A major issue with tubes is the output transformer, especially in SE use. In a single ended circuit the quiescent state plate current is going through the OPT, in a PP configuration the plate current is flowing in the opposite direction through each half. In a PP setup the magnetic field goes through the zero point on each cycle, for SE the magnetic field varies but is "biased" above the zero point. I am not sure if this is in SE's favor or PP's! But it surely means that the SE OPT must be more substantial and designed to avoid saturation. This is an interesting article https://www.x3mhc.no/dokumenter/SE-v-PP-Part2.pdf written by an SE fan
I do have an Elekit 2A3/300B amp (basic) that I now use to drive the Harbeth P3ESR speakers in my BlueOS "Kitchen" system. I had given my VTL85 to my daughter and had been using an NAD amp - OK for cooking by - but the Elekit, with 2A3 tubes makes for rather good listening.
Building that kit made me contemplate designing and building a fully balanced 65N7/300B amp. Over a couple of months I did just that. I'm using switch mode power supplies in a separate chassis (in rural VT power voltages are rather unsteady and the thoughts of designing and building regulated supplies put me off). Also separating the PS from the Amp keeps any AC from the amp chassis and allows me to move it all - at 80 years old I have limitations. When the amp is on and with no input signal the speakers are totally silent - ear 1" from driver silent.
Looking at load lines, and other designs, gave me values for the 6SN7 anode load resistor and their cathode bias resistors, similarly for the 300B cathode resistors. Because the LA4 pre has balanced outputs there is no phase splitter, just a single 6SN7 amp/driver stage, with the cathodes of each "half channel" strapped.
The MeanWell supplies have very low ripple (the WORST is 7 millivolts RMS on the supplies for the 6Sn7s). These supplies are adjustable +/- 10% so I was able to set the heater/cathode voltages to precisely 6.3 and 5 volts. Similarly I could play with the B+ to get to a good place in the load line simulator for the 300Bs.
It is the simplest possible circuit, zero feedback. No SS components in the amp. This also means that there is no protective circuitry so I do not leave the amp on when I am not nearby!
I used rather good components throughout (Vishay metal film or wire-wound resistors and Vishay caps except for Jensen coupling caps - the only cap in the circuit path). OPT's are Lundahl. Tubes are WE 300B (these had been in my VAC 30/30) and TungSol 6SN7GTB. Wiring is Kimber throughout - an old PBJ provided the connection from the Neutrik XLRs to the 6SN7 grids. Heaters/direct cathodes Kimber twisted pairs, not really necessary as it's DC.
The chassis were made by Protocase - a fun design activity in its own right. The physical component layout was done in Adobe Illustrator (from my Graphic Arts/paper engineering work for my wife, she designs holiday cards for MoMA.)
This amp is wonderful to listen to at moderate levels, note: I listen exclusively to "classical" music (from Josquin to Shostakovitch - and perhaps Gorecki). In particular, solo piano recordings are REAL, from the impact of the hammer to the varying overtones of the decay it is as though I was there in the recital hall, every nuance of the performance is presented. Similarly human voice, especially a capella, is magical, the Rachmaninov All Night Vigil is glorious and ethereal - I am in the chapel. The bass is solid, tuneful, controlled (though it does not hit you in the gut), the treble pure and the mid-range delicious.
If you do build your own, kit or otherwise, do invest in a good soldering kit (Weller e.g.), use 63/37 solder and have a "solder sucker" and solder wick available. Jameco is a great source for that sort of kit.
