Cascodyne Architecture

History:
The vacuum tube (or valve) was invented in 1904 by John Ambrose Fleming, and quickly became an essential component for electronic devices, remaining predominate through the first half of the 20th century until being largely superseded by the transistor. It remains the preferred technology for certain applications, such as the magnetron used in microwave oven, some high-frequency amplifiers, and, of course, musical instrument amps and high-end audiophile components.

As their wide-ranging utility was recognized, large-scale manufacturing developed, beginning with Mullard and RCA in the early 1920s, and followed shortly thereafter by Mazda, Phillips, and others.

In order to promote the use of their products, these manufacturers published many example circuits, and these formed the basis of commercial devices based on the technology, such as radios, amplifiers, televisions, and so forth.

Innovative entrepreneurs such as  Leo Fender studied these designs and ultimately turned them into commercially viable products. Hence, if we adhere to the smug pronouncement of our supercilious web 'expert', quoted on my home page, Leo Fender was just another hacker, and nobody this side of John Ambrose Fleming has done anything innovative whatsoever.  This, of course, is utter nonsense.

Fender's genius was his ability to take a novel, costly, and only partially understood technology, and turn it into a commercially viable, reliable commodity, perfectly positioned for a well-defined market.

Tube Amp Architectures:
Virtually all modern tube amplifiers, and the vast majority of analog solid-state ones as well, are based on one of three architectures, all pioneered (commercially, at least) by Leo Fender. Indeed, most digital modeling amps are simply computer simulations of these same constituent components:

1. Basic single-ended (e.g. Fender Champ)
    Common-Cathode Triode Gain Stage --> (volume control) --> Single-Ended power stage.

2. Tweed Push-Pull design (e.g. Fender Tweed Bassman)
    Common-Cathode Triode Gain Stage --> (volume control) --> DC-Linked second gain stage /Cascode Follower --> Tone Stack --> Phase Inverter --> Push-Pull power stage.

3 'Blackface' Pushs-Pull design (e.g. Fender Super Reverb)
    Common-Cathode Triode Gain Stage --> (volume control) -->  Tone Stack --> second gain stage --> Phase Inverter --> Push-Pull power stage.

Soon after Fender's introduction of the tweed Bassman, Jim Marshall, working in the UK, used that architecture as the basis of his own line of amplifiers, mainly switching to more readily-available valves (UK KT-66s in place of US-made 6L6s), and solid-state rectifiers, which delivered a more punchy, aggressive tone. Since Fender moved on to his 'Blackface' architecture shortly thereafter, whereas Marshalls still embody that early Tweed Bassman topology, we commonly refer to that as the Marshall Architecture.

Schematic Haven is a treasure trove of schematics for every popular guitar amplifier ever made, and is an incredible resource for anyone interested in the history and technology of musical instrument amplification. I encourage readers to peruse their content, and can say with confidence that you will find that virtually all amplifiers, from the original Fenders, to the feature-laden Mesa Boogies, to the exotic Dumble, all fall neatly into one of the aforementioned categories. Sure, some of them may add dozens of tricky switches and relays to add gain stages, or revoice tone stacks, or change signal routing, but if you analyze any of the resulting configurations carefully, you will find that they all boil down to one of these three basic architectures.

Tube Types
The vast majority (though not all) commercial amps use the same basic tube types: one or more dual triodes (usually a 12A*7) for the preamp, and one or more pair of power pentodes or tetrodes in the power section. 

One of the first to deviate from this formula was VOX, who employed a single EF86 small-signal pentode in the first (and only) preamp stage of their AC-15 amplifier.

The impetus here was completely financial: the pentode offers far higher potential gain than popular triodes, and by using the single EF86, they were able to replace two or more triodes, along with their associated support circuitry.

Pentodes have some unique traits, some desirable, some not. In addition to their aforementioned gain potential, they can also deliver increased touch sensitivity, and wonderfully complex overtones.  On the negative side, their increased internal complexity makes them sensitive to external vibrations (known as microphonics), particularly when used in combo amps with close speaker proximity, and when configured for maximum gain; both of which were factors in the AC-15. As a result, VOX ultimately redesigned that amp to use a more conventional multi-triode configuration.

When subsequent designers tried  to use the EF86, they usually just copped the first-stage design from the VOX AC-15, without really understanding how the pentode operates, with similar results. 

Phase Inverters
By their nature, push-pull power sections require two input signals, identical in every respect except for being 180-degrees out of phase with respect to each other. The electronic component responsible for splitting the single signal from the preamp into these two requisite counterparts is known as the phase inverter.

There are many varieties of phase inverter, including some power sections that are self-inverting (very inefficient), but the two most popular topologies are the Long-Tailed-Pair and the Cathodyne.

The long-tailed-pair or Schmitt phase inverter is the most widely used, and is extremely flexible.  It utilizes two complementary triodes, arranged like so:

For a detailed explanation of its operation, see Valve Wizard's site (from which I swiped this illustration).

A simpler alternative is the Cathodyne, used most notably in the renowned tone machine, the Fender 5E3 Deluxe.  One obvious benefit of the cathodyne is the fact that it requires only one triode, rather than two, as shown here:


Once again, Merlin Blencowe (Valve Wizard) does a great job of explaining its operation.

The basic cathodyne shown above does a good enough job in low-gain amplifiers, but when pushed, can create tones that range from pleasantly rude to downright obnoxious.

The above link explains the reasons in great detail, as well as describing methods of avoiding it.  

Another limitation of the cathodyne is that it inherently creates less gain than a long-tailed-pair style of phase inverter.  This isn't an issue when driving smaller power tubes, such as the EL84, or the 6V6 used in the original Fender Tweed Deluxe, but can be problematical with higher power tubes, such as the 6L6GC, the EL34, 6550, or KT88.

However, combined with another triode gain stage, the net output exceeds that of the long-tailed-pair, while using the same number of elements.  Here is an example of a DC-Coupled Cathodyne that obviates the aforementioned unpleasant overdrive effects:


In his excellent book, Designing Tube Preamps for Guitar and Bass, 2nd edition, Blencowe describes a further refinement, known as bootstrapping, which uses the cathodyne to force even more performance out of the preceding gain stage. I encourage anyone interested in that depth of explanation to acquire this essential book.

Before examining my own evolution of the phase inverter topology, we need to go back to the very beginning of the signal chain and talk about an electronic structure never before used in a commercial musical instrument amplifier: the cascode.

The Cascode

Virtually all commercial amplifiers use some variation on the common cathode gain stage described earlier, employing either a triode or a pentode. Some, most notably the excellent Matchless Chieftain, use a pair of triodes in parallel, like this:
This topology can be regarded as a single triode, with double the current flow-characteristics, and is largely tangential to the current discussion, other than to observe that, depending on configuration, it can develop increased gain, or greater clean headroom, but not both simultaneously.

The cascode ( or totem-pole circuit), on the other hand, delivers increased gain, greater clean headroom, wide bandwidth, and in addition, incorporates many of the desirable properties of a pentode (notably, touch sensitivity) without the inherent drawbacks. (The name "cascode" was coined in a paper written by Frederick Vinton Hunt and Roger Wayne Hickman in 1939, in a discussion on the application of voltage stabilizers. They proposed a cascode of two triodes (the first one with a common cathode setup, the second one with a common grid) as a replacement for a pentode, and so the name may be assumed to be a contraction of "cascaded triodes having similar characteristics to a pentode".

Very popular in television tuners, the circuit consists of a common emitter stage feeding into a common base stage, like so:


So, you may ask, why has nobody else used the marvelous innovation in a guitar amp before? The reason is insinuated in our pompous "amp expert's" diatribe quoted on my home page: The vast majority of so-called 'amp gurus' don't have a clue what they're doing. They simply examine  a bunch of schematics for famous amplifiers, and grab a gain block from this one, a tone stack from that one, a phase inverter from over yonder .... then tweak component values until they're happy with the result. 

Several years ago I had an opportunity to work closely with a widely known 'amp expert'  whose clientele read like the who's who of the music industry. Prior to
admitting me to his inner sanctum, he insisted I sign a long, formal non-disclosure agreement. I looked forward to learning some great new secrets of amp
design! What I discovered were binders of old Fender schematics and layouts, replete with a few hand-written modifications, all of which have been common
knowledge since the advent of the internet. Indeed, watching him try to troubleshoot one of these simple circuits, I quickly realized that he was unable to
even read a schematic!  His approach was to sit down with one of his old Fender board layouts, and walk through the circuit from input to output (the exact
opposite of what virtually all amp technicians recommend, which is to start at the output and work your way forward.) If distracted, he would angrily start all
over at the input jack.... In fact, he had inherited all of these 'secrets' from a true master of amp modification, long dead, who will remain unnamed, for whom he had worked, possibly as an apprentice, but more likely as a lackey.

Be that as it may, I present here my own variation of this interesting tube topology, which forms an intrinsic component of all of my designs:

(The 6DJ8/ECC88 is a dual triode specifically designed for use in cascode architectures. It is in current production (I don't design anything around obscure, obsolete components), but is readily available as high-quality, American-made NOS units at reasonable prices. These are the units I use in my builds, wherever possible.)

Conceptually, this topology allows  each triode to operate well within its intended range, while together delivering more gain, with full bandwidth, than a single triode could accomplish alone.

Inasmuch as this is basically a clean high-gain stage, this might be a propitious juncture to pontificate on my attitude toward clean versus dirty amps...

We all love the tone of a balls-to-the-wall cranked tube amp, and when I was 16 or 17 years old, my attitude was 'If it ain't a Marshall, it ain't shit', and "I don't even need a volume control -- just on and off, thank you very much. (And that's why I now have to rely on a pair of hearing aides to even watch television.) Volume aside, unless totally buzzed out fuzz is the only sound you ever need, then you probably appreciate the need for something other than a one-trick pony.

I design amps for my own needs, and the way I prefer to play. Yeah, I do like to rock out, but I also want to be able to play some clean jazz, at a volume adequate for playing with a loud drummer, and that means having clean headroom available all the way to the speakers, when needed.

Look at it this way: If you make tons of distortion at your first preamp stage, there's no way you're going to clean it up later in the signal chain.  There are many ways and many places to make a good overdriven sound, but if you don't have a good, multi-dimensional, nuanced clean tone to start with, you've got nowhere else to go.

You'll see what I mean as I work forward through my architecture.

Tone Stacks
I use three basic tone stacks, none of which are my own innovation, other than in terms of specific components and voicings.  The simplest is a variation on the BMP (big muff pi) tone control, which is a combination of a low-pass and a high-pass filter, set up to sweep in opposite directions. Unlike most other simple tone controls, which are almost always just treble-cut circuits, the BMP has minimal interaction with the volume control, and alternately cuts-bass/boosts-treble, and vice versa. I usually build the circuit directly on a push-pull potentiometer, so that pulling the knob out takes the entire control out of the circuit, for a substantial volume boost.

The second tone stack I use, again voiced specifically for each amp, is the James/Baxandall design, popular in high-end stereo equipment, but also used in many vintage Ampeg designs. This tone stack has far less loss than either a Marshall or a Fender tone stack.  It's a two-knob (bass/treble) design that is incredibly versatile, and allows for steep band-pass and band-cut (mid scoop) configurations, as well as all other common frequency sculpting features. 

The Third circuit, and my personal favorite, is one adapted from the super-toneful Fender Brown-Face 6G4-A Super. This tone stack was unique among Fenders, and relied upon an unusual 350k treble control with a 70% tap, and allowed a far greater treble boost than other designs.  I redesigned the circuit to avoid the need for that hard-to-find tapped control, and also modified it to include a mid control which, when dialed up to '10' yields the exact same response as the original 2-knob implementation.

You can see these in the schematics for my various amp designs.

Back to That Phase Inverter ....
Now that you understand the conceptual operation of the cascode, it will be easier to follow my phase inverter topology.  I do use several variations, each applicable to a particular amp design, but the variation used in the JPB-45 is one that contains all of the essential features:



As mentioned above, in my smaller amps I use a single dual-triode, mainly due to space constraints, but when possible, this is my preferred configuration.

In this application, the EF86 is tuned for very conservative gain, which eliminates the noise and microphonics problem, maximizes the tonal complexity when overdriven, and still provides sufficient amplitude to drive the DC-coupled phase inverter. Moreover, the bootstrapping forces the pentode to operate at peak efficiency at all frequencies.

The phase inverter is essentially a cathodyne, modified for the most toneful overdrive characteristics, but using a cascode in place of the usual single triode. This provides output sufficient for driving the largest power tubes.

The Gain control (more on this shortly) determines how hard that huge, clean signal from the front-end cascode pummels the EF86. Cranked, it will handily drive that pentode into the sweetest overdrive imaginable, however, back it off and you can get a clean tone out the back end at large-venue volumes. This is made possible by my ...

No-Penalty Volume Control
Master-Volume (MV) controls have a rather evil reputation amongst tone aficionados, some of it well deserved, some not. The ones placed in front of the phase inverter, such as were implemented in some Silver-Face CBS Fenders is indeed an abomination, and negatively impacts the overall tone even when turned up to the max.  

A preferable, but still imperfect approach is represented by the cross-line design used in the Matchless Chieftain, as well as many other quality amplifiers:
This simple arrangement is situated after the phase inverter and before the power tubes, and relies on the fact that the two signals it connects are 180-degrees out of phase.  When two identical but out-of-phase signals are combined they completely negate each other, resulting in ... silence.  With the 500k potentiometer at its maximum, very little signal is passed, resulting in minimal (but still some) attenuation.  When I use this kind of control (which I do in fixed-bias amps), I typically employ a push-pull control so that in the normal (in) position, the MV is completely out of the circuit. When attenuation is desired, the user must pull the knob to engage, and then adjust to taste.

Wherever practical, I tend to favor cathode-biased amps, for various reasons discussed elsewhere, one of which is that cathode biasing allows to use of this extremely elegant, absolutely lossless (master) volume arrangement.



Here, a dual ganged potentiometer replaces the essential grid-reference resistors, which have to be there anyway! This is only possible in cathode-biased power amps because the grid-reference resistors connect to ground, whereas in a fixed-bias amp they are connected to the negative DC bias supply.

The egotistical 'amp expert'  cited above, in one of his rare moments of 'innovation', implemented a similar scheme, but on a fixed-bias amp,
totally unaware that placing DC voltage (positive or negative) on a potentiometer is a definite engineering 'no-no', for a variety of functional
and safety reasons....

This pretty much covers the essential elements of my Cascodyne Architecture.  Yes, I freely admit that i stand on the shoulders of giants, but if anyone can point out another commercially-built amplifier that remotely resembles this design, I will consider myself roundly chastised.

In any case, I love the way it sounds, I love the way it feels, I love its versatility, it's what I build, and its what I use.
Comments