More How

Building a replica of a well-known classic from a set of detailed plans would pose some challenges for me, but I've thought through it and feel confident I could pull it off without too much trouble, especially since it's been done and documented by many a DIY'er (such as TubeGeek here). As I was drawing up the circuit and layout for a modified design with an optional cascading gain stage and other enhancements, it was clear that I just didn't know enough to make intelligent and informed choices about how best to do things. Advice from a seasoned builder on one of the forums pointed me in the direction of Randall Aiken's website on amplifier design, and to Richard Kuehnel's website and books on the same subject. I refer to both with some regularity, and expect I'll do so until I'm done. Michael also sent me one of Richard's books on pre-amp design, which has helped me test and validate some choices with respect to my modified input stage. (For those having a look at the links, the math is easier than it looks, once you familiarize yourself with the basic concepts and architecture of amp circuits.) Also, a Texas Instruments circuit simulation program (TINA) referenced by a user on one of the forums has also been handy to examine the theoretical output and response of parts my circuit across the audio frequency spectrum - and has helped me to make what I think are good guesses about what resistor and capacitor values will work best in the customized parts of my circuit, which at this point looks like (click pic to open):

Here are results of some of the TINA simulations I've done in order to compare the output of my circuit at each pre-amp stage, to some of the classic cascading triode pre-amps (the Marshall JCM-800, the Friedman BE-100, and the Soldano SLO-100). The plots show the gain and frequency response at each stage following the input stage (each labeled "V#") up to the phase inverter (PI), which splits the signal in two, reverses the phase of one of them, and sends each to a pair of output (power) tubes. The labels for each of the curves are a bit confusing (and inaccurate in places), but you can see how each amp increases the signal voltage in different ways, at each preamp stage, to both distort it and get it up to a level appropriate for their respective power sections. 

 

Since one of my requirements is lots of tone shaping options, I've decided to implement two separate tone control circuits , or "tone stacks", optionally switchable (as you may have already concluded from the schematic). There is a nifty tone stack simulator available on the Duncan Amps website, and lots of other very useful design tools. It helped me arrive at some likely values for my tone circuit components, but it can't tell me how MY signal will look coming through, since my signal is already altered by the preceding circuits and tubes. 

So I also used TINA to get a sense of how my tone circuits will affect my signals. Below. my two input channels are referred to as "5F6" and "JTM", and you can see from the lower curve in each plot that, the signals are characteristically different coming out of first pre-amp stage, prior to passing through the second tube and tone stacks. My aim is to implement the tone stack as designed in the original 5F6A (labeled "TMB" in the plots below), plus add the second circuit (labeled "James" ... more on the James circuit here) to allow shaping of my tone more toward that of a Fender Twin Reverb (Blackface), or an Ampeg SVT.  Here are some sample results of the many simulations I did to figure out whether my selection of component values is doing what I hope:

Also, TINA has a virtual oscilloscope that has helped me arrive at some component values to manage signal gain between stages when the amp is operating in the cascaded channel mode, passing the amplified signal from one channel back into the input of the other, adding one more gain stage to the overall circuit. Below you can see the clean signal trace (bottom-most sine wave) coming through the first gain stage, then a mildly distorted signal after the second, and finally a significantly distorted signal after the tone stack.

 

No amount of simulation or modeling can perfectly replicate the behavior of the real thing, but I think it can get me into the ballpark. Before I call it done, there is little doubt I will have to experiment with many component values and possibly their placement in relation to other parts.

Here is the final layout I had prior to starting the build (click pic to open). I started with Doug Hoffman's circuit layout, and Rob Robinette's chassis layout, and made modifications from there.