Thanks Ralph. I agree. My amp has 14db of NF and "low'ish" output impedance: JA measured low-end output impedance of 1 ohm off the 8 ohm tap and .55 ohms off the 4 ohm taps. Converted into output voltage db variation as a function of FR, JA measured +/- .8db off the 8 ohm tap and about half that amount off the 4 ohm tap. See https://www.stereophile.com/content/audio-research-reference-150-power-amplifier-measurements One poster said that the sonic flavor added by my amp's output impedance is akin to a tone control. Crude analogy, but I see where he is coming from. However, how many speakers, even if driven by voltage source SS amps, do not have ruler flat FR plots. Add room characteristics, more coloration. So, IMO, we go back to the old sage advice. Go with what sounds good to your ears. One more thought. I get your comments above that any amp that uses NF inserts a measure of high order harmonic distortion into the signal path, which sounds harsh and bright to the human ears. Not sure if the 14 db used by my amp generates a lot of high order harmonic distortion, but to my old untrained ears, I still like the sound. |
An amplifier having "constant power" aka "constant wattage"
characteristics will, for a given input voltage, tend to deliver less
output voltage at frequencies for which speaker impedance is low, and
more output voltage at frequencies for which speaker impedance is high.
That will result in loosely approximating delivery of constant power
into those varying impedances. Most tube amps fall into that category,
to a loose approximation. How loose that approximation is will depend on
both the output impedance of the particular amplifier (which varies
widely among different tube amps), and on how the impedance of the
particular speaker varies over the frequency range. @almarg Just as a bit of clarification on this bit: any tube amp can be made to act as a voltage source if you can add enough voltage feedback. If the amp is zero feedback it will behave more like a power source. Further muddying the waters is the simple fact that just because an amp doesn't double power as impedance is halved does not mean that it can't behave with a constant voltage characteristic. Many tube amps do exactly that. In the 1950s before the idea of voltage driven speakers caught on, it was seen as a good thing to make sure that the amp had low distortion as well as the correct damping factor for a given speaker. Now some open baffle speakers need a damping factor of only 1:10 and you did read that correctly: where the amp has 10x the output impedance as opposed to the loudspeaker. OTOH some speakers do need 20:1 to sound right. To allow this to occur, in the 1950s a number of amps were made with a 'Damping Control' which balanced voltage feedback against current feedback. So at one extreme the amp behaved with constant voltage; at the other with constant current, and in the middle where the voltage and current feedback were in equilibrium, constant power. BTW its worthy of note that this idea has nothing to do with solid state vs tubes. IME/IMO, while the Voltage Paradigm has taken over decades ago, the simple fact is that there are speakers out there that don't and **can't** work under the voltage rules- specifically ESLs, a number of magnetic planars, a good number of horns, 'full range' drivers and certain box speakers that designers noticed that they sound better with tubes than solid state... I think the damping control isn't a bad idea at all; if you want plug and play, the damping control is more likely to do the job for you. |
@ieales
"I note that the OP accomplished the selection in a day or two. In decades of tube amp ownership, I've never found it to be so cut and dried, often taking many weeks / months to come to a determination"
Funny you should say that. Sometimes you have to digress a little. Im no different.
Here is my observation so far.
Moving the Olympica iii's to the 8 ohm taps definitely makes the system sound more lively, but at a small cost. I use subs in my system and initially I listened with the subs on. With the subs off its apparent that on the 8 ohm connections without the subs playing as someone earlier stated the bass response is not quit as tight. Not that big a deal to me because obviously the subs make up for that.
I have also noticed a slight decay in the sound stage however. This could just be an adjustment with speaker placement, however that is no small task. It has taken me months of dialing these speakers in to where I think they sound perfect so for this fact alone that could be a deal breaker for me. Now with that said the system does seem to have a little more airiness which I like!
Along those lines definitely a little brighter sound, but not to the point that its annoying so far. No sibilance that I can detect on recordings that I think would reveal that. My room is rather flat so the brighter sound really is not enough to make me switch back if I were just measuring that only. With that said however I have to admit the speaker is not quite as natural sounding as it was for sure. One of the the things SF is good at is they're natural sound quality in this line of speaker anyway, and I have to admit that on the 8 ohm taps they sound a little top heavy, but not really in a bad way if that make sense.
For me at this point its about trade offs. While the systems sounds a bit more lively, and airy which I like, but loosing some of the lower octave quality, and sound stage quality. Could I get used to it and keep things the way they are probably , but being honest sound staging, and a natural sounding speaker is of high priority in my case. I think I would have to go back through speaker placement tweaking to make a good determination, and for me I don't think I want get in a cycle of speaker placement doing compare and contrast between the two taps. I did not intend on this being a laborious exercise.
I will report back and let you know what I decide in a few weeks. Excuse my fickleness for lack of a better word, but I guess that's how this works in this hobby.
-Keith
|
I don't really like to use the phrase "constant voltage" here. I think that the term "voltage source." is more clear. In the industry the two are pretty well interchangeable. Compare how the three WELL REGARDED amps react to the impedance curve of
the simulated load: JC-1 barely, MC501 some and Ref150 plenty. Amps do
not have a constant output impedance across the power band. Coupled to a
non-constant loudspeaker impedance, the combination of the two are a tone control.
Looking
at the phase diagram for the SF Olyiii, we see it leading the 500 to 2k
region and abruptly lags @ ≈3k. Given a selection of transformer amps,
some may sound best on 4Ω, others on 8Ω [2Ω/16Ω anyone?] and some not at
all depending how their impedance reacts with that of the loudspeaker IN THE SYSTEM.
Combined with program phase and level anomalies one has the makings for a Sisyphean task of finding an ideal solution. Ironically it was MacIntosh and ElectroVoice that led the charge back in the late 1950s to the idea of driving speakers with an amplifier that acts with a constant voltage behavior, i.e. Voltage source. This was accomplished by simply having enough feedback that the output impedance was low enough that constant voltage was possible **with most speakers available at the time**. As time went by, lower output impedances became possible and as a bit of a self-fulfilling prophecy, necessary. The problem with this sort of approach has to do with how humans perceive sound pressure, which is done by sensing higher ordered harmonics. While feedback is well-known to suppress distortion, its also well-known to add higher ordered harmonics of its own through a process of bifurcation. This is why we are seeing SETs and other amplifiers that have no feedback and relatively high output impedances, since: 1) you can make a speaker designed for a high impedance output in the amp and get very nice linear response 2) in a nutshell as pointed out above, making an amplifier to act as a voltage source isn't foolproof- its a "Sisyphean task". So after 60 years we still don't have the plug and play flat frequency response that was the goal so long ago. A pragmatic observer might easily come to the conclusion that we never will. |
I have already made that determination on which way sounds better. I note that the OP accomplished the selection in a day or two. In decades of tube amp ownership, I've never found it to be so cut and dried, often taking many weeks / months to come to a determination. I find myself switching back and forth all the time Whist I have marked levels on program for decades, I settle on a transformer tap for a particular speaker for a selection of well recorded material and be done with it. One might argue that if one continually switches, neither is ideal. Examine Fig 1 response curve here in https://www.stereophile.com/content/mcintosh-mc501-monoblock-power-amplifier-measurements-0 for an AutoFormer MC501 and Fig 1 here https://www.stereophile.com/content/audio-research-reference-150-power-amplifier-measurements for a Ref150 and this https://www.stereophile.com/content/parasound-halo-jc-1-monoblock-power-amplifier-measurements for a SS JC-1 Now examine the schematic and impedance/phase diagrams for Stereophile's simulated load here https://www.stereophile.com/reference/60/index.html. [Oh, how I wish Stereophile would run 'transformer' amps into the simulated load from all taps.] Compare how the three WELL REGARDED amps react to the impedance curve of the simulated load: JC-1 barely, MC501 some and Ref150 plenty. Amps do not have a constant output impedance across the power band. Coupled to a non-constant loudspeaker impedance, the combination of the two are a tone control. Looking at the phase diagram for the SF Olyiii, we see it leading the 500 to 2k region and abruptly lags @ ≈3k. Given a selection of transformer amps, some may sound best on 4Ω, others on 8Ω [2Ω/16Ω anyone?] and some not at all depending how their impedance reacts with that of the loudspeaker IN THE SYSTEM. Combined with program phase and level anomalies one has the makings for a Sisyphean task of finding an ideal solution. |
@stringreen
yeah I am past that part. I just find the topic interesting. I have already made that determination on which way sounds better.
Thanks |
Use whichever way it sounds better to you
|
Got a notification of a new q&a paul at PS audio put out (I guess today) Low and behold.he touched on this very topic. What a coincidence. You can see it here if you care to. https://youtu.be/4Q7RjCxHVsE |
Obviously having at least a decent understanding of this helps when trying to match speakers to amplifiers. These are typically not topics that are discussed when making purchases from a dealer. At least I have never had this conversation, but maybe others have. |
|
I don't really like to use the phrase "constant voltage" here. I think that the term "voltage source." is more clear.
Meaning, the voltage is a function of the input voltage, not the speaker impedance.
For what we call a power amplifier, this is usually around 20x. That is, 1 V in = 20 V out. 0.1V in = 2 Volts out, so long as the speaker’s impedance is at or above the rated speaker impedance of the amplifier.
|
Thanks Al and Erik. Good stuff. Thanks. For the info. Very interesting conversation. |
Hi Keith, First, it’s probably worth defining these terms for the benefit of others who may read this thread. An amplifier which provides "constant voltage," aka one that acts as a "voltage source," is one that for a given input voltage provides an output voltage that does not vary significantly as a function of speaker impedance, assuming the amp is operated within its capabilities. Hence for a given input voltage it will deliver significantly more power at frequencies for which speaker impedance is low than at frequencies for which speaker impedance is high. (Power delivered into a resistive load equals voltage squared divided by resistance). Most solid state amplifiers fall into this category. An amplifier having "constant power" aka "constant wattage" characteristics will, for a given input voltage, tend to deliver less output voltage at frequencies for which speaker impedance is low, and more output voltage at frequencies for which speaker impedance is high. That will result in loosely approximating delivery of constant power into those varying impedances. Most tube amps fall into that category, to a loose approximation. How loose that approximation is will depend on both the output impedance of the particular amplifier (which varies widely among different tube amps), and on how the impedance of the particular speaker varies over the frequency range. Which characteristic is better depends mainly on the particular speaker. The majority of speakers these days are designed with the expectation that they will be driven with voltage source amplifiers. But of course many are designed to work well with tube amps of various kinds. And some are suitable for use with either type. (The Daedalus speakers I use are an example of a speaker that is equally happy with either type, and that versatility is made possible by the fact that Daedalus speakers have a very flat impedance curve, i.e., their impedance does not vary very much over the frequency range). If you haven’t seen it, Ralph (Atmasphere) has a good paper on this subject at his website: http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php Best regards, -- Al |
You don't exactly get constant wattage, or constant voltage.
What happens is that with high impedance amps and low impedance speakers is you can loose output power (i.e. dynamic range).
By having a 4 Ohm tap, your 75W tubes still put out 75W. Otherwise, they might only put out 60 @ 4Ohms using the 8 Ohm tap.
This example is exceedingly soft. :) I suggest you find some Stereophile reviews to examine this more specifically.
|
@almarg
Al,
just curious. What would be the benefit of using an amp that delivers constant voltage vs constant wattage? Is one type typically better than the other in regards to SQ? Or does it really all boil down to component matching?
-Keith |
Hi Bruce (Bifwynne), McIntosh uses autoformers providing multiple output taps on many of their solid state amplifiers, and as far as I am aware they are unique in that respect. The rationale is presumably to make life easier on the output stage when low impedances are driven, by reducing the amount of current the output stage and the power supply must provide, and reducing the heat they will generate. One benefit of that presumably being a reduction in distortion. Their literature also talks about the autoformers providing benefits relating to speaker protection, including protection against DC being applied to the speakers in the event of a malfunction in the amp. My understanding is that an autoformer (as opposed to a transformer) can’t block DC, but perhaps the near zero impedance it would present to the output stage at zero Hz (i.e., at the "frequency" of DC) would trigger some other protection mechanism in the amp if that were to occur, and cause the amp to shut down. A consequence of the autoformer approach is that those amps will provide the same rated maximum power capability when a 4 ohm load is connected to the 4 ohm tap, and when a 2 ohm load is connected to the 2 ohm tap, as when an 8 ohm load is connected to the 8 ohm tap. As is typical of many tube amps, and in contrast to just about all other solid state amps which of course will increase their power capability significantly as load impedances decrease, until some limit is reached. And while this amp (each monoblock) weighs 93 pounds, I’d imagine that a solid state amp not using an autoformer, and that is capable of providing 600 watts into 8 ohms and far more into lower impedances, and that is comparably designed in other ways including robustness, would weigh a lot more due to the higher currents and additional heat sinking that would be required. Just curious about the use of autoformers in this application and why the OP perceives that there is a sonic difference between the 4 ohm and 8 ohm taps. The best answer I can provide is probably what I said in the last paragraph of my first post above, beginning with "So I would expect..." I think Ralph may have posted several years ago that a DF of over 8 or maybe 10 isn’t all that important. More specifically, what Ralph has said in some past threads is that "there is no speaker made that requires more than 20:1 for a damping factor." In the OP’s case, I am surprised that his DF isn’t higher since his amp is solid state. Just a guess, but I wouldn’t be surprised if the resistance of the autoformer winding is the main reason the DF is less than in most other solid state amps. Best, -- Al |
@almarg
Perfect. This is a subject that I just don’t have a lot of experience with and I tend to stay away from tweaks that involve input/output current. And I try to stay within manufactured recommendations so I’m just trying to do my diligence to protect my investment. However the difference in SQ so far is compelling enough for me to validate the change is safe.
Thanks again everyone. Very interesting conversation. |
Hi Al,
I've been checking in from time to time. Actually, I've been waiting for you to post something because I always learn something from your responses.
The OP's post caught my interest because he tapped [pun intended] into a topic that you, Ralph and I posted about several years ago. Perhaps you could expound on why the OP's solid state amp would be using autoformers at all. In the usual case, solid state amps have very low output impedances. Hence, their damping factor should be high.
I think Ralph may have posted several years ago that a DF of over 8 or maybe 10 isn't all that important. In the OP's case, I am surprised that his DF isn't higher since his amp is solid state. Is there some design feature relating to the output transistors that makes the use of autoformers necessary?
Further, in the OP's case, you posted that his amp is rated at 600 watts and his speakers have a reasonable 88 db sensitivity. So what is the issue? Just curious about the use of autoformers in this application and why the OP perceives that there is a sonic difference between the 4 ohm and 8 ohm taps. I am a little surprised.
BIF
|
No, you won’t hurt anything by using the 8 ohm taps.
In some other situations, involving amps that are much less powerful and/or speakers that are much less sensitive, the reduction in the maximum power capability of the amp that would presumably result from using a tap that does not correspond to the impedance of the speaker at most frequencies (especially in the bass and mid-bass regions, where lots of energy is typically required) might be conducive to driving the amp into clipping. Which in turn can be harmful to tweeters.
However, given the "conservatively rated" 600 watt capability of your amp (when load impedance corresponds to the designation of the tap that is used), and given the 88.3 db/2.83 volt/1 meter measured sensitivity of your speakers, and given the "Power Guard" function provided by your amp (which reduces its gain if and when necessary to keep the difference between output and input from exceeding 0.3%), that risk would appear to have zero chance of arising in your particular situation.
Regards,
-- Al
|
Ok thanks for all the great replies. However now that we have cleared up that I’m using SS amps. Has the consensus changed any? Will I cause any damage to either the speaker or the amp by using them on the 8 ohm taps? I think this is my main question/concern.
Thanks |
:-)
Best,
-- Al
(For others who may not be aware, Duke's statement just above was made famous by "Roseanne Rosannadanna")
|
"I’m afraid I must offer another correction, this being a matter of considerable significance... It was actually a different character, "Emily Litella," who concluded her commentary on Saturday Night Live with the words "never mind." "
Well, it just goes to show you. It’s always something.
|
Hi Duke,
I'm afraid I must offer another correction, this being a matter of considerable significance :-)
It was actually a different character, "Emily Litella," who concluded her commentary on Saturday Night Live with the words "never mind." As I'm sure you realize, Emily Litella and Roseanne Rosannadanna were both characters portrayed by the amazingly talented Gilda Radner, who very sadly is no longer with us.
Best,
-- Al
|
@almarg wrote: " I believe some of them have been made without the realization that the OP’s monoblock amps are solid state."
Oops.
As Roseanna Roseannadanna would say: "Well... nevermind."
Duke
|
Lots of excellent and informative responses above, except that I believe some of them have been made without the realization that the OP’s monoblock amps are solid state. And the multiple output taps are created via autoformers, the result being that the taps have higher damping factors and lower output impedances than pretty much all tube amps.
Specifically, the MC601 has a specified damping factor of "greater than 40," which if accurate would at least in theory mean an output impedance from the 8 ohm tap of <0.2 ohms, and <0.1 ohms from the 4 ohm tap, and <0.05 ohms from the 2 ohm tap. All of those numbers will result in the amp essentially behaving as a voltage source with this and most other speakers.
So I would expect in this case that interactions between amplifier output impedance and variations of speaker impedance over the frequency range are very minimal contributors to the sonic differences the OP hears in going from one tap to another. The major contributors would seem likely to involve how the sonics of the amp itself (and perhaps also the amp’s maximum power capability) are affected by differences in loading of its output stage that depend on which tap is selected.
Regards,
-- Al
P.S: Bruce (Bifwynne), great to see you posting again!
|
Erik wrote: "What huge spike [in the impedance curve]??
"That’s very typical for a speaker with a 2 way crossover. What I’m impressed with is how flat the rest of the speaker impedance curve is." The impedance curve can matter a great deal when it comes to matching amplifiers and speakers. Most amps approximate a constant voltage source (and most speakers are designed for that) but many tube amps (including McIntosh) approximate a constant power source; that is, the wattage they put out doesn’t change much as the speaker’s impedance changes. Over most of its range that speaker’s impedance looks like about 5 ohms, but then between say 1.5 kHz and 3 kHz the average is more like about 12 ohms. If the amp approximates a constant voltage source, then assuming 1 watt into 5 ohms, it will only be putting about .42 watts into 12 ohms. But if the amp approximates a constant power source, when it’s putting 1 watt into 5 ohms it is also putting approximately 1 watt into 12 ohms. The SPL difference between .42 watts into 12 ohms, and 1 watt into 12 ohms, is about 3.8 decibels. This is enough to spoil the frequency response OR - in this case - enough to fix it! You see, the measurements show the Olympia III to have a roughly 3 dB dip centered between 1.5 kHz and 3 kHz. (In the real world, the actual SPL difference between the two amplifier types would probably be less than the math in this post predicts, because in the real world amps don’t quite behave as true voltage sources nor true power sources. BUT the trend described above is valid in my experience, and explains why sometimes tube amps get blamed for bad sound when the real culprit is poor amp/speaker matching.) Duke |
barnettk, I own an ARC Ref 150SE tube amp. Same issues have come up for me over the years. I find myself switching back and forth all the time. Right now, I am using the 8 ohm taps. Both of us have power paradigm amps. As a result, the voltage output will change as a function of the speaker's impedance curves. IMO, your speaker's have pretty moderate impedance and phase angle plots. I surmise that your amp has little trouble driving the speakers. What would be interesting to know is what the output impedance of your amp is off the 4 and 8 ohm taps. That would give you a clue as to how much the amp's output voltage with vary as a function of changes in the speaker impedance. For example, the output impedance of my Ref 150SE is pretty low. As a consequence, the output voltage varies by +/- .8 db off the 8 ohm tap and about half that off the 4 ohm tap. See John Atkinson measurements here: https://www.stereophile.com/content/audio-research-reference-150-power-amplifier-measurementsPoint being that your speaker output will be flatter off the 4 ohm tap and a little brighter off the 8 ohm tap because your amp will response to speaker impedance bumps with more output voltage. The reverse is the case with respect to speaker impedance dips. So the bottom line is that what you hear will sound flatter and bass will be slightly tighter off the 4 ohm tap and a little more lively off the 8 ohm tap. Btw, I mention that bass will be tighter off the 4 ohm tap because I surmise that your amp's output impedance is lower off the 4 ohm tap. As a result, the amp's damping factor will be higher, thereby controlling the woofers a little better. But the real bottom line is that you use whatever taps sound best to YOU!! |
Thanks for the response Duke and Erik. Right I am aware that speakers change impedance at different frequencies. I was only concerned because I changed from the 4 ohm to 8 ohm outputs. Which is why I looked for the data in the first place before changing it permanently. I too was quite surprised to see how flat the impedance curve was. the amps run very cool, but I guess that’s not unusual for McIntosh Amps, but I’m sure the rather consistent impedance helps also.
Anyway. Listened most of the day today to various samples of music and so far I do like these speakers on the 8 ohm taps. I think that is where they will stay for a while at least.
Thanks again
-Keith
|
What huge spike??
That's very typical for a speaker with a 2 way crossover. What I'm impressed with is how flat the rest of the speaker impedance curve is.
|
" What I am concerned about is the huge [impedance] spike at 3khz..."
Notice that this impedance spike (which is closer to 2.5 kHz) corresponds with a frequency response dip, when driven by whatever voltage-paradigm amp the NRC uses. So an amp like your McIntosh, which is more of a power-paradigm amp (outputs constant wattage rather than constant voltage) fills in that frequency response dip somewhat. This is a really nice synergy.
Duke
|
You won't damage the speaker. The feedback in the Mac will handle that impedance spike just fine.
|
Thanks. I plan to keep it on the 8 ohm connections for about a week and further evaluate which of the two I like the best but from what I heard last night I think I like the way they sound on the 8 ohm tap. I was a little concerned about the possibility of damaging the speaker. I don’t really drive them that hard so it should be fine.
Thanks again |
Keith, yes it is likely that the speakers will sound brighter on the 8 ohm tap and it might also be more sibilant. Depends of your room. But you are not going to hurt anything running them on the 8 ohm tap. It is all about what sounds most natural to you. But, if female voices and violins hurt too much you might go back to the 4 ohm tap.
It is not unusual at all for speakers to have wide impedance swings. Going too low bothers amps the most in term of stability. Higher and the amp just has less work to do.
Mike |
