> copying it, use tubes and readjust the circuit
"readjust"?
I don't have all JH's products memorized, but I suspect you speak of either a 990 with frills or some similar opamp-ish affair made with solid-state.
Here's the deal. There are no P-type vacuum tubes. The output pin of a tube is "always"(*) more positive than its input. You can't require that both input and output be near ground without some major compromise.
Fer eggample: input grid near ground, previous plate near +100V. You can rig-up a -300V supply and use a "voltage divider" to deliver about 3/4 of the signal voltage to the grid at (we hope) about the right DC voltage. However, the resistor divider gives terrible power transfer. While we often pretend a tube needs no drive power, in fact the grid capacitance limits the bandwidth so a resistor divider gives a serious loss of HF response. There is a strong temptation to stick a little cap on the top resistor. An alternative is to use a neon bulb (or Zener) to give better transfer, but noise may be a problem. Any such scheme hurts gain.
When you turn to solid-state: now P-channel devices are possible and common. You start near DC ground, say with N-type. Its output is near the V+ rail, so we put in a P-type device to give gain and get an output more DC-negative than its input.
Direct-coupling with near-zero-DC input and output is possible, sometimes easy, when you have two polarities of active device. It is mighty darn difficult when you only have one polarity. The compromises needed to make it "work" tend to overwhelm the design.
I'm all in favor of you studying it. But I think any good sound you hear in the M-1 is due to JH and probably Deane Jensen and others tweaking similar designs for several decades, not anything as "obvious" as omitting caps.
(*) OK, there are exceptions to prove the rule. A Pentode can be screen-driven, and can have the plate output less-positive than the screen input. Voltage gain is low and current gain is very low: it is a crummy amplifier. It is sometimes used in DIY output stages. The low gain tends to be linear, but I think novelty is the main attraction.
It is also possible to work a tube positive grid with a low plate voltage, lower than grid voltage. But you quickly reach a point that both voltage and current gains become losses.
You can stick batteries in series with the output to level-shift. Current drain may be quite low, so life could be long, but obviously a headache. And batteries tend to get noisy with age, and the fix is.... a capacitor.
