Measurements (from 2026)
After I reviewed the Audio Research SP-10 tubed preamplifier for English magazine Hi-Fi News & Record Review in May 1984 (footnote 1), I bought the review sample (serial numbers 54227013, preamp, and 94226012, power supply). I brought it with me when I moved to the US two years later and, after having it adjusted to operate on 115V wall voltage rather than the UK's 240V, I used it as my reference preamplifier for a few years. However, it spent most of the subsequent decades in my storage unit, only occasionally spending time in my system.
In August 2026, Stereophile's editor, Jim Austin, asked if I still had the SP-10? He was thinking about taking a listen, so I set the preamplifier up in my test lab to check it was still working. After making sure that all the tubes were correctly installed, I crossed my fingers, powered up the 42 year-old preamp, and after letting it run for an hour or so, I performed a basic set of tests with my Audio Precision SYS2722 system.
Looking first at the SP-10's performance as a line preamplifier, all the inputs and outputs preserved absolute polarity, ie, were non-inverting. The SP-10's single-ended input impedance was specified as 50k ohms; I measured 58.5k ohms at 20Hz and 1kHz, and 46k ohms at 20kHz. The output impedance, specified as 1000 ohms, was 1297 ohms at 20Hz, 1147 ohms at 1kHz, and 1121 ohms at 20kHz.
Despite its age, my Audio Research SP-10 offered generally very good measured performance with both its line and phono inputs, offering low distortion.—John Atkinson
Footnote 1: Anthony Cordesman's Stereophile reviews of the SP-10 can be found here.
Fig.1 Audio Research SP-10, Low gain, frequency response with volume control set to the maximum at 1V, into 100k ohms (left channel blue, right red, 1dB/vertical div.).
The SP-10's gain can be set to High or Low with a front-panel switch. In the Low setting, the gain was 12.6dB, left channel, and 12.3dB, right channel. This slight channel imbalance can be seen in fig.1, which plots the SP-10's frequency response into 100k ohms (blue and red traces) with the volume control set to the maximum. The response is flat in the audioband and down by 3dB at 150kHz. Setting the gain to High increased the right channel's output level by the specified 12dB, but the left channel's by 7.5dB; I'm not sure of the reason, but this won't be an issue with line-level sources used with the Low Gain setting.
Fig.2 Audio Research SP-10, Low gain, spectrum of 1kHz sinewave, DC–1kHz, at 2V into 100k ohms with volume control set to the maximum (left channel blue, right red, linear frequency scale).
Channel separation was only fair at 75dB at 100Hz, reducing to 30dB at the top of the audioband. With the volume control set to the maximum and the gain set to High, the wideband, unweighted S/N ratio (ref. 1V output) was 67.6dB (average of both channels), which increased to 74.5dB, left, and 82.2dB, right, when the measurement bandwidth was reduced to the audioband, and to 77.6dB, left channel, and 84.5dB, right when A-weighted. The Low gain ratios were around 6dB greater in both channels. Spectral analysis of the Audio Research's low-frequency noise floor with the preamplifier set to Low gain and outputting a 1kHz tone at 2V with the volume control set to the maximum (fig.2, blue and red traces), revealed that while supply-related spuriae at 60Hz and its odd-order harmonics were low in level, the random noise components were higher than I have found with modern solid state preamplifiers, at –110dB, right channel, and –118dB, left.
Fig.3 Audio Research SP-10, Low gain, THD+N (%) vs 1kHz output voltage into 100k ohms.
Fig.3 plots the THD+noise percentage in the Audio Research preamp's Low Gain mode against voltage into 100k ohms with the volume control set to the maximum. Actual distortion lies below the noisefloor up to 3V output, where it lies at just 0.007%. It increases above that voltage, reaching 0.1% at an extraordinary 60V!
Fig.4 Audio Research SP-10, Low gain, THD+N (%) vs frequency at 3V into 100k ohms (left channel blue, right red).
Fig.5 Audio Research SP-10, Low gain, spectrum of 50Hz sinewave, DC–1kHz, at 2V into 100k ohms (left channel blue, right red; linear frequency scale).
Fig.6 Audio Research SP-10, Low gain, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 2V into 100k ohms (left channel blue, right red; linear frequency scale).
The outputs' THD+N percentage into 100k ohms (fig.4) was consistent with frequency from the low bass through to the mid-treble region, but rose in the top octave, particularly in the right channel (red trace). The distortion signature was primarily the third harmonic at a very low –96dB (0.0015%, fig.5), though the second harmonic was higher in the left channel (blue trace) than in the right (red). Intermodulation distortion with an equal mix of 19 and 20kHz tones was very low. The difference component at 1kHz lay close to –100dB (0.001%, fig.6).
Turning to the Audio Research preamplifier's two phono inputs, to minimize hum I connected a wire from the ground terminal on the preamplifier chassis's back panel to a grounding post on the Audio Precision analyzer. The phono inputs can be used with a front-panel switch set to "Normal" or to "Bypass." The latter connects the phono stage output directly to the volume control and output stage. Gain with the volume control set to the maximum was a high 72dB, which will probably be too much for moving magnet cartridges.
I performed all the phono input testing in Bypass mode. Their input impedance can be set to 0 ohms , 30 ohms, 100 ohms, 10k ohms, and 47k ohms. The measured input impedances were all close to the nominal values. The only exception was the 47k ohm setting, where I measured 52.2k ohms at 20Hz, 45.2k ohms at 1kHz, and 32.6k ohms at 20kHz. The phono inputs preserved absolute polarity.
Fig.7 Audio Research SP-10, phono input, response with RIAA correction (left channel blue, right red) (0.5dB/vertical div.).
The SP-10's RIAA equalization was accurate up to 20kHz (fig.7), though with a slight rise in the low bass and an ultrasonic rolloff reaching –1dB at 80kHz. The left channel's output (blue trace) was 0.26dB higher than right channel's (red trace), probably due to the tubes' age. (I last replaced all the 6DJ8 tubes in the early 1990s.) Channel separation (not shown) was similar to that of the line inputs.
Fig.8 Audio Research SP-10, phono input, Bypass mode, spectrum of 1kHz sinewave, DC–1kHz, at 1mV input with volume control set to the maximum (left channel blue, right red, linear frequency scale).
The phono input's unweighted, wideband S/N ratio, measured with the input shorted to ground, was a relatively low 42.5dB (average of both channels), ref. 1kHz at the nominal moving coil reference level of 500µV. Restricting the measurement bandwidth to 22Hz–22kHz increased the ratio to 45.5dB in both channels, and inserting an A-weighting filter gave a ratio of 62.5dB. Spectral analysis (fig.8) indicated that the SP-10 phono input's noise floor was dominated by very-low-frequency random components. Anthony Cordesman did note in his 1984 review that the phono input's noisefloor was higher than he expected; he recommended that the SP10 be used with moving-coil cartridges with a nominal output of 2mV or higher.
Fig.9 Audio Research SP-10, phono input, Bypass mode, spectrum of 1kHz sinewave, DC–10kHz, for 1mV input (left channel blue, right red, linear frequency scale).
Fig.10 Audio Research SP-10, phono input, Bypass mode, HF intermodulation spectrum, DC–30kHz, 19+20kHz for 10mV peak input (left channel blue, right red, linear frequency scale).
The phono input's harmonic distortion (fig.9) and intermodulation distortion (fig.10) were very low in level. The phono stage overload margins were astonishingly high, at 32dB ref. 1kHz at 500µV. With the volume control set to its maximum, the output level at 1% THD+N was close to 90V!
Footnote 1: Anthony Cordesman's Stereophile reviews of the SP-10 can be found here.
