Amplifier power requirements vs type of music, peak/dynamic power vs sustained power...

[Xulonn]

[Regarding 70's amps designed for "peak Power"] ...simply ran out of steam when really pushed and things got ugly, fast.

This goes beyond the discussion of the Crown XLS2502 power amp, so I started a generic thread on the topic.

I would like to hear from some of our resident technical experts about genres of music and variations in amplifier power requirements. I'm not going to get into the "how much power is needed for tweeters" argument. However, in my many years of playing with audio, I have never bothered to learn to read and fully comprehend waterfall and other complex speaker test graphs, but I assume that some of them show power and current vs. SPL and frequency. ("Speaker graphs and how to read them" would be a great topic for a member with expertise in the area to post a tutorial.) '

My lifetime of listening to music - acoustic and amplified, live and recorded - leads me to believe that the sustained "music power" levels of amplified instrument music, up to including sub-genres of heavy metal and bass-thumping pop styles, would be a poorer application for amps that are designed to provide additional "clean" peak power for actual short-term peaks. Acoustic music (vocals and acoustic instruments) and recordings with a high dynamic range would seem to be a better application for "music power" reserve type designs. Bi-amping also makes sense to me in some situations, and I consider subwoofers - active or passive - to be a variation of bi-amping.

Intuitively, it seems to me that a fraction of a second of higher power demand wouldn't be much of a problem, but once you get up to a second or more, one would really need "sustained" high power capability rather than a reserve. But this is complicated - and eliminated by simply using a very powerful low distortion amplifier that is less likely to blow your speakers than a weaker amp driven into massive distortion. My Classé 70wpc class AB with its inaudible (from a couple of feet away) mechanical power transformer hum is perfect for my needs, and drives my little and very bass-limited Paradigm Atom v6 bookshelf speakers to very loud levels in a small room.

Another factor regarding amplifier performance that you may remember from previous conversations here, is that no-reserve amps that supposedly "track" demand and provide power as needed - like the old Carver Cubes and Bob Carver's other "Magnetic Field" power amps - perform miserably under sustained loads, and indeed, mine went into thermal shutdown during a party at my house while driving a pair of ADS-L810 speakers.

Bottom line for semi-technical audiophiles, as long as there is power for sufficient bass demands of the music one listens to, it seems to me that a modestly powered amp with good "reserve power" make sense only for acoustic and other "low-power" music.

 

[pjug]

I wish clipping indicators were more common. Without this, it is such a mystery whether an amplifier has enough power, unless you go with extreme overkill.

 

[CDMC]

Your are conflating some issues. Addressing the various issues:

1) Headroom- Even the most compressed music still has short term peaks. Generally, 20ms is considered a short term peak. Even with a very compressed recording, you can expect at least 6db peaks, which is 4x continuous power. More realistic is 10db, which means 10x continuous power. Even moderately dynamic recordings can have peaks into the 20db range, which represents 100x the the continuous power.

2) Basic Power requirements- The power needed is a function of the listening level, speaker efficiency, and room size. Take the same room and two different speakers, a small dynamic speaker with extended bass resulting in 84db/watt efficiency and a large Klipschorn with 101 db/watt efficiency. To get the same 105db peak level from those speakers at 10 feet listening distance, the small dynamic speaker will require 300 watts per channel while the Klipschorn will require 6 watts. 105 db is loud, lets say you listen to 95db peaks max, you still need 30 watts for the small dynamic, while the Klipschorn only needs .6 watt.

3) High headroom amplifiers. Carver TFM amplifiers were not a high headroom design. They were a Class G rail switching amplifier which were designed to run cool relative to their power rating, by running on low voltage rails until they needed the higher voltage rails. Like any amplifier, they would shut down if pushed too hard too long. The design itself was very reliable (excluding the early M400 Cube) and used by many touring bands for years for sound reinforcement as they were far lighter and put out less heat than traditional Class A/B amplifiers. Class D amps have since improved on and replaced Class G in this role. What were designed as high headroom amplifiers were the Proton D series and NAD Power Envelope series amplifiers, both of which had more than 6db of short term headroom. All of the foregoing are different designs than Carver's Sunfire Amplifiers which used what he termed a tracking downconverter to vary the voltage rails and keep them at 6 volts above the music signal.

 

[LuckyLuke575]

While this may be overly simplistic, I gauge this point with where the volume knob is versus the listening volume one wants to be at. If you're at 4/10 on the volume amp and your at max listening volume, then you're good. In general, the amp should have a lot more power than what you want to do with it (budget permitting of course).

 

[CDMC]

While this may be overly simplistic, I gauge this point with where the volume knob is versus the listening volume one wants to be at. If you're at 4/10 on the volume amp and your at max listening volume, then you're good. In general, the amp should have a lot more power than what you want to do with it (budget permitting of course).

This in no way determines if you have sufficient headroom. You can have a preamp with a lot of gain (audio research, conrad johnson) coupled with an amplifier that has a lot of gain (i.e. .5 v for rated output) resulting in a combo that can go into clipping when driven by a standard 2v dac at 1/4 on the volume control. Likewise, you could have a passive preamp and an amplifier that needs 5v to get rated power that will never clip when driving by a standard 2v DAC with the volume turned up all the way.

I should also add that when using a traditional potentiometer, you ideal want the sensitivity to be that your primary listening volume is where the knob is at 1/2 volume or higher. This is because potentiometers tend to not track linearly at lower levels leading to channel imbalance. This is difficult to achieve in practice because preamplifiers are designed to have 10+db (and in many cases 20db) of gain to make sure there is sufficient gain when using a phono-preamp. While the reference level for a consumer CD or DAC running through RCA is 2V (4V if XLR) it is not uncommon to have phono-preamps that are only putting out 200mv (.2 volt) or 10 db less. So where a digital source may have plenty of output to run a passive preamp that needs only to cut the input signal, analog sources often need a boost.

 

[Doodski]

While this may be overly simplistic, I gauge this point with where the volume knob is versus the listening volume one wants to be at. If you're at 4/10 on the volume amp and your at max listening volume, then you're good. In general, the amp should have a lot more power than what you want to do with it (budget permitting of course).

You're mistaking the potentiometer taper for power output relative to the position of the volume.
http://www.resistorguide.com/potentiometer-taper/

 

[NTomokawa]

If you're at 4/10 on the volume amp and your at max listening volume, then you're good.

Learn something new every day!

 

[Xulonn]

Good points. Thanks for the contributions to the discussion. My interest in starting this thread was to continue to broaden and deepen my knowledge, and perhaps help some younger members learn more about this critical aspect of putting together speaker-based audio systems.

But I will comment that a modern digital music system using a combination of a computer audio application (Daphile in my case) digital volume control from a keyboard, plus the electronic rotary volume control of my Topping DX7s DAC makes the old "clock position" metaphor much less useful. It is still useful to know that you are not "cranking it all the way up" for even fairly loud listening sessions.

My many decades of playing with audio has allowed me to "experiment" with many different mid-priced combinations of speakers and amps, and even the Carver M400 Cube + ADS L810 combo sounding pretty good. A McIntosh Mac 1900 receiver (50wpc) worked well with a pair of JBL S8 (LE15A woofer + 175 midrange + 075 tweeter) systems in 16 cubic foot ported cabinets in a fairly small room. (Actually, those speakers were gross overkill for my rental house!) Years later when we had a big living room with 18' peaked ceiling, my system included a 3.5wpc 2A3 SET amp driving a pair of Klipsch Forte II speakers at ~100dB sensitivity. It worked for me. For now, my Classé 70wpc works fine driving the Paradigm Atom bookshelf speakers in a really small room

I'm considering getting what will probably be my last set of speakers because of my age - a pair of 44" tall, 88dB sensitive Salk Songtowers that will give me a fair bit more bass extension and satisfy my passion for beautiful wood cabinetry. I have a feeling that the Classé will be sufficient for those as well. (I miss the deeper bass in some of my electronic music, and don't want to deal with integrating a subwoofer or two into my system.)

 

[RayDunzl]

Since most of us listen to digits, and since those digits have already been limited to a defined range, with a few exceptions for intersample-overs, or misguided EQ boosts, I'm not sure where these surprise demands for headroom come from...

 

[Doodski]

Good points. Thanks for the contributions to the discussion. My interest in starting this thread was to continue to broaden and deepen my knowledge, and perhaps help some younger members learn more about this critical aspect of putting together speaker-based audio systems.

But I will comment that a modern digital music system using a combination of a computer audio application (Daphile in my case) digital volume control from a keyboard, plus the electronic rotary volume control of my Topping DX7s DAC makes the old "clock position" metaphor much less useful. It is still useful to know that you are not "cranking it all the way up" for even fairly loud listening sessions.

My many decades of playing with audio has allowed me to "experiment" with many different mid-priced combinations of speakers and amps, and even the Carver M400 Cube + ADS L810 combo sounding pretty good. A McIntosh Mac 1900 receiver (50wpc) worked well with a pair of JBL S8 (LE15A woofer + 175 midrange + 075 tweeter) systems in 16 cubic foot ported cabinets in a fairly small room. (Actually, those speakers were gross overkill for my rental house!) Years later when we had a big living room with 18' peaked ceiling, my system included a 3.5wpc 2A3 SET amp driving a pair of Klipsch Forte II speakers at ~100dB sensitivity. It worked for me. For now, my Classé 70wpc works fine driving the Paradigm Atom bookshelf speakers in a really small room

I'm considering getting what will probably be my last set of speakers because of my age - a pair of 44" tall, 88dB sensitive Salk Songtowers that will give me a fair bit more bass extension and satisfy my passion for beautiful wood cabinetry. I have a feeling that the Classé will be sufficient for those as well. (I miss the deeper bass in some of my electronic music, and don't want to deal with integrating a subwoofer or two into my system.)

View attachment 44389

Your Classé Audio 75 w/ch Class A will work wonderfully with these transmission design towers. The cherry wood looks gorgeous.
http://www.salksound.com/model.php?model=SongTower

 

[CDMC]

Since most of us listen to digits, and since those digits have already been limited to a defined range, with a few exceptions for intersample-overs, or misguided EQ boosts, I'm not sure where these surprise demands for headroom come from...

There are not surprise demands, we know the maximum level possible. We are however slightly misusing the term headroom in this thread, referring to peaks above average level as headroom rather than the absolute maximum level above average as used by recording engineers.

 

[CDMC]

Good points. Thanks for the contributions to the discussion. My interest in starting this thread was to continue to broaden and deepen my knowledge, and perhaps help some younger members learn more about this critical aspect of putting together speaker-based audio systems.

But I will comment that a modern digital music system using a combination of a computer audio application (Daphile in my case) digital volume control from a keyboard, plus the electronic rotary volume control of my Topping DX7s DAC makes the old "clock position" metaphor much less useful. It is still useful to know that you are not "cranking it all the way up" for even fairly loud listening sessions.

My many decades of playing with audio has allowed me to "experiment" with many different mid-priced combinations of speakers and amps, and even the Carver M400 Cube + ADS L810 combo sounding pretty good. A McIntosh Mac 1900 receiver (50wpc) worked well with a pair of JBL S8 (LE15A woofer + 175 midrange + 075 tweeter) systems in 16 cubic foot ported cabinets in a fairly small room. (Actually, those speakers were gross overkill for my rental house!) Years later when we had a big living room with 18' peaked ceiling, my system included a 3.5wpc 2A3 SET amp driving a pair of Klipsch Forte II speakers at ~100dB sensitivity. It worked for me. For now, my Classé 70wpc works fine driving the Paradigm Atom bookshelf speakers in a really small room

I'm considering getting what will probably be my last set of speakers because of my age - a pair of 44" tall, 88dB sensitive Salk Songtowers that will give me a fair bit more bass extension and satisfy my passion for beautiful wood cabinetry. I have a feeling that the Classé will be sufficient for those as well. (I miss the deeper bass in some of my electronic music, and don't want to deal with integrating a subwoofer or two into my system.)

View attachment 44389

The Salk cabinets are beautiful. I have a pair of Song 3s in walnut sitting in my living room.

 

[Blumlein 88]

There are not surprise demands, we know the maximum level possible. We are however slightly misusing the term headroom in this thread, referring to peaks above average level as headroom rather than the absolute maximum level above average as used by recording engineers.

In recent recordings peaks above average level is about 6 db.

If your amp will hit peak possible level without choking, then you aren't missing anything. If the specs are accurate on the Salks, your amp would probably reach 106-109 db SPL at 1 meter cleanly. That would be slightly reduced at the listening position probably. That amount usually satisfies most people.

Can't say I understand the hairsplitting about headroom. In digital, we have a fixed max hard limit. Headroom is that minus the average signal level. Things are less hard-limited in analog signals, but aren't all that loose.

Way back when, and still in principal most recording people wanted 20 db headroom over average level. Few final recordings have that much headroom these days. If your amp and speaker play say 105 db without being over-driven or sounding wonky, and 105 db peaks are enough for you then what more is there to want?

Classe amps aren't super low thd amps. Claims are .1% or -60 db which is probably enough. I've owned Classe 15 and 25 (and for that matter DR3 VHC) amps. Same thd spec more power and current.

 

[pjug]

Since most of us listen to digits, and since those digits have already been limited to a defined range, with a few exceptions for intersample-overs, or misguided EQ boosts, I'm not sure where these surprise demands for headroom come from...

I suppose you could go through the numbers that way and make sure the system does not clip. Or maybe you can then know know when clipping is possible if you know exactly how much attenuation is applied with digital volume control. But i use multiple sources and analog preamp, and have enough gain to easily drive my amp well into clipping. In that situation it is a guessing game.

What I have done is put a scope on my speakers with dynamic music playing as loudly as I would ever want it. That seemed to indicate no issue with clipping, not even close. So I seem to have more amp than I really need.

 

[jsrtheta]

...My interest in starting this thread was to continue to broaden and deepen my knowledge, and perhaps help some younger members learn more about this critical aspect of putting together speaker-based audio systems.

Some older ones, too.

 

[RayDunzl]

So I seem to have more amp than I really need.

That's my solution, too.

 

[KSTR]

In active studio monitors, notably ones uses for tracking and mixing with almost fully uncompressed material, you never can have enough headroom. It is not overkill to have 300W for a 10W tweeter and 1200W for a 60W midrange.... and even then you'll need soft clippers on the woofer and mid channels in order to achieve "transparent" clipping (which still should be indicated). All this places high demands on the input/preamp/crossover stages (S/N-ratio especially), and of course thermal limiters are mandatory to protect the drivers. The amps (class-D with SMPS) are as well not rated for continuous duty and don't need to be rated any higher than the thermal rating of the associated driver (with a derating factor of at least two for both the power and the time constant).
For consumer active speakers, the constraints may be relaxed a bit but not as much as one would would expect from the low crest factor of today's recordings. The crossover also needs some voltage headroom (6...12dB) for transient signals, a fact that is seldom understood and often neglected and results in unexpected clipping with real music signals that didn't show up with (swept) sine measurements....

I won't comment on passive speakers with any arbitrary amplifier as I consider this a design failure right from the start unless a carefully adjusted loudspeaker management system is used to protect the speaker and (properly selected) amp while allowing high transient powers.

 

[restorer-john]

Can't say I understand the hairsplitting about headroom. In digital, we have a fixed max hard limit. Headroom is that minus the average signal level. Things are less hard-limited in analog signals, but aren't all that loose.

That said, in real world usage terms, going from one recording to another recording, track to track and adjusting the volume on the fly to give a comfortable level, you really have little idea what headroom is remaining, unless you have a volume control and entire system that is calibrated to 0dB (flat out) that equals close to 0dBFS and full power at the speaker terminals.

The reality is most people and most systems will have considerable extra front end gain, over and above what will drive their amplifiers to full (clipping) power. (they have got used to a volume control that 'peaks' around 10-11 o'clock or maybe 2 o'clock at the most) As such, high dynamic headroom amplifiers are useful and arguably a very valid way of ensuring that reserve without paying the big bucks for continuous power.

A big power amplifier with 200W/ch@8R, ~350W/ch@4R and a few dB of clipping headroom into 4&8R is perfect for just about any speaker, music and level of intoxication IMO.

Basically, to determine how long the high dynamic power (commutating rail) amplifiers can deliver their ~4-6dB headroom, you use a toneburst and vary the length until the rails collapse and the burst squashes on the scope. The dynamic rating was IIRC 20mS (EIA), NAD claimed up to 500mS and Proton claimed into the 1-2000mS on their big amps. But both of them also looked at a time integrated IV to protect the OP Trs from SOA violations as well as monitored HS temperature and PT frame temperature. They were cautious in case a pub rock band tried to use them...

I'll throw a few early NAD brochures on the scanner for this thread.

 

[snapsc]

@Xulonn

Back in the day I owned the ADS L810 series 2 powered by a Carver M1…now I have a pair of Salk Veracity STs powered by a 2Cherry….and yes, you are going to love the craftsmanship of the woodworking performed by Jim Salk's team! The sound ain’t bad either. The transmission line will give you a deep, tuneful low end that is easy to integrate with a sealed subwoofer if you ever decide it is needed.

The Salks are 6 ohm nominal, 4 ohm minimum….so you should probably consider their sensitivity as being closer to 86 db as I’m pretty sure the 88 db is based on 2.83 volts @ 1M which would be about 1.5 watts.

As for power and headroom: If you plug your data into this website calculator, you will see that with 2 loudspeakers placed within 4 feet of the rear walls and powered by 8 watts, the SPL at the listening position is 91 db…and with a 70 watt amplifier you are close to the 10x headroom for peaks.

 

[Frank Dernie]

The Devialet has a calibrated volume control with 0dB producing full power from a full digital signal. In fact the max power is programmable but I onlt reduce it if I am using it with my 109dB/watt horns.
The volume control allows positive settings so it would be possible to clip it. My normal listening level is between -30dB and -15 dB depending on the recording compression so I know I am not clipping the amp.