Linn LK1 preamplifier & LK280 power amplifier Measurements
Starting with the LK280, this raised a healthy 82.6W at 1kHz at the onset of clipping into an 8 ohm resistive load with one channel driven, almost doubling to 158.9W into 4 ohms, and revealing the excellent power-supply regulation. (The power transformer mechanically hummed quite loudly at this latter level, however.) The LK280 is non-inverting, and its input impedance was slightly higher than specified, but still low at 3830 ohms, necessitating use with a preamplifier capable of driving low-impedance loads. The sensitivity was also slightly higher than spec, with 990mV required for maximum output, implying 28.5dB voltage gain rather than the specified 26dB. The output impedance was low at 0.06 ohms, 0.07 ohms, and 0.1 ohms at 20Hz, 1kHz, and 20kHz respectively, though there was a little DC present on the output, –27.9mV (left) and +10mV (right).
At a 1V output level, the '280 was ostensibly flat throughout most of the audio band, though there was a slight droop at the edges, the response being 0.75dB down at 21Hz and 20kHz. Investigating the transient behavior of the power amplifier with both rectangular pulses and raised cosine waveforms—Alvin Gold had reported on an inherent instability in the amp's predecessor, the LK2—revealed nothing out of the ordinary other than the effect of the bandlimiting.
The LK1's line stage was found to be similarly band-limited, as can be seen in fig.1. The upper of the two traces is the response via the Aux input, measured at the preamp's output terminals, which is 1dB down at 25Hz and 31.5kHz. The lower trace is the MM phono response, which shows the ubiquitous European low-frequency IEC roll-off, being 1dB down at 40Hz and –3.9dB at 20Hz. This bandlimiting of the phono section also shows itself in fig.2, the response of the LK1 to a 5mV-amplitude, 500Hz squarewave, fed into the MM input via an inverse-RIAA network. The response via the MC input was identical at the high end, but had rather more bass extension than the MM input, not reaching –1dB until 26Hz rather than 40Hz. The intention of this restricted bass response is to reject disc rumble and warp-induced spuriae, and this the LK1 will most definitely do, the MM input being 21dB down and the MC input 16dB down at 6.3Hz, the center frequency of LP warp information. It is ironic, therefore, that the Linn LP12 turntable has sufficiently low rumble and the Linn Troika/Ekos cartridge/arm combination a sufficiently well-tuned LF resonance that this intrinsic filtering is probably unnecessary.
Fig.1 Linn LK1, frequency response at 1V into 100k ohms (from top to bottom at 20Hz): line stage, MM phono stage (1dB/vertical div., right channel dashed).
Fig.2 Linn LK1, MM phono stage, 500Hz squarewave.
The MM input sensitivity was higher than spec, 4.33mV at 1kHz giving 500mV out with the volume control at maximum. The MC stage was 20.6dB more sensitive, requiring 388;mV at 1kHz for nominal output. MC residual noise was inaudible, even at maximum volume levels, so I didn't bother measuring it. (However, it is essential to position the LK1 well away from the LK280 power amplifier if 60Hz hum is not to be injected into the signal.)
I did measure phono-input overload margins, however. The maximum input levels into the MC stage, assessed by eyeballing the waveforms on a 'scope, were 560µV at 20Hz, 5.2mV at 1kHz, and 56mV at 20kHz, ie, at least 20dB margin across the audio band for a 500µV/cm/s cartridge, which is about average performance. The MM input was less good, however. Although 9.4mV at 20Hz and 52.3mV at 1kHz represent a good and an average margin, respectively, approximately 250mV at 20kHz is not so good, representing just 14dB of headroom when the RIAA pre-emphasis is taken into account. Although recorded music never has 20kHz energy at +14dB, clicks can have, and it is possible that the subjective annoyance of such spuriae will be worsened when the LK1 is used with highish-output MM cartridges.
A note on the 20Hz overload via the MM input: rather than clip, at 9.4mV input the waveform suddenly featured a sharp glitch on the rising slope, as can be seen in fig.3. Increasing the level further produced a second glitch at the analogous point on the falling slope. As the lack of a 5-pin XLR connector meant that I couldn't assess disc-input overload by looking at the waveform at the tape-out sockets, I couldn't, therefore, immediately diagnose whether this glitch was occurring in the MM circuitry or in the line stage. Backing off the volume control reduced the level of the waveform without removing the glitch, implying that it was the MM circuitry producing the effect. This was confirmed by the fact that the line stage by itself failed to reveal a similar problem. It also didn't occur via the MC input.
Fig.3 Linn LK1, MM phono stage, 20Hz overload waveform.
The LK1 appears to have only a limited line-stage gain of 7dB. With the LK280 input sensitivity of 990mV, this means that 443mV will be required into the Tuner, Aux, or Tape inputs to drive the combination to maximum output, which might be on the high side for some tuners or cassette decks. The maximum output level at 1kHz measured 5.724V RMS, this presumably the maximum swing of the switching ICs. Those sensitive to absolute phase effects should note that the LK1's line stage is polarity-inverting.
The output impedance of the LK1 measured low, at 71.5 ohms, but there was some mystery about the preamp input impedances as these were nothing like the specification would suggest. The line-level inputs all measured 99.1k ohms rather than 9k ohms. Looking inside indicated that these impedances appear to be set by a resistor array—this clearly marked 100k! The phono input impedances were also too high, at 99k ohms, MM, and 220 ohms, MC. A puzzle. Obviously, my sample of the LK1 wasn't tested by a computer!
Finally, I checked the action of the volume control, which promises both high precision and good channel matching. The steps vary across the control's range of action, being much less than 1dB at the top—the first 1.6dB of attenuation takes 10 steps—between 0.25dB and 0.5dB for most of the rest of the range, increasing to 1.5dB, 1.8dB, 2.5dB, 3.5dB, and 6.1dB for the bottom five steps. Channel matching was within 0.2dB over much of the range, but worsened at the bottom of the dynamic-range "window," the final four settings being 0.25dB, 0.5dB, 0.8dB, and 1.4dB adrift, respectively. Frankly, this is academic, the LK1's volume control performing to a far tighter tolerance than the mechanical pots fitted to the majority of preamps I have auditioned. In contrast to most other preamps, the volume control doesn't cut the signal off completely. I measured a maximum attenuation of 46.5dB (left) and 47.9dB (right). The Mute button, however, does provide a deep mute to silence, rather than the 20dB or so extra attenuation often found elsewhere.—John Atkinson