This review page is supported in part by the sponsor whose ad is displayed above
Reviewer: Srajan Ebaen
Source: Zanden Audio Model 2000P/5000S
Preamp/Integrated: ModWright SWL 9.0SE; Bel Canto Design PRe2; Eastern Electric MiniMax; AudioZone PRE-T1 silver & copper [on review]
Amp: Decware Zen Taboo; 2 x AudioSector Patek SE; Canary Audio CA-308 [on review]; Fi WE421A [on review]
Speakers: Zu Cable Definition Mk1.5 with new external bass attenuator (retro-fitted)
Cables: Zanden Audio proprietary I²S digital cable; Stealth Audio Indra; Zu Cable Varial [on review]; Cerious Technologies interconnect [on review]; Zu Cable Ibis; Crystal Cable Reference power cords; Z-Cable Reference Cyclone power cords on both powerline conditioner
Stands: Grand Prix Audio Monaco four-tier
Powerline conditioning: 2 x Walker Audio Velocitor S
Sundry accessories: GPA Formula Carbon/Kevlar shelf for transport; GPA Apex footers underneath DAC and preamp; Walker Audio SST on all connections; Walker Audio Vivid CD cleaner; Furutech RD-2 CD demagnetizer; WorldPower cryo'd Hubbell and IsoClean wall sockets
Room size: 30' w x 18' d x 10' h [sloping ceiling] in long-wall setup in one half, with open adjoining living room for a total of ca.1000 squ.ft floor plan and significant 'active' cubic air volume of essentially the entire (small) house
Review component retail: ca. $,8000, details TBA
 

When, more than half a century past their introduction, a modern tube amp manages to secure a US patent covering two different items; when said amp offers 600 watts of peak power and is claimed to reduce distortion as load behavior worsens into territories usually off-limits to valves; when this amp is rumored to have had SACD mastering engineers at Telarc pronounce it hands-down superior to big Brystons and Krells in the 30-90Hz bass range; when this amp's forest of tubes doesn't exceed four robust KT88s per channel on a stereo chassis; when the asking price for this beast clocks in well below $10K... well, then my ears prick up despite being a lower-power aficionado of the SET and SEP persuasion. When a reader introduces me to this amp in the first place, notions of synchronicity arise to suggest more than coincidence. Curiosity and responsibility intersect. Clearly, a phone call to the designer is in order.

After talking to Terry Tekushan, it turns out that I have, in fact, seen this amplifier before. I've even taken its photo [left] - at the HE 2003 show in San Francisco. It enjoyed "forthcoming" status at InnerSound, then still operated by founder Roger Sanders. Quoting from my show report: "InnerSound of electrostatic speaker fame launched its iTube CCDA amplifier which, naturally, is fully endowed to deal with dastardly low-impedance reactive loads such as are presented by electrostatic loudspeakers. Duh. At 150 watts into 8/4 ohms, this amplifier uses massive custom-designed output transformers, fully differential circuitry and a servo offset bias control to avoid the negative feedback effect of cathode-resistor auto bias (often reducing transformer capacity by up to 1/3). This servo is said to eliminate DC offset on the primary altogether. It also constantly monitors and compensates for bias drift to 1/10th of a
millivolt, making the need for matched pairs a thing of the past. A two-stage soft-start circuit and quads of KT88 or 6550 per channel complete the basic specs." I even remember the teaser ads in Stereophile. A tube amp designed specifically for punishing electrostatic loads. A mysterious new circuit apparently in defiance of convention. Something novel in a category that tends to excel at repackaging yesteryear's circuits.


Alas, this white elephant never materializes. Or so I believe. What really happens is simple. A subsequent takeover and change of ownership at InnerSound compels Tekushan to rethink his pending collaboration. He decides instead to launch his proprietary design under its own brand. Voila, Western Reserve Audio Design from Lakewood/Ohio. A simple genesis for what clearly isn't a simple design. Actually, call it design simplicity that like all strokes of brilliance turns out to be simple to an extent that fails to explain why nobody has thought of it before. The patent confirms that if someone did ponder it earlier, they never managed to work out the gnarly gremlins to advance from theory to practice. Nobody has authored a fully functioning, reliable and fairly priced commercial product that incorporates this particular circuit. Until now. Drum rolls, please. After all, a drive-anything tube amp in a stereo chassis that limits the paralleling of output devices to quads should be welcome news to the power players among audiophiles - owners of Thiels, Aerials, Soundlabs, Dynaudios and other speaker designs that insist on low-impedance stability and stiff reserves. Perhaps the Western Reserve name is more than just a nod to its Ohio roots? How to test this particular load invariant feature with my friendly in-house transducer chappies would of course pose a challenge. Unless I perhaps wired two pairs of Gallo References to the amp in parallel to halve their impedance and then rock the Casbah?


Excerpted from Terry's White Paper, here are the relevant details of the WRAD 300. "The 300's first amplification stage handles most of the voltage gain and phase inversion through common cathode coupling with an essentially grounded lower grid. It consists of a push-pull differential pentode pair (two EF86s per channel with helically wound filaments for AC hum rejection), fully voltage and current regulated through an unconventional (though deceptively simple) cascaded transconductance tube regulator design. This constant current differential topology offers tremendous voltage swings, wide bandwidth and 6dB of headroom. Together with the careful selection of the capacitor composition and the inherent constant current nature of pentode tubes, this results in a circuit that is linear, low distortion, perfectly balanced and free of the harshness often associated with pentodes that are carelessly applied. The cathodes of the two pentodes are direct-coupled and returned through high resistance to a regulated negative signal ground source. The lower grid is essentially grounded, allowing the input signal to communicate with the lower tube. More importantly, it allows the voltage feedback from the output terminal of the amplifier to communicate with the circuit in exactly the same way as the input signal. This improves distortion and stability particularly with reactive speaker loads."


"The driver stage consists of two 12BH7 dual triodes operating as a push-pull long-tailed pair for maximum voltage output, gain and driver balance. The grids of this stage are direct-coupled to the respective plates of the pentode differential amplifier. The push-pull triode sections are individually cathode biased to assure linear operation regardless of the matching of either the EF86s or the individual sections of the 12BH7. They are AC coupled to one another rather than driven direct. The output of the driver stage plate circuits are then capacitively coupled to the output tube grid circuits in the conventional manner."

"The output stage consists of two pairs of KT88s configured in push-pull parallel ultralinear class AB1. The tubes operate below the class A operating current but are not cut off as in class B operation. At no point does current flow through the grid circuits. The output tubes are operated in fixed bias, i.e with a negative grid supply for current control. One pair remains fixed, the other pair is variable through the automatic action of the servo offset control until perfect DC balance is attained. The tubes are referenced to ground through a maximum current limiting resistor which also serves to develop the reference voltages for the servo circuit. The audio signal component of the output stage is referenced to ground through the cathode/current limiting resistors but also between the push-pull phases of the output tubes. This allows for maximum gain from the power stage, aids in maintaining signal balance, further rejects noise and distortion and limits the maximum current in the output tubes in the event of excessive grid leakage in a power tube. The output transformers are custom designed over-built 500-watt units (24 lbs ea.) conservatively rated to minimize distortion and maximize power delivery particularly in the low frequencies. Secondary connections are available for both 4 and 8 ohms."


"The servo offset control is essentially an error correction amplifier using the cathode current of one pair of output tubes as reference. It automatically adjusts the output tube bias to ensure the most linear operation of both the output tubes and the audio output transformers. This circuit monitors the current flowing through one set of tubes and biases the other set to a current equal to the first. The circuit itself is voltage and current protected to allow for a lifetime of reliable service even in the unlikely event of a catastrophic tube failure. The circuit is sonically transparent and improves the long-term performance of the amp (due to the aging of the tubes and periodic replacement) and in the short term (due to the drifting of tube currents during warm-up and day-to-day variations that inevitably occur). The circuit's main elements consist of a dual operational amplifier that compares the output tube voltages monitored from the cathode circuits. The op-amp then acts on the base of a transistor that changes the bias voltage on the tubes whose currents will be matched to the non-variable or reference tubes. The process is continuous but damped so it doesn't attempt to react to transient music signals. This circuit is perfect for the tube amp owner who is not inclined to frequently monitor output tube balance (many amps do not even give you this provision and hope matched tubes will be used and stay matched). It is also perfect for the audio buff who checks these things frequently but is always fighting that nagging feeling that something may be a little out of adjustment. This circuit essentially eliminates any magnetization (net current flow is zero) of the ultralinear output transformers, allowing them to respond only to the actual audio signal rather than any distortions created by unequal current draws from each end of the two halves of the transformers' primaries. It also compensates for unequal tube aging, mismatched tubes and bias drift as a function of thermal behavior during a single listening session. The servo ICs are protected against over current and also employ an over-voltage clamping circuit in the event of outright tube failure."
"The 300 uses a modest amount of global feedback to improve linearity and speaker damping. Due to the linearity of the circuitry and transformer characteristics, only 12dB of feedback is applied, among the lowest of any push-pull amplifier extant. High levels of feedback can give amplifier performance figures that look very good on paper but in practice give unpredictable results on difficult speaker loads and output levels that induce clipping of the amplifier. Once clipping occurs, feedback becomes useless at best or destructive to sound quality and stability at worst. The feedback network consists of one simple resistance from the output transformer secondary winding (speaker output) to the lower grid of the differential amplifier. In this manner, the feedback is taken to a point of equal impedance as the input signal - a mirror image in fact... High frequency or phase compensation is not used in the feedback loop at all. The reason is simple: while feedback phase compensation looks good on test instruments, a real speaker with inductance, capacitance and speaker cables can reflect information back to the amplifier in unpredictable ways that may include high frequency information that is shifted enough in phase that the negative feedback as compensated may begin to look like positive feedback, creating momentary
instability or oscillation. The 300's feedback circuit is simply not affected by such events. High frequency compensation occurs at an input pad before the first stage and a mild Miller network in the output stage"


"The 300's power supply is a low impedance type with a custom-designed power transformer whose winding technique gives excellent inherent regulation without the common hysteresis and power sag effects. This eliminates the need for extremely large capacitors and their lengthy charge times. The requisite filter storage then is determined by the effective impedance of the transformer, the idle current of the circuit and the total power consumed under full-load output conditions, a key factor why this amp's modest Class AB1 idle current (appr. 30 milliamps) provides completely unfettered dynamics. The power transformer can deliver very high peak current without appreciable sagging of the high voltage supply. The output stage pulls its power from the wall outlet more effectively."

"The power supply makes extensive use of non-polar polypropylene filter caps to improve high frequency RF and crosstalk rejection over electrolytics. The sensitive EF86 pentode plate supply eschews electrolytic filtering altogether and is controlled by a single 6DJ8 as a passive current-mode cascaded voltage regulator. (One section of this dual triode supplies the plate circuits of the EF86s, the other their screen grids). A soft-start circuit begins with stand-by, then charges the capacitors and finally applies full high voltage to minimize stress on the tubes, rectifiers, capacitors and relay contacts. In the case of power interruptions, the amp automatically reverts to standby, preventing hot restart trauma."


"In conclusion, the benefits of the regulator circuit in conjunction with the pentodes are as follows: perfect phase inverter output symmetry with equal plate load resistances; low distortion, wide bandwidth; and excellent transient response. Feedback affects both phases of the inverter output equally for greater overall stability. The resultant tight control of the EF86 characteristics makes the sound less susceptible than an unregulated circuit when vintage Amperex EF86s are swapped for, say Teslas or Sovteks."


But there's more still - a unique power factor correction/damper circuit in the output stage related to delta circuits in 3-phase AC circuits. The damper circuit consists of series cathode resistors from each phase to ground (the cathodes being capacitor-coupled) and communicates to the load through the output transformer and through the transconductance of the power tubes to their respective coupled cathodes. This is said to improve headroom, clipping symmetry and power delivery into off-impedance loads. This output load correction network is also responsible for the low distortion into highly reactive loads. This damper circuits appears twice in each channel's circuit and in modified form in the Class A cathode-biased driver stage. The WRAD 300 delivers 150wpc into 8/4 ohms and offers 600 watts peak. Distortion remains below 1% at full output. Frequency response is given as 10Hz - 40KHz +/- 1dB. Input impedance is 220K, power consumption at idle 275 watts or 925 BTUs. Dimensions are 18" wide x 8.5" tall x 17" deep and weight is a colossal 98 pounds. What's it sound like? That, my friend, we shall both know soon enough. The amp's on the truck to Taos ...
PS: Trading anecdotes of tube amp R&D (I have a few from my days at Mesa), Terry sent me the following e-mail. He's always had the utmost respect for Roger Sanders who, at the beginning of their interactions about Terry's new amp, wasn't yet fully in agreement with a "tube amp as a legitimate high performance modern product". That's easy to understand considering Sanders designed reactive electrostatic speakers that, in general, eliminate tube amps from the equation in more ways than one. [The only other tube amp I know that for example can drive the big SoundLabs is the Wolcott - Ed.]:


"One prototype was given the long-term test. It played music 24/7 for 6 months. No excess heat, no problems. Other tests included the intermittent power outage test, the surge test, the brown-out test, the unstable voltage test, the open output test, the shorted output test and the small child "it's fun to turn it on and off" test. Unfortunately a few of these tests were inadvertent but it passed all of them.


My favorite was the "arc & burn" Innersound test. Roger had sent a pair of very early production series I Eros. Actually, I bought them. Roger said to me, "Well, Terry, you can't blow those panels, especially with a tuuuube amp." He always put tube amp in a somewhat derisive tone. Well, I had one of the amps running the top (the ESL part) and was playing music entirely too loud - but I had a point to prove. I went to the next room to vacuum the rug (it was way too loud in the same room) and before I knew it, there was a terrible distortion to be heard. Being paranoid, I thought, "Crap. I blew something on the amp." This was fairly early in the development process. I ran into the room to see the wall behind the left speaker lit with an unsteady blue light, smoke rising from the speaker. The insulation on the edges of the metal stators had broken down and was allowing a sustained area of arcing to develop. It was quite violent. And smelly. The amp was playing through this whole mess like nothing was happening.


I called Roger who was unconcerned about the panel. He asked, "What did the amp do?" "It played through it." "Did it blow its fuse?" "No..." "Did the tubes arc or glow funny?" "Nope." "Did it get hot or smell?" "Nothing. Everything was normal." Roger was impressed. This wouldn't be the last time Roger would be pleasantly surprised by some aspect of this amp's performance. You see, Roger was never a tube amp guy. As a matter of fact, he was hostile toward them. These things would happen and he would say "very unusual." I would hang up the phone after one of these conversations and say, "Phhhht. So there!" Infantile, I know, but it seemed satisfying at the time." I'd say it's still satisfying to owners of hard-to-drive speakers today, to learn about a tube amp that was specifically designed for them.