Hi, shame on me for keeping silence for so long. I missed you All. As always - good reading on this thread!
I found this note in my files on Bob Carver, while his name was mentioned above:
[...]
In 1972, Robert Carver, then president of Phase Linear Corporation, published a paper outlining Phase Linear’s research into sonic differences between power amplifiers. This was by way of explaining why his company had launched a model producing what seemed an outrageous 350 watts/ch. Back then, it was described as a ‘700 watt’ amplifier by adding-together the power capability both channels, or else by citing ‘peak’ power.
Re-appraising power
In his 1972 paper, Carver cites a series of measurements which drive home a fact that was then less well known than it should be: that the ‘loudness’ capability of a power amplifier can be substantially affected by the supply rails’ regulation and voltage. This is at least one way in which, for a given loudspeaker, the highest undistorted sound level (whether experienced or measured) doesn’t necessarily fol- low the amplifier’s continuous rated power.
PSU dynamics
Carver noted that this was particularly true if the amplifier’s power supply was fairly ‘soft’, i.e. one exhibiting significant voltage sag on continuous drive, thanks to an undersized transformer, or reservoir capacitors. For example, an amplifier rated for a continuous 100 watts (into 8 ohms) with a badly soggy supply dropping from +65v to +45v on load, could sound louder on music passages, than a 200 watt amplifier (again into 8 ohms), with tightly regulated rails of + 57v. Although Bob Carver wasn’t the first person to stumble across the effect, judging by what he implemented, he clearly glimpsed consequences that others didn’t. Subsequently, the effect was recognised as ‘dynamic headroom’.
Dynamic headroom
For domestic Hi-Fi in the US, the IHF (Inst. of Hi-Fidelity) defined dynamic head- room nearly two decades ago, in their IHF 202 specification. It’s the maximum average (or short term ‘rms’) power available for just 20mS every 500mS, beyond the rated continuous power. Sometimes it’s called ‘burst power’. Since 1987, there’s been documentary evidence of something amplifier designers and audiophiles have long suspected: That music’s peak SPL requirements can extend in bursts to 300mS and beyond, with a 20% duty cycle, ie. lengthy transient bursts can recur every 1 to 2 seconds.
[...]