Power Amp Output Questions

[Guest105]

No, it is not. I don't know how you extrapolated your conclusions regarding Krell, and I don't know what you mean about free lunch. As should have been apparent from reading this thread, Krell's "under-rating" is most likely due to their interest in being able to present their amp as capable of "doubling down" as impedance is halved.

You are confusing the concept of "dynamic headroom" with the notion of "headroom".

Dynamic headroom refers to the ability of an amp to deliver power above its maximum continuous power rating for some very brief period of time. There are no standards for dynamic headroom ratings.

Headroom refers to the amount of non-cclipping power available beyond a given continuous output level. In my earlier example, which I will not repeat as it appears to have little effect on your understanding, I referred to an amp running at 1 watt of continuous output. If its maximum continuous output capability is 100 watts before clipping, you would have 99 watts of headroom (approximately 20dB).

On a related note...you referred earlier to the 30dB dynamic range of classical music. I think you are confused about this as well. The dynamic range is MUCH greater than 30dB. While some recordings of some rare pieces might have dynamic peaks 30dB greater than the "average" sound level, I think most experts would say that headroom to accommodate transient peaks of 20dB is generally more than adequate for a home audio system. Personally, I target 15-18dB at normal listening levels.

I would ask you to carefully re-read posts 114 and 118, and then my following conclusion:

If you've determined that you need a 200 watt amp to provide appropriate headroom (in your listening space, with your speakers, with the type of music you listen to and at the loudest volume you will want to listen to on an extended basis), you can buy a an amp rated at 50 watts with 6db of dynamic headroom, or an amp rated at 200 watts with no dynamic headroom. With the 50 watt amp, you will not know if the 150 watts of "dynamic power" can truly be delivered, for how long they can be delivered, and at what level of distortion they will be delivered. Simply put, you will be running that amp at its ragged edge of capability. Not so for the 200 watt amp.

 

[Superampman]

No, it is not. I don't know how you extrapolated your conclusions regarding Krell, and I don't know what you mean about free lunch. As should have been apparent from reading this thread, Krell's "under-rating" is most likely due to their interest in being able to present their amp as capable of "doubling down" as impedance is halved.

You are confusing the concept of "dynamic headroom" with the notion of "headroom".

Dynamic headroom refers to the ability of an amp to deliver power above its maximum continuous power rating for some very brief period of time. There are no standards for dynamic headroom ratings.

Headroom refers to the amount of non-cclipping power available beyond a given continuous output level. In my earlier example, which I will not repeat as it appears to have little effect on your understanding, I referred to an amp running at 1 watt of continuous output. If its maximum continuous output capability is 100 watts before clipping, you would have 99 watts of headroom (approximately 20dB).

On a related note...you referred earlier to the 30dB dynamic range of classical music. I think you are confused about this as well. The dynamic range is MUCH greater than 30dB. While some recordings of some rare pieces might have dynamic peaks 30dB greater than the "average" sound level, I think most experts would say that headroom to accommodate transient peaks of 20dB is generally more than adequate for a home audio system. Personally, I target 15-18dB at normal listening levels.

I would ask you to carefully re-read posts 114 and 118, and then my following conclusion:

If you've determined that you need a 200 watt amp to provide appropriate headroom (in your listening space, with your speakers, with the type of music you listen to and at the loudest volume you will want to listen to on an extended basis), you can buy a an amp rated at 50 watts with 6db of dynamic headroom, or an amp rated at 200 watts with no dynamic headroom. With the 50 watt amp, you will not know if the 150 watts of "dynamic power" can truly be delivered, for how long they can be delivered, and at what level of distortion they will be delivered. Simply put, you will be running that amp at its ragged edge of capability. Not so for the 200 watt amp.

I didn't make a declaration as to why. How is my guess not as good as yours?
And yea, 'free lunch'. Go over the thread. Figure it out.

Headroom, dynamic headroom. I understand the difference without your help. The only piece of the puzzle I was missing was 'duration'. Hence my conclusion.

Are you sure about the dynamic range of a typical classical music recording? 30db represents a thousandfold power requirement increase from the quietest to the loudest passages.

 

[whoaru99]

There is dynamic range, then there is peak to average ratio/crest factor. Sometimes "dynamic range" is used to describe both; right, wrong, or indifferent.

Dynamic range is, generally, the ratio of the most quiet sound to the loudest sound, say 80dB. This is different than the ratio of the average level to the peak or PAR/crest factor. PAR/crest factor may be 20dB for example, but is typically less. The latter is generally the consideration when discussing amp power inasmuch as planning how much power you need to cover the peaks while playing at any given average level.

 

[Guest105]

There is dynamic range, then there is peak to average ratio/crest factor. Sometimes "dynamic range" is used to describe both; right, wrong, or indifferent.

Dynamic range is, generally, the ratio of the most quiet sound to the loudest sound, say 80dB. This is different than the ratio of the average level to the peak or PAR/crest factor. PAR/crest factor may be 20dB for example, but is typically less. The latter is generally the consideration when discussing amp power inasmuch as planning how much power you need to cover the peaks while playing at any given average level.

Yes...dynamic range, correctly, is the ratio of the quietest passage to the loudest. For classical recordings, this can be in the neighborhood of 100dB or more. Here is the introductory paragraph of a technical article on the subject:

Dynamic-Range Issues in the Modern Digital Audio Environment*

LOUIS D. FIELDER, AES Fellow

Dolby Laboratories Inc., San Francisco, CA 91403, USA

The peak sound levels of music performances are combined with the audibility of noise in sound reproduction circumstances to yield a dynamic-range criterion for noise-free reproduction of music. This criterion is then examined in light of limitations due to microphones, analog-to-digital conversion, digital audio storage, low-bit-rate coders, digital-to-analog con version, and loudspeakers. A dynamic range of over 120 dB is found to be necessary in the most demanding of circumstances, requiring the reproduction of sound levels of up to 129 dB SPL. Present audio systems are shown to be challenged to yield these values.​

And here is the link to the paper:

http://www.zainea.com/Dynamic range.htm

Headroom is not an absolute value, as we have both made clear. Rather, it is the power available for transient variations above the average output level, which in turn is defined by the maximum continuous output of the amp, the speaker sensitivity and the listening volume.

As for Superampman's comment regarding the Krell amp:

I didn't make a declaration as to why. How is my guess not as good as yours?

The difference between the Krell's rated and actual continuous power output is NOT headroom, and it is NOT dynamic headroom. As explained above, headroom cannot be defined by the amp manufacturer...it is situational. It is not dynamic headroom because this term applies to power ABOVE the continuous power output capability of the amp (a spec which can be determined by the manufacturer, albeit using there own, vs standardized, criteria). This is not the case in the Krell example.

Hopefully this explains why his conclusion is definitely not correct, whereas my speculation as to their rationale may be.

Incidentally, the difference between 1dB and 30 dB does not represent a thousand-fold increase in power. It is only 16x.

 

[Hyperion]

Think in absolutes all you want but that's silly in this regard. What's the purpose of testing to the point if you destroy the amp? I can see that, maybe, in development but in terms of practical use effectively useless information because it means destroying the amp to achieve it.

It's kind of like those high current specs you see some amp mfg spec. Who cares if the amp can produce 100A peak current into effectively a short circuit? It means nothing from a practical sense except a number that looks good on paper.

Just a minute ! you are trying to define/measure Headroom / Dynamic Headroom or whatever - which are specifications which take the output of an amplifier FAR beyond it's continuous ratings. When I suggest that to find out what this rating might be for a given amplifier, you may take an amplifier beyond it's design ratings and potentially destroy it finding out what its rating is, (or not), you suggest this is "silly" ! - certainly on-topic, if nothing else.

 

[Guest105]

Arrghh. Headroom is NOT, repeat NOT, an amplifier specification. It is a variable amount of available power, calculated as the difference between an amplifier's maximum non-clipping output and the average output level of that amp when driving a particular set of speakers at a particular volume.

The upper boundary of headroom is the amplifier's maximum non-clipping output. In an amp with a stiff power supply, thiis would be the same as the amplifier's maximum continuous power rating (as opposed to "FAR" beyond it).

"Dynamic Headroom" IS a specification quoted by some manufacturers, and it can be measured. However, the measurement thereof has not been standardized. One manufacturer might measure it with no regard to distortion, another might measure it with no regard to duration, etc.

In any case, maximum output is limited by rail voltage.

 

[fred soop]

Yes...dynamic range, correctly, is the ratio of the quietest passage to the loudest. For classical recordings, this can be in the neighborhood of 100dB or more.
.....
Incidentally, the difference between 1dB and 30 dB does not represent a thousand-fold increase in power. It is only 16x.

Whoa. Don't confuse dynamic range with headroom (peak to average). We have been looking at headroom which would be peak to average. Full dynamic range covers levels well below average and would have to consider the S/N of the system.

A 10 dB increase in power is 10 times the power, so a 30 dB increase IS 1000 times the power. For voltage, rather than power, a 20 dB increase is 10 times the voltage. 16 times the power is 12 dB. 16 times the voltage is 24.1 dB.

 

[Guest105]

Whoa. Don't confuse dynamic range with headroom (peak to average). We have been looking at headroom which would be peak to average. Full dynamic range covers levels well below average and would have to consider the S/N of the system.

A 10 dB increase in power is 10 times the power, so a 30 dB increase IS 1000 times the power. For voltage, rather than power, a 20 dB increase is 10 times the voltage. 16 times the power is 12 dB. 16 times the voltage is 24.1 dB.

Whoa back to you. How have I confused dynamic range with headroom? Please let me know. I think I've been quite clear (and quite accurate) in my descriptions of dynamic range, headroom, dynamic headroom and dynamic peaks (aka crest factor). What, specifically, have I got wrong? I've provided definitions which seem to be consistent with my reference texts on the subject, and with your definitions, above.

In several threads, superampman has resisted the notion that headroom is best achieved by running the amplifier at an average output low enough that the remaining power is able to accommodate transient peaks without clipping (or with minimal clipping). He seems to be fixated on the notion that headroom is a function of the amplifier's design, and that an amp with no "headroom" (his term, not mine) would be representative of inferior design. I think it''s pretty clear that he is actually referring to dynamic headroom, since by definition the headroom of a particular amplifier cannot be stated as a specification.....as I've said over and over again, it is a situationally derived variable. And his earlier comment about the 30dB dynamic range of classical music - which evoked my response regarding the actual dynamic range for classical music to be as great as 100dB or even more - almost surely was a mistaken reference refering in actuality to crest factor (dynamic peaks).

And yes, you are of course correct as regards power and 30dB change. Not sure how I made such a silly error.

 

[whoaru99]

Just a minute ! you are trying to define/measure Headroom / Dynamic Headroom or whatever - which are specifications which take the output of an amplifier FAR beyond it's continuous ratings.

Not quite sure what you're getting at?

In most cases, the dynamic headroom of an amp isn't too far beyond its rated power. That said, there are a couple potential exceptions.

1. An amp with sandbagged rated power.
2. An amp like some of the NAD (or was it Proton?) that had unusually high dynamic headroom ratings in the 6 or 7dB range.

Other than those and similar, the dynamic headroom of most amps isn't that far beyond its continuous rating. If one is relying on that to keep you out of trouble you don't have enough rig for the gig.

 

[Hyperion]

I'll keep out of this discussion from now on, I shouldn't have entered it in the first place. If I have an area of expertise this isn't it.

 

[Superampman]

Hey, I'm certainly not averse to learning and understanding the nature of our gears' technology. That's why I joined this forum. As I said, my point of view is indelibly burned into my brain(so to speak). Most of my involvement here is in the form of deference. Some of the things I 'know', clearly need to be unlearned. There's no goal to win an argument here for me.

So, getting back to the Krell, why is it not possible that from a practical standpoint, they quietly built 'spurious' headroom into their amp knowing for all intents and purposes that the consumer would perceive the additional performance as such? After all, they make no reference to it. Just asking.

Incidentally, I think I did mention somewhere that my Stewart Meridian 105 monos are 100w with 7db headroom. But I didn't mention they'll spank the Krell in sq.:yes:

 

[Guest105]

Yes, and my Sansui AU-919 has 20dB of headroom (when its average output is at 1 watt). And 0dB of headroom when it's putting out 123 watts (which is its maximum output before clipping)

Or did you mean that your Steward Meridian 105 monos are rated at 100 watts continuous, with 7 dB of dynamic headroom above that? And if the latter, did they specifiy the duration and/or distortion of this "reserve power"?

As regards the Krell amp....the difference between its rated continuous output and its maximum continuous output is not spurious (which I take to mean "false" or "fake"). It exists, and it is not "dynamic headroom". Please note that dynamic headroom refers to power available beyond the maximum continuous output capability of the amp. Dynamic headroom by definition is NOT continuous power (whereas the difference between the Krell's rated and actual output IS continuous power).

 

[Superampman]

Okay, 100w min. continuous, 7db dynamic headroom. That's the published information. No duration or distortion spec.

I'll reword my question about the Krell. Why is it not possible that from a practical standpoint, they built the amp to accommodate consumers' misconceptions about headroom and dynamic headroom? This way, both the informed as well as the misinformed are going to be happy.

 

[Guest105]

Sure, that's probably what they were thinking.

Many folks don't understand that to avoid/minimize clipping, you must keep an amp's average output at least 9-12dB below its rated/maximum output capability (ideally, 15-18dB) to accommodate transients (dynamic peaks) in the signal.

I suspect that not having a firm grasp of this concept, along with the easily-confused distinction between headroom and dynamic headroom, leads to many misconceptions.

I find that many folks also struggle with the notion that an amp with a fully regulated power supply and no "dynamic headroom" is superior in design to one with a loose power supply and lots of dynamic headroom. Stated another way, a stiffer power supply is superior to a looser one.

 

[Superampman]

The late Dr. Marshall Leach couldn't have said it better in reference to the original iteration of the very limited production Leach Superamp Mono of which I have a pair. Seemingly endless power. I've never been able to cause the clipping indicator led to come on. I suspect it's because input sensitivity is 2.4v for a conservative 300w. Fabulous amps.

 

[fred soop]

Whoa back to you. How have I confused dynamic range with headroom?

Actually, you have NOT confused dynamic range with headroom. The reason I mentioned it is that the discussion has been concerned with headroom or lack thereof, including peak to average ratios, but adding dynamic range to the mix will only serve to add further confusion and "dilutes" the original points that you are trying to make.

 

[fred soop]

I find that many folks also struggle with the notion that an amp with a fully regulated power supply and no "dynamic headroom" is superior in design to one with a loose power supply and lots of dynamic headroom. Stated another way, a stiffer power supply is superior to a looser one.

Strictly speaking, yes. However, the power amplifier itself is a regulator of sorts and there is no reason to add that complexity to the power supply. The main thing is that the power supply be able to drive the amplifier to its full rated power, not counting any inflated short term ratings, whether it is regulated or not. Self and Cordell both agree on this in their power amplifier design books. In fact Self explains much of this at length in Chapter 26 of Audio Power Amplifier Design, sixth edition. Since it is copyrighted material, I can only offer brief summaries in this thread. Turns out, however, that pages 620 through 623, covering most of the major points, is available in the Amazon preview.

Actually, if the supply sags a bit when putting out the additional current for a 4 ohm load, the voltage is still more than sufficient (lower than required for the 8 ohm load) and results in slightly less heat dissipation in the output transistors as a result of the lower voltage. While that sag will prevent doubling down, is doubling down really necessary? Do we really need double the power just because the speaker happens to be 4 ohms?

I did find some info regarding the <bogus> headroom spec from some ancient (1978) IHF data: The test signal is a 1 kHz sine wave that is increased by 20 dB for 20 msec every 500 msec. It is also listed as a "primary" specification meaning that it should be published for all amplifiers. Sounds to me that it is only marginally less bogus than the completely inflated "peak power" that was used years before that.

 

[Guest105]

I think that the discussion has been focused on two issues...1. headroom, and 2. dynamic headroom. However, dynamic range was introduced earlier when it was suggested that classical music had a range of 30dB. This suggests that the poster was confusing dynamic range with dynamic peaks/crest factor, so I addressed this closely related issue as well.

I'm still not sure that I've been able to clearly make the point that headroom and dynamic headroom are two closely related but distinctly different concepts. I went down this road in reaction to Superampman's conclusion that any amp without significant dynamic headroom is inferior...and his assertion that I must be mistaken in my recommendation that the best means of assuring adequate headroom is to assure that one has an amp capable of continuous output which is 15-18dB greater than its average output (which is, of course, a situational value).

In any case, I've officially given up! All I could do would be to repeat what I've already posted, which would of course be a waste of energy.

 

[Guest105]

Strictly speaking, yes. However, the power amplifier itself is a regulator of sorts and there is no reason to add that complexity to the power supply. The main thing is that the power supply be able to drive the amplifier to its full rated power, not counting any inflated short term ratings, whether it is regulated or not. Self and Cordell both agree on this in their power amplifier design books. In fact Self explains much of this at length in Chapter 26 of Audio Power Amplifier Design, sixth edition. Since it is copyrighted material, I can only offer brief summaries in this thread.

Actually, if the supply sags a bit when putting out the additional current for a 4 ohm load, the voltage is still more than sufficient (lower than required for the 8 ohm load) and results in slightly less heat dissipation in the output transistors as a result of the lower voltage. While that sag will prevent doubling down, is doubling down really necessary? Do we really need double the power just because the speaker happens to be 4 ohms?

I did find some info regarding the <bogus> headroom spec from some ancient (1978) IHF data: The test signal is a 1 kHz sine wave that is increased by 20 dB for 20 msec every 500 msec. It is also listed as a "primary" specification meaning that it should be published for all amplifiers. Sounds to me that it is only marginally less bogus than the completely inflated "peak power" that was used years before that.

Fully agreed. I was only trying , simplistically, to make the point that "dynamic headroom" is NOT necessarily an attribute...and may in fact be more representative of a less robust power supply than of a high-quality amp.

 

[Superampman]

I think that the discussion has been focused on two issues...1. headroom, and 2. dynamic headroom. However, dynamic range was introduced earlier when it was suggested that classical music had a range of 30dB. This suggests that the poster was confusing dynamic range with dynamic peaks/crest factor, so I addressed this closely related issue as well.

I'm still not sure that I've been able to clearly make the point that headroom and dynamic headroom are two closely related but distinctly different concepts. I went down this road in reaction to Superampman's conclusion that any amp without significant dynamic headroom is inferior...and his assertion that I must be mistaken in my recommendation that the best means of assuring adequate headroom is to assure that one has an amp capable of continuous output which is 15-18dB greater than its average output (which is, of course, a situational value).

In any case, I've officially given up! All I could do would be to repeat what I've already posted, which would of course be a waste of energy.

FWIW, you could not have been more clear. I appreciate your effort.