Micro Brown Bassman-Princeton hybrid build

[cdemike]

After wrestling with some hum issues in an older build, I decided to just start over with a clean slate. The old amp is a 6G2 Princeton using an ECC99 push-pull output stage, but it was a really early build of mine and there are a whole host of grounding issues that I haven't been able to resolve, and I have been flirting with both a 6G6 Bassman build and the idea of a switchable harmonic/amplitude tremolo system; this seemed like a pretty good opportunity to tackle all the above. I didn't really love the ECC99 output section as a model for "big amp" sounds, since the triodes didn't seem to sound that similar to pentodes when driven, so I'm ditching them for a dual EF80 section inspired by Rob Robinette's mico amp series. The build is keeping one channel similar to a Brown Princeton, since it's my wife's amp mainly, and she really likes that circuit. The trem circuit is based off Steve Luckey's Trem-O-Nator circuit and uses the Supro-style oscillator for a wider range of speeds.

There are a few things I wanted to get a sanity check on before buying parts:
1. I'm not sure if the use of a transformer phase inverter is a good idea here. After thinking about ways to cut down on complexity, I realized that small audio transformers cost about the same as a single 12AX7, and I don't really want it distort anyway. Maybe it'd be better just to use another MOSFET since I'm already going down that road with a SS oscillator? It seems like the transformer helps cut down on overall complexity in the signal path, but this is my first time designing around an inter-stage transformer.
2. Putting the bridge rectifier on the heaters seems like it might allow me to elevate the heaters pretty easily while also providing a voltage source to drive the tremolo transformer (this build will not hum....). The old build I'm cannibalizing used an ANTEK toroidal PT which has no center tap on the filament winding, so I wanted to see if this would also eliminate the need for an artificial center tap. I currently have a humdinger pot installed, so no worries either way, but it seemed like this was a two-birds-one-stone situation unless I'm misunderstanding something.
3. I eliminated the third gain stage from the 6G6 design since the harmonic tremolo stage read to me like a repurposed LTPI, which I know adds some gain, albeit less than the unbypassed 100k/1.5k gain stage it would replace in both the 6G6 and 6G2. Because of the gain discrepancy, I adjusted the second gain stage of the Bassman channel to compensate and moved the treble control to follow the second stage rather than the third gain stage. I didn't see a reason why this would cause a major difference in the amp's sound, but there's a part of me that's wondering if I'm thinking about this too simplistically.
4. Because EF80s have much lower headroom than 6L6, I adjusted the grid resistors to be about 10% as large as in the original schematic thinking it'd get me close to scaling the signal so the output section sees a proportional signal as in the full-sized amp. 20k seemed to be right since Rob Robinette's site mentioned there should be a roughly 90% reduction in signal strength, and based what I've gathered from the EF80 datasheets that seems to be true. Really this last piece is to see if others have built with EF80s or similar small output tubes and have had success with other approaches or values.

Of course if there are any other red flags involved with the circuit as it stands now, I'm happy to hear it!

[cdemike]

Been thinking this design over and making some progress. I've been mining rootz's thread on his 6G2 build (https://ampgarage.com/forum/viewtopic.php?t=36749) to learn as much as I can. There are a lot of solutions to design problems I've identified in my build which mostly pertain to keeping the resulting amp sounding as close as possible to the original circuits as I can but with the added functionality. I'm also working around space constraints in the chassis, which has meant I'm not sure I can stay with a 6G6 as the non-Princeton channel (the Princeton channel is a mandatory item, since it's my wife's amp and she really likes the 6G2). I considered a resistive mixer to combine the signal from the first 6G2 gain stage with the signal from the 6G6-B's treble pot at the fourth triode in the bass channel's signal chain. I decided against that since it'd result in a lot of interactivity between the 6G6 treble control and the 6G2 tone and volume controls. So I next considered a common anode mixer, but I have concerns about that (specific questions at the bottom of the post). The space constraints led me to realize I could switch the amplitude trem from an optocoupler style circuit to a bias-wiggle circuit in order to both simplify things and get a swampier-sounding trem.

I'm also likely abandoning the transformer phase splitter approach since implementing it seems to add more complexity than it tries to solve. So I'm currently drawing up schematics using the MOSFET-based oscillator from the bottom Sluckey's trem-o-nator page to drive a cathode follower which then feeds a phase splitter (either 12AX7 or LND150-based, depending on whether the broader preamp design leaves a triode free). The cathode follower's output could then be switched to connect at the junction of the output section's grid leak resistors like in an Ampeg J-12B (https://sluckeyamps.com/RCA/Ampeg_J12B.pdf).

I was hoping to get help on two things:
-Is there a way to get a common-anode mixer to behave more like the center-biased 100k/1.5k gain stage it'd replace? I'm less concerned about the loss in gain since my output section will be a lot easier to overdrive than 2x 6L6s, but I've encountered circuits that sound fuzzy because of the overdrive characteristics of the common anode mixer.

-I'm not sure I understand the operating characteristics of the harmonic trem's mixer stage. If I re-conceptualize that stage's shared 4.7k/2uF share cathode as a split cathode arrangement, I think it'd be 9.4k cathode resistance with a 1uF bypass capacitor. That'd create a somewhat similar operating point to a JCM800's cold clipper, which seems to run counter to Fender's general effort to maximize headroom. The cathode bypass capacitor would be an important difference, but the stage would remain biased pretty cold, reducing headroom, right? If that's the case, is there a significant downside to center-biasing with an 820R cathode resistor?

Thank y'all!

Schematics for reference:
6G6-B Bassman: https://schematicheaven.net/fenderamps/ ... _schem.pdf
6G2 Princeton: https://schematicheaven.net/fenderamps/ ... _schem.pdf
6G5-A Super (reference for 5-triode harmonic trem): https://schematicheaven.net/fenderamps/ ... _schem.pdf

[cdemike]

Maybe this would be better to ask under Technical Discussion?

Edit (5/19/24): here's a schematic with the design as it stands so far. Still wondering if this makes sense in terms of adjusting the splitter/mixer stage for the harmonic trem. I thought maybe keeping the common anode mixer stage unbypassed might in a weird way make it behave more like a vanilla gain stage by increasing the output impedance.

[cdemike]

Did a bunch of digging and thanks to some help from the Amp Garage community in the tech section, I've refined the design. It's definitely striking me as too complicated to work in the bass channel of any brown/blonde Bassman, so I've opted to work in switchable tone stacks with options to pick between the tweed/G62 style and the late brown style with the tapped treble pot. Since the tweed/6G2 style is much less lossy and cuts significantly less midrange, I may try to put it on a relay with the option to foot switch it as a kind of "lead boost" once the amp gets built. I'm sure there will be teething issues once it's built that require tuning and revision, so I'm trying to keep things beyond the original scope of the project to a minimum. The design I'm planning to build consequently has some major changes from before:
-It struck me that I didn't really need to bend over backward to have channel switching work via separate input jacks. This has felt like the Gordian knot of the project, and Sluckey made a post in the Hoffman amps forum regarding a similar project with similar complications pointing out that switching doesn't need to operate only via the jacks... Definitely grateful for all the resources available, especially when since I clearly needed to think out of the box regarding the channel switching. So that's now managed via a DPDT switch.
-With the switch simplifying things greatly and freeing up a bunch of space, the triode shortage resolved, making the trem circuit planning much more straight-forward. The new oscillator schematic builds on a post Martin Manning made in rootz's thread posted previously in this build thread. With the amount of loading in either trem circuit, it seemed like it might allow more flexibility in fine-tuning to use the cathode follower after the LFO in order to maintain signal strength.
-With the amp being a low-wattage build using a triode output section rather than beam tetrodes, I wanted to see if I could work a pentode into the preamp to compensate for some the difference in overdrive character. I've also been very interested in Tubenit's work on the 5879-based clipping circuits, so this seemed like an opportunity to explore that in addition. It obviously won't make the amp sound like it runs on 6V6s or 6L6s, but it seems like it might be closer. The values I used were chosen based on the RCA 5879 datasheet with an eye toward moderate headroom.
-Thanks to maxkracht's help in the tech section, I think I better understand the harmonic trem's operating point. I did some digging in a few threads following Revibe builds and the colder operating point seems important both in terms of the EQ as well as in terms of the functionality of the trem, since the trem effect operates by alternately oscillating the bias on each side of the mixer tube. So the cold bias impacts the depth and rate of the trem as well as contributing to the EQ of the mixer stage's output since the bypass cap will sound different at different operating points. With the 5879's significantly higher output impedance, it shouldn't be terribly difficult to tune this with split plate load resistors on V2 if needed. The low headroom does seem to explain why Fender used split plate load resistors in the original harmonic trem circuits.
-Since the PT I have (ANTEK 50VA unit with 240V secondaries) can supply significantly more current than the amp needs, I wanted to find a way to build a source of compression into the circuit (recognizing that sag will sounds and feel different than other kinds of compression). The cathode follower after the mixer section seemed like a reasonable way of accomplishing that, especially since the chilly bias on the mixer has been observed in a few Revibe threads I've come across to have low headroom like I suspected; since the mixer stage may function like a cold clipper, the DCCF following will impact the same side of the wave-form and hopefully contribute to the same production of even-order harmonics.
-Given the smaller output of this amp vs 6V6s or 6L6s, a LTPI seemed like overkill and potentially setting the output section up to receive far too much signal, especially in light of the switch to a 5879 as the second gain stage. A cathodyne seems to make more sense in this context, and since the PI will be receiving a much hotter signal than in an 6G2, I wanted to keep the PI relatively stable with an effort at minimizing fizziness. So the approach I took to to cathodyne used the Paul C 5E3 mod as the starting point and incorporated Merlin's suggestion of a big grid stopper as extra insurance. The switch in PI design created a separate problem in terms of the NFB loop since neither the 6G2 nor the loops from the bigger brown amps that used LTPIs would work. Since the pentode was included with the aim of having it contribute to some of the overdrive character, it didn't seem like a good place to insert NFB even though that would be the closest analogue to the 6G2 schematic. Likewise, inserting it in the trem mixer seems like it may unnecessarily interfere with the harmonic trem's function. The simplest solution may be to have a gain stage following the mixer and insert NFB there like most cathodynes, but that would make the signal strength reaching the output section even greater than with a LTPI (another motivator behind the decision for the DCCF following the mixer). I remember the Marshall JTM1 used a cathodyne with the NFB being fed in at the PI's grid, though that cathodyne was self-biased, but it seems like this approach might work.

I'm hoping to start ordering parts since I'll have some more time during July to spend working on it, so I wanted to post the schematic and see if this design makes sense. This schematic looks pretty different than either of the amps I used as a starting point and also seems pretty different than a lot of the schematics I've seen, so I'm somewhat nervous that I'm way out in left field with this schematic, so I'm really hoping to get input before I order the parts.