Tube Power Amp Reviews

I believe that it is de rigeur to evaluate amps with a variety of speaker loads in order to flush out their performance capabilities in as many contexts as possible. To that end I also included the following loudspeakers in my listening tests: the Koval-modded old Quad ESLs, the Celestion SL600s, the Acoustat Spectra 22s, and the MartinLogan Sequel IIs. Some of these loads are not only pretty insensitive but also quite reactive. In all cases, however, the RM-9 proved to be a competent and uncomplaining partner—never at a loss dynamically, and clipping gracefully when pushed against the wall. The bass registers were well-delineated, with plenty of detail and pitch definition. Not quite the precision and extension of solid-state bass, but most impressive nonetheless for a tubed amp. The mids were detailed, but not in an over-analytical fashion. Massed voices, for example, were clearly resolved. You could not only place the sopranos, but the soft male voices in a mixed chorus were plainly audible. Midrange textures were liquid and suave in the best tube tradition, though not overly, romantically lush. The first sample possessed a sight glare in the upper mids, which as far as I can tell is not representative of the design and may very well have been caused by a sample defect. One of the output tubes would cut off intermittently, apparently due to a defective tube socket. As I jiggled the tube I could reestablish high voltage.

The second sample of the amp provided by Music Reference has been working just fine for several weeks, and in no way could be described as bright. The timbral accuracy of the RM-9 is very good—more faithful to the source than the Quicksilver, for example. During a sonic shootout between these two amps, the RM-9 distinguished itself by imposing less of its sonic signature on the reproduced sound. Daniel, my 13-year-old son and sometime assistant, had a much easier time recognizing the essence of his mother's voice during an administration of the Lesley Test when the RM-9 was in the chain. The upper mids and presence regions were well-behaved, displaying the requisite amount of sweetness on violin overtones, and very good brass bite. The treble was a bit on the soft side, smooth and liquid. Yet treble transients were very well resolved, without attendant zip or harshness.

The area where the stock RM-9's performance fell relatively short of my expectations was in its portrayal of the soundstage. It was good all right in terms of focusing instrumental outlines and delineating the dimensions of the soundstage. Its retrieval of hall reverb was quite good too, and if you're anything short of a devout tubeophile, you'll probably find nothing amiss here. But for me, it simply lacked the spaciousness, the 3-D flair and image palpability of, say, the Prodigy 150 Monos, as well as the utterly transparent soundstage of a Class A amp. Some of the soundstage drama was plucked away by the RM-9.

This is also a quiet amp; no hum or much tube noise to speak of. So I was quite surprised, following a 45-minute power outage, when the RM-9 developed a hum problem all of a sudden. Could it be that the Power Company screwed up? Defeating the ground pin on the RM-9's three-prong plug and reversing the hot and neutral pins in the AC outlet immediately cured the problem. So here's something else for you to worry about: your local Power Company screwing up mains polarity.

At this point in the proceedings, I was staring at a likable amp, eminently listenable, without any significant weaknesses, yet one I just could not fall in love with. A competent amp all right, but without a performance aspect to really get excited about. It lacked the midrange incisiveness I've become addicted to: that combination of transparency and spatial resolution that commands your undivided attention because it mimics the live experience so closely.

Enter the KT88
Initially, I scrounged together whatever KT88s I had on hand to retube the RM-9. Remember, we're talking a lot of KT88s. I had two matched pairs of the Gold Aero variety that I was saving for a rainy day. And by raiding the Quicksilver Monos I was able to raise the necessary quota of output tubes. Later, Music Reference provided me with a pair of matched quartets of Chinese KT88s, which I used for most of the listening.

It was quite a sight to behold! Count them, eight glowing KT88s. Midrange textures were definitely richer and lusher now, and a tad darker as well. Upper bass was fuller and not as tight when compared with the stock version. But the most notable difference was in terms of imaging. Palpability of image outlines went from good to impressive. And soundstage transparency and spatial resolution improved a few notches as well. It was now much easier to visualize hall dimensions and to retrieve hall reverb. Image outlines were consistently palpably focused. What I'm describing is serious goose-bump–inducing capabilities in these respects. Julianne Baird (The English Lute Song, Dorian 90109) singing through a chain that included the Sequel IIs, the Wadia 2000 Processor, and the Arcam Delta 170 transport, gave me an inkling of the best CD sound I've ever heard. There she was, coherently focused in space, without a trace of fuzz through her upper octaves, and that delicate interplay of direct and hall sound clearly resolved. In general, reproduction of the mids and upper octaves was smooth and edgeless, without fuzz or grain.

The RM-9 again established its dynamic strengths. Even with the power-hungry Sequel IIs it had no trouble at all eliciting a linear, uncompressed dynamic range, expanding from soft to loud and affirming its ability to clip gracefully when really pushed hard.

Summary
Unquestionably, as AT&T is fond of saying, the KT88 version of the RM-9 is the right choice, easily worth the extra $300. For a modest increase in price, you essentially get a stereo amp with imaging capabilities that rival that normally afforded only by monoblocks. There are no serious sonic weaknesses. The bass is not as tight and extended as that available from solid-state quarters. But it is competitive with the bass performance of any tube amp I've heard to date. The RM-9 clearly loses out to the VTL 300 monoblocks in punch and authority, but is just as strong in bass detail and definition. The mids and treble are texturally liquid and suave. There's no fuzz or edge accompanying treble transients. Timbral accuracy is another strong suit of the RM-9; it imposes very little coloration on the reproduced sound. The treble is a bit soft, the upper bass a tad loose, and the mids could be accused of being a bit dark-sounding—at least with the KT88 version. But, all in all, the RM-9 does very little to interfere with the essence of the music.

Let's not forget the extended tube life and potential reliability of the design. I say potential, because reliability implies a track record, which in the case of the RM-9 I know nothing about. But at least from the several months I've lived with it, it has impressed me as the sort of amp that could take a lickin' and go on tickin'.

Both Vulcans and Humans would be pleased with the RM-9. It combines good looks, logical execution, and the sort of musical involvement we humans very much appreciate. In its KT88 incarnation, the RM-9 is worthy of a strong Class B recommendation.