Official Site of Pritchard Amps

Audio Myths and Truths

Throughout the research that developed Pritchard amplifiers, Pritchard had to evaluate two areas accepted audio knowledge and accepted engineering knowledge. Engineering has failed to produce amplifiers acceptable to musicians. Pritchard amplifiers work fabulously because they have proprietary and patented design features that are unblemished by incorrect notions of accepted common knowledge. Pritchard has found and patented solid state circuitry that exaggerates the nature of tubes and their circuits.

Common audio knowledge itself is often based on mythical misinterpretations of facts that are poorly presented and overly simplified. Even when concepts are properly presented, common knowledge over simplifies because it cares little for such minor inconveniences as being completely correct. On the other hand common knowledge is occasionally accurate. Consider the following examples:

The common knowledge of the difference between tubes and transistors is simple: Tubes produce even harmonics; transistors produce odds; and even harmonics are better than odds. These concepts are probably incorrectly derived from Russell Hamm’s famous paper “Tubes Versus Transistors - Is There An Audible Difference?” (Journal of the Audio Engineering Society, May 1973), which may be found on the following link ( http://www.gprime.com/proaudio/tubes/tubes.htm ). There are two reasons for these mistaken impressions, insufficient depth in reading and insufficient disclosure in this famous paper.

“Tubes produce even harmonics.” may come from section titled “1. Tube Characteristics.” in “Tubes Versus Transistors - Is There An Audible Difference?”. The statement is “The outstanding characteristic is the dominance of the second harmonic..” This point is driven home further by the Figure 4 which shows a dominate second harmonic and a fourth harmonic becoming larger than a diminishing third harmonic. However, further reading at the end of the same section and paragraph finds that triodes can have the same third harmonic dominate nature as pentodes (probably due to differences in grid-cathode bias). Consequently, we can deduce that “Tubes produce even harmonics.” is false or at least misleading because tubes produce all of the harmonics. This is stated clearly approximately two-thirds of the way through this section: “However, the major characteristic of the tube amplifier is the presence of strong second and third harmonics, sometimes in concert with the fourth and fifth, but always much greater in amplitude. Harmonics higher than the fifth are not significant until the overload is beyond 12 dB. These characteristics seem to hold true for wide variations in circuit design parameters.”

The possible dominance of the third harmonic in a tube circuit puts the common knowledge “The difference between tubes and transistors is that tubes produce even harmonics while transistors produce odd harmonics.” into the realm of great myths.

There is also an unstated assumption within “Tubes produce even harmonics”. It is that the tube behavior is independent of the circuit it is used in. The best circuit counter example to this notion is the push-pull stage, which more or less cancels even harmonics and augments the odds. A perfectly balanced push-pull vacuum tube stage will produce no even harmonics. Thus, circuitry surrounding the tube under test is quite important, crucial to be more precise.

“Transistors produce odd harmonics.” may come from section 2 of “Tubes Versus Transistors”: “The distinguishing feature is the strong third harmonic component.” This point is further emphasized by Figures 8 and 10. However, there are two points of issue. First, the following statement is “All harmonics are present, but at a much lower amplitude than the third. When the overload reaches a break point, all the higher harmonics begin to rise simultaneously.” The second statement, however, brings up another issue from Hamm’s paper: “Both amplifiers shown have single-ended inputs and push-pull outputs.”

The last stage of an amplifier handles the largest signals and consequently determines much of the amplifier’s distortion characteristic. Push-pull amplifiers, as noted above, are well-known for their cancellation of even harmonics. However, there is an unresolved problem because the push-pull may be unipolar with an output transformer as in tube amplifiers or may be bipolar and emitter follower. In the latter case, the output nature is more determined by the next to the last stage. Unfortunately, the paper provides no information on the amplifier types or specifics on their circuits.

“Transistors produce odd harmonics.” is effectively countered by Pritchard’s work published as “The Tube Sound and Tube Emulators” in dB Magazine where he disclosed a tube emulator microphone preamplifier that produces the same, or at least a very similar, harmonic structure as the triode microphone amplifier depicted in Figure 4 of Hamm’s “Tubes Versus Transistors”. This counter example shows that even harmonic production is not limited to tubes and solid state circuits can produce even harmonics as well as odd harmonics. In fact, this paper discloses Pritchard’s solid state tube emulator microphone amplifier that has approximately the same harmonic structure as Hamm’s Figure 4 two-stage triode amplifier.

“Transistors produce odd harmonics.” is also countered to the trained eye by the patented XGPA “plate” characteristics. The distances along the diagonal “load line” between the “plate” curves are not constant, but increasing from right to left. This creates even harmonics.

“Even harmonics are better than odd harmonics.” probably comes from the description of the harmonics in Russell Hamm’s “Tubes Versus Transistors”. The second harmonic is described as adding “body” and “punch”, while the third is “covering” or “blanketing”. There are two problems with these concepts. The second harmonic, and even harmonics in general, are formed by processes that also produces a DC bias. That bias generally upsets the output stage, and in tube amplifiers, can saturate (at least partially on each cycle) the output transformer. Next in push-pull stages, the third and other odd harmonics as well have a phase relationship with the fundamental that does make a difference. If one phase produces waves that are more square (clipped) and the other produces waves that are more triangular (expansive). The later comes from the increasing gain characteristic typically found in vacuum tubes and not found in bipolar transistors in their power operating region. Further, output transistors are rarely operated common emitter or source for stability reasons. They are usually operated as a follower for large current gain and the inherent feedback and low output impedance. Since “Tubes Versus Transistors” considered only microphone amplifiers and amplifiers with significant amounts of feedback, it is quite unlikely that Hamm encountered the expansive odd harmonics.

If even harmonics were truly better than odd harmonics, why then do the overdrive preamplifiers of many guitar amplifiers use significant resistors in series with the control grids that nullify grid conduction bias shifting that create even distortion harmonics?

The formulation of even and odd harmonics is different. Even harmonics are created by treating the top portion of the wave differently than the bottom portion. Even harmonics are present to the extent that the top and the bottom portions of the wave are not mirror images. Odd harmonics are created by treating the top an bottom halves equally, thereby making them mirror images. Third and higher odd harmonics affect the middle of the wave but not like the top and bottom. Checking, this is quite reasonable because single ended amplifiers have even harmonics as well as odd harmonics, whereas push-pull amplifiers tend to have only odd harmonics or significantly lower even harmonics due to cancellation effects in push-pull circuitry.

If even harmonics were truly better than odd harmonics, why then do the overdrive preamplifiers of many guitar amplifiers use significant resistors in series with the control grids that significantly reduces the grid conduction bias shifting that create even distortion harmonics?
“Tubes are slower than transistors.” is misleading. The first impression it raw speed or frequency response. Since tubes are quite capable of frequencies far beyond audio, namely radio, television, and beyond, the tube circuit must be the determining factor. But then the time delay through audio amplifiers is so small, just millionths of a second, it is humanly undetectable. Consequently, the phenomenon that created this characteristic is elsewhere in the circuitry, namely the treatment of the attack.

During the development of Pritchard amps, one prototype was deemed too slow. Perhaps for the first time a solid state amp was actually slower than a tube amp. A simple change to the patented circuitry made it “quicker”.
“Transistors distort too fast, faster than tubes.” Although Russell Hamm shows solid state microphone amplifiers distorting faster than their vacuum tube counterparts in his previously cited paper, his Figure 2 shows the 12AX7 distorting faster than a 2N3391A, which is faster than 5879 vacuum pentode. From this and circuit differences in section 1, one can readily deduce that this characteristic is circuit dependent as well. Feedback suppresses pre-clipping distortion at the expense of rapidly rising distortion post clipping. Further, driving a tube with a DC coupled low-impedance signal generator is quite different than driving a tube with a capacitively coupled, at least medium impedance source. The capacitive coupling allows grid conduction to create a bias shift and that creates even harmonics in that stage. The higher impedance creates grid conduction clipping.

Pritchard amplifiers are designed with enhanced pre-clipping harmonic distortion and little feedback, virtually none in the treble region, so that distortion occurs over a wide range and is consequently “slower” and consequently more useful in playing in the edge region between clean and dirty. Pritchard amplifiers are also designed with grid conduction emulation to produce the bias shifted even harmonics.
“Tube amplifiers, even the same amplifier, sounds different in Europe than in the United States.” This is factual because the output stage of tube amps are very much like an amplitude modulator of AM radio. The power supply ripple modulates the amplifier signal in this structure. The typical solid state amplifier does not have this capability for a variety of reasons. The ripple in the United States and other locations with 60 Hertz power is 120 Hertz while the ripple in Europe and other locations with 50 Hertz power, is 100 Hertz - hence the difference.

Pritchard patented circuitry to produce ripple modulation in solid state amplifiers so that Pritchard amps would have this notable tube amplifier characteristic. This structure is also the source of fat; the type of fat that makes the tone fuller as opposed to having more bass.
“Tube amplifiers try to respond to the guitar loudly but falls back.” This is a creation of a sagging power supply. While there are other potential contributors to sag, the tube rectifier has been the most recognized. Pritchard Amps do not use tube rectifiers because the voltages are so low, instead special circuitry exaggerates the sag so much that the 60 pre-clipping watt amplifier produces an estimated 180 watts Peak Distorted Power. For this and other reasons, Pritchard amps really cut through.
“Point-to-point wiring is superior to other forms.” is usually, but not always better than the super-neat bundled wiring approach and is also better than poorly designed printed circuit boards as well. It is better than the bundled wiring approach because the wire-to-wire capacitance is much lower, and consequently, the potential for parasitic feedback is much lower. However, a well designed printed circuit board has only one problem when compared with point-to-point wiring: increased wire-to-wire capacitance.

Pritchard amplifiers have a distinct printed circuit board advantage because the XGPA technology has a much lower impedance than tubes (about 1/10th) everywhere except at the amplifier input and output. The lower impedance readily handles the closer spacing and higher capacitances of printed circuit boards. This can be seen readily by the stability in an amplifier that has higher gains, is physically smaller and where the orientation of the output stage is centered within the chassis (for better heat dissipation).
“Class A output stages are better.” is derived from the engineering notion that all distortion is undesirable. True Class A delivers less distortion in push pull stages by eliminating the crossover problem or stretching it over the entire operating region, and having the push-pull tubes compensate each other - at least for even harmonics. The problem is that the engineering notions generally do not reflect a musician’s artistic desires. Lower distortion is not as ideal as full-bodied tone. As compared in the discussion of the implications of harmonics in “Tubes Versus Transistors”, The Class AB stage produces the previously unidentified “expansive odd harmonics”. The dominate “expansive” third harmonic is quite unlike the “covering” or “blanketing” third harmonic, described in “Tubes Verses Transistors” - they are much more artistic and desirable. They contribute to touch sensitivity, giving more response than expected at high levels. They are also a source of punch.

True Class A has no sag or sag's psycho-acoustic benefits, since the power supply current never changes significantly.

True Class A also produces less power than Class AB for the same output tubes. In fact, it is theoretically impossible for most amplifiers claimed as Class A to be true Class A. The following table shows the approximate maximum True Class A output power for various push-pull pentode-connected tubes (triode-connected tubes produce less).

Power Tube	Push-Pull Pair Output Power	Push-Pull Quad Output Power
6BQ5 / EL84	10.2 watts	20.4 watts
6V6GTA	11.9 watts	23.8 watts
6L6GB	18.7 watts	37.4 watts
6CA7 / EL34	21.3 watts	42.6 watts
5881	22.1 watts	44.2 watts
6L6GC	24 watts	48 watts

True Class A, push-pull in particular, also has virtually no sag since it draws virtually constant current from the power supply. The theoretical proof is below Analysis of True Class A Push-Pull.

“Amplifiers are supposed to replicate their inputs without any embellishments.” is a concept followed when convenient, but ignored in general. (See “Is Objective Correct?” ) This concept is strictly ignored in guitar and instrument amplifiers because they are expected to be warm, fat, full-bodied, resilient, an alive.
“Output transformers are an important part of the tube sound.” is partially correct. Tests on a Fender transformer show that the frequency response goes well beyond audio probably because the primary and secondary windings are interleaved. At the low end of the spectrum, the transformer has a saturation characteristic usually excited by the DC component of even harmonics. This characteristic can be emulated readily.
“All Tube Amp Characteristics are good and necessary.” is mostly correct, but not totally correct. The investigation of the various stages found that the severe overdrive of the output stage produces unpleasant and masking high-order harmonics due to crossover distortion. Consequently this rather undesirable feature was omitted from Pritchard Amps designs. Similarly, the modern boutique amplifiers use substantial grid resistors in the preamplifier overdrive circuits to minimize bias shifting due to grid conduction. Additionally, some designers are moving away from 12AX7's to pentodes also probably to minimize overdrive bias shifting.
“Better distortion requires more gain.” is an illusion created by the name “high-gain” distortion. The character of distortion is provided by other means since it still lets the amplifier breath, i.e. compress and relax.
“Emulators do it all.” is a marketing exaggeration to suggest that a solid state circuit does all that tubes do. Unfortunately, in virtually all cases, such emulators are about equivalent to a car body lacking a drive train emulating transportation. The approximation of a single operational facet has been and probably still will be adequate for the emulator label. Solid state has been “just like tubes” for two reasons, the miss-application fo engineering view of linear behavior, and the gross exaggeration of the term “emulation”.

Tubes Versus Transistors

While Russell Hamm’s paper has pointed out the differences in microphone amplifiers, a broader question needs to be answered: What are the differences between tubes and transistors?

Vacuum triodes are practically unique. Their plate characteristics rise from zero current at ever quickening rate versus voltage, i.e. the plate resistance becomes lower as the current rises. Although vacuum triodes do have a constant current region, it is sufficiently out of the normal operating range that it can be ignored. On the other hand, the remaining devices under consideration, pentodes, bipolar transistors, and field effect transistors of all types, exhibit much higher (often approximated as infinite) plate, collector, or drain resistance than vacuum triodes. (Pritchard patented, solid state circuitry that creates the vacuum triode plate characteristics.)
Of the various devices, only junction field effect transistors and vacuum tubes have grid (gate) conduction in the upper currents of their operating regions. This makes junction field effect transistors much more like vacuum pentodes, not like triodes, if one ignores screen grid effects as most think.
Grid conduction, unless driven by a low impedance source, controls the output of triodes around its maximum negative excursion. This characteristic, depending upon the grid-cathode bias, gives triodes their rather unique second harmonic dominance as Russell Hamm found.
The output of other devices is limited by the saturation characteristic. In audio power pentodes, saturation characteristic looks like a 200 ohm resistor. This value goes up drastically as the output comes out of saturation. This is the prime reason that the second harmonic rises after the third in pentode circuits.
The bias requirement for small vacuum tubes is a smaller percentage of the power supply voltage than for small transistors and the variance in vacuum tube parameters seem to be smaller. This feature combination implies less need for feedback and a much more controlled distortion characteristic.
The bias requirement for vacuum tubes is in the same order as the input signal. By comparison, field effect transistors have bias requirements substantially larger than their signal requirements. This makes the biasing of these transistors much more critical than for vacuum tubes, and consequently, the bias circuitry can have some adverse effects on the signal when it distorts.
The wide ranging semiconductor parameters demand feedback to limit production variations. While feedback is quite compatible with the engineering paradigm, it is not compatible with overdrive. Although the feedback suppresses pre-clipping harmonic production, it enhances post-clipping harmonic production, bringing up objectionable high-order harmonics rapidly.
The use of substantial feedback and the high collector or drain resistance allows substantial ripple in the power supply and consequently solid state output stages tend to have short attack durations.
The gain characteristic of power pentodes and field effect transistors rises with increasing current, but drops at high currents in bipolar transistors.
Any potential that might exist in solid state output stages for compression and ripple modulation are banned by the engineering amplifier paradigm: replicate the input without any embellishments. Pritchard has patented the solid state implementations of these features.
There are no negative complements to vacuum tubes as there are PNP transistors and P-channel field effect transistors to complement NPN transistors and N-channel field effect transistors. The availability of complements can change design approaches.
The circuit design approaches for tubes and transistors differ to accommodate the differences in their properties. Circuits are the way they are because they are simpler that way.

Pritchard amplifiers implement an exaggeration of the wonders of artistic amplification. It is obvious that this knowledge has had a huge impact upon the quality of the artistry and the extent of the operational versatility. There are amplifiers available with good tone. There are amplifiers available with substantial versatility. However, Pritchard Amps uniquely offer both in a single package.

Analysis of True Class A Push-Pull

The following analysis of push-pull Class A output stage. Putting aside harmonic production by the tubes, the plate currents of push-pull tubes 1 and 2.

I₁ = I_b (1 + S sin (Tt))

I₂ = I_b (1 - S sin (Tt))

where

I₁ is the current of plate 1

I₂ is the current of plate 2

I_b is the plate bias current

S is the signal level

sin (Tt) represents the sinusoidal signal of frequency T over time t

Notice that for true Class A operation neither I₁ nor I₂ is supposed to become zero. Since the sine function ranges between -1 and 1, S must be smaller than 1.

Notice also that the power supply current is I₁ + I₂ = 2 I_b and does not vary with the signal level, S. Consequently, the power supply does not sag in response to signal level. When the harmonics are included in the equation for I₁ + I₂, the even harmonics have some effect, but it is minor and much less than the change of power supply current for Class AB and much, much less than the change of power supply current for Pritchard amps.

Notice too, that at zero signal, S, the output tubes each carry the bias current and this is limited by the plate power dissipation of the output tubes. This sets this limiting relationship:

P_dmax $ V_S I_b

where

P_dmax is the maximum power dissipation of the tube

V_S is the power supply voltage

I_b is the bias plate current

Continuing, the two plates are connected to a load by a transformer. But for this analysis, let the load be from plate to plate. Thus, by center-tapped primary transformer characteristics:

I_L(2 N) = I₁ N - I₂ N

substituting, adding, and dividing by 2N

I_L = (I₁ + I₂) / 2 = I_b S sin (Tt)

where

I_L is the load current as seen in the transformer primary

N is the number of turns of each half of the center-tapped transformer

The next problem is to set the load impedance. If one were to design a transformer, this would set the turns ratio from the primary to the secondary. However, to simplify the problem this will be the load as reflected on the primary. The sizing is done by the maximum clean signal is limited by the supply voltage, V_S, and the various voltage losses, V_loss, which are dominated by the minimum plate voltage for current I_b. But first, the plate-to-plate voltage, V_PP, is related to the primary load impedance R_L

V_PP = R_L I_b

then V_S - V_loss $V_PP / 2 (center-tap and power supply connection to plate voltage)

thus V_S - V_loss $R_L I_b / 2

The RMS output power, P_RMS, is now

P_RMS = V_PP I_b / 2 or R_L I_b² / 2

Since V_PP / 2 #V_S - V_loss #V_S

Then P_RMS = V_PP I_b / 2 #(V_S - V_loss) I_b #V_S I_b # P_dmax

Thus, maximum power output of a push-pull pair of tubes is less than the maximum dissipation of one of the tubes. Since V_loss is about 15% of V_S, the maximum output of a pair of push-pull Class A tubes is about 85% of the plate dissipation of one tube. Of course, this is doubled for a push-pull quad. The table above provides these values for popular output tubes.