Is a small SET amp powerful enough (for me)?

[artto]

Most of you who have been around here a while are familiar with my listening room so Ill use it as an example. Its a generous sized space for a dedicated use, around 500 square feet, approximately 4300 cubic feet. The distance from a flanking Khorn (front grill) to the sweet spot listening position is about 15-6.

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

As we all know, in free space, sound pressure level falls off inversely proportional to the square of the distance, which is the same as saying the sound level is halved with each doubling of distance from the sound source. In reverberant spaces this is not quite true as the sound typically doesnt decay quite as fast, especially in smaller spaces that we listen in, and there are other considerations such as the speaker design. But for practical and conservative considerations, lets just give ourselves the benefit of doubt and use the inverse square law. Since my room has substantial acoustical treatment the speakers perform more like they are in free space anyway. Just for the sake of argument, I took some quick measurements using white noise and found that indeed there was approximately a 7.5 dB drop in SPL from 3 in front of one of the Khorns to the listening position which closely corresponds to what would be expected in free space and predicted by the inverse square law.

Now, if a single Khorn, is putting out 104dB at one meter (3 feet), then it stands to reason that, in my room, at a listening distance of 15-6 from that speaker, the SPL will be about 96.5 dB. This was confirmed by measurement.

But wait, we are using stereo. There are TWO speakers, and TWO amplifier (channels). In my case, THREE actually. Adding the second speaker and amplifier adds 3dB back into the equation. 96.5dB + 3dB=99.5dB. The center speaker, most of the time is probably not adding more than 1.5db to the overall equation. Lets be conservative and round that to 1dB. So now 99.5dB+1dB=100.5db, AT THE LISTENING POSITION, with ONE WATT of power per speaker (actually less on the center channel).

Therefore, with two watts per speaker the SPL goes up to 103.5dB, four watts 106.5dB, eight watts, 109.5dB.

No, this is not going to give you concert sound levels that only the likes of John Entwistle or Pete Townsend could appreciate, but its certainly up there with some of the loudest live levels youll ever experience as a listener in anything acoustic, and certainly even much, if not most amplified situations. Pretty much any 300B, 211 or 845 SET amp could handle this. 2A3 might come up a little short (in MY room) but realistically would probably do the job 95-99% of the time. In MY situation, 50 watts on each speaker could easily approach 126dB, at the listening position.

[NOSValves]

Artto,

Really only one way to find out for sure! Try it yourself and see. Heck someone has to have SET amps in the Chicago area that could come by and have some fun with you. I would be very curious to see what you come up with trying the 2A3 amplifier. Of course a breif few hours would be aweful short to make any real solid impressions. If you just want to buy something to try I suggest you check the used market on Audiogon.

Craig

[Parrot]

A few things I see. 

First, the 104dB/M/W spec is a lab measurement. It's not like you're going to get that with any and every recording you're playing in non-lab conditions.

Second, why would anyone be satisfied with having adequate power 95-99% (your estimate) when one can easily have it 100% of the time.

Third, there's a lot more to high fidelity than a dB rating. As I've mentioned before, I can get my 2" built-in old TV speaker to 104dB. Not a pleasant sound.

[sfogg]

How loud do you listen at?

Play a CD at what you consider the maximum volume you listen at. Now without touching your volume control play a test CD that has say a 1kHz test tone at -20dB. Many test CDs will have this.

With that playing measure the SPL at your listening position. Measure with both speakers active and with just one. Add 20dB to both measurements and that is the max. peak level your system would hit at that volume level. Based on that that will give you a better idea on where you stand.

As a point of reference in my old setup with the SE-OTL (rated at 1.5w but more like 1w) I was hitting around 93dB at my listening position (about 12') when they were run full range before the amp started to get nasty sounding. With the LaScala's high passed to my subs (how I normally run them) it could do about 97-98dB before getting nasty. Those were the modded LaScalas with the ES600s (6 ohm load so the SE-OTL put out less power) in them that had 2 or 3dB of insertion loss. With a simpler crossover I'd probably have been around 100dB-102dB before getting nasty.

Shawn

[artto]

A few things I see.

First, the 104dB/M/W spec is a lab measurement. It's not like you're going to get that with any and every recording you're playing in non-lab conditions.

Second, why would anyone be satisfied with having adequate power 95-99% (your estimate) when one can easily have it 100% of the time.

Third, there's a lot more to high fidelity than a dB rating. As I've mentioned before, I can get my 2" built-in old TV speaker to 104dB. Not a pleasant sound.

#1 I used random white noise (Prosonus Studio Reference Disc track # 44) for setting my test levels which in a very general way most resembles music for all practical "testing" purposes. The 104dB/M/W may be a spec lab measurement, but I was also very conservative, rounding both calculations and measurements to the low side.

#2 Thats why my favorite is still the 50 watt push-pull Luxman triodes. At full output, still operating in spec, they can take the speakers, in my room, to well over 120dB.

#3 Duh. Oh yeah, that's right, you haven't been to my place yet [*-)]

[Parrot]

Someday I would like the pleasure of checking out your room and system. I've wanted to for years.

[artto]

How loud do you listen at?

Play a CD at what you consider the maximum volume you listen at. Now without touching your volume control play a test CD that has say a 1kHz test tone at -20dB. Many test CDs will have this.

With that playing measure the SPL at your listening position. Measure with both speakers active and with just one. Add 20dB to both measurements and that is the max. peak level your system would hit at that volume level. Based on that that will give you a better idea on where you stand.

As a point of reference in my old setup with the SE-OTL (rated at 1.5w but more like 1w) I was hitting around 93dB at my listening position (about 12') when they were run full range before the amp started to get nasty sounding. With the LaScala's high passed to my subs (how I normally run them) it could do about 97-98dB before getting nasty. Those were the modded LaScalas with the ES600s (6 ohm load so the SE-OTL put out less power) in them that had 2 or 3dB of insertion loss. With a simpler crossover I'd probably have been around 100dB-102dB before getting nasty.

Shawn

I listen to an extremely wide range of music, so the sound levels change appropriately. With somnething like TOOL or NoFX its crunch time. With the Goldberg Variations or Yo Yo Ma Solo its obviously much more quiet. Actually, over the years I've found I enjoy the sound quality more when the system is turned down. Room acoustics adds its own form(s) of distortion, which, just like speakers and amplifiers, increases as the volume goes up.

I wouldn't advise using a 1KHz test tone for setting levels the way you've described. I do that ONLY when setting the levels of each speaker relative to each other, to exact levels and getting a reference on the level of the center speaker. And even then, its only with the SPL meter right up against the center of the midrange horn (grill) so there is no interference from the room boundaries or objects in the room. With a 1KHz sine wave, in most rooms, you can walk around with the SPL meter and see a wide variation in SPL due to reflections, room objects and mode effects. That's why I used random white noise. Sine waves are very diffcult to use for test purposes in small closed spaces.

[oldtimer]

No, this is not going to give you concert sound levels that only the likes of John Entwistle or Pete Townsend could appreciate,

==============

Remind me again--didn't Townsend go deaf?

He's not deaf, he just always played like he was.

[Parrot]

His hearing was damaged when Keith Moon dynamited his drum kit in close proximity to Pete. 

[sfogg]

I meant the loudest you listen at. Don't care what music it is, just whatever you listen to the loudest. That will tell you what the max SPL you need to hit is. Without knowing that you don't really know where you need to be to satisfy all your listening.

"With a 1KHz sine wave, in most rooms, you can walk around with the SPL meter and see a wide variation in SPL due to reflections, room objects and mode effects."

You will see much less variation using a 1kHz test tone then you would using a signal that includes bass in it such as white noise. Room modes in the bass will alter the levels by very large amounts. +-10dB or more easily. The point isn't measuring around your room. It is measuring at your listening position. That tells you how loud it is where you listen from.

But if you want to using white or pink noise (at -20dB) and do the exact same test I said that is OK too. The potential problem there is if you listening position is in a peak or null it is going to alter your measurement.

Just playing pink or white noise as loud as you can stand doesn't account for peaks in music and will give you a SPL 'target' that is lower then it should be.

The method I'm suggesting will account for the max possible peaks in music.

BTW, pink noise is more representative of music then white noise. White noise has too much energy up top compared to most music.

Shawn

[edwinr]

This where lots of people go wrong, artto. Absolute spl's are a given with the Klipschorn. A few watts WILL give you reasonable spl's. But the real time demands placed on an amplifier by even a speaker as sensitive as the Klipschorn, will tax a very low powered amplifier. I don't think we have equipment that can accurately measure the speed and dynamics inherent in live music. If you ask a 3 watt amplifier to play loudly (95db at your sweet spot?) something has to give. You either have decent spl's, or you have have accuracy. I think the secret to appreciating the quality of low powered SET amplifiers, is to reduce your listening levels even further than techno babble might suggest.

But anyway. Why not have both - decent spl's and accuracy. Or build a very small room... []

[sfogg]

"I don't think we have equipment that can accurately measure the speed and dynamics inherent in live music."

We can though. That is exactly what my measurement will tell him.

Important thing to know... on a CD (or any digital format) there is an absolute maximum signal level. It is 0dB full scale (0dBfs). Nothing can play louder then that. Peaks in the music will get very near 0dBfs or if badly mastered hit 0dBfs and clip. The key point is the peaks can't be any louder then 0dBfs, it is a hard limit.

So, one listens to their system using their preferred CD and turns the volume control up to where ever they listen to their system at maximum volume would be.

Eject the CD.

Don't touch your volume control.

Put in a test CD with 1kHz test tone recorded at 20dB below 0dBfs (-20dB). This test tone is steady state, no crest factor to worry about as with white or pink noise, and this test tone is going to play back at 20dB below the maximum possible SPL level at that volume position on your pre-amp. Steady state tones are easy to measure and since it is at 1kHz room modes won't effect the reading and neither will A or C weighting that most meters will have.

Take whatever that measurement is and simply add 20dB to it to account for the signal playing 20dB below full scale. That number is the maximum peak SPL level.

Shawn

[Duke Spinner]

well... it may get loud enuff ...

but with greatly compromised frequency response ...

[Erik Mandaville]

Shawn:

"With a simpler crossover I'd probably have been around 100dB-102dB before getting nasty."

I have thought about that in the past as it may have applied to your situation.

A pair of SE OTLs was incredible, both in terms of quality and loudness. A single SE OTL was capable of clean output at SPL levels far higher than what we are used to. For movies, the 300B Moondogs were pretty amazing, too.

Erik

[artto]

Shawn, I understand what you're saying now. I must admit that wasn't really my objective at the moment but I can see its usefulness. I was just trying to explain in general terms the relative output and how much power is being used and did some quick measurements to confirm the basic math. I guess the point I was trying to make, in simple terms, is that we often tend to forget that most of us are using more than one amplifier and speaker, in my case three, and that is usually left out of the "total" equation, and that this can easily add 3 to 4.5 dB of total output to the equation. So, instead of using 3 watts (ie: a single 2A3 SET) as the reference for instance, one should also consider looking at it from the perspective of 3x3 watts or 9 watts total. In most cases that is probably enough to do the job, in others it is clearly not. For me it appears to be borderline depending on what I'm listening to.

[edwinr]

You know, arrto. If this SET stuff wasn't so bloody expensive, I'd buy a kit for myself and build it. Just for fun. Trouble is, they charge an arm and a leg - even more for built up stuff. I love mucking around with tubes. I'll be better off with a simple pp kit.