A New Breed of Supertuner from Carver
Carver TX-11 FM tuner.
Dimensions: 17 1/4 by 3 3/4 Inches (front panel), 12 3/4 inches deep plus clearance for connections.
Price: $550. Warranty: "limited," three years parts and labor.
Manufacturer: Carver Corporation, 14304 NE 193rd Place, Woodinville, Wash. 98072.
Except where noted otherwise, all data are for the WIDE IF-bandwidth setting with the Carver NOISE and MULTIPATH REDUCTION circuitry off.
A WHILE BACK, a friend commented that Bob Carver seemed to him more an inventor than simply a design engineer. And we agreed: A Carver product that doesn't embody some spectacularly innovative feature or technical breakthrough is a rare bird.
So in a way, his first tuner is no surprise, even though it's like no other we've ever tested.
What distinguishes the TX-11 is its ability to pull clean, noise-free sound out of weak or multipath-ridden signals that would send you lunging for the mono switch on any other tuner we know of. The Carver doesn't even have a mono switch.
Instead, there are two rectangular push pads, labeled NOISE REDUCTION and MULTIPATH REDUCTION, which when pressed activate an ingenious and very sophisticated audio signal processing circuit that improves the TX-11's effective stereo sensitivity to beyond the generally accepted theoretical limit.
The key to this seemingly magical accomplishment is a fresh look at the process of retrieving independent left- and right-channel audio signals from stereo FM broadcasts, in which they are matrixed to form mono sum (L + R) and stereo difference (L-R) signals. Normally, these are added together in a tuner's stereo demultiplexer circuits to produce the desired left and right-channel outputs.
The problem with this is that only the mono part of a "stereo FM" transmission is actually frequency modulated. The difference signal is amplitude modulated on a subcarrier and is therefore far more susceptible to interference. This makes the requirements for good stereo reception considerably more stringent than those for comparable mono reception. Not only must the received signal be stronger, but it must also be substantially free of interfering reflections (or multipath) from nearby hills, buildings, and so forth. For many people especially those living in cities or rural areas-this means that few, if any, receivable stations are listenable in stereo.
One such unlucky soul was Bob Carver. Thinking about the problem, it occurred to him that the mono and difference signals are ordinarily far more similar than different. The information responsible for stereo localization makes up only a small fraction of the total L-R signal; what remains is ambience information, which is essentially a random-phase version of the L+R component with a slightly different overall spectral balance.
This insight led Carver to develop what might be thought of as an alternative stereo demodulation system (see block diagram). The incoming FM signal is tuned, detected, and de-multiplexed in the usual fashion, yielding left- and right-channel audio signals (L and R). But instead of going directly to the tuner output terminals, the audio signals first pass through a matrix circuit, which adds and subtracts them to reconstitute the original L +R and L-R signals. (The separate L and R signals are pre served, however, for use when the noise and multipath reduction switches are released or when the tuner decides they are clean enough that there would be no audible benefit from further processing.) The relatively noisy and distorted difference signal goes to what Carver calls a leading-edge detector, which separates out the small amount of transient information necessary for proper stereo localization.
How much of the L-R survives this process is determined by the switch setting—NOISE REDUCTION, MULTIPATH REDUCTION, or both--with the first retaining the most and the last keeping the least. In any event, most of the L-R is discarded, along with much distortion and interference.
Meanwhile, the comparatively clean L+R is passed through a phase-scrambling circuit that randomizes it to simulate the ambience information that the leading-edge detector has thrown away. The resulting ambience signal is mixed (according to the instantaneous amplitude of the L-R) with the localization signal from the leading-edge detector to form a composite quasi-
difference signal, which is then added to the original L+R signal, creating clean, quiet left- and right-channel signals that can be substituted for the original low-quality L and R.
Purists might cavil that this is not "true" stereo, and in a certain pedantic way, they would be right. Although separation is quite good with any signal confined to a single channel (because of yet another special circuit) and on transients, it is considerably diminished with typical steady state inputs. And the ambience signal, though statistically very similar to the real thing, is not identical. But given that stereo is just an illusion anyway, we're inclined to view such objections as spurious.
Direct A/B comparisons between the Carver tuner's regular and processed stereo modes as well as between the Carver tuner in processing mode and another high-quality tuner in regular stereo revealed only marginal differences other than a marked disparity in noise and distortion on weak signals. Anyone who didn't know some thing special was going on in the Carver would assume that it was operating correctly in normal stereo and that the other tuner was either substandard or defective.
In any case, the bottom line is that if the TX-11 gets a strong, clean signal, its special circuits automatically stay out of the way; if it doesn't, they will make it perform almost as well in stereo as most tuners would in mono, and therefore sound much better. You can't lose.
===========
Stereo sensitivity (for 50-dB noise suppression) 37 1/2 del at 98 MHz, with 0.41% THD+N (38 1/4 dBf at 90 MHz; 37 1/2 dBf at 106 MHz)
Stereo sensitivity with Carver circuitry on (for 50-dB noise suppression)*
NOISE REDUCTION
24 1/4 del at 98 MHz
MULTIPATH REDUCTION
23 1/4 del at 98 MHz both 22 del at 98 MHz Mono sensitivity (for 50-dB noise suppression) 12 del at 98 MHz Muting threshold 27 del Stereo threshold 5 del Stereo SMI ratio (at 65 del) normal 70 1/2 dB
NOISE REDUCTION
On 70 1/2 dB MULTIPATH REDUCTION on 69 1/2 dB
NOISE and MULTIPATH REDUCTION on 67 1/2 dB Mono S/N ratio (at 65 dBf) 77 dB
CAPTURE RATIO
SELECTIVITY (alternate channel)
HARMONIC DISTORTION (THD+N) Carver circuitry off stereo at 100 Hz 0.21% at 1 kHz 0.12% at 6 kHz 0.18% Carver circuitry on (stereo) NOISE at 100 Hz 0.36% at 1 kHz 0.36% at 6 kHz 0.22% both at 100 Hz 0.25% at 1 kHz 0.21% at 6 kHz 0.24% 1.1 dB 443/4 dB mono
0.18%
0.053%
0.12%
MULTIPATH
0.30%
0.30%
0.23%
STEREO PILOT INTERMODULATION
normal
0.13%
NOISE REDUCTION on 0.06%
MULTIPATH REDUCTION OR 0.05%
NOISE and MULTIPATH REDUCTION on 0.03%
IM DISTORTION (mono)
AM SUPPRESSION
PILOT (19 kHz) SUPPRESSION
SUBCARRIER (38 kHz) SUPPR.
0.043%
61 3/4 dB
71 dB*
109 3/4 dB
============
We hasten to add that the TX-11 is no slouch in normal stereo, either. Diversified Science Laboratories' measurements document very good to outstanding performance in every respect. There are, however, a few minor oddities in the data, stemming from the fact that in a couple of respects our test sample is no longer representative of cur rent production. Carver has increased the bandwidth of the narrow IF filters, so we have not included any of the data accumulated with that setting. He has also incorporated a modification developed in the course of DSL's testing that should improve the already superb 19-kHz pilot rejection by about 20 dB. Neither of these changes is of great significance.
Switching in the Carver circuits makes major differences only in the stereo quieting curve, which becomes dramatically steeper, and in the measured (not audible) separation, which deteriorates somewhat. As the steeper quieting slope would imply, sensitivity is quite a bit enhanced-by about 13 to 16 dB, depending on whether NOISE REDUCTION, MULTIPATH REDUCTION, or both are selected. Two of the three set tings degrade the ultimate signal-to-noise ratio very slightly, but not enough to be of any audible consequence.
What does matter is that switching in the Carver circuitry enables us to get clear stereo reception with a short piece of zip cord on stations that defy conventional tuners connected to a standard dipole antenna.
Carver says that NOISE REDUCTION optimizes the leading-edge detector's response for weak signals and provides the least improvement in sensitivity, while MULTIPATH REDUCTION optimizes it for squelching multipath interference and gives a somewhat greater sensitivity gain. The two together are supposed to help with really severe reception problems and to provide the highest sensitivity. In practice, we find all of the possible settings effective, but usually less different from one another than we expected. In most cases where we have felt need of the processing, we have preferred to keep both buttons pushed in.
This is also a very attractive and easy to operate unit. As befits its exceptional signal-grabbing capacity, there are sixteen station presets-the most we've ever seen each with its own LED indicator. (Regular tuning-station search or manual step ping-is in 100-kHz steps and is reason ably fast.) And there is a six-element LED signal-strength meter covering a well-chosen range. The muting threshold is set so that you will get adequate quieting on any stereo station received with the Carver circuitry on. Or you can defeat it, for reception of very weak signals, by going from auto to manual tuning. The stereo threshold seems a little too low, however, to assure adequate quieting on some very weak stations, for which mono reception would be more appropriate. The only other things we can find to complain about are the front-panel legends and indicator lights, which are sometimes difficult to read (especially from a distance), and the use of a binding post and collar for 75-ohm coaxial antenna connections, instead of a more convenient F connector. Carver's user manual is excellent, with especially good coverage of antennas. And bolt-on ears are available for those who prefer rack-mounting.
On the test bench and in the home, the TX-11 is an extraordinary tuner. For people who live in good reception areas, it is, for all practical purposes, as good as any other top-grade tuner we've looked at (and less expensive than many of them); for those who don't, it is by a wide margin the best tuner we have tested to date.
------
New Equipment Reports
Preparation supervised by Michael Riggs, Peter Dobbin, Robert Long, and Edward J. Foster.
Laboratory data (unless otherwise noted) supplied by Diversified Science Laboratories.
------
(High Fidelity, Jan. 1983)
Also see:
Carver TX-11 FM Stereo Tuner (AUDIO magazine, Dec. 1982)
Teac V-95RX bidirectional cassette deck (Equip. Report, Jan 1983)