"The only difference between 50 watts and 100 watts is 3 decibels".
While this is mathematically and electronically technically true it's a little too "black & white".
Agree, that's why in my responses to such questions I typically referred to bench tested outputs. For example, you can compare the bench test data obtained by the S&V magazine, or Audioholics. If both units are tested under the same or very similar conditions you can have a valid argument of the difference between 50W and 100W.
Many times we are not considering anything more than the manufacturer's suggested specifications. I say suggestion because as we are all aware, well most of us, that the manufacturer's published specifications are often done under entirely different conditions and don't universally apply to all products in the same way.
I would give you basically the same response as above. In addition, there is a
little more validity if we compare the specs of products of the same manufacturer, such as comparing the output of the Marantz SR5010 and the SR7010. To come up with a better argument, I would still refer to bench test data.
One product advertises 50 watts per channel and another advertises 80 watts per channel. Both have the same quantity of channels. The argument I usually see in regards to comparing such products is that the difference in power would be negligible because the wattage difference is less than double the output difference is less than 2dB and not readily noticeable. Now assuming that both products reach their specified ratings at exactly the same distortion levels and going beyond that produces the same level dynamic compression and/or failure then yes, the difference is negligible between these two products.
The problem with this argument is that we do not know how these two products will behave after they've exceeded their manufacturer specifications. Maybe one of the products shuts down or distorts so horribly that it's unbearable while the other continues to produce sound that doesn't have noticeably audible artifacts.
This is valid but only to a point. Let me give you an example where this point of yours can be irrelevant. In my two channel system, I typically use 0.1 to 0.25W when sitting 9 to 10 ft from the speakers. 99% of my music collection contain dynamic peaks <20 dB. I have yet to see my peak indicators reaching 30W per channel. Most of the time the highest peaks would stay below the 15-20W mark 100% of the time. So even if there are rare momentary peaks that my meters and/or VU meters failed to display, I know a 50W or 100W rated amp should make no difference for me. As such how they behave when being over driven pass their clipping point is relevant. In fact, I have try this system with my very old Denon 3805, and a few other power amps; and am convinced that the 3805 has more than enough power for this particular system.
The transistors have the ability to output instantaneous current far exceeding their specifications, and often times they do. When your power supply is out of juice, it's out.
If you define "out of juice" to mean just exceeding the rated output, then you are not quite correct (your last sentence). If it was true, HK, NAD kind of AVRs would have yielded superior two channel driven outputs relative to their D&M and Yamaha counterparts. Search HTM, S&V and HCC's bench test data and you will find that more often than not the opposite was true. I can only guess that HK and NAD amps are typically amplifier(electronics) limited whereas D&M, Yamaha etc., are power supply limited. As EE, I prefer the latter, but it all depends on the applications.
Power transformers typically have excellent overload capacities. When overloaded, the output voltage of a power transformer will dip more, say instead of dipping 1 to 2% at rated output, it may dip 10% or more when severely overloaded, but it can still deliver current much higher than their rated current, and that would work well with help from the capacitors. Well design power supplies that also include large capacitors will have very good sustained slightly (say 10 to 15%) overload capability and much higher short term overload capability. Transistors circuits also have internal resistance and they do have to dissipate a lot of heat especially when driving reactive loads so they are also current limiting to certain extent. The high instantaneous current you refer to is exactly that, instantaneous, i.e., very short duration peak current. They don't have the sustained overload capability that power transformers do have. I am quite sure modern AVR's electronic protective circuits are designed mostly to protect the power transistors . Power transformers, again, are more robust by nature and typically can be very well protected by just fuses.
In industry, high power solid state drives are often protected by high speed fuses specially designed for protecting semiconductors. On the other hand, power transformers are quite often protected by fuses that are slow acting to take advantage of their sustained overload capability. Google power transformer fuses characteristics, look at the Time/Current graphs and you will see how much abuse transformers are expected to take. Just one last example, the rule of thumb of a 3 phase induction motor's start up current is 6X its rated full load current. A transformer that is rated to match the motor it powers in a "direct on line" starter situation must be able to handle that kind of current for many seconds otherwise it just can't do the job. Sorry, I can keep on going but I should stop.
Anyway, I really do agree with your sentiment.
