High value, high efficiency speakers for SET amps

04-22-15: Mapman
Atmasphere,
I think you are assuming that the speaker is full or near full range with your numbers. Smaller speakers with less low end extension MUST require less power to achieve a certain level, all else the same. The lower the frequencies attempted, the more power is needed, all else held constant. I think that is basic physics. So I think my argument is sound that one way to coax more out of a few good watts is to defer on or even filter out the lowest frequencies if needed one way or another at some point. That allows your watts to go further.

I would put it that deep bass extension, efficiency, and compactness of cabinet size trade off against one another in the design of a speaker. Sacrificing deep bass extension and/or making the cabinet larger (not smaller) will work in the direction of enabling the speaker to have greater efficiency.

So I agree that in choosing a speaker having significantly greater efficiency than the De Capo, that also meets Rebbi's stated price point and size preferences, much or all of the bottom two octaves or so will probably have to be sacrificed.

However, the efficiency of a given speaker is what it is. (Although that is not to say that it is what the specs say it is :-)) And provided that the efficiency number is defined in a way that is representative of the speaker's performance across the range of frequencies it is capable of reproducing with reasonably flat frequency response, a speaker having an efficiency of 102 db/1W/1m, to use Ralph's example, will produce within that frequency range an SPL of 102 db (+/- frequency response variation) at a distance of 1 meter when provided with 1 watt (neglecting room effects), regardless of what the speaker's deep bass extension is.

Regarding filtering out the deep bass at a point upstream of the amplifier, yes, that of course would make a given number of watts go further, but of course implementing that filter without compromising the SET magic is another matter, especially at a modest price point.

On the other hand, though, how far the 2 watt figure Ralph mentioned will go in conjunction with a given speaker efficiency will of course vary considerably depending on the volume preferences, listening distance, and room size of the listener, and perhaps even more so on the dynamic range of the music that is being listened to. And it seems clear in this case that Rebbi is generally quite happy with the performance of his SET amplifier with the 86.7 db/2.83 volt/1 meter De Capos. My guess, therefore, is that a speaker which honestly gets into the mid-90's/1 watt/1 meter, while also not compromising anything about the De Capo's sonics which he finds appealing, would make for an excellent upgrade. I have no specific suggestions to offer, though, in the stated price range.

A probably unnecessary reminder, also, that efficiency and intrinsic sonic quality are not all that have to be considered. Ideally impedance should be high, and should not dip down to low values at any frequency, and should not have severely capacitive phase angles (especially in the lower part of the spectrum where lots of energy is typically required, and especially at frequencies coinciding with impedance minima). And the less variation of impedance as a function of frequency the better. Speakers not meeting those criteria would not only entail increased risk of unsatisfactory results, but would tend to give results that are not even consistent among different SETs, due to differences in impedance interactions.

Just my $0.02. Best regards,
-- Al

My sincerest condolences as well, Rebbi. And thanks for the beautifully put reminder about what really matters in life.

Best,
-- Al

Charles,

A question for you that I'm genuinely curious about, which relates to this discussion. A while back you indicated that you planned to order Chesky's CD remastering of the 1962 performance of Dvorak's New World Symphony, Jascha Horenstein conducting the Royal Philharmonic Orchestra, which I had recommended. If you have done so, how well were your Franks + Coincident speakers able to handle the dynamic peaks on that recording? And also, of course, what were your impressions of the sonics of the recording?

Best regards,
-- Al

Rebbi, thanks for the nice words and for the clarification about size and positioning. Regarding:

04-22-15: Rebbi
As for lower end extension, my old Merlin TSM-mmi's were only rated down to 55 Hz, and I found them so lacking in punch and dynamics that I sold them after a couple of years. The De Capo's, rated down to 42 Hz, taught me that I needed a speaker to have at least a decent taste of low-end punch or I'll get bored and the music loses its foundation. Some of the candidates, like the Lore 2.0, are rated to extend quite lower than the De Capo - down to 30 Hz in the case of the Lore 2.0, which should be very satisfying in my room.

I have no specific knowledge of the Lore, but a general word of caution about speaker bass extension specs: My perception has been, generally speaking, that those numbers stand a considerably greater chance of being both misleading and meaningless than sensitivity or efficiency numbers. For several reasons, including differences in the +/- db tolerances different manufacturers may use; differences in the distortion percentage the number may be based on; differences in the volume level the number may be based on (for a given distortion % and a given +/- tolerance bass extension can be expected to narrow at high volume levels); differences in the extent to which the specified bass extension takes room effects into account; and in cases where speaker impedance is significantly lower at low frequencies than at higher frequencies the possibility that the number may be based on the ability of solid state amplifiers to increase power delivery into lower impedances. And as you've no doubt seen it is common for bass extension to be specified without any of those things being indicated.

Good comments by Mapman just above, btw, some of which relate to these same points.

I note regarding the De Capo, btw, that although it is spec'd down to 42 Hz as you indicated, the measurements I linked to in my previous post show 3 db of rolloff at a bit above 60 Hz, and 6 db of rolloff at a bit above 50 Hz. And that is at a modest input level of 2.83 volts, roughly corresponding to 1 watt. In EE circles, btw, bandwidths are most commonly specified on the basis of 3 db of rolloff.

Best regards,
-- Al

04-26-15: Lewinskih01
I guess my bottom-line-question is whether the vanishing distortion at lower output level that is a trait of SETs is also a trait of SEPs, like the Dynamo?

Hi Lewinski,

Ralph can most likely provide the best answer to that question, but pending his response I'll say that the vanishingly low distortion of SETs under small signal conditions is presumably due to a combination of two things: Absence of crossover distortion, resulting from the single-ended operation, and the inherent linearity of good triodes, especially under small signal conditions.

The Dynamo, being single-ended, will provide the first of those benefits, so the question becomes the degree to which the linearity and distortion characteristics of an EL34 operated under small signal conditions can approach those of a good directly heated triode such as the 300B. I don't know the answer to that question, but I think it's worth reflecting upon the fact that a lot of very highly regarded amplifiers over the years have used the EL34 in a push-pull configuration, crossover distortion included.

Best regards,
-- Al

04-30-15: Rebbi
So it makes me think that when looking for a speaker for this kind of amp, it makes some sense to look at brands that are not only high-efficiency and flat impedance, but also otherwise designed with single ended operation in mind. Again, I don't know the physics and engineering of this, but there must be other factors involved in designing a speaker specifically to be "SET friendly."
Almarg, feel free to chime in here. :-)

I don't know either, Rebbi. And in the absence of empirical inputs from users and/or the manufacturer, I doubt that it could have been predicted based on its published and measured characteristics that the De Capo would have any particular synergy with SETs (within the maximum volume capability of the combo). Although as you realize the De Capo's relatively benign impedance characteristics at least do not exclude the possibility.

Some very speculative guesses, though, as to things that might be contributing factors:

1)Although I haven't seen any frequency response specs or measurements on the Kit1, I would suspect that it has a more limited bandwidth at both ends of the spectrum than most push-pull amps. Perhaps that helps to keep frequencies out of the De Capo that don't have much if any audible significance but that it might not be able to handle comfortably (i.e., without intermodulation or other effects that could result in those frequencies affecting audbily significant frequencies).

2)Perhaps the design of the De Capo is such that it particularly benefits from the enhancement of dynamics that SETs will tend to provide, as Ralph/Atmasphere has explained is a consequence of how their distortion characteristics vary as a function of signal level.

3)As shown in the NRC measurements, the De Capo's impedance has a sharp rise to about 30 ohms in the vicinity of 70 Hz, where a lot of energy is often required. Relative to its rated power capability, a SET will be able to deliver more power into that high impedance than most push-pull amps, especially solid state amps.

As I say, though, those are all just speculative guesses as to what might be contributing factors, and I don't think that this particular synergy (within the maximum volume capability of the combo) could be either ruled out or ruled in "a priori." So I agree with the comment by Charles just above.

Best regards,
-- Al

One more possible contributor to synergy between Rebbi's speakers and SET amplification, adding to the possibilities listed in my previous post:

4)I note that John Atkinson states in his measurements of the older de Capo i version of the speaker that "the speaker's low-frequency tuning is actually a little on the overdamped side." If the same holds true of Rebbi's BE version, its sonics could presumably benefit from the low damping factor of a SET.

Best regards,
-- Al