I'm gonna go ahead and open up a potential can of worms.
This question seems to be regurgitated quite often and in my opinion with unsatisfying conclusions in the discussion.
Q: What do power ratings of a driver mean and how do they translate to picking a suitable amplifier?
Let's take a B&C driver for example, let's say a 12" 12NDL76.
It was tested to withstand 400W pink noise at crest factor 2 (6dB) for 2 hours (at this c.f., peak power is 4xRMS)
Now what does this mean? A lot of times I see interpretations that rely on crest factor, sounding something like 'well then it definitely can run 400W music material with c.f. 4 (12dB), because that is less compressed. But not 400W c.f. 1.4 (3dB, sine wave!), that would kill it!'.
However, don't we measure Power RMS here? AKA the driver saw 400W RMS over 2hrs, and not 400W peak power of that pink noise, which would mean only 100W RMS at 6dB c.f.. In the latter case they would have to list 100W in the specs, not 400W. But when we start orienting ourselves on peak power values, then the whole idea of performing 2h tests wouldn't make any sense.
When we talk about RMS power, the crest factor shouldn't matter - RMS is RMS. But then again, going forward with that logic, I can't seem to make 100% sense of the way speaker driver specs are being published and how this relates to the amplifier I'll need to run that driver hard.
Q: And then there is the mystery of amplifier specifications. Does Physics work here?
To me, amplifier power output ratings unfortunately seem to suffer from a comparable lack of logic.
Here again, we find the crest factor being of importance. But why? I understand that a lower crest factor is inherently a more difficult task for an amplifier, but shouldn't that be reflected in an accordingly smaller maximum rated power output?
Let's take the LabGruppen PLM 12K44 for example. You can find the thermal spec sheet online.
On the line under 120V, 30A mains for example says that each of the four channels can deliver 'rated power' of 3000W to a 4Ohm load. At the top of the document it says the measurements are done at 1/8 signal duty. In this case 2817W are being fed into the amp, about half is dissipated into heat, leaving 1502W available for all four channels. Generously assuming 100% of that goes into the speakers, that is 375W/ch. And that must be RMS power! And if we want to scale that up to 1/1 signal duty we indeed obtain 8x375W=3000W. So far so good.
But can we actually get that 1/1 signal duty? For example I would like to power 4x 4Ohm dual 18" 2200W continuous on one of each channel. This demand seems to collide with the statement of the 1/8 signal duty being the most extreme sustained normal operation as stated in the spec sheet? And how in the world would the amp even deliver the rated RMS output at 3000W/ch when the input at 1/8 signal duty is already at 24A!
When we assume a crest factor of 12dB for the pink pseudo noise, then the peak power in this case is about 16xRMS, which is 16x375W=6000W. it seems to me that the rated power in reality is somewhere in between the RMS power and peak power of the test signal, 375W < 3000W < 6000W. It also seems that it is connected to the crest factor of the test signal. And I would expect the rated power therefore to be reduced for lower c.f..
The power spec sheet indeed reveals that the 3000W/ch refers to a burst power (1 kHz, 25 ms burst power @ 150 BPM, 12 dB Crest factor), so it is indeed more of a peak power rating rather than RMS. So 1/8 signal duty is somewhat misleading as it suggests, 1/1 is possible.
Other reputable pro amplifier manufacturers seem to benchmark their rated power output by burst power capability as well.
Speaker driver and Amp together
To stick with this example, with my dual 18" rated at 2200W, 4Ohm, I really would think the 12K44 would be a great match with its 3000W/ch at 4Ohm. But following real RMS values, my 2200W sub would not be properly powered. And this goes against the typical advice which goes like 'get an amp that has 1.5x rated power of the drivers RMS value'. So how does this fit advice in here?
Anyway, looking forward to your thoughts on this.
This question seems to be regurgitated quite often and in my opinion with unsatisfying conclusions in the discussion.
Q: What do power ratings of a driver mean and how do they translate to picking a suitable amplifier?
Let's take a B&C driver for example, let's say a 12" 12NDL76.
It was tested to withstand 400W pink noise at crest factor 2 (6dB) for 2 hours (at this c.f., peak power is 4xRMS)
Now what does this mean? A lot of times I see interpretations that rely on crest factor, sounding something like 'well then it definitely can run 400W music material with c.f. 4 (12dB), because that is less compressed. But not 400W c.f. 1.4 (3dB, sine wave!), that would kill it!'.
However, don't we measure Power RMS here? AKA the driver saw 400W RMS over 2hrs, and not 400W peak power of that pink noise, which would mean only 100W RMS at 6dB c.f.. In the latter case they would have to list 100W in the specs, not 400W. But when we start orienting ourselves on peak power values, then the whole idea of performing 2h tests wouldn't make any sense.
When we talk about RMS power, the crest factor shouldn't matter - RMS is RMS. But then again, going forward with that logic, I can't seem to make 100% sense of the way speaker driver specs are being published and how this relates to the amplifier I'll need to run that driver hard.
Q: And then there is the mystery of amplifier specifications. Does Physics work here?
To me, amplifier power output ratings unfortunately seem to suffer from a comparable lack of logic.
Here again, we find the crest factor being of importance. But why? I understand that a lower crest factor is inherently a more difficult task for an amplifier, but shouldn't that be reflected in an accordingly smaller maximum rated power output?
Let's take the LabGruppen PLM 12K44 for example. You can find the thermal spec sheet online.
On the line under 120V, 30A mains for example says that each of the four channels can deliver 'rated power' of 3000W to a 4Ohm load. At the top of the document it says the measurements are done at 1/8 signal duty. In this case 2817W are being fed into the amp, about half is dissipated into heat, leaving 1502W available for all four channels. Generously assuming 100% of that goes into the speakers, that is 375W/ch. And that must be RMS power! And if we want to scale that up to 1/1 signal duty we indeed obtain 8x375W=3000W. So far so good.
But can we actually get that 1/1 signal duty? For example I would like to power 4x 4Ohm dual 18" 2200W continuous on one of each channel. This demand seems to collide with the statement of the 1/8 signal duty being the most extreme sustained normal operation as stated in the spec sheet? And how in the world would the amp even deliver the rated RMS output at 3000W/ch when the input at 1/8 signal duty is already at 24A!
When we assume a crest factor of 12dB for the pink pseudo noise, then the peak power in this case is about 16xRMS, which is 16x375W=6000W. it seems to me that the rated power in reality is somewhere in between the RMS power and peak power of the test signal, 375W < 3000W < 6000W. It also seems that it is connected to the crest factor of the test signal. And I would expect the rated power therefore to be reduced for lower c.f..
The power spec sheet indeed reveals that the 3000W/ch refers to a burst power (1 kHz, 25 ms burst power @ 150 BPM, 12 dB Crest factor), so it is indeed more of a peak power rating rather than RMS. So 1/8 signal duty is somewhat misleading as it suggests, 1/1 is possible.
Other reputable pro amplifier manufacturers seem to benchmark their rated power output by burst power capability as well.
Speaker driver and Amp together
To stick with this example, with my dual 18" rated at 2200W, 4Ohm, I really would think the 12K44 would be a great match with its 3000W/ch at 4Ohm. But following real RMS values, my 2200W sub would not be properly powered. And this goes against the typical advice which goes like 'get an amp that has 1.5x rated power of the drivers RMS value'. So how does this fit advice in here?
Anyway, looking forward to your thoughts on this.
Your workings are correct, however many people are fooled by the ratings of the new fangled digital amps with their multi-kW ratings in a casing the weight of a six-pack. Forget about real electrical RMS ratings, these are burst values as you have discovered, typically for 20mS at 1kHz. The output of a '5200W' per channel Powersoft amp on continuous test dropped to 270W after 15 minutes. It's very much like the old inflated 'Peak Music Power' ratings of the '70s of which Tandy were particularly fond; you know, 4" bookshelf speakers 250W PMP. Well now it's back again so that manufacturers can advertise a 1 Unit amplifier with 12kW output using only 2kW input power or whatever. If only... Generally, digital amps are ok for subs with ordinary music but obviously less capable with near-continuous bass-heavy genre. I use the biggest conventional amps I can find (about 290kG of them in total and 6 x 2kW for the 18s, 4 x 3kW for the 21s), but they are becoming increasingly scarce these days and I predict that they will have all but disappeared on the brand new market by the time this decade is out. The bottom line is to buy the biggest amps you can reasonably afford - you can always turn them down. I've used 1k Real RMS Watt amps on 80W HF drivers for years with no failures. These new-style ratings do actually fairly well match the spectral power required for actual music, but I'm still not convinced about their suitability for extremely bass-heavy signals, when the ratings should be taken with a pinch of salt. All is not lost though; your '2200W' drivers would likely melt after a few minutes at a real 400WRMS, since AES ratings also yield an inflated number as they are also 'real music' based numbers. I'm sure that the marketing men will argue that it's all about Real World conditions. I believe it's just the same old big numbers ****-waving exercise. Things ain't as simple as they used to be!
Forget Peak power, burst power, all marketing BS.
The 400W rating (if they are not lying) is real, RMS is RMS and waveform does not matter, it´s thermal/heating power, integrated by voice coil and surrounding parts.
In any case, it´s "translated" to sinewave power which heats up voice coil (lamp bulb filament - coffee pot heater - etc.) to the same temperature.
Crest factor here basically means since Pink Noise is a Musical Signal representation , and by normal program dynamic range some peaks WILL be clipped if amp is played loud, then they added some peak clipping to approximate more Real World conditions. Not bad.
Don´t try to extrapolate much from that, nor calculate peak power, etc. , I always used my simplistic "back to the basics" approach and reasonably kept speakers alive on stage situations, for decades now.
Again, see above, stick to thermal (RMS) power and you´ll both be safe yet not under-use what you have.
Tight limiting is the norm; if not and amp is allowed to clip freely as is often the case with MI and even DJ applications, or sometimes driving bandpass subwoofers who often don´t care about clipping because "anyway box response kills clipping harmonics" , then RMS is "squarewave clipped signal", up to 2X the "clean sinewave RMS" rating, consider that when selecting speakers.
You bet!!!!
Always!!!!
Although some people gets angry when I write that simple Truth 😉
Call it Marketing Dishonesty and we are talking 😉
See above.
ALL of this is Math Gymnastics, intended to cheat/mislead/lie to you.
Don´t read more than this basic Truth there.
Horoscopes use VERY complicated Astral Charts, which LOOK impressive, are chock full of Astronomical "Tech" terms, impossible to navigate except by well trained "Sage people":
Does it make them any truer?
NO!!!, just babble and nonsense.
Which are accurate for burst sinewave music programs ONLY.
What kind of Music do you listen to? 😉
I find Lab Gruppen and similar amp ratings useless cr*p, just like that, and trust only "traditional" no nonsense ratings.
Anything else is useful only to awe the gullible.
Completely inapplicable here where speaker RMS power can be measured (at least the spec is straightforward) but amp spec is useless galimatias ... how can you compare both?
I can´t.
If anything, test the amp under same conditions as speaker: how much pink noise power, crest factor 6dB can it supply into rated load for 2 hours?
THAT is the REAL value.
I think we can all agree that amplifier specs are an utter mess to dig into.....😕
I do match amp's output capability to active multi-way driver sections.
I think it is essential for subs, and sometimes necessary for low-mids. Mids & CDs are usually not an issue.
My goal is to have each section reach the overall desired speaker SPL capability, and have no dynamic clip limiting in any section.
I use 18dB as the crest factor for music, which may be a little low across the frequency range. Meyer's M-Noise is an attempt at pink noise that better replicates music, and crest factor rises on the top end.
In sizing amps, I work backwards from an amp's peak voltage capability, and see how much wattage that implies.
For instance, the PLM 12K44 has a peak voltage spec of 194V.
For subs, i use a crest factor of 12dB, which brings down average (RMS, whatever) voltage to 48.5V.
For a 4 ohm load, using P=V^2/R, I get about 600 Watts.
That's the most wattage i want the the amp to deliver into a 4 ohm sub........ without potential clipping.
Of course that ignores the actual impedance curve which needs a little scrutiny.
It also ignores most folks may never hear the difference with a sub being pushed harder and clipping..
But me, I'd rather add more subs and stay clean.
To my ears, avoiding sub clipping matters far more than group delay, so commonly scrutinized. ymmv.
here's a very good amp power rating article...... https://www.prosoundtraining.com/2011/12/13/inside-amplifier-power-ratings-part-3/
And Keele's classic tone burst test ....https://www.audiosciencereview.com/forum/index.php?attachments/assessment-of-loudspeaker-system-short-term-maximum-output-pdf.53408/
I think that tone burst test defines the driver & amp combination needed for each driver section, to simultaneously reach maximum linear SPL....
and truly clean sound, at whatever desired SPL, be it high or low.
I do match amp's output capability to active multi-way driver sections.
I think it is essential for subs, and sometimes necessary for low-mids. Mids & CDs are usually not an issue.
My goal is to have each section reach the overall desired speaker SPL capability, and have no dynamic clip limiting in any section.
I use 18dB as the crest factor for music, which may be a little low across the frequency range. Meyer's M-Noise is an attempt at pink noise that better replicates music, and crest factor rises on the top end.
In sizing amps, I work backwards from an amp's peak voltage capability, and see how much wattage that implies.
For instance, the PLM 12K44 has a peak voltage spec of 194V.
For subs, i use a crest factor of 12dB, which brings down average (RMS, whatever) voltage to 48.5V.
For a 4 ohm load, using P=V^2/R, I get about 600 Watts.
That's the most wattage i want the the amp to deliver into a 4 ohm sub........ without potential clipping.
Of course that ignores the actual impedance curve which needs a little scrutiny.
It also ignores most folks may never hear the difference with a sub being pushed harder and clipping..
But me, I'd rather add more subs and stay clean.
To my ears, avoiding sub clipping matters far more than group delay, so commonly scrutinized. ymmv.
here's a very good amp power rating article...... https://www.prosoundtraining.com/2011/12/13/inside-amplifier-power-ratings-part-3/
And Keele's classic tone burst test ....https://www.audiosciencereview.com/forum/index.php?attachments/assessment-of-loudspeaker-system-short-term-maximum-output-pdf.53408/
I think that tone burst test defines the driver & amp combination needed for each driver section, to simultaneously reach maximum linear SPL....
and truly clean sound, at whatever desired SPL, be it high or low.
Only two things matter
1) That your amplifier has enough power (Watts) and your speaker enough sensitivity (dB SPL @ 1Watt, 1metre) to achieve the dB SPL you desire at your listening position
2) If your amplifier grossly exceeds the power handling of the speaker, and there's an excited monkey at the volume control knob, or your crossovers are misconfigured for higher frequency drivers, this vastly increases the chances of the speaker driver being blown.
1) That your amplifier has enough power (Watts) and your speaker enough sensitivity (dB SPL @ 1Watt, 1metre) to achieve the dB SPL you desire at your listening position
2) If your amplifier grossly exceeds the power handling of the speaker, and there's an excited monkey at the volume control knob, or your crossovers are misconfigured for higher frequency drivers, this vastly increases the chances of the speaker driver being blown.
It doesn't. The whole idea that you must have an amplifier that exceeds the speakers rated power is an unscientific myth. The thinking that a clipping amplifier results in blown speakers, also a myth. Too much power being sent to the speakers is what blows them up. Clipping can increase power at higher frequencies, due to the sharp edges of the clipped waveform. The effect however is really minimal on actual musical content, your ears will be hurting from the clipping distortion before the tweeters are. If you have a 100Wrms amp, a 100Wrms tweeter and a properly functioning crossover, you're going to have to try really hard to blow up the tweeters. Driven to a full square wave, a 100Wrms amplifier can theoretically produce 400Wrms if the powersupply is up to it, but you're at 50% THD and your ears will be bleeding if on musical content.
Last edited:
The anwer to the OPs uestion is “Not much”. At least in home hifi.
First there is no standard for rating power so you cannot compare manufacturer to manufacturer.
And power ratings for amplifiers are also kind of nebulous.
You want suffiicent power, with well behaved clipping behaviour, to play as loud as needed. how an amplifier clips can make way more difference than how much power they output.
dave
First there is no standard for rating power so you cannot compare manufacturer to manufacturer.
And power ratings for amplifiers are also kind of nebulous.
You want suffiicent power, with well behaved clipping behaviour, to play as loud as needed. how an amplifier clips can make way more difference than how much power they output.
dave
Utterly. RMS Power of RMS Voltage?
Work on the voltage, to ensure the speaker can receive unclipped peaks of the >6 dB crest factor transient signals which up all music. That includes even the heaviest, compressed, sine wave basslines of electronic music, once they are put through a high pass filter - and since every single speaker ever made is an acoustic high pass filter, then we're all good.
Some good reading:
https://www.merlijnvanveen.nl/en/study-hall/168-don-t-judge-a-booklet-by-its-cover
https://www.merlijnvanveen.nl/en/study-hall/191-crest-factor-part-1
https://www.merlijnvanveen.nl/en/study-hall/192-crest-factor-part-2
https://www.doctorproaudio.com/content.php?2260-loudpeaker-power-handling
http://www.synaudcon.com/site/wp-content/uploads/2014/01/L1107_Amped_Up.pdf
https://www.prosoundweb.com/cross-p...-designed-to-drive-a-loudspeaker-or-a-heater/
And a useful tool:
https://www.prosoundtraining.com/2016/07/19/power-amplifier-calculator/
Dissipated power by a Vrms across a resistor gives W, not Wrms
RMS means that the value was calculated using a mathematical operation called Root Mean Square.
Apply RMS to voltage, current and power.
You will see that applying RMS to power does not give coherent results with the RMS voltage or current.
The error is very common, even done by manufacturers of amplifiers and loudspeakers.
But there is no reason to constantly repeat it.
When you apply science and maths to audio, stay scientific.
Calculating the RMS of an AC power waveform over time gives a meaningless value. You can do it, but it's not relevant to physical reality. DC is a different matter, but that's not usually a concern for a practitioner choosing transducers or amplifiers to amplify music or speech.
Both versions of the AES standard, and the IEC standard, use RMS Voltage to determine the average power applied to the transducer during testing.
Likewise, our amplifiers are not really 'power amplifiers' - they're increasing the voltage of the waveform given to them. Hopefully, without modifying it in any significant way.
Burst tests are commonplace because they show the peak in clipped voltage output available in an amplifier without needing much current to flow. That gives us the headroom available for short transients, which are very common in all music and speech signals.
Long-term sine wave tests into a
resistive load can be considered as the 'torture' test. It shows the continuous current output capability of an amplifier, but in reality, we rarely encounter this in audio because of the reactive and inductive nature of loudspeakers. Both the transducer and the cabinet they're in.
Then there's the pink noise testing, which is where the power draw under 1/8 load value in many amplifier spec sheets comes from. It's an approximation of lightly compressed (or slightly clipped) broadband musical content at typical crest factors.
Both versions of the AES standard, and the IEC standard, use RMS Voltage to determine the average power applied to the transducer during testing.
Likewise, our amplifiers are not really 'power amplifiers' - they're increasing the voltage of the waveform given to them. Hopefully, without modifying it in any significant way.
Burst tests are commonplace because they show the peak in clipped voltage output available in an amplifier without needing much current to flow. That gives us the headroom available for short transients, which are very common in all music and speech signals.
Long-term sine wave tests into a
resistive load can be considered as the 'torture' test. It shows the continuous current output capability of an amplifier, but in reality, we rarely encounter this in audio because of the reactive and inductive nature of loudspeakers. Both the transducer and the cabinet they're in.
Then there's the pink noise testing, which is where the power draw under 1/8 load value in many amplifier spec sheets comes from. It's an approximation of lightly compressed (or slightly clipped) broadband musical content at typical crest factors.
Hi kyleneuron, first thanks for all the instructive links earlier.
Must say your quote above doesn't make sense to me.
RMS of an AC waveform, according to my laymen's understanding, represents the heating power of an equivalent DC voltage.
So for me, that has relevant meaning. And I find RMS measurements and values useful for determining sensitivity and setting thermal limiters, for drivers in multi-ways.
I measure the RMS voltage of an amp's output on each driver, using pink noise as the stimulus, with all pass-band processing in place.
A good voltmeter can average complex waveforms over time, to give an RMS time-integrated voltage. (How accurate is debatable, and a subject of how much money was spent on the meter, but for the purpose i'm describing, sufficient accuracy is within reach of a hobbyist.)
Over the same time period of the average RMS voltage measurement, an SPL Leq is taken. Which of course can easily be equated into sensitivity.
Imo, i achieve real world full-bandwidth integrated sensitivity that can be compared to manufacturers' specs, and know how much to de-rate the published sensitivity specs (almost inevitably Lol) . As well as de-rate manufacturer's specified max SPL based on mathematical power handling extrapolation.
For thermal limiters, i use either AES or IEC torture test wattage ratings, and calculate the voltage implied for the nominal impedance given.
And set that as limiting voltage, figuring my thermal limiters kick in over a period of seconds, which is way under any torture baking periods.
So, the idea of RMS voltage is useful again.
I know a lot of knowledgeable folks say amps aren't really 'power amplifiers....but i think they are both voltage amplifiers and suppliers of sufficient current to sustain the voltage amplification into given loads.
So yeah, they may be linear voltage amps, but needing the necessary current reservoir makes them power amplifiers too....and of course of various wattage ratings.
I mean, there's amps optimized for low-Z and for high-Z. And further optimizations within the low-Z range (eq 8 ohms vs 2 ohms).
I conclude amps are voltage amps, but with different power ratings for different loads, and their wattage capability into various loads ends up being separates them the most.
So do I! If you re-read my post, both myself and the standards used in audio rely heavily on RMS ratings - of the voltage for the AC waveform.
This is a good idea, and one I do myself. It’s enshrined in the ‘loudspeaker toaster test’ used by Pat Brown of SynAudCon, and how the MIV (Maximum Input Voltage) is commonly determined for loudspeaker GLL data used in room modelling too.
Yep, you’re reversing their calculation. Which is possible, because the standard is based on RMS voltage - not RMS power - the latter of which you haven’t stated a use for, yet 🙂
I have access to several power amps and looking at some speakers for use as MAINS for my band. Speakers I'm looking at are JBL JRX 225 rated at 500 -1000 watts @ 4 ohms and Peavey PV215 rated at 700 watts program, 1400 watts peak @ 4 ohms. The amps are QSC types; GX5 and RMX 2450 which both can drive 700 watts @4 ohms.
None of the JBL or Peavey ratings make sense to me so would I have a problem driving either of these speakers with these type amps? BTW, I'm not going to ever drive these amps to max.
None of the JBL or Peavey ratings make sense to me so would I have a problem driving either of these speakers with these type amps? BTW, I'm not going to ever drive these amps to max.
Brent,
The amps shouldn't be a problem.
The JBL JRX 225 roll off below 100 Hz. The low "E" on a guitar is 82Hz, low "E"on a bass is 41Hz.
If you were driving the amp to half power, 350watts, and adjusted the low end of the guitar up just +3dB, it would hit 700 watts.
Forget about flattening response below ~80 Hz, the speakers will just flap.
If you were to try to flatten it's -10dB response at 60Hz with EQ, the amp would use 10 times the power as at 100Hz.
Say the amp is cruising along at 1/8th power, 87.5watts, add +10dB =875 watts, the amp is overdriven (clipping) by 175 watts.
The GX5 "GuardRail™" circuitry, not so much.
The amps shouldn't be a problem.
The JBL JRX 225 roll off below 100 Hz. The low "E" on a guitar is 82Hz, low "E"on a bass is 41Hz.
If you were driving the amp to half power, 350watts, and adjusted the low end of the guitar up just +3dB, it would hit 700 watts.
Forget about flattening response below ~80 Hz, the speakers will just flap.
If you were to try to flatten it's -10dB response at 60Hz with EQ, the amp would use 10 times the power as at 100Hz.
Say the amp is cruising along at 1/8th power, 87.5watts, add +10dB =875 watts, the amp is overdriven (clipping) by 175 watts.
The QSC RMX 2450 clip limiter will help when you do.
The GX5 "GuardRail™" circuitry, not so much.
It's remarkable that the terms "RMS power" and "RMS watts" are still being used (and have only ever been used for audio power amplifiers), solely because they were mistakenly used by a government agency 50 years ago.
https://www.n4lcd.com/RMS.pdf
https://www.n4lcd.com/RMS.pdf