Power Tube Tone

[K O'Connor]

Hi Guys

Vacuum tubes aka valve, allowed the electronics age to begin and grow into something indispensable in our modern lives. For most things, newer technologies have taken over from tubes, but there are still places where tubes are either the only possible gain element to use, or they are simply preferred for their unique characteristics. For example, microwave ovens use a magnetron tube to generate the microwave energy used to stimulate food to cook it. Add a wave guide and a capacitor and this is the heart of every microwave oven regardless of what the front panel looks like. Very high-power radio transmitters are still tube-based although modular solid-state system are working their way into this realm.

For audio amplification, tubes are still used by many manufacturers, hobbyists and music enthusiasts even though dramatically lower distortion can be attained using solid-state circuits. Nominally, a playback system is not supposed to change the sound, but every listener assembles a system to his own liking, and this is because our individual hearing is unique as is our aesthetic sense of how things should sound. Tubes are considered to be almost "organic" in how they handle music, which can be explained in scientific terms of their structure and resulting nonlinearities. The varying transfer curve of what the output looks like for any given input level gives the tube a "pleasant" character, and makes it doubly useful for musical instrument applications where the amplifier is "the other half of the instrument".  MI is about tone creation, so we want the character of the tubes to shine through, and to be able to combine these characters to achieve our sonic goal.

Fortunately for both hifi and MI, there are a lot of power tube types that share a common base type and pin-out, making them more or less plug-and-play compatible - with the necessary check of the idle condition to assure safe operation. The sound of each tube is unique, but we can group them all based on their sonic texture, which strongly follows their internal structure.

The first broad strike of distinction of power tube tone is based on the use of beam-forming plates. This was an RCA development in the 1950s, and it resulted in a reduction of distortion contributed by the tube. Tubes using beam-forming plates are still called "pentodes" or "tetrodes" as there are still five or four elements within the tube, respectively, but the screen grid is replaced by plates. There is no need for a critical alignment of the various grids, and the electron stream is focused towards the plate.

Beam-forming plate tubes:
6CA7
6L6GB/GC
5881
6550
7027
7481
7491
8417
These tubes have a neutral tone with no specific emphasis of frequencies. Compared to the types below, this group may sound "clean", or "less bright", or "less harsh".

"Kinkless tetrodes" is Mullard's term for their power tetrodes with critically-aligned  grid wires. This group includes the famous KT-series:
KT-66
KT-77
KT-88
KT-100*
KT-120*
*Devised well after Mullard went out of business and the trade name was bought by Mike Matthews of Electroharmonix. So, the structure may not be true to the legacy?

Kinkless tetrodes tend to have a higher distortion than the beam-forming plate tubes, with this distortion being perceived as "thick", "rich", "brittle", or "harsh", "muddy". Personally I find this group to have a muddy tone, but it is a huge component of the distinctive Hiwatt sound.

Structural pentodes have grid wires for the control grid, screen and suppressor. As a power tube there is some critical alignment necessary. This group includes:
6BQ5
6V6
6973
EL-34
EL84
These tubes tend to have a "brittle", "bright" or "harsh" tone, except for 6V6 which is either  "creamy" or "muddy". Being a 9-pin miniature base type, the 6BQ5, EL-84 and 6973 have no substitutes per se, as all have the identical tone.

Each group has variants of many of the tube types within it and there are no doubt other tubes not mentioned here.

Note also that the tones described are for the wiring connection using the tube elements as generally perceived to be "standard", where signal is applied to the control grid, the screen is tied to a fixed voltage, and signal is taken from the plate through an output transformer. The character of each tube will change when alternate operating modes are used, such as triode wiring or the use of ultralinear taps on the OT.

Triode tone is generally perceived to be "clean", but can be deemed "dark" when compared to a reference tone that contains more distortion (as with KTs or structural pentodes / tetrodes). Structural triodes and heater/cathode triodes have a tone similar to the beam-forming plate pentodes.

Ultralinear tone is somewhere between triode tone and the innate tones of the various tube types. UL connections are intended to reduce THD within the tube and the OT yet exhibits a compression effect of its own, representing a gain nonlinearity with signal swing. For most musicians, Fender's dabbling with UL connections turned many players and hobbyists away from this operating mode, but their sonic achievement resulted more from other factors of the rest of the circuit design.

In hifi, many of the "tubes with character" are revered, and this depends a lot on the circuit the tubes are used in and the interactions between the power amp in total with the speakers. Many hifi PAs use unstable circuits with too many gain stages within a single feedback loop. The classic Williamson falls into this category despite achieving good results for the day. A design that squashes the tube characteristics to achieve remarkable results is the MacIntosh Unity-Coupled OT design. Both Williamson and MacIntosh paid great attention to the winding of the output transformer, with the latter being a break from tradition combining both plate and cathode drive, with the bonus of bootstrapping the screens and the driver tube. Audio Research followed along in the plate + cathode drive and often used cathode feedback to the output tubes to provide better linearisation.

MI power amps using tubes fall into two basic designs based on the splitter circuit used. For both, the overall effect of a net low-gain means that the character of the power tubes is well out in the open, and substituting other tube types lends different textures to the amplifier. This means that the player does not have to be limited to the tone of the stock tube set, provided there is a means to adjust individual tube bias (as with our Bias Mod Kits).

As TUT3 demonstrates, the famous Ampeg SVT (Super Valve technology) and V9 amps used an unstable, overly-complex power amp circuit that required many band-aids to make stable. Fender copied this circuit for use in all of its current 300W tube bass amps. As is the usual result of such copying, all of these amps tend to "eat" power tubes unless we make the simple modification recommended throughout the TUT-series.

[soundmasterg]

Hi Kevin,

You said this >> 

The first broad strike of distinction of power tube tone is based on the use of beam-forming plates. This was an RCA development in the 1950s, and it resulted in a reduction of distortion contributed by the tube. 

Don't you mean the 1930's? After all the 6L6 came out in 1936 I believe and it had beam forming plates from the beginning right?

Also, in my experience the 6973 doesn't sound anything like an EL84 and sounds closer to a 6V6 but maybe that's just me. 

Also, I was wondering if you have heard of vacuum transistors? It's a development by NASA and the AMES research center among other places. I've been following the progress on these over the last 10 years or so, and they've recently figured out how to make them bi-polar, but the interesting thing that I found about them is that the curves looked like tubes/valves. I find myself wondering if these ever get produced somewhere if they can be used in audio to good effect. I can share some links if you want that I've collected about these over the years.

Greg

[K O'Connor]

Hi Guys

Greg, you must be older than you look

I was just going by a 20+ year old memory of an old Wireless World article I photocopied at the library, which seemed like it was hot on the heels of the tube coming out. In retrospect, it was probably more of RCA et al shedding light on the electron flow patterns within beam power tubes, focusing on the 6L6.

I was briefly enamoured with the 6V6 sound, but it was not compatible with more complex preamp tones. Randall's Line Backer series of amps used 6V6 output stages, and of course Jim Kelly used a quad in his amps. In the Kelly amps the quad tended to even out some of the 6V6 distortions resulting in a clearer tone, but still warm. Those small 9-pin tubes never appealed to me.

Yes, the "semiconductor tubes" or "vacuum jfets" are interesting. I believe they were developed towards the goal of making computers smaller and faster. At the microscopic level one could argue that everything happens in a vacuum. I'm not sure if it was a sideways development or interesting offshoot from the path that lead to quantum computers? All of it is certainly beyond the soldering skills I have

[soundmasterg]

Hi Kevin,

I'm in my early 50's now.....getting old.

I recalled something saying the 6L6 has been in continuous production somewhere around the world since 1936, and I know that has beam forming plates, so that is why I was saying the 30's.

Some of the small 9 pins sound closer to a 6V6 than an EL84 like the 6973 or the 6CM6 for example. I like the chime and nice mids of the EL84 but I quickly start missing the bass that other tubes have like the 6L6 or even the 6V6 or 7591. Actually my favorite power tube is probably the 7591/7868 because it has almost as much gain as the El84 for a power tube, but it has a more even frequency response. Since it has a lot of self gain, you don't need to hit it with a large gain out of the preamp, and if you use a suitable phase inverter, then it can really overdrive the power tubes. In an amp I made for myself, I used a LTP circuit for the phase inverter, but I used a 12BZ7 instead of a 12AX7 to hit the power tubes really hard. It has the same gain as the 12AX7 but half the output impedance so it drives the power tubes twice as hard. They're pretty cheap too for NOS. They can be microphonic, especially if used in V1, but in the PI they seem ok.

Here are some links for the vacuum transistors for anyone who might want to know more about them. I think they were made primarily for space applications as they don't have any issues with radiation like traditional transistors do (FETS and BJT's) and the first prototype was able to transmit into the lower middle of the terahertz gap which is beyond transistors, and from what I understand, tubes can still get a bit higher, but it is impressive. The first prototype also had no degradation of performance up to 200 degrees C, which is very impressive. They figured out in the last couple years how to make them bipolar also, which greatly expands their use possibilities.

I've worked at Intel in the last decade (until I was laid off anyway) and have some experience with traditional computer devices. They've almost reached the physics limits of MOSFETs in processors, so there will have to be something new to continue Moore's Law, and these vacuum transistors are one of the possibilities. Of course I'm interested to see if they would work for audio.

Greg

https://spectrum.ieee.org/introducing-th...of-nothing

https://pubs.acs.org/doi/10.1021/acs.nanolett.6b04363

https://engineeringcommunity.nature.com/...transistor

https://physicsworld.com/a/nasa-scientis...ransistor/