Watts! How many do we need?

The success of the Pass 25 and other small-bore tube amps Indicates that's plenty for most applications. But remember this - speakers are current driven devices and benefit greatly from amps acknowledging that fact.

Most modern speakers are less than 8 ohms and so are quite current hungry. Also, loudness is logarithmic, so twice the power is only 3dB louder - a just noticeable difference. To sound twice as loud requires 10X the power. Also remember the limits of your speakers. KEF LS50s 84dB @ 1W/1M and 100W limits means they are limited to 104dB maximum output. A big JBL 4367 is at least 10dB more efficient and will handle 200W peaks for a max output of 117dB.  The KEFs are meant for near field listening only a few feet away, the JBLs meant for a large room with the listener 10-15 feet away, so it all must be worked as a system, not just components in isolation. As for me, my 87dB Monitor Audio S300 7G and a solid 60W amp in my 12X16 room is plenty. In other systems, 200W was just adequate, and in another a 40W tube Mcintosh MC240 was fabulous. Lots of valid answers, just don't ignore physics.

My Speakers are Harbeth C7ES-XD at 86db sensitivity. They are rated at up to 150 watts. Their impedance is 6 ohms with the lowest at 5.8 ohms. So an easy drive. But I've been told they like power. The Accuphase P4600 is rated at 150 watts per channel at 8 ohms and 300 watts at 4 ohms and the amp is probably capable of delivering 225 watts per channel at 6 ohms. I have them positioned in a well furnished carpeted room about 7 feet apart and 30" out from the front wall very slightly toed in. I sit about 8-9 ft from them. I have the meters set to freeze at the highest output and they have never exceeded 1 watt so far. Now I don't listen at some of the levels some of you do. My sound pressure meter never exceeds 75 db and is mostly between 60 and 70 db. If the meters are accurate I'm in Class A all the time and the amp is not breaking a sweat.

atmasphere wrote:

If you need over 100 Watts to make your speaker really sing, you have a problem-the speaker might be criminally inefficient unless you are in a very large room.

I would agree, but practically speaking it’s hardly as much low efficiency as it is a difficult load caused by the passive crossover. A lot of amp power can be wasted here, sometimes forcing even several hundred watt and PSU-sturdy amps to their knees. Making matters worse though we mostly see the combination of low efficiency and difficult load, whereas conversely the combo of high efficiency and easy load - not least via active configuration and higher impedance - will make a given amp sound substantially better for a given SPL.

Low eff. in addition to difficult load is a sonic bottleneck that to some can’t be ignored, while to others it’s the only thing they know. To my ears it’s not unlike listening to speakers covered by a blanket - the music just never really frees.

The more power you need, quite often the harder it is for the amplifier to sound like real music. Most higher powered amps I’ve seen simply don’t, although they are pretty good at sounding like electronics.

That’s a popular notion, and I assume not without merit, but as you implicitly indicate there are exceptions. Both due to the specific amplifier design and because my actively configured high efficiency speakers - i.e.: high eff. in the entire frequency range, incl. the subs - present such an easy load to the 3 similar amps, each of them frequency limited to their respective driver segments and independent of the others’ load, the amps are given ideal working conditions and seeing their potential more or less maxed out.

To explain: a 625W amp (8 ohms) given only a ~620Hz on up signal driving a 111dB horn/compression driver combo coupled directly to its terminals with a close to pure ohm load will be cruising along with very low distortion - even at deafening levels. If it’s already a good design, and it is, it will see its performance envelope fulfilled in a way no passive, low efficiency speaker iteration with a single amp covering the entire frequency range can equal.

As others have said, meters, even digital ones, are s-l-o-w. Even on digital recorders, the meters are too slow to see "transient peaks", which is one reason manufactures recommend recording with the "peaks you do see" topping out at -12dB. This allows enough headroom for those transients such that no clipping or compression takes place.

So yes, having more watts can be a good thing - as long as the slew rate of the amp is capable of keeping up with the music and offering those louder transients that might last less than 1/100th of a second, such that the music sounds as "real" as possible.

So, when you are looking at meters, you are more or less looking at what the continuous output power is, averaged over a significant fraction of a second, therefore yes, you will see that most of the time the amp is putting out 1W to 10W at even "loud" volumes. 

Most of the time, I'm using 1-2w listing to the radio. If I want to rock out, or make the house vibrate I use around 50-60w this is out of a 200w amp. At 100db (system can do well over 100db without clipping) at 60w, there will be 100+w peaks.

Since it takes double the power to produce 3db gain, it's good to have all the power, and use it when you want it. Generally a 200w amp will sound better at 50w, then a 50w amp at 50w. Plus staying that far under the amp's peak power produces much cleaner power to the speakers. 

Most of the time, we will all be happy with a 20w amp, but I would never buy anything under 100w