Why would we want to reinvent that particular wheel if one can use an inexpensive 48v to 1.5 volt XLR transformer adapter?
I am not sure what you mean by the phrase "inexpensive 48v to 1.5 volt XLR transformer adapter," do you have a link to something in particular?
In general transformers are not inexpensive when compared to the price of a couple of transistors and a handful of resistors, and when you are very space constrained as in this application, transformers are never smaller than a couple of transistors.
Do we know for sure the capsules can take higher voltage
The capsules do not have to "take" any voltage, they are pre-polarized electret capsules, meaning the backplate has a captive charge, so they provide their own polarizing voltage.
A previous post pointed out that these Audio-Technica microphones have a small PCB directly behind the capsule with a 2SK660 JFET device. See here for one data sheet (Renesas in this case, although I think multiple manufacturers have built compatible devices):
Renesas 2SK660 datasheet
Note that the maximum drain to source or gate to drain voltage is 20V, so this device could safely run with 10V across the JFET without any concern.
Most of the recommendations, e.g. using the Schoeps style circuit, would put much less than that across the JFET. The traditional Schoeps balanced buffer has a 12V zener diode providing 12V voltage for the circuit, and ~2K resistors in the drain and source of the JFET biased at around 1mA. This JFET would typically have around 0.25mA bias current at 0V gate-source (Idss range of 60 to 500 uA), which would drop around 0.5V across 2k Ohms, so the drain would be at 11.5V and the source would be at 0.5V, give or take. You could scale up the resistors proportionally to the reduced bias current, so e.g. 10K resistors in source and drain (a circuit configuration called a phase-splitter), which would then drop 5V across the resistors, for 7V at drain and 5V at source, which is probably a good range across the JFET (2V drain-to-source).
a result of a poor differentiation between phantom and bias voltage
Phantom power refers to the scheme of providing power for microphone buffer circuitry on the same wires that the audio travels, so that a microphone will not need either an internal battery, or separate wiring for power.
The standard is 48V connected to two 6.8k Ohm resistors which are then each connected to the two signal wires to the microphone. The shield provides the return current for the power. Any current provided to the microphone will cause a voltage drop across those resistors (which are in the pre-amp), so the microphone never actually has to deal with 48V, but exactly how much less than 48V depends linearly on how much current the buffer circuitry uses.
Bias
voltage in the context of a microphone would typically refer to the voltage applied through a high resistance to charge an externally biased condenser capsule. That is not applicable in the case of an electret condenser capsule, because the electret film contains captured charge and does not need any additional external source of charge in order to function.
In a general sense the term bias refers to to the quiescent (i.e. even with no signal) operating conditions of a circuit, so in a buffer circuit using transistors you have to design the circuit to set the bias
current appropriately through each transistor to keep the transistor near an optimum operating point. There are various ways to achieve that which are discussed in various texts describing transistor circuit design and which you can see in all the variations of buffer circuitry devised by the various microphone designers through the years.
I understand these capsules have internal FET and so that speaks of bias, not phantom, voltage application to me.
I hope you can see from the discussion directly above that yes, you have to be concerned about the bias conditions of the FET, but that voltage and current has to come from somewhere, and that somewhere is the phantom power provided to the microphone by the preamp, so they are not really separate concerns. The job of the microphone designer is to decide how to take the voltage available at pins 2 and 3, and provide a useful circuit which has a very high input impedance to present to the capsule (which is where the JFET comes in), and can drive the output pins with a lower impedance. For noise rejection purposes the impedances from both pins 2 and 3 should be very close to each other when referred back to pin 1, whether or not both pins are driven with microphone signal.
All the previous discussion was regarding various ways that aim might be achieved, and fit into the space available in the 822 microphone body.
