Is all that clever electronics keeping everything in check a bit like driver aids on a car though. Some people value their ABS, traction control, lane assist, stability control etc etc, and other people buy 80s porsche 911 turbos, lotuses and TVRs. Its a different experience, is either better? I cannot belive anything unnecessary in the signal path or control circuitry is actually improving sonic performance. Maybe that relay is giving you a 99% safety increase, a 5% cost increase for a 1% sonic performance decrease, and that is a trade some people appreciate and some will not tolerate. A friend of mine was very happy recently to build a new wiring harness for a race car and save 800g and made getting the engine out possible without removing all the wiring harness, to most that is just pointless, especially when the cost of materials was more than I paid for my last car, but to her it was absolutely worth it. We all have different priorities.
The bit about the relay I can go with yes but the rest of the analogy, whilst I can see it, is a little tenuous.
Any SS power amp type
can have switch on and off thumps. It matters not whether it was designed in 1965 or 2021 nor does it matter if it's class A, A/B or D etc. Note I said can and not will.
In the vast majority of SS power amps the speaker is suspended between two voltage rails, one positive with respect to ground and the other negative. A typical 100W amp would use + and - 50V with respect to ground, 100V total.
This output midpoint is obviously kept at 0 Volt for no input signal and moves up and down in sympathy with the musical waveform to drive the speaker.
As I've banged on about so many times before, negative feedback (NFB) controls virtually every parameter of an amplifier (and hence is barely ever mentioned on hi fi forums
) and this includes the keeping of the output at 0V for no input. Off topic but when an amp "goes DC" NFB loses control and in this case the output pegs at either 50V or -50V either pushing the cone right out or sucking it right in and meanwhile putting 312W of DC through the voice coil which quickly fries!
At the moment an amp is switched on the output will try to peg at +50V or -50V (in this case) until the NFB loop gains control of the situation and pulls the output to 0V by "servo action" (which is what NFB is).
This is what switch on thump is. On switch off the reverse can happen... ie the voltage in the smoothing capacitors gradually discharges and often a point is reached where there is not enough voltage remaining to keep the NFB going but there is still say 10 or 20V left in the caps which discharges through the speakers making a switch off thump... or even odd noises and whistles as the NFB loses control and the amp becomes unstable.
How loud the thump is and what it sounds like vary enormously with circuit topology.. completely symmetrical, "mirror image" designs tend to pull equally in each direction at switch on and can have virtually no noise at switch on, usually a slight click only. Capacitor-less designs similar as there is nothing to charge up so it all happens so quick all you hear is a slight click. Other topologies can vary from a slight click to a loud crack to a slower thump depending on all sorts of factors which affect the speed at which the NFB gains control and the voltage reached at the output during this time. Even such aspects as the tolerance of the smoothing caps can come into it as if one is say 11000uF and the other is 10000uF then the smaller one will take less time to fully charge... same at switch off time in reverse...
Capacitor coupled output amps tend to be the worst as the capacitor has half the supply voltage on one side and the speaker on the other... At switch on the cap has to charge to half supply voltage and this happens through the speaker giving the often horrendous thump and speaker cone sucked in or pushed out during this time. In some units such as the Quad 303 it is possible to use another capacitor as part of the design which limits the speed at which this charging occurs and so all you get is a "flup".
Now with all the often contradicting factors in the design of an amplifier and with the need to get low distortion, low noise, decent slew rate, wide bandwidth etc etc etc, the issue of what happens during the first 0.05 to 0.3 seconds after switch on is well down the list of priorities! In fact it is only even considered when it could cause damage to the output transistors by them both switching fully on at the same time and destroying themselves... this is rare and beyond the scope of this however.
Sometimes there is a happy synergy where it so happens that fitting a larger capacitor in a part of the circuitry can slow down the charging (in a cap output amp) so there is virtually no switch on thump and the 303 is a good example of this. However what nobody in their right mind does is design the entire circuitry around avoiding a thump in the first 0.1 second after it's switched on and then worry about distortion, frequency response, hum and buzz, damping factor etc etc etc as an afterthought!!
So, it thumps or it doesn't to one extent or another and if it is really bad then a relay can be used at the output, usually doubling as speaker protection relay, which delays connecting the speakers up until the thump has passed.
The relay is usually powered from a separate supply with very small smoothing caps so that on switch off it disconnects the speakers before any switch off thump happens.