Conrad-Johnson Premier 17LS line-stage preamplifier Sam Tellig on the 17LS2

Sam Tellig wrote about the Premier 17LS2 in October 2004 (Vol.27 No.10):

If you own a mediocre preamplifier, it will compromise the sound of your entire system. If you substitute a great preamp, it will likely transform the sound of your entire system—making a much bigger difference, say, than upgrading your power amp or buying a better CD player.

You can also give your entire system what Antony Michaelson, of Musical Fidelity, describes as "the ineffable magic of tubes." Just put a tubed preamp in the equipment chain. That's what Bill Conrad and Lew Johnson, of Conrad-Johnson, recommend.

Designing a great preamp isn't easy, and buying a great preamp doesn't come cheap. Bill Conrad once told me that producing a preamp, for him, is the greatest challenge in hi-fi. You're dealing with low-level signals. There's so much that can go wrong—and be audible. Noise is an issue—noise from the AC line, from the power supply within the unit itself. Not to mention tube noise, if the preamp is tubed. (Tubes, when they're quiet, are ideal for amplifying low-level signals. You can't top tubes for linearity, especially in a single-ended triode circuit.)

The rewards, though, are worth it.

In 1996, Conrad-Johnson set the hi-fi world on its ear with the ART Preamplifier—or line stage. (You needed a separate ART phono stage.) ART stood for Anniversary Reference Triode. The preamp's gain came from one simple "composite triode" circuit. No feedback. Now discontinued, the ART Series 2 was last seen selling for $15,995.

Four years after the original ART, in March 2000, Conrad-Johnson followed with the Premier 17LS line stage, which offered much of the performance of the ART at a fraction of the cost. (Brian Damkroger reviewed it in the May 2001 Stereophile, Vol.24 No.5.) C-J loaned me a 17LS when I met their MV60SE and Premier 140 power amps.

Wait till you hear the new version, Conrad-Johnson's Tor Sivertsen told me. "It gets that much closer to the ART."

The 17LS2 was launched in November 2003. It currently retails for $5500—that's $1000 more than the original LS17 when it debuted in March 2000.

Why the price increase? You can blame those new 2.0µF Teflon capacitors, of which the Premier 17LS2 has 10. "The Teflons are better at resolving detail and preserve spatial information, but cost something like three times as much as the polystyrene caps used in the original Premier 17," Lew Johnson said.

What, exactly, is a capacitor?

"A capacitor is a device to store energy," Lew explained. "Think of two items in your kitchen drawer—a piece of conductive tin foil and Saran Wrap, which serves as a dielectric, or insulator. Producing a high-quality capacitor is not so easy—especially a large one with a value of 2.0µF. Conrad-Johnson worked with their supplier to develop the necessary film and winding technique.

"The bigger the capacitor, the greater the rejection rate," Lew continued. "It's really not practical to make a [non-electrolytic] capacitor with a value greater than 2.0µF."

For the Premier 17LS2, Conrad-Johnson improved the power supply. The input power-regulator circuit, which is shared by both channels, now uses costly Vishay resistors for better high-frequency characteristics, according to Lew. Each channel's final power-supply regulator has been fitted with higher-current devices to roughly quadruple the regulators' current capabilities. The output impedance has been reduced by a factor of four, increasing the channel separation and lowering the noise floor. Lowering the noise floor increases the signal/noise ratio. Less noise means greater dynamic range.

Bill Conrad, Lew's business partner, once said that there is no such thing as a perfect part. Every part has some effect on the sound quality, so the fewer parts, the better. Keeping the parts count low makes it easier to select each part for the best sound quality without having to pinch pennies. (Bill and Lew are former Federal Reserve economists.) Simple circuits tend to sound better than complex ones. And you can't get simpler than a triode.

C-J's ART was an example of two-channel simplicity at its best. It used a single active-stage composite triode circuit, which we'll chat about with Lew in a moment. The key point is that the ART's entire amplification circuit—all the gain—was a single active stage with no buffers and no feedback. The original 17LS, and now the 17LS2, use a more affordable version of the ART circuit. So, for that matter, does the new ACT2 line stage ($12,000, vs $15,995 for the now-discontinued ART Series 2).

If you want to hear the limitations of the 17LS2, you should listen to the ACT2, Lew suggested. I think he means that you don't know what you're missing until you hear it. I'm almost afraid to listen to the ACT2—I might want to buy it! Perhaps I can persuade Marina's ex-husband to take her back. (I laugh my evil laugh: Marina's ex reads Stereophile.)

On the whole, I find it difficult to write about line stages, especially inexpensive ones. Solid-state line stages often impart a metallic glare to the music. A FET is a FET and a tube is a tube. Less costly tubed line stages can sound noisy. Even worse, some of them have been known to pass a small amount of DC on to the power amp. Power amps don't like that—especially solid-state ones.

So if great line stages are expensive and inexpensive ones tend to be problematical, why not get rid of the active line stage altogether? After all, I can run the McIntosh MDA1000 digital processor straight into my McIntosh MC275 power amp with no preamp at all. (That's because the MDA1000 has its own volume control and additional gain. I wish more DACs offered these features.)

There are times a so-called "passive preamp" works out pretty well. ("Passive preamp" is an oxymoron: If a device is passive, it doesn't amplify, and vice versa.) On many occasions, I've gotten good results with my Purest Sound Systems P500 straight into a power amp or amps. It's still a favorite, though I don't know if it's still available.

But a passive preamp can compromise the sound in ways the listener might not be aware of. Nothing may be added, but something might be subtracted, such as dynamic range or a flat frequency response. Besides, even with a passive unit—perhaps especially with a passive unit—the parts quality is critical. The volume control, the switches, the internal wiring, the output jacks—all can and do affect the sound quality. Which is why you see some expensive passive devices on the market.

Another approach is to buy an integrated amplifier. You save money right away: no need to surrender several hundred dollars to the wire bandits for a costly interconnect from preamp to power amp. An integrated saves space, too.

With a single-chassis integrated, your preamp and power amp are at the same ground potential, which could be beneficial, reducing noise and thus improving dynamic range. Not without reason do some inexpensive British integrateds—those from Creek and LFD Mistral, for instance—leave some listeners gobsmacked as much for what they don't do as for what they do do.

Furthermore, an integrated amplifier can be designed to eliminate the active preamp section entirely, by building enough gain into the power amp's input stage and appropriately tweaking the input sensitivity. No need to take into account the interconnect from preamp to power amp: there is none. No need to allow for different volume controls: the volume control is a known quantity.

Conrad-Johnson is taking this approach in its forthcoming (maybe at your dealer now) Premier 150CA solid-state "control amplifier," rated at 150Wpc and expected to retail for around $7000. All the gain comes from the power-amp circuit. The Premier 150CA is a power amp, with switching and volume controls. (The term "integrated amplifier" may be a misnomer when there are not two amplifiers to integrate.)

"Gosh, Lew, I just saved $5500 by not buying a Premier 17LS2. That means my Premier 150CA cost me only a marginal $1500." Or thereabouts. "Isn't this the cost-effective way to go?"

"You're right," laughed the former economist. "The new Premier 150CA will be a cost-effective solution for many customers. But, as you know, there are those who will stick with separates, for various reasons. Maybe they love their tubed monoblock amplifiers, for instance. Besides, in practical terms, you're limited to how big you can build an integrated amp. I don't see making an integrated as big and powerful as our Premier 350 power amp, with 350Wpc. So some people will still want separates."

The Premier 17LS2 looks identical to its predecessor—same chassis, same styling. It offers five inputs with two "external processor loops," which are twin sets of inputs/outputs for use with a tape deck or external processor. Intriguingly, you can use the 17LS2 as a purist two-channel preamp for listening to music through your left and right speakers; and you can use it in a surround system, at the push of a button, when you watch movies in six channels, or whatever.

It's easy. You connect your DVD player's left- and right-channel audio outputs to the 17LS2. When you want to play movies, you engage one of the external processor loops. This sends a unity-gain signal to your surround-sound processor. You control the volume via your surround-sound processor. Meanwhile, your left- and right-channel signals go through the 17LS2. When you want to listen to music again—presumably in two channels—you just change the 17LS2's input to CD and the external processor loop automatically disengages.

The 17LS2's level control has more than 100 settings and, like the input selector, it operates via microprocessor-controlled relays. A network of precision, low-noise, metal-foil resistors adjusts the gain and the balance control. All functions are accessible from the very substantial remote, including the balance control, which is not addressable from the 17LS2's front panel. About the only thing missing is a mono switch. All inputs and outputs are single-ended RCA. There are two sets of RCA outputs.

Lew described the 17LS2's circuit as a single-ended class-A triode circuit. "There's no circuit simpler than a triode," he observed. "There is no additional cathode follower and there's no separate buffer stage. The low output impedance and the buffering are provided by the preamplifier's voltage amplifier itself.

"We created a super triode—or a composite triode—by paralleling multiple triodes from the four Sovtek 6922 tubes. Each of the tubes is a twin triode. Four triodes are paralleled for the left channel and four triodes are paralleled for the right. They are put together in parallel to make one big triode. In effect, you have a single triode per channel."

Why doesn't everybody do this?

"For a couple of reasons. First, this is a zero-feedback circuit. That's part of why the preamplifier has the dynamic shading that it does. The price you pay for this is that performance becomes critically dependent on the component parts."

There is no margin for error.

"One of the things that feedback does for you is diminish the dependence of the circuit on each of the parts in the circuit."

Feedback covers up crimes!

"Yes, it does," Lew laughed. "Can I use that? Feedback covers up crimes in the sense that feedback reduces distortion, noise, and dependence on each resistor and capacitor in the circuit. Ultimately, the performance of the circuit depends on two resistors, and everything else is corrected, in theory, by the feedback.

"Without feedback, every part of the circuit affects the performance of the whole circuit. You can hear differences when you change each part. The higher the quality of the parts, the less feedback you need. If you use very-high-quality parts, you don't need any feedback at all. Every part has to be carefully selected, including each of the four Sovtek 6922 double triodes."

"Sovtek," by the way, usually means the Reflektor factory in Seratov, Russia. In the 17LS2, these tubes sit inside a cutaway compartment recessed into the front panel. The compartment is covered by a removable grille with a waffle open front and a perforated solid top. It's possible that the unit sounds better with the waffle grille removed. At any rate, the top of the grille gets rather warm, meaning the tubes will run cooler when they're not being...ah, grilled.

"The power supply is especially critical," Lew continued. "The power supply is in the signal path and the circuit itself provides no rejection of power-supply noise."

What does reject the noise? Those expensive 2µF capacitors.

"Capacitors do two things," Lew explained. "They store energy and they block DC voltage." C-J uses capacitors in both capacities. "The output coupling capacitors take the signal from the plate of the output tubes and block the DC voltage because your amplifier doesn't want to see DC voltage. Yet you want the capacitor to pass low frequencies. The larger the value of the capacitor, the lower the frequency it will pass. We use two of those 2µF capacitors per channel, in parallel.

"The other 2µF capacitors are used in the regulated power supplies. One is used in the preregulator, or the input power supply. Two others are used in the band-regulated power supplies for each channel. These regulated supplies isolate the channels from each other and isolate everything from the AC line. The outputs of these power supplies also have an energy-storage reservoir—again using 2µF capacitors."

"So, Lew, why do you need a line stage at all?"

"Primarily, to buffer the volume control from the cable that connects the preamp to the power amp."


I love to play the technological illiterate, which I am. In one ear, out the other. I have to write it down fast.

"A volume control—any volume control—is, by nature, a relatively high-output-impedance device. All it is is a variable resistor. If you chose a volume control that didn't have a high output impedance, then it would present too heavy a load for the source that's driving it, such as a CD player. Your CD player probably doesn't want to see a source impedance below 10 to 15k ohms. If you use a 10 to 15k ohm potentiometer, then you might have a 5k ohm source impedance when you actually set your volume control. This forms a low-pass filter with the interconnect cable, rolling off the high frequencies."

It's not just the highs you might lose, it's the dynamics. This is what I meant earlier when I said that a passive device might not add anything to the signal, but it could subtract. The other reason you might need a preamp is that you need the extra gain to drive your power amplifier to full output.

"Even if the high-frequency filtering effect is inconsequential," Lew said, "you can create problems loading down the volume control when the input impedance of the amplifier is low relative to the output impedance of the source."

I used the Premier 17LS2 mainly with C-J's Premier 350 amplifier driving a pair of Triangle Celius ES loudspeakers (new version). I compared it with my Purest Sound Systems P500 "passive preamp." All interconnects were relatively short. CD came from various sources. I also used my reference Parasound JC-1 monoblock amplifiers, which don't like tubed line stages that pass even minute amounts of DC. The C-J Premier 17LS2 got on famously with the Parasounds.

I was immediately impressed by what the 17LS2 did, compared with the original 17LS. I didn't have them side by side, or even that close in time, and the systems were different—all I can say is that the 17LS2 hit me in a way that the original 17LS, good as it was, didn't quite manage to do.

Compared to "passive preamping," I heard, from various CD sources, improved dynamics, both macro and micro. The music flowed better, played with a greater sense of ease. Harmonics were richer, fuller, less electronic. I was, in a word, seduced.

So this is what the Conrad-Johnson ART and ACT2 are about.

Well, yes, apparently. As I said, I'm almost afraid to audition the ACT2 for fear I'd have to buy it and trade in my wife. Never before, in my own system, have I heard a line stage make so much difference. The Premier 17LS2 seemed to be doing so much. I couldn't resist telling Lew Johnson.

"I would say that it's not doing things, but rather the Premier 17LS2 is not getting in the way. I don't think it's adding dynamics, for instance; it's not impeding them. It allows things to happen. It has the ability to pass along fine variations in dynamics."

This is the "microdynamics" stuff I was talking about. It's subtle, but, along with spatial resolution, it's one of the things that can make reproduced music almost literally come alive.

Like the 17LS, the Premier 17LS2 inverts phase or polarity. If your power amp doesn't invert phase, you'll need to reverse your speaker leads at the power amp or at the speakers. The 17LS2 comes with a spiffy power cord that was made for medical equipment.

"I don't like to admit it, but it affected the sound quality," Lew said.

At least you won't need to buy a power cord from the wire bandits.—Sam Tellig

2733 Merrilee Drive
Fairfax, VA 22031
(703) 698-8581