recording was engaged, a level difference appeared at the head outputs: one channel was at half the RMS voltage of the other, indicating a bias-related issue. Since the deck uses a single oscillator for both channels and the erase head, I checked it first—it was fine. That pointed to the bias circuit, in this case HX Pro.
In the TD392, HX Pro is implemented with discrete components rather than the later NEC IC that became the de facto standard.
This HX Pro circuit consists, in simplified terms, of two operational amplifiers (one quad and one dual) that extract the high-frequency portion of the signal (5–18 kHz) and use it to generate a control voltage for a transistor that drives a step-up transformer and effectively acts as a variable resistor relative to the incoming bias signal. Don’t worry if this sounds complicated — it really isn’t; here’s the schematic.
Each channel has its own HX Pro circuit.
Voltage checks revealed significant differences, so I first suspected the M5218P IC and replaced it—but the problem remained.
I then examined the transformer-drive transistors and found one behaving very strangely: it went into and out of short circuit twice—something I’d never seen before—before ending up permanently shorted. A nearby protection resistor had also gone open as a result.
I found replacement parts in my spares, installed them, and the circuit came back to life—no overheating, equal levels on both channels, and correct voltages at the record heads.
It took several hours to fix, but this is a hobby—no pressure. If I couldn’t fix it, I could always throw it away, but I didn’t want to. I wanted to give it a new life.
I removed and washed the front panel, lubricated the record-level knob mechanism, and cleaned the potentiometer, which had been intermittent.
The TD392 also boasts some excellent specs: a frequency response of 20 Hz to 22 kHz within ±3 dB on all tape types—something very few decks could achieve. Measurements confirmed it reached 21–22 kHz at the −3 dB point relative to 1 kHz.
TD392 IN PRACTICE
In everyday use, the TD392 performs very well. The control layout is clear, though you need to memorize the small buttons. The Tape/Source monitor button could have been more prominent. What I really missed was illumination—for the meters, counter, and cassette well.
Rec Mute would have been better implemented as Auto Space (inserting a 5-second pause), but here it’s the older type: recording stops only while the button is held, without affecting transport operation.
SOUND
Sonically, the TD392 is excellent and very similar to the CD491—highly musical. The sound is soft yet powerful. Like most top-tier decks, it delivers strong bass that makes mid-range decks sound thin by comparison. The bass is deep and warm, almost turntable-like, and the soundstage is surprisingly wide for an older cassette deck.
High-frequency detail is excellent, HX Pro is unobtrusive, and vocals are very pleasant. It’s difficult to hear a difference between Source and Tape when monitoring. Headroom is set so that recommended maximum recording levels are +1 dB for chrome and +3 dB for normal tapes.
Newer Harman models, and newer cassette decks in general (Yamaha and others), tend to have a somewhat cleaner sound, but often one that is also harder and more soulless. The TD392 is not muddy—don’t get me wrong—it is simply different and does not push that kind of clinical cleanliness that makes it sound as if you were playing a CD rather than a cassette deck. It does not try to imitate the compact disc; it is a story of its own.
And… I like it very much. Getting used to its sonic character is easy and quick. The somewhat ascetic design of this unit does not justify its price until you hear how it records—at which point all doubts simply vanish. A truly fine machine, for serious enthusiasts.
Comparing it to the newer TD4500, which I described in the previous text, I could say—this is only my humble personal opinion—that if you want an uncompromising copy that comes very close to a 1:1 original, the TD4500 is an excellent deck. However, if you want more emotion, a sense of powerful bass, and the charm of older decks, the TD392 is the right answer. Listening to it is pure enjoyment, and the TD392 would positively surprise me every few minutes in some aspect of the sound.
In contrast to the clean, aggressive, and excellent TEAC V-7010—which reminds me of a young, energetic martial arts fighter—the Harman is like an old but experienced boxer, able to assess the situation correctly at first glance. Its sound is something I could easily live with, and one that truly grows on you.
CONCLUSION
