Class A (in this case this tube amp) means the supply current is always higher then the maximum current that is to be delivered.
This determines the current in the output stage.
At optimum bias setting (around 13.5 Volts) the G1/G2 is capable of delivering 12 Volts peak-peak (= 6V peak = 4.3V eff in 1k resistor)
the output resistor 47 Ohms + impedance of HP (lowest = 16 Ohms) means the maximium current that can be drawn = 6 V/63 Ohms = 95 mA.
The current in the ouput stage in this case (the G1/G2) is dictated by the glowing current.
This needs to be around 350 mA and this is spread over the 2 channels (this is where the crosstalk comes from) so about 175 mA is needed.
More than enough for the maximum drawn peak current (95 mA) for a 16 Ohms HP.
Opamps are NOT in class A (when loaded with Low Ohmic things like cans in C'moy alike topologies) and thus operate in class AB.
The output stage of an opamp needs a certain minimum current (a few milliamps) the rest of the opamp current is used for the other transistors/fets/parts inside the IC (usually a LOT of parts)
so class AB (opamps and most SS designs) draws little current.
When the supply voltage is the same (let's assume) then the power (P in Watts) is U (voltage in volts) x I (current in Amps).
example: class A current 150 mA, voltage 24 Volts = 3.6 Watts (all dissipated in HEAT)
class AB current 2mA , voltage 24Volts = 0.05 Watts (again all dissipated in heat)
the power of a class A amp (in this case) is 72 times that of an AB type amp, hence the difference in temperature.
For the record if you load an opamp high-ohmic (much higher then HP's) it will also be in a class A state.
there are 'tricks' to let it stay longer in class A too.
Now for the specific Indeed.
current in output stage = approx 165 mA and is determined by the LM317 + resistor.
Voltage across MOSFET (when bias set at 14V and a IRF630 or IRF510 is used = (D-S voltage) = 14V.
voltage across LM317 = 2.5 Volts (U source - U filament tube - Uref (1.25 volts)
voltage across filament = 6.3 volts.
They all carry the exact same current of 165mA so power is known.
MOSFET = 2.3 Watts.
LM317 = 0.4 Watts.
Now I don't know much about cooling ;D but the LM317 cooling fin is specified at 40oC (or 40K)/Watt.
so 0.4 watts x 40 = 16oC increase in temp from the outside temperature.
Here there is a snag as this is specified in free air convection which is NOT the case in these amps as they are close to a warm tube (that heats the surrounding air) AND there is no free air flow in the amp.
It remains sorta trapped between the PCB and top plate (removing the perspex top plate should HELP) and increasing cooling surface too (this lowers the K/W ratio)
so the increase in temp will be higher than the calculated 16oC .
When the air temp that surrounds the LM317 is around 30 oC the fins would become 46oC.
The same calculations can be done for the MOSFETs.
the cooling fin is specified at 15oC/W and the power the FET dissipates = 2.3 Watts (talking about Bravo/indeed/miridiy amp NOT G2 although it has the same current the K/W of the fins are lower in the G2) 34oC temp increase.
As these fins are close to the outside (G1 not in case of the G2) the free air flow is pretty much present and only very slightly obstructed by the PCB so outside temp = 20oC, increase = around 35oC means a temp of 55oC for the MOSFET's
increasing the airflow lowers the temperature AND increasing the surface area decreases the temperature simply because the thermal resistance (K/W) is lowered.
The dissipated power remains the same, the temperature will alter.
Hope this technical explanations clarifies some...
The high temps are not much to worry about.
The effects of this will not be seen for 6 to 10 years of use.
In the coarse of time the solderjoints of the LM317/MOSFET and tube will be going bad because of the high temperatures.
Resoldering after a few years MIGHT be needed.
Do keep in mind that although the parts are specified to still work within parameters at high temperatures the lifespan will be shortened, the higher the temp the more dramatic the effect.
Cooler parts simply have a longer lifespan.
This 'trick' (too high temp on PCB's) is often used by designers on purpose to shorten the lifespan of commercial equipment (TV/DVD/audio).
Also placing electrolite caps next to hot components is used to purposely shorten the lifespan of electronics. (TV/SMPS)