Most high performance output stage designs I've seen so far use the triple emitter follower in various configurations. While this design performs well, it has some difficulties to deal with. Maintaining a constant DC bias operation point with changing temperature is difficult. Also, electrical stability is at the edge. The double emitter follower is super stable, but by far does not offer the high level of performance. I wonder why output stages configured as diamond buffers aren't used more often. I did a bit of research that I share in an extra article. The diamond buffer is more stable electrically than the triple emitter follower. Both can be tamed by adding shunt compensation at the input of the stage. From a thermal point of view, the driver transistors compensate the power transistors without requiring typical VBE multiplier arrangements.
Diamond buffer variants
This is the well-known basic Diamond Buffer circuit, a unity gain buffer often used in OP-amp output stages.
This is a slightly improved variant of the circuit with reduced voltage swing across the driver transistors and lower distortion due to local feedback.
Some benefits of the diamond buffer presented here: The constant current sources do most of the heavy work, while the driver transistors are held at fixed VCE. This reduces all kinds of non-linearity of the driver transistors and fast, low voltage and low power transistors can be used. The triple emitter follower has all transistors of all three follower stages subject to very large VCE voltage swing variations and therefore high influence from Early effect. The constant current sources employed in this design are derived from my research I detailed here on my website. This type of CCS has very high PSRR, does not overshoot on transients and due to being cascoded, the power transistors can be cooled on the main heat sink without having influence on the standing current. As reference for the CCS, I used two diode connected transistors here since this has slightly lower voltage headroom requirements than the LED referenced CCS. In any case, this driver module needs to be run from power supply rails that are higher than the output stage rails. Diodes D5 to D8 are for overload management. I don't expect D7 and D8 to ever conduct. Since they are held at constant voltage, they do no harm in any way. Diodes D5 and D6 are very important as they shunt excess input voltage to the output node. While this is not nice for the front-end driving the buffer, this avoids saturation of the buffer, which would not recover gracefully. This is just a fallback measure since all front-ends of the modular amplifier system are designed to avoid saturating themselves or subsequent stages. R11 and C1 are just in case the front-end does not offer provisions for a shunt compensation network. By all means, this should be installed on the front-end PCB. C2 and C3 allow push-pull operation. Only one of them is to be installed. The bias of the output stage transistors is set by resistors R9 and R10. This is a bit inconvenient as both resistors need to be the same value and also impact DC offset. DC conditions are what I see as the main uncertainty that needs to be evaluated once the driver module is built an hooked to a BJT output stage. I'm very curios how the module performs on the bench. In case it doesn't perform well, I can quickly swap in a different driver module and end up with a very different amplifier. While the schematic looks fairly complex, the actual signal path is very short.
It goes without saying that all stand-offs and transistors are on the mounting hole grid of the chassis. The driver transistors are mounted to the main heat sinks in order to sense the temperature and adjust the output stage bias accordingly. And the cascode transistors of the CCS are mounted to the main heat sink as well since they dissipate significant power, while the rest of the CCS does not.