Single electron tube vs single transistor audio power amplifier in class A operation with no negative feedback

At your risk. Potentially lethal voltage. 

Take all safety precautions. Additionally a suitable fuse should be used in series with any battery in order to protect from  a potential fire hazard or skin injury.

An additional precaution in the heft hand inside the back pocket. To the best of my knowledge this is a military rule of thumb in order to avoid an accidental electric current flowing through the heart.

It looks like the triode (or pentode triode connected) starts to win.

So far attempts to increase power output using bipolar transistors give a loud sound but there is some kind of grit, not a very clean sound.

For example the Darlington MJ1000, operated at Vb=12V, Ic=400mA with 21mH to 22mH output transformer with air gap.

Or 2N3055 operated at Vb=12V, Ic=140mA (must softer better sound, sweeter)

So far perhaps the best sound from transistors on these experiments is the Pleiades 2N3053 amplifier, one 2N3053 transistor for each channel with bias resistor from collector to bass (the examples above are biased from Vb to base). Operating conditions Vb=24V, Ic=12mA, output transformer 2.5Kohms to 8 Ohms. It is low power and it sounds fantastic on loudspeakers and perhaps this explains why the Neve 1073 using the 2N3055 at low power is reputed to sound so good.

Higher Vb experiments on the above amplifier gave louder sound at the expense of quality. What is left to do is to repeat the experiment with bias from Vb to base rather than from collector to base.

What about the electron tube?

Well the DL94 triode connected at Vb=90V, Ia=10mA, output transformer 2.5 KOhms to 8 Ohms and the EF183 triode connected at Vb=120V, Ia=20mA, output transformer 5 KOhms to 4 Ohms, both sing with depth, deep bass, clean smooth sound, hard to describe in words. On these amplifiers a step up input transformer of 1:7 is used to drive the control grid instead of a voltage amplifying triode! Bias is normally from cathode resitor bypassed by a good quality Elna magenta color Starget electrolytic. The DL94 was also operated with no resistors or capacitors and a Vc battery instead for bias through the input transformer.

Signal path of above experiments:

CD, for example Flashdance Soundtrack - Sony CD Walkman line out - Pleiades one electronic device power amplifier in class A with no negative feedback - Marantz 5-C for some of the latest transistor experiments and Celestion Ditton 150 for all the rest