Contrary to popular belief, the Martin Logan electrostatics have powered woofers that let you hear more bass if the amplifier has better low end punch. Please refer to the stereophile review of the montis where they can validate what I am stating here. The bass improves and sounds as if you went to the back of the speaker and turned the dial up a few notches. Regarding the fact that the Martin-Logan Renaissance and Montis speakers incorporate powered woofers, yes, I would certainly expect that their bass performance would vary significantly depending on the amplifier that is used. Just as bass performance can be significantly affected by a preamp or a source component, even though those components provide no power to the speakers. However, since those speakers draw nearly negligible amounts of current and power from the amp at deep bass frequencies, and also given that the amp being used will not be exposed to any back-emf from the woofer, I would be cautious in extrapolating from these results expectations about what kind of bass performance can be expected with non-powered speakers. Also, as George mentioned the impedance of these and many other electrostatics descends to very low values in the upper treble region, about 0.5 ohms at 20 kHz in the case of these particular speakers. With that impedance being highly capacitive (and therefore more demanding of the amp) in much of the treble region. While most non-electrostatic speakers have impedances that either rise or remain relatively constant in the treble region, and that usually become increasingly inductive, rather than capacitive, at high frequencies. So while I extend my congratulations to the OP for a job well done, and for providing us with a great deal of valuable information, I would be cautious in extrapolating from results with these two speakers expectations as to how an amp would perform if it were to be used with non-powered and/or non-electrostatic speakers. Regards, -- Al |
Some thoughtful and excellent posts by Viber6, IMO.
In fairness to Techno_dude, though, while I certainly wouldn't go as far as to say that ARC "tries to make their tube gear sound like SS," I believe that a major contributor to his perception is the fact that ARC amps tend to use more feedback and consequently have lower output impedances and higher damping factors than many and perhaps most other high quality tube amplifiers.
The interaction of that relatively low output impedance (for a tube amp) with a speaker's variations of impedance as a function of frequency will tend to have tonal consequences that come relatively close to those that would result with a solid state amp, nearly all of which have output impedances that are near zero and inconsequentially small relative to speaker impedances.
Stated another way, ARC tube amps tend to act more like voltage sources than most other high quality tube amps. Meaning that for a given input voltage to the amp it will come closer to maintaining a constant voltage into varying load impedances than most other high quality tube amps (and therefore delivering more power into low impedances than into high impedances), as long as it is operated within its maximum voltage, current, power, and thermal capabilities.
Regards,
-- Al
|
A point to be aware of regarding the possibility of pairing a Spectral power amp with the Ref10 (or any other) ARC line stage or preamp: All currently produced Spectral power amps, and some of their older models, have input impedances of 10K. (A few of their older models have input impedances of 100K). All or nearly all ARC line stages and preamps have a load impedance specification of 20K **minimum.**
Regards,
-- Al
|
4425 9-18-2018
All ARC preamps are really not designed to drive subwoofers except possibly REL.
The line-level input impedance of many and I believe most REL subs is 10K, well below ARC’s usual load recommendation of 20K minimum for their line stages and preamps. The input impedance of the power amp, that would be paralleled with the input impedance of the sub for one of the two signals in the preamp’s balanced pair of output signals if the power amp is connected via XLR and the sub via RCA, would further reduce the 10K load impedance. The speaker-level connections provided by REL subs would work well with ARC power amps, but there is an important caveat. ARC fully balanced tube power amps typically have their circuit ground connected to the 4 ohm tap, not to the common tap. So in the case of a fully balanced ARC stereo tube amp the black ground wire of a REL sub’s speaker-level input should be connected to the 4 ohm tap. The red and yellow signal wires should be connected to the 8 ohm tap of the corresponding channel, or alternatively to the 16 ohm tap if 16 ohm taps are provided. In the case of monoblocks I would consult with ARC as to what to do with the black wire. On another note, Bill_k makes a good point about the reference in the Merrill literature to GHz speeds. As far as I can see none of the information provided at their website makes clear what parameter is in the GHz range (and for that matter I wasn’t able to even find a clear statement that it applies to the switching frequency), and a specification of signal bandwidth is not provided. Regards, -- Al |
Thanks, Bill. The one reference to GHz speeds I see at that link is a statement that "the Gallium Nitride Transistors can operate in the Gigahertz range." I'm sure that is true, but of course that statement provides no indication of how fast the transistors are switched in the amp.
Best regards,
-- Al
|
Viber6, yes, it appears to me that GaN transistors are used in the output stages of the new Merrill amps. And consequently they are directly in the signal path, and have nothing to do with the power supply. However, it also appears that the amps are operating in class D, or some variant thereof. (In addition to the power-to-weight ratio of the amp being strongly suggestive of that, the designer specifically states so at the link Bill provided). Which means that the transistors in the output stage of the amp are being switched, at high speed. I’m pretty sure that Bill sees it the same way.
And although the transistors themselves are capable of being switched at GHz frequencies, that says nothing about what speed they are actually being switched at. All I would infer from the writeups is that the signal bandwidth is likely to be higher than the bandwidth of all or nearly all other audio amplifiers, probably including the Spectrals.
Regards,
-- Al
|
As usual Bill_k is of course entirely correct in his comments about class A.
I would add that the reduction in amplifier temperature that occurs when a class A amp is delivering a lot of power to the speakers is unlikely to be noticeable, for two reasons.
First, a class A amp has low efficiency, meaning the ratio of max power out to AC power in. For example the Pass Xs150 monoblocks are rated to provide an output of 150 watts max while drawing an AC input of 700 watts. In that example the amp will never be dissipating (converting into heat) less than 700 - 150 = 550 watts, and will be dissipating significantly more than that most of the time.
Second, since speakers convert just a very small fraction of the power that is supplied to them into sound, with the rest of that power being converted into heat, if that amp is supplying 150 watts to a speaker, and therefore dissipating only 550 watts by itself, most of the 150 watts the amp is not dissipating will be converted into heat anyway, by the speaker.
In the case of a class A amp, therefore, it is easy to quantify how much power is converted into heat and injected into the room. To a very close approximation it is equal to the amp’s specified power consumption. For a pair of Xs150s that would be 1400 watts; for a pair of Xs300s that would be 1800 watts.
Also, btw, how hot an amp may feel if it is touched on its heatsinks or elsewhere has no direct relevance to that, and an amp that may seem cooler to the touch might actually be putting more heat into the room than one that seems to run hotter, depending on the design of the heat sinks and other aspects of the design of the amps. What matters is how much power the amp consumes. To illustrate that point, consider the difference in surface temperature between a 100 watt light bulb and an amp consuming 100 watts.
To answer Viber’s question, the output transistors or tubes in a class A design conduct a bias current (which is the current the output devices conduct when no signal is present) that is greater than the maximum amount of output current the amp is designed to be able to deliver. When a signal is present some of that current is diverted to the speaker. In a class AB design the output transistors or tubes just conduct a small bias current, and the presence of a signal causes them to conduct larger amounts of current, most of which is sent to the speaker.
Regards,
-- Al
|
Thank you for the nice words about my post, gentlemen. With respect to the last paragraph of my post I should clarify that in the case of class A tube amps having output transformers the output tubes themselves will of course not conduct more current than the amp is capable of supplying to the speaker. The output transformer will step up the current supplied by the output tubes substantially, while stepping down the voltage supplied by the tubes correspondingly, such that the amount of power sent out of the transformer (approximately equal to voltage x current) will be approximately the same as the amount of power sent into it by the tubes (neglecting minor power losses that occur in the transformer itself).
Best regards
-- Al
|
I agree with Viber’s comments about damping factor.
Usually damping factor is defined on the basis of an 8 ohm load. Assuming that is the case here, and assuming the spec is accurate, a damping factor of 5000 means that the amp has an output impedance of 8/5000 = 0.0016 ohms. While the combined resistance of the two conductors of say a 10 foot 10 gauge speaker cable is 0.02 ohms, 12.5 times as much as the output impedance of the amp.
So with that speaker cable the damping impedance presented to a speaker by an amp having a damping factor of 5000 will be 0.02 + 0.0016 = 0.0216 ohms. Not much different than the damping factor limitation imposed by the cable itself, and still vastly smaller than the impedance of the speaker (that ratio is what matters), especially in the bass region.
So IMO it is fair to say that a damping factor in the thousands is overkill by a wide margin, and if such an amp has better bass control than an amp having a much lower damping factor the reason is something else.
Also, btw, noted designers Roger Modjeski (Music Reference) and Ralph Karsten (Atma-Sphere), while admittedly being designers of tube amps having damping factors that are far lower than the damping factors of nearly all solid state amps, have both stated in threads here that no speaker in existence needs a damping factor greater than the low double digits.
Best regards,
-- Al
|
If connecting the Lampi directly into the Block Audio amps proves to be superior to having the Ref10 in the path, as the comments by WC’s wife would seem to suggest, it wouldn’t surprise me if a major contributor to that is the fact that the Block’s input impedance (10K unbalanced/20K balanced) is exactly at ARC’s 20K **minimum** load recommendation for the Ref10 (which btw is also their recommendation for most of their other preamps and line stages). And that assumes WC was using balanced connections between the Ref10 and the Blocks; the Block’s 10K unbalanced input impedance would be even worse.
An output impedance spec or measurement, or alternatively a minimum load recommendation, doesn’t seem to be available for the Lampi Pacific, but given that it uses a power tube in its output stage, as well as a high quality coupling capacitor, I would think it likely that it could handle the 10K load it would see with its unbalanced-only outputs without difficulty.
Regards,
-- Al
|
|
WC, I may have missed your mentioning it, but are the Wireworld cables you use to connect the Ref10 to the Blocks balanced or unbalanced?
If they are balanced, perhaps one factor (probably among several) contributing to what the Ref10 brings to the table when inserted between the Lampi and the Blocks is simply that the Blocks may sound better when provided with balanced inputs than when provided with unbalanced inputs. I note that the manual for the Blocks specifically recommends the balanced inputs "for best sound performance." Which of course cannot be done with the Lampi connected directly to the amps.
And if they are unbalanced, trying balanced cables between the Ref10 and the Blocks at some point would certainly seem to be in order. Not only because it may be more optimal for the Blocks, but also because it may be beneficial with respect to what I mentioned earlier about the marginal impedance compatibility between the Ref10 and the Blocks.
Regards,
-- Al
|
To add to the foregoing comments about the benefits a preamp can provide, without a separate preamp some DACs will not be able to drive some power amps to full power.
The Lampi Pacific is a case in point, when paired with some of the power amps that have been mentioned. It has a specified maximum output voltage of "6 V pp" (i.e., 6 volts peak-to-peak), which is 2.12 volts based on the usual "RMS" (root mean square) way of defining such things. The Pass XS300 that has been discussed, like most Pass amps, has a specified gain of 26 db. While the XS300 is capable of 300 watts into 8 ohms and 600 watts into 4 ohms, a gain of 26 db applied to an input of 2.12 volts means that the Pacific DAC would not be able to drive the XS300 to more than 225 watts and 450 watts, respectively, into those impedances. And that would be with the volume control at max, and with a recording whose peaks reach very close to "full scale," i.e., to the maximum possible digital value. Some recordings are engineered such that they fall several db short of doing that, which would mean that more than half of the XS300’s power capability could not be utilized with those recordings, even with the volume control at max, if it were directly driven by the Pacific.
Regards,
-- Al
|
A symphony I recommend to those who have not had much exposure to classical music is Prokofiev's short (~ 15 minute) "Classical Symphony," a work that IMO is just about impossible to dislike. Here is an excellent video of an excellent performance: https://www.youtube.com/watch?v=p5jL6Ma9tvkRegards, -- Al
|
Jafox 2-16-2019
Other than playing with the shield, if there is one, tying it to the ’-’ line at one end of the cable or the other, or perhaps a "network" on one end or the other, designating a speaker cable as "directional" sounds more like hype than anything. Perhaps a resident technical guru here like almarg could inject some value here. Thanks for the mention, John. I agree with your statement 100%. I say that despite a variety of "explanations" of wire directionality that have been proffered by some manufacturers and certain frequent participants in cable-related threads here and elsewhere, relating variously to crystal structure, how the wires are drawn in the manufacturing process, the direction in which noise may "prefer" to propagate in the wires, the breakin effect asserted in the statement by Nordost that was quoted above, etc. As I see it, when it is not practical or possible to obtain a quantitative perspective on explanations that may be offered for a claimed effect, that would provide some indication of whether or not the effect may be great enough in degree to have a reasonable chance of being audibly significant, there is virtually no limit to the explanations that can be conjured up and asserted. And consequently such explanations amount to speculation at best, perhaps tinged with conflict of interest in the case of some manufacturers. As an aside, btw, I **can** envision that a symmetrically designed cable conducting digital audio signals might sometimes exhibit directional properties. Digital audio signals have significant frequency content at tens of MHz, and higher in some cases, which conceivably could mean that minor mechanical differences in how the wires are soldered or otherwise attached to the connectors at each end might result in differences in the degree and timing of reflection effects that occur at those RF frequencies as a result of impedance mismatches between the the components and the "characteristic impedance" of the cable and connectors. Which in turn, depending on which way the cable is connected, might in some circumstances affect the waveform of the signal that is received by the destination component to a degree that contributes to timing jitter at the point of D/A conversion. Also, kudos for suggesting that any such findings be verified by re-trying the "wrong" direction, which would reduce the possibility that a perceived difference may have resulted from extraneous variables such as differences in the warmup state of the equipment, reseating of the connectors, etc. My perception over the years has been that it is all too common for audiophiles to attribute perceived effects to the wrong variable, as a result of inadequately thorough methodology. Especially when it comes to effects that are seemingly implausible. IMO. Best regards, --Al |
Viber6 2-16-2019
Yes, I have found that thinner speaker cable yields a more treble oriented balance.... Maybe almarg can help with some sensible science to back up our listening impressions, rather than the pseudoscience of many cable designers. Thanks for the mention, Viber, but that involves so many variables, at least potentially, that I would not want to speculate as to why that may be true in many cases. Those variables would include not only the characteristics of the cable, such as resistance, inductance, capacitance, length, "characteristic impedance," the thickness and dielectric absorption characteristics of the insulation, skin effect (the importance of which, btw, though not necessarily insignificant, tends to be over-hyped by many manufacturers, as can be shown by calculation), but just as importantly the interactions of the cable characteristics with those of the speakers (especially their impedance characteristics over the frequency range), and with those of the amplifier (especially its output impedance, how much feedback it uses, and perhaps its bandwidth). Also, given all of those potentially significant variables and system dependencies I would be cautious in extrapolating general conclusions about such things from individual experiences, even if those experiences are particularly diverse. Not saying it can’t or shouldn’t be done, just that I would be cautious in doing so. Best regards, -- Al |
Viber6 2-17-2019
I was just asking if you could take a simple cable like zip cord with
known specs at different gauges, and approximately predict that bass
frequencies are reduced more than high freq. You can make up some
numbers for some amps or speakers you know, so you can estimate the
effect, or just use a pure resistor as a model for a speaker (Maggies
are said to be a pure resistive load). If this is still not possible
for you to predict this, I thank you and respect your insight. No, I still wouldn't want to make any such predictions. Also, given the many variables and unknowns I cited in my previous post, as well as the fact that perceived tonal balance can be affected by more than just frequency response (e.g., by low level distortions of various kinds), I would expect that any correlation between gauge and perceived frequency response/tonal balance would be a loose one at best. Best regards, -- Al
|
