TS > LS> Well, how well does the crown 10kw amp old up 
TS > I don't recall.  Note that it's rated at 2/3 of an ohm, not TS > 8 ohms. 
 Obviously they're good enough for the target
TS > customer mil contractors.
 LS> Quite stable, more stable than the Crest Audio 1000x series.
Note that those Crown amps are audio amps, but the intended use is NOT to 
feed audio to speakers.  It's to see what might fail from resonant or other 
undamped vibration in electronic or mechanical structures.  
 LS> As you know, or should know, the impedance of a 
 LS> speaker is a reactive load, never stays at one 
 LS> impedance.  I could visually see that happen over the 
As you might have noted by now, I work with RF as well as audio.  In tuned RF 
circuits, it's more normal to use an impedance bridge, vector voltmeter, or 
network analyzer to tune critical matching points in a transmission system 
for matched resistive elements of impedance, and to make reactive elements 
negligible at a carrier frequency and minimal and symmetrical around it.  
In speakers you do not have a reactive load.  You have a load some element of 
which is reactive, operating over a wide enough band and with transient needs 
such that it's not practical to use matching networks.  
 LS> woofers at 800 watts RMS near clipping from my amps 
 LS> (8 ohms, dipping to 3 ohms at 50 hz, a characteristic 
 LS> that Yamaha would not explain).
Many speakers tend to have a resistive and inductive reactance component that 
contribute about equally to the total Z.  In a woofer, the resistive 
element's value is approached as you go down in frequency approaching DC, 
where the X(L) drops to zero.  The speaker when in use acts as a 
motor/generator, and often has a dynamic change in impedance based on its 
mechanical construction and loading from the enclosure, those mechanical 
factors changing the electrical load presented to the amp based on voice coil 
position and motion.  
You're unlikely to see full specs published on most speakers, as the details 
can get a bit complex.  With solid state design, it's easiest to just use 
overkill damping factors and assume a good amp can provide current wasted on 
reactive load demands that don't transfer real power.  OTOH, resistive 
loading by a voice coil may transfer more energy to heat as magnetic fields 
are generated to move the voice coil than reactive current loading.  There's 
a certain amount of art to transducer design, not readily definable in 
scientific terms.  
 LS> Current gain from the amp, not the wire!.  My amps can carry 100 amperes 
Modern solid state amps are designed as voltage gain circuits, whose outputs 
are nominal zero ohm current sources.  If such amp designs were based on 
current gain, the output would bouce all over the place based on load 
impedance variations.  Gain is the ratio of in/out.  In solid state 
electronics, the Hfe of a transistor is the most common application of 
current gain.  Some industrial telemetry uses current loops for noise 
immunity.  That's not the term you apparently mean here.  It sounds like you 
mean to say peak and sustained current sourcing capacity.  
 LS> per channel continuous, 200 amperes peak.  I tested 
 LS> this out about 20 times a year when the speaker wires 
 LS> short circuited and then melted the insulation on my 10 guage cabling.
Unless you defeated normal AC line protective devices, how did you sustain 
100 amperes?  
TS > #6 or 8 wire would likely survive.  #1 Cu would be required TS > for 
that current under NEC if it were house wiring.
TS > Allowable voltage drop under NEC translates to undesirable TS > damping 
factor degredation for audio.
 LS> Hmmmm.  NEC standards do not come into play on speaker 
 LS> cables, unless you are *certain* about the current 
 LS> gain coming out of the amp *steadily and continuously*.
NEC standards do apply to speaker wiring as is being discussed, and these 
power levels don't receive low voltage code benefits.  Code regulates cable 
types for various installations, plenum exposures, etc.  As to use of high 
voltage code standards, they're only a guideline for safe wire heating.  
Damping factor would be quite poor at the voltage drops allowed by NEC for 
120 VAC, and so usually you want wire that well exceeds sizes where heating 
is a problem.  
TS > tubes in an audio amplifier.  Using a single pair of
TS > non-custom catalog item tetrodes made by Eimac, it's
TS > possible to design a 2 megawatt audio amp (larger if you TS > parallel 
tube pairs).  With operating headroom, that would TS > be practical for a 2.5 
MW transmitter emitting A3
TS > modulation.  1 MW is the largest standard catalog item TS > transmitter 
made without using parallel amplifiers.
 LS> Tubes are only OK for midrange from 100hz to the highest highs due to 
 LS> transformer saturation I've found in Cary, 
 LS> Counterpoint, Nikko, Conrad Johnson, Carver, CAT 
 LS> amplifiers all made from tubes.  They have no low end 
 LS> kick, but are loud because of their signifigantly 
 LS> lower damping factor, sort of like a receiver 
As we've discussed, and as you seem to have noted from practical experience, 
huge audio amps aren't desirable driving speakers.  There is something 
distinctive about a single amp in the megawatt size range though, and that 
can only be acheived through tube technology.  
Tubes have plate impedances so much higher than most speaker Z's that 
matching is a problem.  That tends to lead to designs using iron, which is 
expensive if of decent quality.  Use of transformers leads to matched 
impedance systems, rather than current source designs.  The latter offer much 
better damping factors, and the ability to force more accurate cone motion.  
A good tube amp works fine for sine waves, though the cost of iron leads to 
many tube amps being less than excellent quality.  Several percent distortion 
levels are typical of what's actually manufacturer in large amps used in 
shortwave transmitters.  
In high end audio systems, I tend to think of most tube applications as BS.  
Maintenance is a nuisance, and current source drivers work better with 
speakers.  If someone wants coloration of saturating iron, that can be 
acheived at a line level in lower maintenance electronics.  
 LS> bass).  For the professional setups i do with my 
 LS> sound system, i would rather run 1 amp per speaker 
 LS> cabinet like i always do because i notice reactive 
 LS> EMF problems are affecting the clarity and the 
 LS> transience of the speakers in large setups.  
That's why I was questioning your apparent preference for several KW per unit 
amps.  
 LS> Front loaded woofer systems are so old fashioned and 
 LS> inefficient (woofers mounted on the front panel of 
 LS> the speaker box, the old fashioned way, no pro setup 
 LS> uses these anymore to project sound over a crowd but 
 LS> are better suited for monitors).
"Front loaded" with respect to acoustics has to do with horn loading versus 
box designs which only provide acoustic loading to the rear of the driver.  
Typically this results in some upper low and lower mid controlled 
directionality, as it creates some horn effect.  The common alternative is a 
Theile aligned small ported box, or some variant, which is easier to 
transport but may be beamier and accept some lower mid distortion products in 
exchange for a size/efficiency compromise.  
Carpentry, plastic molding or metal casting, or fastening hardware are really 
a separate subject from basic acoustic design.  One design concept I like, 
rarely used, is the transmission line cabinet as used to be the basis of 
Irving Fried's systems.  
Terry
--- Maximus 2.01wb
---------------