VTL 100W Compact monoblock power amplifier
Nowadays, musical instrument amplification is a whole new, apparently Japan-dominated world, with MIDI this and FX boxes that, and, apart from classics like the Fender Twin Reverb which soldiers on regardless, there is hardly a tube to be seen. Yet to this baby-boomer, it seemed that the whole British music explosion of the '60s was powered by just one thing: the EL34 high-power pentode.
With the exception of the Fender, which I believe used 6L6s, and the Vox AC30, which used EL84s, it was the EL34-powered Marshall and Vox guitar amps that fueled the Beatles (till they were seduced by transistors around the time of Rubber Soul), the Stones, the Who, the Animals, Hendrix, and Cream. Take my Vox AC100, to which I had graduated after using first a Marshall 50, then a Marshall 100. Packing four push-pull EL34s into a tiny chassis to put out 100W, it ran hot enough to fry eggs on, yet, apart from having to replace tubes every six months or so, when they became sufficiently microphonic to reproduce your voice if you shouted loud enough (and we all had to shout loud enough when the average PA power was only 100W), it was the most reliable and the most musical amplifier I have ever used.
I gave up playing the bass professionally when I joined Hi-Fi News & Record Review as what was in effect the office junior in 1976. Though I continued to play regularly, the one career inexorably took over from the other and the Vox, long since retired, disappeared into the bowels of my mother's garage when I moved to the States. The hankering for tube sound stayed with me, however, despite an almost total dependence on transistors for playing records, and when Vacuum Tube Logic's David Manley offered to lend me a pair of his 100W Compact Monoblocks, I didn't need to be asked twice.
The "Compact" series of monoblock tube amplifiers from VTL, available with 50, 65, and 100W output ratings, differs from the De Luxe models, such as the 300W Mono reviewed by J. Gordon Holt last month, in that they have somewhat smaller power supplies and lack the more expensive models' regulated B+ supply for the input tubes. Otherwise the circuitry is very similar. The input signal is taken to the grids of both halves of a Sylvania 12AT7 twin triode running in class-A, the output being taken from the two plates connected together. Though the voltage gain remains unchanged with this arrangement, the transconductance doubles (doubling the change in plate current for the same grid voltage change), lowering the noise level and halving the tube's output impedance. The amplified signal is then fed via a series capacitor to another 12AT7 (this the equivalent 6201 tube in my samples), the two halves of which operate as a long-tailedpair phase-splitter; ie, this stage outputs identical in-phase and antiphase signals from the two plates. These are taken, again via coupling capacitors, to the output stage, which consists of four EL34s, two "pushing" and two "pulling" the primary of the output transformer.
David Manley echoes my old Vox in opting for "fixed-bias" operation for his output stage, whereby a separate winding on the power transformer is used to derive an independent DC voltage supply to maintain the output tubes' grids at a negative potential with respect to their cathodes. (The alternative is to stand the cathodes on resistors, the so-called "cathode-" or "self-bias.") According to Manley in his book The Vacuum Tube Logic Book (footnote 1), the disadvantages of fixed biasinga slightly increased circuit complexity and a need for a heftier B+ power supplyare more than outweighed by its giving greater power with lower distortion from the same tube lineup, coupled with its bestowal of longer life on those tubes.
There is another downside to fixed bias, only obliquely referred to by Manley, in that if ever the separate bias supply failsvery unlikelyor is cancelled out by, say, very low frequency signals, the control grid rises to ground potential, becoming positive with respect to the cathode, and the tube first glows cherry red, then fails, perhaps with catastrophic effects on the rest of the circuitry.
With traditional high-power amplifiers running with fixed bias, such as the classic GEC design handbook's KT88-based 100W design (footnote 2), it was recommended that a cathode resistor be automatically switched into circuit if the bias supply went down, allowing the output tubes to stay alive. The VTL designs have no such provision, though to be fair, neither do any other modern fixed-bias amplifiers of which I am aware. However, it explains why the VTL warranty only covers tubes if a VTL preamplifier is used. If the preamp develops DC on its outputpossible with a servo-type, direct-coupled, solid-state designor very-low-frequency oscillationsas can happen with a regulated single-rail tube preamp when the regulation starts to fail (footnote 3)the pulse from the former and the AC from the latter will be passed to the VTL's output stage by the design's extended LF response and will swamp the fixed bias supply, leading to tube failure.
The EL34s in the VTL 100W Compact Mono each have a standing bias of between 26 and 29mA, which should be sufficient to minimize any crossover distortion, the tubes running then in what is termed class-AB1 mode (the "1" meaning that no grid current flows). VTL actually refers to it as class-A1, meaning that for a significant proportion of the amp's dynamic range (said to be 66%), the tubes are effectively working in class-A. The bias measurement for each tube is via color-coded sockets on the front panel.
The EL34 being a pentode, the designer has a choice about the biasing of its screen grid: either to connect it to the B+ voltage to give conventional pentode operation (footnote 4), or to use the "Ultralinear" connection thought up by David Hafler and H.I. Keroes in 1951. (Manley points out, however, that this was predated by D.T.N. Williamson's use in the UK of the very similar "partial triode" operation in 1948.) In this topology, the screen grid is connected to a tap on the output transformer primary winding somewhere between the appropriate end, where the plate is connected, and the midway point, which is connected to the HT supply. When this screen tap is between 30% and 45% of the way toward the center, the tube's output impedance is lowered compared with pentode operation, enabling it to more easily hang on to the output transformer, and the output power is roughly doubled compared with triode operation. The penalty is said to be a slight rise in distortion unless the two tubes pushing and pulling the transformer primary are carefully matched.
Footnote 1: 1988 Price $10, including postage and handling, from the address given in the Specifications Sidebar.
Footnote 2: GEC, now one of Europe's largest defense electronics companies, was the parent company of the now defunct M-O Valve Company, manufacturers of the classic KT series of beam pentodes. Co-authored by MO-V's Graham Woodville, published in 1957, and entitled An Approach to Audio Frequency Amplifier Design, this slim volume was the bible for those who participated in the first UK tube revival about 10 years ago.
Footnote 3: This happened with my SP10, supposedly contributing to the failure of a Krell KSA-100 power amplifier. Unlike a conventional dual-rail solid-state preamp, where such fluctuations on the voltage rails will have a minimal effect on the audio signal due to the circuit's common-mode rejection, a single-rail design can't help but pass them on to the output.
Footnote 4: Audio Research uses a variant on this mode in their all-tube designs, whose screen grids are connected to separate, regulated HT supplies. Their recent hybrids, however, go a different route altogether, though one which faintly echoes the Ultralinear topology, by connecting the screen grids to the outputs of power MOSFETs carrying the audio signal floating on HT.