Well it's nice to know that you shouldn't be so paranoid about it. And proof being in the pudding, this Sorado head I just got looks like it was run for who knows how many years from the 8Ω tap into a 4Ω cab, actually measures something like 3.6Ω. Although I tried both outs, and [I thought] it was louder out of the 4Ω out, so the head has to work harder to achieve the same volume level, hence heating up?
Another testament to the durability of tube amps in general and Sunn amps specifically; I have an ex pro touring 2000S rig that was gigged
hard and put up wet for a decade or so, and the
whole time it was driving two 4 ohm cabs for a combined nominal impedance of 2 ohms using the 4 ohm tap.
When I got the rig the amp tested at factory spec., no harm done.
What you are actually measuring is the voice coil's DC resistance, not AC speaker impedance. Both measured in ohms, but two different values.
Impedance is referred to as "nominal" when dealing with speakers, because it's sort of an
average impedance to current flow, if you look at a driver's response curve there is also generally an impedance curve with it as well. If you look at one of those impedance curves, you will notice that an 8 ohm driver is only
actually at 8 ohms impedance for a small window of time/frequency.
Try the Eminence website's speaker spec. sheets to see what I mean.
It's not really so much a case of the amp "working harder", it's more of an efficiency issue. A tube amp running at matched impedance is "balanced", and performs as it was designed to.
However, given the fluctuating nature of speaker impedance there needs to be a rather large degree of load flexibility designed into the circuit and/or with the chosen components in order for it to function in practical application, so a 100% nominal impedance mismatch isn't really pushing the boundaries as much as it might seem by looking at the speaker's advertised
nominal impedance.
Of course, little of this is really relevent with non transformer coupled transistor amps, in which the only clear and present danger is running too
low of a load impedance. If the impedance gets too low for the transistors to dissipate the energy effectively, it will melt down like Chernobyl.
There was a hip phrase going around here a while back about people "running at 2", which was a reference to running 2 ohm loads on their transistor amps for maximum output.
That phrase was used like some badge of honor.
I'm not sure where it came from, but IMO it's more like a "badge of ignorance". The speaker response graphs I referenced earlier will attest to the fluctuating nature of impedance curves, and 2 ohms is
WAY too close to a dead short (kind of like a cyanide capsule for transistor output stages) to be messing with, even if the amp is 2 ohm stable.
But, as usual I digress...
Again, the search feature is your friend; these subjects have been covered here ad nauseum over the years...