Final Laboratory Music-4 phono preamplifier, Music-5 line preamplifier, & Music-6 power amplifier Measurements

Sidebar 3: Measurements

Each of the three Final Laboratory components was powered from its battery power supply for the measurements, the single-box DC-5 for the preamps, the two-box DC-6 for the power amp; when appropriate, I list the battery voltage.

The Music-4 phono preamplifier offered a maximum voltage gain of 49.6dB, making it suitable for use with moving-magnet cartridges or high-output moving-coils. (The battery voltage was indicated on its meter as "±18V.") The '4's input impedance was around 500k ohms. This is far higher than normal for a phono preamp and as Art Dudley noted, might lead to excess top-octave energy with low-output MCs, though it's fair to note that such cartridges will never be used with the Music-4. The circuitry didn't invert absolute polarity. (The switch marked "±" on the top panel selects the power-supply voltage polarity for the meter, not absolute phase, as might be thought.)

The Music-4's output impedance at 1kHz was 106 ohms. It was even lower at 20kHz, at less than 2 ohms, but at 20Hz the impedance rose to a whopping 17.6k ohms. As a result, the Music-4's frequency response will be very dependent on the input impedance of the line preamplifier. Fig.1 shows that the low frequencies roll off below 1kHz (!) with a low 1000 ohm load, and that, even with the 10k ohms typical of many solid-state preamps, the bass is down 3dB at 28Hz, which will give the device a rather lean tonal balance. The bass actually tips up slightly into 100k ohms, which suggests that it will be flat into an impedance approximately that of the Music-5 line preamp.

Fig.1 Final Music-4, RIAA error into (from top to bottom): 100k, 10k, 1k ohms (1dB/vertical div.).

Consistent in the fig.1 traces is a shallow but broad depression in the mid-treble. This might make the sound a little recessed, but in some systems it might also be perceived as improved image depth. Note the rising ultrasonic response. This is generally due to a series-feedback RIAA topology, which is best for ultimate signal/noise ratio, but levels off at unity gain above the audioband rather than following the RIAA de-emphasis curve down to zero. Again, the Music-4 should not be used with low-output MCs that might put out fairly high levels of ultrasonic spuriae on things like record ticks.

The Music-4's wideband, unweighted S/N ratio was respectable at 65dB (ref. 5mV input at 1kHz), this improving to 75.4dB(A) when A-weighted. Channel separation (not shown) was good at better than 80dB across the audioband. With the gain set to its maximum, the Music-4's overload margins were okay, at 16.9dB at 1kHz, equivalent to an output level of 11.4V. The margins improved slightly, to 18.3dB, at the band edges, and could also be increased by reducing the gain with the trim controls.

Overall, the Music-4 phono preamp's measured performance doesn't appear to offer anything special. Its very high source impedance at low frequencies and the dependence of its bass response on the loading make it a poor choice for use with line preamps other than Final's own Music-5.

The Music-5's voltmeter, too, indicated "±18V" for the measurements. The input impedance at 1kHz measured slightly higher than specified, at 61k ohms. However, the difference is inconsequential. With the gain-trim and volume controls set to their maximums, its voltage gain was 20.6dB. The volume control's unity-gain setting was 8:00. The trim controls could reduce the overall gain dramatically, giving, at their lowest settings, an attenuation of 49.3dB.

With these gain-trim controls set to their maximums, the Music-5's frequency response was flat across the band (fig.2), regardless of load impedance. The output impedance in this condition was extremely low, at below 1 ohm at all frequencies. However, as the trim controls were rotated to give less gain, the Music-5's source impedance increased dramatically. With the controls set to 12:00, the impedance reached an extraordinarily high 67k ohms in the midrange, and was still 10k ohms at 20kHz. As a result, there was a large degree of interaction between the Music-5's frequency response and the input impedance of the power amplifier. This can be seen in fig.3, which shows that a flat response is obtained only with a 100k ohm load. With a load of 1k ohms, the response is down 10dB at 10kHz—resulting in very audible muffling of the unit's highs.

Fig.2 Final Music-5, frequency response into 100k ohms (top) and 600 ohms (bottom), with volume control and gain trim at maximum (0.5dB/vertical div., right channel dashed).

Fig.3 Final Music-5, volume control at maximum, gain trim control at 12:00, frequency response into (from top to bottom): 1k, 10k, 100k ohms (5dB/vertical div.).

Noise levels were low and the channel separation (fig.4) was basically good, though afflicted with the usual capacitive coupling, which increases the level of crosstalk with rising frequency. With the trim controls set to give maximum gain, the Music-5 could deliver more than 12V into loads from 10k to 100k ohms (fig.5), and still more than 10V into a load of 1000 ohms (fig.5). But with the trim control set to 12:00 (bottom trace in fig.5), the maximum output voltage is drastically curtailed, to 1.6V, though the noise and distortion levels below clipping are also dramatically reduced.

Fig.4 Final Music-5, channel separation with volume control and gain trim at maximum (R-L dashed, 10dB/vertical div.).

Fig.5 Final Music-5, distortion (%) vs output voltage into (from right to left at 3%): 100k, 10k, 1k ohms with gain trim at maximum; 100k ohms with gain trim at 12:00.

This interaction between the trim-control setting and the measured distortion can also be seen in fig.6, which plots THD+noise against frequency under various conditions. The top pair of traces were taken at 7V into 600 ohms, the middle pair at the same voltage but with the load increased to 100k ohms. The bottom pair of traces were taken at 1V into 100k ohms, but with the trim controls set to 12:00. Though the measured THD is low, common to all three pairs of traces is an increase in the distortion percentage at high frequencies, which suggests limited gain-bandwidth product on the part of the circuitry.

Fig.6 Final Music-5, THD+noise (%) vs frequency at (from top to bottom): 7V into 600 ohms and 100k ohms with gain trim at maximum, and 1V into 600 ohms with gain trim at 12:00 (right channel dashed).

X