03-29-14: Bifwynne
Perhaps Ralph or one of the other tech members can speak to how much SPL variation will result if the "wrong" type of amp is used. Perhaps, for discussion purposes, we should assume the MLs were designed to be driven by a SS "constant voltage paradigm" amp and the "wrong" type of amp is a Power Paradigm amp. Hi Bruce, I did some quick calculations for the OP's Aerius, based on John Atkinson's measured impedance curves shown here. It should be kept in mind that the impedance characteristics of the Aerius are significantly less demanding than those of many other Martin-Logan speakers. The extremes are a 25 ohm peak at 46 Hz, with approximately a 0 degree phase angle, and a 2 ohm minimum at 20 kHz, with approximately a -20 degree phase angle. To simplify the calculations I ignored the effects of the -20 degree phase angle. For a given input level to the power amplifier, a solid state amplifier acting as a voltage source will put approximately 11 db more power into the speaker at 20 kHz as it would at 46 Hz, as long as it is operated within the limits of its power capabilities. For a given input level to the power amplifier, a tube amp having an output impedance of 2 ohms will put approximately 6 db more power into the speaker at 20 kHz as it would at 46 Hz, as long as it is operated within the limits of its power capabilities. If we assume per your question that the speaker is designed to provide flat frequency response when driven by a solid state amplifier, the speaker's response when driven by that tube amp would therefore be rolled off by 5 db at 20 kHz, relative to its response at 46 Hz. That difference certainly figures to be audible, although not necessarily objectionable in many circumstances. The difference will of course be significantly greater with amplifiers having 3 or 4 ohm or higher output impedances, and with many other Martin Logan speakers. What seems likely to often be more significant, however, is simply the ability of the amp (regardless of whether it is tube or solid state) to cleanly generate enough power at deep bass frequencies, into the high impedance. Obviously that factor is highly dependent on the dynamic range of the recordings that are listened to, as well as on the listening distance, room size, and volume preferences of the particular listener. Best, -- Al |
Bruce, thanks. Youre welcome! If ZEROs are used to double the ML's apparent impedance plot, would that change your calculations? It would narrow the gap somewhat. The 11 db figure for the solid state amp wouldn't change, but the 6 db figure for the tube amp having a 2 ohm output impedance would increase to 8 db. So the difference would be reduced from 5 db to 3 db. Also, maximum power capability for the solid state amp would probably be cut in half, with distortion performance being improved as long as the amp is operated within that reduced power range. I'm not sure that it's possible to meaningfully generalize about what would happen in those respects in the case of a tube amp, with this kind of wide variation of speaker impedance as a function of frequency. I believe it would depend on the particular amp, and whether it is OTL, SET, push-pull, high or low feedback, etc. Did the ML's phase angle plots factor into your calculation? If so, directionally, how so (if phase angle is negative or positve)? I intuit that harsh phase angles can constrict the SOA of an amp (SS or tube). I am not clear of the impact on sonic coloration, assuming the amp is operating within its respective SOA. As I indicated, the phase angle was approximately zero at 46 Hz, and I ignored the -20 degree phase angle at 20 kHz in order to simplify the calculation. For purposes of a rough ballpark calculation that seems reasonable. It might not be, though, in the case of other electrostatics where phase angles may be more severely capacitive (i.e., negative). Off the top of my head Im not completely certain what the directionality of the coloration effects would be in those cases, assuming the amp is capable of handling the increased difficulty of the load, and supplying the necessary increase in current. For a given impedance magnitude, such as 2 ohms in this case, I believe that a severely negative phase angle would mean that the resistive component of the impedance (which is the component that can absorb power and convert some of it to sound) would be smaller. I believe that would tend to increase sensitivity to output impedance differences between amplifiers, and therefore exacerbate the difference between the solid state and tube amp calculations I provided. As I say, though, Im not entirely certain of that. You ask some tough questions :-) Best, -- Al |
George, where do you see an indication of a MEASURED damping factor of 4.7?
Regards,
-- Al |
George, the 4.7 number appearing on that page is a specification, not a measurement. On the same Manley page an output impedance of 1.5 ohms is specified, which is confirmed by TJN's measurements in the review you linked to. On the same Manley page, and also in the manual, the statement "optimized for 5 ohms" appears. Maximum output power, input sensitivity, and frequency response are all specified for a 5 ohm load. The 5 ohm power rating is higher than the 8 ohm power rating, for the same distortion level.
I see no reason to assume that the 4.7 number is accurate and the 5 ohm number is not. As I said, the converse seems much more likely.
Regards,
-- Al |
Hi Unsound,
The Manley Snapper has a single output tap, which is described as having been optimized for a 5 ohm load. Its output impedance is specified as 1.5 ohms, as George indicated. Its specified damping factor of 4.7 is slightly inconsistent with that (8 ohms/4.7 = 1.7 ohms; 5 ohms/4.7 = 1.06 ohms), but is in the same rough ballpark.
You might be recalling having read something in which the author confused the output impedance of a tap with the load impedance the tap is optimized for. A not uncommon error.
Best regards,
-- Al |
George, if I may be a bit pedantic as well, although multiplying the 1.5 ohm specified and measured output impedance by the specified damping factor of 4.7 ohms equals 7 ohms, both the manual and this page at the Manley website clearly state that the output is optimized for a 5 ohm load. As I indicated the numbers are not quite consistent, but I would be inclined to think it more likely that the damping factor spec is inaccurate than the 5 ohm figure. In part because I note that the damping factor number does not appear in the manual, but the 5 ohm figure does. And also because maximum power ratings are provided only for 5 ohms and 8 ohms, not 7 ohms, with the 5 ohm capability being higher than the 8 ohm capability. Regards, -- Al |
Correction to my previous post: delete the word "ohms" which appears immediately following the number "4.7"
Regards,
-- Al |
Ralph and I and others discussed his views on the relation between feedback and output impedance in this thread. It essentially comes down to a matter of terminology. Regards, -- Al |
Excellent post, Bruce, with which I agree. I also agree with the first two-thirds or so of the paper George referenced, although I would point out that the author has taken the reasonable step of simplifying the circuit analysis he presents by only taking into account the resistive component of the load impedance. I would take exception, however, to much of the last part of his paper, beginning at the point where he states that: Loudspeakers are not constant power devices. Loudspeakers must have a constant drive voltage to provide a constant acoustic output with changing frequency. IMO (and I feel safe in saying in Ralph's opinion as well), that statement is incorrect and misleading. I would agree with it if it referred to the majority of loudspeakers, rather than being expressed as applicable to all loudspeakers. Many electrostatics being notable exceptions to his statement, IMO. The fundamental error he makes, IMO, is that he assumes that with all speakers flat frequency response in (in terms of voltage) corresponds to flat frequency response out. That will be true, at least approximately, for the majority of speakers these days. However it will not be true for a substantial number of speakers, including many electrostatics. As Ralph has pointed out many times, it comes down to the intention of the designer. Consider the classic Quad ESL-57 George referred to earlier, which is revered to this day (impedance curve shown here). It was designed before solid state amplifiers existed! And it is very successfully used with tube amplification by many audiophiles to this day. Although in fairness I'll say that I recognize that **at least among solid state amplifiers** the vintage Mark Levinson ML-2 which George referred to is recognized as being a particularly synergistic match, and that combination (with two stacked Quads per channel) was the heart of the highly regarded HQD system produced by Mark Levinson's original company (together with a large Hartley woofer and a Decca supertweeter, with a total of six ML-2 monoblocks driving the two channels in a triamped configuration). Regards, -- Al |
Unsound and Bruce, as you may have noted TJN's article that Unsound referenced addresses how impedance interactions affect frequency response at the speaker's INPUT terminals. In itself that says nothing about what the speaker's acoustic output will be like in response to those inputs.
I have no knowledge as to whether optimal frequency response of the acoustic output of the Martin-Logan Aerius would result from a flat frequency response at its input, or from a frequency response at its input that is non-flat in some manner. But see the comments in my previous post regarding the Quad ESL-57, and take a look at its impedance curve that I referenced.
Best regards,
-- Al |
04-09-14: Unsound
Unless there is some sort of internal eq or the impedance actually compensates for the drivers-crossover/speaker systems inherent deviations from flat frequency response (something that has been done to some degree in crossover designs from other manufactures, but something I've yet to hear attributed to ML designs), I'd hazard a guess that it might be reasonable to extrapolate that the sound output to somewhat mimic the impedance/frequency graphs provided. As I indicated I have no specific knowledge of the behavior of Martin-Logan speakers. But keep in mind that application of a voltage which is constant as a function of frequency will result in power delivery which progressively increases as impedance decreases. (In saying that I'm oversimplifying a bit by ignoring phase angle effects). As you'll realize, what a speaker basically does is to convert some fraction of the electrical power supplied to it into acoustical power. And as Ralph pointed out earlier, the efficiency of an ESL (power out vs. power in) does not decrease in step with impedance as frequency increases, but instead remains pretty much constant as I understand it. So in the absence of specific indications to the contrary I see no reason to expect flat frequency response in (in terms of voltage) to result in flat frequency response out. As I see it the usefulness of TJN's measurements is that they provide insight into the DIRECTION in which tonality will be affected as a function of amplifier output impedance. But they provide essentially no insight into what output impedance will be optimal. Best regards, -- Al |
Great posts by Bruce (Bifwynne), with which I fully concur. 09-13-15: Georgelofi
... amps that cannot deliver current at those low loads ... cannot give a flat frequency response into those types of loads, especially ones that dip down to 1ohms. This is true, but I would emphasize the word "into." The frequency response characteristics of the signal at the input terminals of a speaker that will result in flat frequency response in the acoustic output of the speaker will depend on the design of the particular speaker. As Bruce said, it is "important to know whether the ESL was voiced to be driven by a SS or tube amp." And in that regard it is worth noting that the Quad ESL57 was designed before solid state amps existed. Although admittedly, as I believe you (George) mentioned earlier in this thread or in another similar thread, the vintage Mark Levinson ML-2 solid state amp in particular, rated at only 25 watts or thereabouts into 8 ohms but capable of supplying huge amounts of current into low impedances, is considered by many to have been a good match for the ESL57. While at the same time that speaker has been and still is used with tube amplification by many audiophiles. Regards, -- Al |
Again, though, as I indicated in my previous post George's comments mainly address frequency response at the amplifier output/speaker input. And flat frequency response at those points does not necessarily mean that the acoustic output of the speaker will have flat frequency response.
Ralph (Atmasphere) stated above that "in an ESL, the efficiency is fairly constant despite the impedance." ("Efficiency" referring to acoustic power out vs. electrical power in). I believe that is generally true. But even if we assume that the efficiency of a given ESL is just a little bit closer to being constant across the frequency range than it is to mirroring the impedance curve, then the minimal variation of amplifier output voltage as a function of load impedance that is characteristic of almost all solid state amps (assuming they are operated within the limits of their voltage, current, power, and thermal capabilities), and hence the increase in power delivery that will occur as impedance decreases, will result in greater frequency response variation in the acoustic output of that ESL than would result with a tube amplifier (operated within its capabilities) whose output impedance is some relatively high value, and whose output voltage therefore varies significantly as a function of varying load impedance.
Apologies for the long sentence :-)
Regards,
-- Al
|
Ralph, take a look at the impedance curve for the CLX, shown at the bottom of this page. Could it be that Tsushima1's negative experience with the Zero was the result of the speaker's extremely high impedance at low frequencies being multiplied 2 or 3 or 4 times, resulting in the tube amplifiers he used running essentially unloaded at low frequencies? Or, at best, running into impedances at low frequencies that were non-optimal for the output taps provided on the amplifiers? Best regards, -- Al |
