This review page is supported in part by the sponsors whose ad banners are displayed below

Where the SIT transcends tubes again is at the frequency extremes. Whether it's due to greater bandwidth and thus phase linearity, lower distortion, lower noise or the lot I don't know. But if you believe that a direct-heated triode owns treble purity, meet this SIT. Triode treble will sound comparatively veiled. In the bass the advantage seems to be a function of raw current. Here the transistors outplay the bottles. This improves articulation and grip and includes wiriness or steeliness when required. This short overview of the stereo SIT provides us with a general roadmap. It also lists certain highlights which deserve closer scrutiny. While this SIT platform does bridge transistors and tubes, it doesn't clone the latter. It remains distinctive. Depending on your triggers and their sequence and what music you fancy, you still might prefer valves.

For art music—the voice + piano or girl with guitar archetype—triodes really are eminently suited. Transition to a Flamenco ensemble against full symphonic backdrop however [Juan Carmona, Sinfonia Flamenca, Harmonia Mundi] and the SIT should retain superior sorting, clarity and energetic projection. I still own four valve amps. Two of them are 300B SETs. Since the review of the F5 I've listened mostly to FirstWatt transistors. Where the J2 was a refinement over the F5—less brash, more suave—and as a power JFET a quasi preview of the static induction transistor, the SIT-2's climb up the FirstWatt ladder has clearly skipped a few rungs. Its difference from the precursors is greater. How would the SIT-1 push that envelope?  

SIT-1 20kHz square wave

"The SIT device allows the possibility of arbitrarily deciding on the specific harmonic structure of the amplifier's distortion by setting the relative values of 2nd or 3rd-order harmonics. The knob on the meter is useful for a number of things. As a concept you will probably find that the area to the left of the center of the green zone is best for low-impedance loads or when you are looking for a richer 2nd harmonic character. If the load impedance is greater than 8 ohms, you will tend to put the meter needle on the right side of the green zone. Both SIT-1 and SIT-2 can be driven in 'direct mode' where the source component looks directly at the gate of the SIT without other active components in-between This is desirable but in that case the input impedance is 10Kohms plus the capacitance of the gate which is greater than the 1.000pF of the IHF standard. A preamp or source component with an output impedance of 1.000 ohms or less should have no problem. Passive preamps and some tube designs with very high output impedances however will limit bandwidth which could be a performance issue. To address this there is the option of a simple P-channel Jfet follower on the input. This raises the input impedance to 100.000 ohms and about 10pF. Anything will drive this without problems. The distortion of the P-channel Jfet is incredibly low and doesn't alter the sound in any significant way but assures no impedance interaction and a bandwidth to 500kHz."

"I think you will be surprised by the adjustability of the sound on the monos. They don't have any more power but the effect is worth the price." On how this entire SIT adventure began: "A couple of years ago I was talking to Jeff Cassidy at SemiSouth. He mentioned that at one time they had made a special run of static induction transistors on some kind of military/industrial contract; and that one of their technical people had remarked that they were nearly ideal for use in audio amplifiers. 'Really', says I, 'Do you have any of them left over?' No, the customer had scooped them all up but they would be interested in making more. The price of a small run was astronomical—at least to me—and I spent quite a bit of time pondering the risk. Then I wrote the check and didn't look back."...

Shifts. During the late 90s, the fascinating concept of shiftable 2nd and 3rd-order harmonic weighting in power amplifiers had perhaps been unwittingly exploited already by Randall Smith's very fresh Mesa Engineering Baron. Its cleverly trademarked TandemState Imaging™ was marketing speak for switchable triode pentode operation plus adjustable NFB. The three pairs of 5881 per channel could be individually set to run in triode or pentode. This allowed for previously seen all triode or pentode operation but also a very novel two-thirds mix either way. That obviously changed power output whilst the feedback control added altered output impedance. Unlike the SIT-1, the Baron thus did not isolate its THD shifting feature. It joined other variables and got mixed in with their unpredictable interactive effects.
In 2011 the Audiopax Maggiore 100 monaural amps from Brazilian auteur Eduardo de Lima introduced six single-ended triode-strapped KT88 amps per channel where two paralleled banks of three amplifiers are connected in series at the outputs. Amp duos one and two run pair-matched output transformers which differ between the pairs. Amps 3 and 4 run dissimilar transformers which are additionally different from pairs 1 and 2. This makes for four distinct OPT versions per channel.

The six bias controls are referred to as TimbreLock™. They manipulate the amp's THD behavior. Via load-line shifts these adjustments manipulate the distortion interaction with the speakers. The recommendation is to have the LED-set bias of the two matched pairs track each other. The values for the central pair as the wild cards meanwhile may be deliberately offset to produce the desired 'kinky' effect.

Versus FirstWatt's simple variable resistor behind the bias control and an ultra-minimal power circuit behind that, Audiopax's need for output transformers plus their ambitious brief for 100 watts of pure class A1 triode power without global feedback required a wildly more complex approach. Now add Swiss manufacture with its exclusive €60.000/pr sticker.

Window of THD shift between allowable lowest and highest bias values
for two standard load impedances

These two isolated examples redefine Nelson's appreciation for simplicity by stark contrast. Relative to the above graphs; relative to the clever TandemState Imaging and TimbreLock suggestions of audible soundstage and tone color effects; and mindful of the fact that descriptive depth eludes any catchy term on principle - how exactly would the SIT-1's particular bias adjustments manifest? That after all is the primary attraction over the SIT-2 (though having discrete power supplies per channel is part of the mono equation). I honestly didn't quite know what to expect. And that's the most exciting thing about assignments for which one has no real precedent. Live and learn. Or is that learn and live?