After years of using the Expressive Technologies SU-1 and SU-2 step up transformers I come across to a fact this morning that the Expressive was reviewed by Robert Harley in June, 1992. I read the Expressive’s preamp review but I never seen this one. I will not comment on the Robert ’s founding and just provide the review. I just mention that something was very sick in Robert Harley’s mind when he loaded the SU-1 transformer to Vendetta Research phonostage. Vendetta is MC-level corrector with maximin input impedance of 200R. What the hell he anticipated to hear loading his cartridge with 200/625=0.32R. It was .32R!!!!

I reprint the article here in order to preserve the writing for posterity as who know tomorrow the Stereophile will be sold to a big real-estate agency and they will be publish review about colonial houses.

Expressive Technologies SU-1 moving-coil step-up transformer
By Robert Harley  (Stereophile, June, 1992)

http://stereophile.com/phonopreamps/692su/

What's this? A review of a $3000 moving-coil step-up transformer in this digital day and age? Yep. Although the market for such a product is small, the fact that the Expressive Technologies SU-1 step-up transformer enters previously uncharted state-of-the-art territory warrants these pages of editorial space. Furthermore, LP playback appears to be alive and well at the upper end of the high-end spectrum, a market segment addressed by the SU-1 (footnote 1).

Before getting to the SU-1, a few words about Expressive Technologies are in order. In many ways, they are like the small tweaky outfits that started the high-end audio business; they have a zeal for achieving the best possible sound without regard for user convenience or what the product costs. Unlike shaky garage operations, however, Expressive Technologies builds their products to the highest mechanical standards and fit'n'finish. Further, because the company is a labor of love by three audiophiles who have achieved financial success in other fields, Expressive Technologies is likely to be around for some time.

The SU-1 moving-coil step-up transformer is an offshoot of Expressive Technologies' research into designing a very ambitious electrostatic loudspeaker. Part of that design effort included building a step-up transformer for the electrostatic panel. The knowledge gained in transformer design was applied to stepping up a moving-coil cartridge's output signal. Incidentally, the SU-1 step-up was designed by a Bell Labs researcher who holds a PhD in analog circuit design.

The SU-1 step-up is one of only two products offered by Expressive Technologies. The other is the IC-1 interconnect. Because virtually all SU-1s are sold with two pairs of IC-1 (for reasons described later), this review is in essence an assessment of both products. In addition, I will comment briefly on the IC-1's sonic qualities when used with other products.

The SU-1 is definitely not what one expects a moving-coil step-up transformer to look like. At 19" wide, over 5" high, and weighing 35 lbs, the SU-1 is clearly a serious product. The rear panel holds two pairs of gold-plated RCA jacks and a ground terminal. There isn't much to talk about inside the unit; the SU-1 is a black box that contains two other black boxes—the left- and right-channel transformers. Build quality is excellent.

A few other technical notes: the SU-1 works best with low-output moving-coils, especially those with output voltages of less than 0.5mV. The relatively low-output (0.3mV) AudioQuest AQ7000 cartridge I used was a good match for the SU-1, while many of the newer high-end moving-coils have output voltages of 0.18mV, making them good candidates for use with the SU-1. However, cartridges with a high output impedance (above a few ohms) should be avoided. Not all phono stages work well with the SU-1; the phono stage input impedance should be high (47k ohms is ideal). Phono preamp gain is also a critical factor in how much difference the SU-1 can make to the system. High-gain phono stages don't need the stepped-up voltage, and some may be overloaded by the SU-1 when driven by even moderate-output moving-coil cartridges. The ideal conditions for the SU-1 are a very low-output moving-coil and a 47k ohms input impedance moving-magnet phono stage.

Listening

The SU-1's effect on the musical presentation varied from the dramatic to the subtle, depending on the phono preamp used with it. I'll start with the dramatic, heard when auditioned with an Audio Research SP11 Mk.II. The combination of the fairly low-output AQ7000 and the moderate-gain SP11 really benefited from the SU-1. First, the dynamic contrast was greatly improved, with a wider range between loud and soft. The background was "blacker," and high-level transients were sharper and more lifelike. In addition, there seemed to be greater resolution of dynamic gradation, the music taking on finer degrees of dynamic shading. Overall punch and slam were greatly improved; inserting the SU-1 in the phono chain was like giving the system a shot of steroids. Drums had more impact (especially kick and snare), contributing to a greater feeling of life.

Because much less gain was needed in the SP11's line stage with the SU-1, the noise floor was noticeably lower. Soft passages were more clearly delineated, rather than having a light hiss superimposed on them. This was especially apparent with records cut with low signals and music with very wide dynamic range. This improvement alone may be worth the price of admission.

I also found the SU-1 increased the "see-through" transparency of the presentation. There was a greater clarity that provided a clearer view into the soundstage. Soundstage depth, width, and focus were all improved by the SU-1. The impression of instruments and voices hanging in three-dimensional space was heightened, as was the feeling that a slightly opaque veil had been removed from between me and the music. The SU-1 was the antithesis of murky and congested.

The SU-1 also made instrumental textures more liquid and smooth. The treble, in particular, was more gentle and relaxed. Despite the softer presentation, there was a distinct impression of hearing more detail and information. The SU-1 uncovered another layer of minute detail and subtlety in the music. The enhanced detail, however, tended to reveal the subtle fabric of finely woven textures, rather than merely emphasize etch and grain.

The bass also benefited from the SU-1, becoming rounder and more liquid, with a greater naturalness. Acoustic bass had a warmer, less sterile and stultified character. Pitch definition improved, allowing greater resolution of individual notes. This was especially apparent during Ray Brown's playing on the Bill Evans record Quintessence (Fantasy F-9529). The instrument had a "fuller" sound, and was just more natural and right. Overall, the SU-1 step-up and IC-1 interconnect did wonders for the AQ7000/SP11 combination.

Next up was the Mod Squad Phono Drive. With both moving-magnet and moving-coil inputs, and front-panel selectable loading, the Phono Drive provided a wider range of conditions. I compared the sound of the AQ7000 going directly into the moving-coil input vs the SU-1 in the signal path connected to the moving-magnet input. This is a typical real-world condition; the SU-1's voltage gain replaces the additional circuitry of an active moving-coil stage.

Like my impressions with the SP11, adding the SU-1 was a revelation. The Hales Signatures seemed to have another octave of bass extension. The greater weight and power in the bass was not a subtle improvement. Transient leading edges were much more sharply defined, giving the presentation a greater sense of immediacy. The drums and percussion on the LP Roland Vasquez and the Urban Ensemble (Arista/GRP 5002), for example, took on new life and drive.

The SU-1 also had a dramatic effect on soundstage depth and apparent hall size. The presentation became deeper and more spacious, with a feeling of instruments floating in three-dimensional space. On the Robert Lucas album Usin' Man Blues (AudioQuest AQ-LP1001), for example, the space surrounding the harmonica and vocal bloomed to give a wider and more open perspective. There was a greater sense of the loudspeakers falling away and being replaced by images hanging in space. The sound was just more palpable and real with the SU-1. After hearing the SU-1, it was not easy to go back to listening without it.

To see what the effect of the SU-1 would be on an excellent MC stage, I used it to drive John Curl's Vendetta Research SCP2B (footnote 2). Despite the fact that the Vendetta SCP2B is designed to accept a moving-coil output without a step-up device, it was not overloaded by the SU-1 when driven by the relatively low-output (0.3mV) AudioQuest AQ7000 cartridge. However, as the Vendetta has more gain than the SP11 Mk.II, the improvement in S/N ratio was less meaningful. The Vendetta was quiet with or without the SU-1. I didn't hear the dramatic increase in dynamic contrast with the SCP2B that I did with the SP11, and with the Vendetta's variable input impedance adjustment set too low, the sound was rolled off in the treble and lacking life. The presentation was also a little on the lightweight side, the opposite character heard with the SP11. Not a recommended combination.

I must point out that I was unable to audition the SU-1 with any interconnects other than the IC-1 because of high hum levels with other interconnects. Whatever the grounding arrangement, I got an unacceptable amount of hum without the IC-1 interconnect (which has separate ground wires).

While I'm on the subject of the IC-1 interconnect, which costs $595/meter pair, terminated, I'll take this opportunity to relate my experiences with it. First, I can say without hesitation that the IC-1 is the best-sounding interconnect I've auditioned. I've had much experience with it over the past several months, and recently experimented with various interconnects between the Mark Levinson No.30 digital processor and the Audio Research LS2. In all cases, the IC-1 was clearly superior. There was a more natural portrayal of instrumental and vocal timbres. The IC-1's crystal transparency provided a more realistic feeling of the actual instruments' sound. There was also more inner detail and greater resolution of fine textures. The IC-1 presented another level of detail and nuance in the music. The bass was full and rich, yet tuneful and articulate. The treble was more laid-back than many interconnects, yet never sounded lacking in life. In short, the IC-1 excelled in every area, especially the first two noted: natural presentation of timbres and resolution of inner detail.

Despite the IC-1's superb sonics, it's not for everyone. Not only is it thick, bulky, heavy, and very difficult to bend, it's absurdly so—like a garden hose filled with ice (footnote 3). If you're going to set up your system once and not change interconnects, the IC-1's unequaled musical characteristics will make the effort worthwhile. Those who, like reviewers, constantly change equipment and interconnects, are cautioned about the decidedly user-unfriendly nature of the IC-1. Why is it the best-sounding stuff always seems to be the most difficult to use?

Conclusion

In suitable systems, the Expressive Technologies SU-1 step-up transformer can dramatically improve the quality of LP replay. The lower noise level, increased dynamics, more spacious soundstage, greater transparency, and more natural presentation of instrumental textures rendered by the SU-1 were nothing short of stunning. In fact, I found it hard to go back to the SP11 and Phono Drive without the SU-1 in the chain. These impressions will not hold true for all phono stages, however. Because the degree of improvement rendered by the SU-1 varies greatly with the phono section, prospective buyers are urged to audition the SU-1 with their preamp and cartridge before making a buying decision.

Would I spend $2950 for the SU-1? Without hesitation. It really was a quantum improvement, one I believe justified by the SU-1's not insignificant price. I should reiterate that using Expressive Technologies IC-1 interconnect in the phono chain is essential, both because of the potential for hum and the synergistic sonic effects of the two products. The cost of two pairs of IC-1 should therefore be factored into the SU-1's price.

If you love LPs and have a low-output moving-coil and a preamp with a high impedance input, the SU-1/IC-1 is a "must hear" product. But I'll warn you: Once you hear your favorite LPs through the SU-1 step-up and IC-1 interconnect, you may not want to live without them.

Description: Moving-coil step-up transformer: Voltage gain: 28dB. Turns ratio: 25:1. Input impedance: 75 ohms. Recommended preamp input impedance: 47k ohms, 100pF. Frequency response: 0.1Hz–250kHz, +0, –3dB. Phase shift: <3° at 20kHz. Group delay variance: <100;us (20Hz–20kHz). Noise: >–100dB (referenced to 0.5mV). Maximum input voltage: 1V. Maximum output voltage: 25V.
Dimensions: 19" W by 5.4" H by 8.25" D. Weight: 35 pounds (net).
Price: $2950 (1992). Approximate number of dealers: 5.
Manufacturer: Expressive Technologies, P.O. Box 6401, Holliston, MA 01746-6401 (1992). Company no longer in existence (2008).

Review System

I auditioned the SU-1/IC-1 combination over the past few months. Phono preamps used in conjunction with the SU-1 included the Audio Research SP11 Mk.II and a Mod Squad Phono Drive, all driving (independently) an Audio Research LS2 line-stage preamp. Turntable was the Well-Tempered with its tonearm modified by LP Labs (see Vol.15 No.1, p.224).

Loudspeakers were primarily Hales System Two Signatures, augmented with a Muse Model 18 subwoofer. Power amplifiers were my long-term favorites, the VTL 225W Deluxe monoblocks, or the two solid-state amplifiers reviewed in April (Parasound HCA-2200 and McCormack DNA-1. Interconnects were the Expressive Technologies IC-1 in the phono chain, Straight Wire Maestro between preamp and subwoofer, and AudioQuest Diamond between subwoofer and power amplifiers. Loudspeaker cable was a 3' bi-wired run of AudioQuest Dragon/Clear (with the VTLs) or 8' runs of AudioQuest Sterling/Midnight (with the two solid-state amplifiers).—Robert Harley

Measurements

When I measured the SU-1, the results appeared to be more related to the test setup than the SU-1's intrinsic performance. Just as there was a large variability in sound quality depending on the phono preamp used, the measurement results were highly dependent on the test conditions.

I initially measured significant rolloffs at the audio-band frequency extremes—down 2dB at 20kHz and nearly 3dB at 20Hz. These rolloffs would be highly audible, yet no rolloff was suggested by my auditioning. The cause of the measured rolloff was twofold. The 25 ohm source impedance (the lowest possible setting) of the Audio Precision System One's signal generator formed a voltage divider with the SU-1's 75 ohm input impedance. Because the transformer inductive reactance presents a much lower impedance at low frequencies, more and more of the voltage is dropped across the source impedance as frequency decreases instead of across the transformer input. Hence the bass rolloff (footnote 1).

The AP System One's relatively high output impedance (compared to a low-output moving-coil, typically less than three ohms and sometimes a few tenths of an ohm for the ultra-low output types) was also responsible for the treble rolloff. A transformer's output impedance is a function of the source impedance driving the primary and the transformer's turns ratio. Specifically, the transformer's output impedance is equal to the square of the turns ratio multiplied by the source impedance. With a source impedance of 25 ohms (the System One's generator), and the SU-1's 25:1 turns ratio, the output impedance thus becomes very high (nearly 16k ohms). Further, the capacitance the secondary winding drives is also a function of the input capacitance: the source and interconnect capacitance are also multiplied by the square of the turns ratio. A few hundred pF at the input results (in the SU-1's case of a 25:1 turns ratio) in several tens of ;uF the output must drive. This high output impedance and high capacitance form an RC filter across the output, dropping high frequencies across the RC network instead of across the load, and rolling off the treble (even with the System One's 100k ohms input impedance).

JA remeasured the transformer, driving it with the Krell KBL line-level preamplifier to get a low enough source impedance (2.2 ohms), though this does make the test signal a little noisy. His frequency response, taken at 5mV input, is shown in fig.1. The solid lines are the left channel; the dashed, the right channel. To show the effect of load impedance, JA used the Audio Precision's inputs set to 600 ohms—the curves that droop above 4kHz—and 100k ohms—the curves that peak above 50kHz, reaching about +8dB. (Into the lower load, of course, the absolute level dropped by 10dB or so.) The measured voltage gain at 1kHz into the 100k load was almost exactly the specified 28dB.

Fig.1 suggests that the SU-1's balance will be highly dependent on the associated components, primarily the cartridge's output impedance, the interconnect capacitance, and the phono stage input impedance. Indeed, this was suggested by the auditioning. Potential purchasers are advised to audition the SU-1 in their own systems before making a buying decision. Basic guidelines, however, can be offered. The SU-1 will work best with low-output moving-coil cartridges (both because of their low output impedance and need for higher gain) and 47k ohm input impedance moving-magnet phono stages. The combination of the AudioQuest 7000 cartridge (0.3mV output and 2.5 ohms output impedance) and the SP11's 47k ohm input impedance was a good match for the SU-1. (The Lyra Clavis and Parnassus have an output impedance of less than 1 ohm and may benefit even more from the SU-1.)

The putative Achilles' heel of step-up transformers is their distortion and overload characteristic. Well, the SU-1 proved exceptional in this regard. Fig.2 shows the THD+noise in the transformer's output for output levels ranging from 0.1V to 50V, equivalent to input levels of 4mV to 2V. The three curves shown are for 20Hz (top), 20kHz (middle), and 1kHz (bottom): in all three cases, the sloping curve from left to right represents the system's noise floor; as the signal rises, the slope down eventually changes to a slope up, indicating that the distortion products have risen above the noise. At 1kHz, you can see that this change happens at a whopping 15V output, equivalent to an input of 600mV. With reference to the standard MC output of 0.5mV, this is equal to an overload margin of 61.6dB! The phono stage input will clip long before the SU-1. Even at the frequency extremes, the "bend" in the THD characteristic occurs at very high input levels, equivalent to margins of 44dB at 20Hz and 31dB at 20kHz. (The reduced margin at 20kHz is due to the RIAA curve pre-emphasizing this frequency on the LP by almost 20dB compared with the level at 1kHz. This factor also makes the reduced margin of the transformer at very low recorded frequencies less of a problem, though it must be remembered that warp information comes through at full level.)

Axel wrote:

…did you experiment with either primary or secondary SUT loading using your various carts? If so would you mind to share any of your findings?

Or course I did, anyone did. I am not sure what findings you would like me to share as they are no different then anyone else’s. I am not a big fun of primary loading, put in this way I see no needs for it with my transformer and my cartridges but fine-tune secondary loading is a very standard practice. My both stereo arms (Ortofone Jubilee and Ortofone SPU Classic with ellipse) are loaded with 57.6R and my mono arm (SPU Classic Mono) is loaded with 42R

Axel wrote:

PS: Not sure if you have any 'knowledge tree' info about MM carts, or are you still of your earlier opinion: "Not good enough"?

I do not use MM cartridges.

Hi Romy,

of course, no manufacturer give one value only (non that I would know), but will spec. a range starting with some value e.g. >10ohm or say >100ohm and either putting a cap at 500ohm, 1k or nothing at all, meaning 47k (if that suits you). Lyra would give a start point of say >100ohm 'normal' (no SUT) and >10ohm for use with SUT.

You say: >> The right values of loading would be derivatives of many things and it is fine that in different installations they might be wary within a sensible limits. <<
Of course also fine, yes. I was hoping to get some specifics from you when looking at a BEST "impedance match" ~ 1:1 with the cart in use, and if you had figured out some ball-park value which then can be used as point of departure for listening test.

The smaller the DCR value of the cart (>0.3mV) the closer is the REAL impedance of the cart (20Hz - 20kHz) to that DCR value, because it will have a VERY small inductance, no?

Axel

Axel wrote:

I was hoping to get some specifics from you when looking at a BEST "impedance match" ~ 1:1 with the cart in use, and if you had figured out some ball-park value which then can be used as point of departure for listening test.

Axel wrote:

OK, this is useful also, in that you have not noted a "pattern" as you say. My point is closer to your suggestion, that I do change cart --- and sometimes buy new ones,  B U T I do not buy other SUT (at least I try to avoid it). I'm currently waiting for the return of my new Ortofon Windfeld (the old one developed a problem with wrong VTA of cantilever) so they give me a new one after MUCH waiting time. (I rings a bell with what your Jubilee did, as did my previous Jubilee!!! also exchange because of split open body). So, I'm getting ready to hock up the new (I hope now correctly working one) with my FR XF-1 type M SUT. With that messed up Windfeld (riding almost on the bottom) I found that primary loading sounded best and 10ohm! Maybe the new one will be happy with it also, but maybe not - I will see. Using primary loading is more 'uncomfortable' because of the resistors having to be soldered into the primary side of the trannie ---- and getting good resistors is also a pain (I use Shinkoh Tantalum, other R sounds crap as you'd call it).

Axel wrote:

In my case I got more resolution with primary loading, BUT it gets very quickly too bright with even small changes, sounding all wrong. The 18k on *secondary* sounded fine, but was missing some more detail that I know from connecting the cart straight with no SUT and using full 78dB gain of phono….