Tube Preamp Reviews

Editor: We offer our appreciation to the LL1 Signature review team: to Robert J. Reina for effectively revealing and focusing on the main sonic qualities of the LL1, and to John Atkinson for his interpretation of the measurement results, which are, as always, without peer.

From the start, Mr. Reina set out to clear specific hurdles in a product of the level of the LL1 Signature, in order to be truly impressed with the preamp, and he thoroughly covered the entire ground by dealing with those issues. He ended up being surprised by finding "three areas in which the [LL1] significantly exceeded the performance of any other tubed line stage" he'd heard. The conclusion he arrives at is based on his intimate familiarity with the music material, which served as a litmus paper for the preamp's ability to reveal various aspects of recorded materials. His closing remarks, about a "stunning design, stunningly executed," and feeling "floored by the several areas in which it achieved levels of performance I hadn't thought possible," are truly gratifying to hear.

Yet again, our sincere thanks to John Atkinson for exposing the essence of the LL1 with his measurements. It is truly an honor to have these tests conducted by a top-notch professional who understands the product and draws accurate conclusions from the measurement results—that is, indeed, a rather rare quality in our times.

Overall, the review is fairly comprehensive and accurate. There is just one point I'd like to address about one measurement: the unweighted, wideband signal/noise ratio. The measured difference of about 43dB between the wideband and the same parameter restricted to the audioband is unrealistic to the point of making it practically improbable. Allow me to enumerate the reasons for this statement.

We keep the data log of our own measurements of each unit we manufacture. Taking 40 production units as a basis for supplying the measurement ranges between the lowest and highest readings, I would like to show these figures in illustration of my point.

Our Measurements of Signal/Noise Ratio (ref. 1V RMS):

Wideband: 85.6–92.5dB
Audioband (22Hz–22kHz): 93.4–99.6dB
A-weighted: 99.5–105.3dB

As one can see, the largest difference between the wideband and audioband readings is 14dB. In any individual unit, the difference would not account for more than 10dB. These variations stem mainly from the distribution of spectral density of each tube's internal noise.

Besides this, going as far back as the design stage of the LL1, I can firmly state that an LL1 prototype exhibiting such a signal/noise ratio would have been rejected outright as a nonviable product. Moreover, even if, for some unaccountable reason, a production unit showed a similarly low measurement, that unit would have never passed our QC.

Taking into account the considerable quantity of these preamps currently sitting in various systems all over the world, and using the testimonials of our customers as well as our own experience with the LL1, it is safe to say that this preamp is dead quiet, even when used in systems with very sensitive loudspeakers. In one such system, comprising the ML3 Signature amps, LL1 Signature preamp, and speakers with 110dB/W/m sensitivity, only a very residual noise—like a barely audible rustling of leaves—can be heard with one's ear inside the speaker horn.

In my opinion, the low value of about 63dB that appears in the review's "Measurements" section has to do with some grounding issues or radio interferences that were present during the measurement process. Something of the sort must have been going on in order to account for this otherwise inexplicable result.

Respectfully—Vladimir Lamm, Lamm Industries, Inc.

Having noted Vladimir Lamm's comments about the LL1 Signature's noise performance, I repeated the signal-to-noise ratio measurements, using serial no.110032, using both my Audio Precision SYS2722 and vintage Audio Precision System One. Both analyzers gave very similar measurements, which were also very similar to what I published in the review.

As before, the measurement conditions were: input shorted to ground, volume control set to its maximum, and ratios ref. 1V output. The separate ground connection on each Audio Precision analyzer was connected to the ground connection on the Lamm power supply; and before running any measurements, I checked that there was nothing wrong with the AC wiring to my test bench.

When examining a products' noise performance, the measurement bandwidth has a significant effect on the result. Measured with a bandwidth of <10Hz–>500kHz, the LL1 Signature's S/N was 63.1dB. Reducing the bandwidth to 22Hz–22kHz gave a S/N of 105.7dB, while switching an A-weighting filter into circuit improved it to 106.7dB. Setting the volume control to its minimum gave a wideband, unweighted S/N of 63.2dB but increased the other two ratios by 3dB.

What was significant was that remeasuring the unweighted ratio with a <10Hz–80kHz bandwidth improved the S/N to 92dB, though the reading was very unstable, ranging from 85dB to 95dB, due to the presence of very-low-frequency, 1/f noise. This suggests that Vladimir Lamm's conjecture—that the test system and preamplifier were somehow picking up RF interference—is correct.

Spectral analysis with a 100kHz bandwidth showed nothing of interest between 20 and 100kHz, but did uncover some very-low-level spuriae between 5.5 and 8.5kHz that shouldn't have been there. The exact frequency and level of these spikes changed slightly every time I took the spectrum. Suspecting interference from the nearby computer that runs the Audio Precision software, I turned off the WiFi, but this produced no change in the noise measurements. I then moved the preamplifier and its power supply as far as possible away from the computer, but again got almost no change in the S/Ns or spectra. So if the LL1 Signature is picking up ultrasonic or RF noise, it is not from the test computer.

I am at a loss, therefore, as to why my wideband, unweighted noise measurements are so much worse than Vladimir Lamm's.—John Atkinson