joepampel

Feb 1, 202313 min read

OG Guitar Amp History

Updated: Feb 7, 2023

Not the last word, but a few hundred on the most popular & I'd argue important circuit topology of all time when it comes to rock and roll.

There is this legend in rock music, it involves import embargoes and a specific Fender amp and this English drummer who decided to make his own copy of it and helped spark a revolution. And that's not exactly wrong, but it's not nearly the whole deal. And neither is this - but I hope I can shed a little more light because the origins of the design go back farther, and wound up in more places and affected popular music immeasurably. We can't hit all of the amps that are based off this fundamental design - they are legion. I would run out of space on this site before I listed them all. But we'll hit some OGs, and look at some similarities and differences.

"Guitar amps were in the tube manual", well no they weren't. What was in the tube manual was how to use a tube - by the folks who made them. It was a sales tool at the end of the day. Electronics vendors still make books full of sample things you can do with their various widgets.

Leo Fender had a background in electronics but couldn't play guitar. He started making guitar amps back in the 1940s as part of K&F with Doc Kaufman. By the time we get to the mid 1950's he's been at it a while, and is running the Fender Electric Instrument Company. By 1950 they have introduced the first solid body electric guitar, the Telecaster, and in 1951 he revolutionizes popular music by introducing the Fender Precision bass - a bass with frets! Shut the front door... Yea, it was a big deal. And so Fender started voicing his amplifiers to work well with his guitars. And his amps were getting steadily better and better, the designs were evolving along with the music they were asked to make.

Over at Gibson they went the other direction; they started by making guitars and mandolins. The Gibson Guitar and Mandolin company was founded in 1902 and invented some monumental devices like the truss rod and the adjustable bridge. It is hard to imagine these things getting invented; they are so ubiquitous and important. It's akin to inventing a C major chord.

By the 1935 they introduced the EH-150 - an electric slide guitar (there was a Hawaiian craze) and the next year they introduced the ES-150, their first electric spanish guitar. As the 30's wore on, they started making more guitar amps as Charlie Christian kicked off a new revolution - the electric guitar as not just a rhythm instrument, but something that could take a lead and had to be heard. (And maybe some someones before him I don't know) So by the mid 50's Gibson had been at it a while too, longer than even Leo. Gibson too, voiced their amplifiers to work well with Gibson guitars; it only makes sense. So there were subtle differences.

As these two leaders competed, they appear to have inspired each other a great deal. You can bounce from design to design and see Gibson borrow this and improve it, and then Fender borrows something else from them and improves it. There is generally no real plagiarism, more like two jazz greats pushing the boundaries as they play off each other. Somewhere around 1954, Gibson introduced a 1x15 combo amp known as the GA-70. It was not any huge departure, but there were some important incremental changes.

The basic design was going to inspire Vox in the UK, and get tweaked by Leo into one of the most popular Fender amps of all time, and then a certain Mr. Jim Marshall was going to roll his own after seeing what Leo did and start a fire over in the UK. And if your author dug back farther, it would probably get still harder to pick out where certain circuit features originated. For example the GA77 from 1953 is nearly the same. But the GA-70 is as good a start as any, and it gets referenced a number of places, including Jim Elyea's heavily researched "Vox Amplifiers, the JMI years".

Here is that GA-70 schematic:

I didn't re-draw it in the more common style because I think it's good practice to see things drawn differently. Gibson didn't like to draw the bulbs around their tubes which makes this a little confusing. The normal convention is to draw the bulb around single section tubes and half a bulb around each section of dual section tubes, with the half bulbs pointing at their other halves.

I have annotated the things we're going to see in the other amps:

12AY7 first voltage amp with a shared cathode resistor. (5F6-A, JTM45)
.02uF coupling caps into 1M volume controls, and a bright channel (bypass cap on volume)
270k mixer resistors (Bassman, JTM45)
12AX7 2nd voltage amplifier ( " ")
12AX7 direct coupled cathode follower (" ")
Tone controls following the cathode follower (bass and treble) This layout shows up in the AC30 Top Boost and the AC50.
Presence control using the negative feedback loop.

The most notable difference is that the GA70 uses a cathodyne phase inverter here where the Bassman would use a Schmitt inverter. They have a very low gain 12AU7 as the final voltage amp and phase inverter, so this is a low gain & high current solution probably targeted at clean playing. Otherwise, this is essentially a Tweed 4x10 Bassman circuit in many important ways. Cathodyne inverters always have a gain of <1.

But first, here in 1954, we also have the Fender model 5E4 Super Amp, a 2x10 combo with nearly the same layout (remember the jazz riffing analogy?)

So 1-7 & 9 - check. All basically there except here there is a 12AY7 for the second stage, and we have an interim tone control circuit and 12AX7 driving 6V6 outputs. But then in 1956, we get some key tweaks:

1,2,3,4,5,6,7... all there, again. It also has the choke input power supply. It is nearly the same amp as our Gibson except it's using a 12AX7 for the last gain stage and cathodyne phase inverter. And now it is also using a 12AX7 in the 2nd tube in place of the 12AY7 in the 5E4 model as well as 6L6G outputs. We are getting bigger and louder.

The tone controls are the one big difference; we don't see the later topology that will conquer the world. Here we have a "tone" control that is acting for treble and an uncommon bass control circuit. The move to the other style circuit not only allowed the Bassman to shine, but with small tweaks, it lived on in all of the classic 1960's Fenders and is still copied today.

One interesting difference besides the odd tone circuit is the local negative feedback on that 2nd gain stage that has been there since '54. After the treble control, there is a wire running to a 4.7M resistor (center top of the diagram - 5M in the earlier model) which connects to the grid of the 2nd voltage amplifier. This reduces gain, reduces distortion and flattens our frequency response a bit. Fender would pull this trick out from time to time to tweak certain amp circuits. It shows up on some 1960's era Fender Bassman amps as well.

To summarize though, the 2 biggest missing bits in the Gibson, and this early Fender, are the later tone stack, sometimes referred to as BMT (bass, mid, treble) which appears almost ready to go in the Gibson but Fender really perfects it.

And the other piece which is probably key since we're talking rock guitar - the move to a Schmitt phase inverter. This gives us more drive for the outputs, more power out and better overdrive when the amp clips.

Don't feel bad for the 5F4-A; it would also show up in the 3x10 Bandmaster and make history when Pete Townsend recorded "Who's Next" with one, driven by a Gretsch 6120.

Fender then tweaks (fixes?) the tone controls, adds a mid control and replaces the cathodyne phase inverter with a Schmitt (aka long tail) phase inverter which has must better gain using a 12AX7 and which can drive the output tubes to a solid 40W, or around 15W more power than the Gibson likely made. Fender moved the choke from the input to the now more standard location between the plate supply and screens. This let Fender use a less expensive choke and also put it where it could do the most good. Value engineering? Sure, why not.

He also replaced the single 15" speaker with 4 10" speakers, or compared to his own Super Amp, he went from 2x10 to 4x10 which is approximately a 6dB increase in volume (equivalent to 4X the power, roughly.)

These early Jensens could not take a lot of power and speaker failures were pretty common. Splitting the load among four of them was a good attempt at better reliability, not to mention better bass response. So the new Bassman was also moving a lot more air, approx 177sq in for the 15" vs 316 sq in for the 4x10 configuration.

Net net the Fender would be louder, punchier and have more gain. You can bet it would bury the competition on stage, and when getting demoed in the music stores.

What we wind up with is this then, the mother of all guitar amps:

The first 7 annotations are the same; we've added:

8. Our new Schmitt phase inverter, using a 12AX7. More gain, more drive. More more more!

9. Our choke which now lives between the plate and screen supplies.

Fender cleverly added a midrange control, and tweaked the circuit a bit to work more intuitively. The Gibson/Vox stack works in an unusual fashion. With the bass and treble "off", you get a very mid-range focused sound, and then can add back bass or treble. Most amps take after the Fender approach where having your knobs set to the mid point is a good starting position and you can add or cut from there. Mid-point on an AC30 Top Boost or AC50 is a pretty dramatic setting with a big mid scoop and a lot of bass. It can be confusing.

And next, the JTM45 that Marshall built initially, you can compare it to the 5F6-A above. You don't need to be a rocket scientist for this one. Sincere flattery going on. You can see it in the photos of the layout below as well.

Drawing (c) Mark Huss, annotated by the author (sorry to mess this up Mark!)

So what do you have? Well, you have some of the 5F6-A improvements; the improved tone stack & the Schmitt phase inverter. You have the choke input filter as on the GA70 & 77. But then you have the deliberate or accidental? changes made in 2 key areas that one could argue launched the "Marshall Sound".

The first stage tube was swapped from a 12AY7 with a mu of 45 to an ECC83/12AX7 with a mu of 100. So Marshall goosed the front end gain by a smidge over 2X.
The output impedance changed from 2 ohms to 16, and this affects the amount of negative feedback, which affects the performance and feel of the amp.

The negative feedback connection is that wire from the speaker out, to the 27k resistor to the 5k presence control. The voltage is divided between the resistor and the control here.

If we assume we still have 40W of output, 40W into 16 Ohms means we have about 25.3 volts and 1.58 amps of signal at the speaker output.

The Bassman is delivering its 40W into a 2 ohm load which means we have 8.94V at 4.47 amps of current.

net net, the Marshall has about 5.7X the raw feedback voltage, or 9dBU more, which is nearly 10X the level (10dBu would be 10X)

(check my math: 0dBu = .7746, 25.29V = 30.27dBu, 8.94V = 21.245dBu, 30.27 - 21.245 = 9.025)

As a practical matter, negative feedback lowers the source impedance of the output stage, which we generally call 'dampening factor' - it's is a proxy for how well the amp can control the movement of the speaker(s). As a player you feel this as "tightness". It also flattens out the frequency response and reduces harmonic and intermodulation distortion.

It works by taking a sample of the output signal with THD artifacts that are coming out of the stage, and send them back to the input of the stage - except out of phase. In doing do, we are 'pre-correcting' for things that will occur later. It's a bit whacky, but it works. It was patented back at Western Electric in the 1920s and it took the patent office 9 years to grant it because they did not believe it would work. (see Harold Black article link below!)

The presence knob works by shunting some highs to ground, thus reducing their level in the feedback signal and in doing so, boosting their gain.

Finally, negative feedback also changes the transition into overdrive. The feedback loop fights distortion both via pre-emphasis (adding out of phase distortion at the input) as well as by lowering the stage gain by the number of dB of feedback you have applied. If your stage gain is 25dB and you apply 5dB of negative feedback, your stage gain is now 20dB.

But when you reach the end of the road - clipping - negative feedback stops working properly. Your gain effectively shoots up by however many dB of negative feedback you had and you transition into overdrive more abruptly than in a circuit without it, or with less of it. Until this point, your feedback level has moved in concert with the output signal. Now the output signal has flatlined - it cannot get any bigger. If you feed the stage even more signal, it will not have negative feedback applied in the correct ratio anymore because there is no headroom. Your signal gets a mid boost, your tightness diminishes. A lot happens fairly quickly.

It is hard to know what the design thinking was around goosing the front end and upping the feedback. But whatever the reasoning, it gave the Marshall product a different feel and sound. This same basic circuit with 4 output tubes would become both the high power Fender Tweed Twin that players like Keith Richards favor, as well as the Marshall Super Lead (aka 1959) whose playing legend runs from Pete Townsend & Jimi Hendrix up through Eddie Van Halen and others. Marshall would increase their punch and tightness even more soon by switching to a solid state rectifier and larger filter capacitors in the power supply, further differentiating their amps. They changed over to EL34 output tubes fairly early on, and made the amps progressively brighter as time went on. This basic topology with a few tone-shaping tweaks along the way ran essentially unchanged (save for adding a master volume in '75 ) until the JCM800 was released in 1981. An amazing run for any consumer product when you think about it. This basic topology is still the root of modern rock amps from a number of manufacturers.

Layouts

In practice, the Fender circuit was laid out as follows in the 5F6-A:

Author photo

And the original JTM45 Marshall was laid out like this:

I think I swiped this pic from a Mitch Colby facebook post! (thank you!)

The bias supply (left side) is laid out differently, but the reset of the circuit layout is identical, down to the filter cap in between the tone stack and phase inverter. Even the controls and switches on the front panel are the same.

Marshall used a better quality board and turrets where Fender had used vulcanized fiber and eyelets. The vulcanized fiber could absorb moisture (and did!) and lead to weird issues. But both systems were rugged and durable and have lasted over 60 years out in the world. That is pretty impressive for a consumer product built to a price point.

Extended family

The Vox AC30 borrowed it's top boost circuit with bass and treble controls from the Gibson GA70 and when the Top Boost circuit is in place, you have effectively the same pre-amp as the other variants we're looking at: gain stage, volume, mixer, gain stage with direct coupled cathode follower and tone stack following with a Schmitt phase inverter.

The big points of departure are interesting as well: the AC30 does not use negative feedback at all, and they used EL84 output tubes for their 30W of output. The prototype apparently used EL34s but they were too tall for the chassis design that Vox wanted to use so they swapped them out. The difference there is the required drive voltage. A pair of EL34s will bias up around -35V so you need a 70V peak to peak signal to drive full output. The EL84 biases in the AC30 around 10V, so a 20V peak to peak signal will drive it to full output. This same basic pre-amp in the Bassman is putting out 90V peak to peak. (-45V bias) the headline being we can definitely overdrive the output section very hard, and even harder since we are not losing any gain to negative feedback, we're really just limited by the power rail here. The amp will tend to be more mids focused and control the speakers less well also due to the absence of negative feedback.

The factory drawing can be a bit disorienting, so I have annotated it a bit to hopefully help out. This is a circa 1969 drawing shamelessly borrowed/stolen(?) from the truly excellent www.voxac30.org.uk site which has a ton of great info on these amps. The basic design had not changed from the prior amps and it was simpler to take this version which incorporates the top boost circuit so you can see things a bit better:

I blurred out the tremulant (vib/trem) circuitry to simply the view. It's most of the chassis.

Just to recap we have:

12AX7/ECC83 first voltage amp with shared cathode circuit
Volume controls, bright channel sometimes has a bright cap (like Bassman & JTM45)
Mixer resistors for the 2 channels
2nd voltage amp (also a 12AX7/ECC83)
Direct coupled cathode follower
Tone stack (very Gibson-y here!)
There is no presence control because there is no negative feedback loop, so no #7
Number 8 is our Schmitt phase inverter, drawn sideways here.
And our filter choke, between the plate and screen supplies.

The AC30 does have a "Cut" control which has roughly the same basic effect as a presence control; it cuts or allows the highest treble to come through. It is located in the output stage and consists of a potentiometer and a capacitor. You effectively "short" high frequencies between the two phase inverter outputs (which are 180 degrees out of phase).

The Vox AC50 takes this topology in a slightly different direction. While they kept the same 3 pre-amp tubes and 2 outputs, they built 2 separate pre-amp channels that are mixed later on. They also substituted a 12AU7 in the first position which is curious to me, but the headline is the key pieces are all here, if creatively mixed around a bit:

AC50 MkII schematic, (c) Chris Devine. Annotations by the author

1 - 9 are all the same, except again there is no #7 because there is no negative feedback used in this circuit. Like the AC30 before it, Vox apparently did not feel the need for it. This tends to give a more mid-focused sound, and a smoother transition into overdrive.

I like how they took the same 3 pre-amp tube positions and created two fully independent pre-amp channels with not just their own volumes, but their own tone controls. The mixer resistors therefore get moved from behind the volume controls until after the tone controls. The other tweak is they separate the cathode circuits of the first stage voltage amp - something Marshall would do as well, to give each channel its own flavor.

By the way, the 2nd stage in all of these amps - the triode followed by the direct-coupled cathode follower has such an interesting effect on things that it even got its own AES paper: https://secure.aes.org/forum/pubs/conventions/?elib=11614

More extended family:

The Traynor YBA-1 "Bassmaster" a few tweaks, and 6CA7 outputs.

We could do this all day.