90db 4 ohm with tubes?

I'd agree the problem here is a speaker with wide load impedance fluctuations coupled to a power amp with high output impedance. My stipulation of good transformer and power supply isn't the solution in this case.NFB is an option for lowering the amp's output impedance but IMO introduces more harm overall than good. Widely fluctuating speaker loads just seems like a bad design choice. Why not just design a speaker with a flatter impedance curve? This would allow more amplifier choice and flexibility.
Regards,

Charles...., you are spot on correct. But to satisfy your own curiosity, check the published bench test reports on a number of the "Big Boy" speakers like Revel Salon 2 and Magico. They have some pretty rough impedance curves and the manufacturers recommend high current - high power amps (i.e., SS) which usually have extremely low output impedances.

It is my anecdotal understanding that there are a few speakers that have relatively flat impedance curves that would make them tube friendly. Your point is really a "bitch" issue that the big tube amp manufacturers need to take up with the big speaker manufacturers.

In my case, I like the sound of my ARC amp, NF and all. Fortunately, through dumb luck, I only recently figured out that my amp can drive my speakers without adding too much "acoustic flavor."

Cheers,

Bifwynne,
A good friend of mine recently bought a pair of the Rockport Altair speakers. They are a nominal 4 ohm load but are driven effortlessly by tube amplifiers and sound wonderful and very natural. For this reason I'll assume they must have a relatively smooth impedance curve and thus aren't restricted to high current SS amplifiers in order to be driven. I wish other designers would follow Andy Payor's example. If it's going to be a lower nominal lmpedance, strive to keep it reasonably flat.
Regards,

I think you will still have issues with frequency response variations due to impedance variations, if you run a large solid state amp. If you run a speaker with a nominal 8 ohm impedance that rises to 16 Ohms, and drops to 4 ohms at times, I'm guessing that may cause up to a 6 db variation in levels at different frequencies, with an amp that doubles its power every time the impedance is halved.

For example, if an 8 Ohm speaker rises to 16 ohms at roughly 12 kHz, the amp (large solid state) will only put out half of the power at the time, causing about a 3 db drop in sound level, at that frequency and time. If the load drops to 4 ohms at 1 kHz, the power will double at that time, giving it roughly a 3 db level increase during that drop to 4 ohms, due to the amp doubling its power at that moment. That's why I'm guessing up to a 6 db variation in this case, if the amp doubles its power that way. Even more of a change in frequency response variations yet, if the speaker impedance varies more than what I just mentioned.

So a potent solid state amp that doubles its power when the impedance is halved, may be less neutral sounding then an amp that can't double its power. I can see this happening. So basically, its still about how each amp reacts to certain loads demanded by various speakers, solid state, or tube.

With my other post, I'm thinking what an amp may do when it just has to power that certain frequency (or frequency range), that has a high or low impedance variation, in comparison the frequencies that are more linear as far as impedance goes. It may be more noticeable during a solo performance too. When the amp has to power other frequencies at the same time, the power to that frequency with a different impedance can be more stable is what I'm thinking.

That speaker mentioned above with the 28 ohm peak, who does know how it will sound when something, maybe even more-so, a solo performance that may have a lot of music at the frequency itself a lot of the time. It may do okay with one amp, have problems on the next. That's one major thing I like when they do reviews, is the part of measuring something for possible problems to be aware of.