B.M.C. Phono MCCI phono preamplifier Measurements
Sidebar 3: Measurements
The MCCI is designed in Germany but manufactured at B.M.C.'s own factory in China. Once I got the MCCI on the bench, measuring it proved trickier than I'd expected, due to the preamplifier's "current injection" input circuit. Unlike almost all other phono stages, the MCCI takes advantage of a moving-coil cartridge's very low impedance, its inherent current-generating capabilities, and its balanced, floating-ground architecture, and arranges for its input stage to operate in current mode. It thus has a very low input impedance (specified as <3 ohms), and will therefore appear to be almost a short circuit to the Audio Precision SYS2700 test system (see www.ap.com and the January 2008 "As We See It"). While this low impedance would not be a problem with MC cartridges, if I connected the AP's balanced output to the MCCI's balanced input, high current would flow and might damage the amplifier.
I discussed the solution with designer Carlos Candeias at last April's New York Audio Show. At his suggestion, I created a balanced interconnect with a 10k ohm resistor in series with the terminating XLR jack's pins 2 and 3. (I matched the two 10k resistors to within 2 ohms.) Pin 1 of the XLR was connected to ground and to the shield in the usual manner, but I then soldered a 10 ohm resistor across pins 2 and 3. The Audio Precision's balanced output stage would thus see a high 20k ohm load, but the voltage presented to the MCCI would be typical of an MC cartridge, but fed from a high impedance to emulate a current source. The cable attenuated the signal level by just over 66dB, a 1kHz signal level of 1.018V resulting in 500µV at the MCCI's input.
On the circuit board for each of its two channels, the MCCI has internal jumpers that let the user select among three levels of gain (Low, High, Very High), standard or Neumann RIAA, a subsonic filter, and three Low End Corrections (Linear, Bass Boost, Bass Boost and Warmer Sound). These jumpers are fiddly and not very well identified. I measured the MCCI with its default settings, though I did examine the RIAA error with and without the Neumann correction.
As supplied, the MCCI offered a very high gain of 71.8dB; this can be reduced in two steps, by 6 and 12dB. It preserved absolute polarity (ie, was non-inverting) with the input and output XLRs wired with pin 2 hot. It wasn't possible to measure the input impedance. The output impedance at 1kHz was 200 ohms from the balanced XLRs, 100 ohms from the single-ended RCA jacks. The single-ended impedance was the same at the frequency extremes as it was at 1kHz; the balanced impedance was slightly lower at 20Hz, slightly higher at 20kHz, but this will not have any practical consequences.
The MCCI is a very quiet phono preamp: its unweighted wideband signal/noise ratio, measured with the input shorted and ref. 500µV at 1kHz, was 83.2dB in the left channel, 71.8dB in the right. A-weighting these ratios gave improvements to 99.9 and 103.3dB, respectively. Channel separation was equally superb, measuring >105dB at 20kHz.
The blue and red traces in Fig.1 show the RIAA error of the left and right channels, respectively. The two channels are superbly well matched in both level and equalization. There is a slight (<1dB) boost visible at low frequencies, which I assume corresponds to the Warm internal jumper setting. The response rises above 8kHz, reaching +0.7dB at 20kHz and +6.75dB at 100kHz. This is the so-called Neumann "fourth pole" modification of the RIAA curve, of which I am not a fan. Fortunately, this can be disabled with the internal jumpers, to give the superbly accurate RIAA response shown by the green and gray traces in fig.1.
In his "Manufacturer's Comment" letter (p.141), the MCCI's designer, Carlos Candeias, mentions that my measurements of the preamplifier's frequency response does not correspond to our measurements at the low end. I did check this measurements, and got the identical result. As I wrote above, I suspected that our review sample had had its internal jumpers set to apply a degree of bass boost.
With the very high gain as supplied, I wasn't surprised to find somewhat limited overload margins. Referenced to the standard MC output level of 500µV at 1kHz, the MCCI reached 1% THD at a level 12.5dB higher than that at 20Hz and 1kHz, both equivalent to an output voltage of 8.4V. However, the overload margin at 20kHz was lower, at just 3.4dB. Reducing the circuit's gain with the internal jumpers will increase the overload margin by the same amount. I recommend using the MCCI set to the lowest gain possible with any particular cartridge and system.
Not only is the MCCI a low-noise preamplifier, it also offers low distortion, its THD+noise typically measuring around 0.067%. Fig.2 shows the spectrum of the B.M.C.'s output while it amplified a 1kHz tone at 2mV input. Even into the very demanding 600-ohm load, the third harmonic, the highest in level, lies at 74dB (0.02%), the second at 90dB (0.003%). I inadvertently measured the MCCI's high-frequency intermodulation with the level close to the maximum the preamplifier can handle at the top of the audioband (fig.3). Even so, though a number of higher-order intermodulation products are visible, with the 18 and 21kHz products lying at 44dB (0.6%), the second-order product at 1kHz is at 64dB (0.06%). Reducing the input level rapidly reduced the levels of all the products.
B.M.C.'s Phono MCCI measures well. I am not surprised that MF liked it as much as he did.John Atkinson