As I do know, very low frequencies require a lot of wattage. Very high ones, not so much at all. But, what is the "wattage curve" if it exists and is there an equation?
My usage would be to determine, for example, how much amplifier wattage I would be saving if my active crossover design did not give a driver/speaker full range but rather just 200Hz and above. Or, any range in between certain frequencies for active bandpass crossovers. Or, instead of feeding 0Hz to xxHz, I cut that to 40Hz to xxHz sent to a subwoofer.
The big question and Why?: What is the wattage "saved" that can then be used to provide power to the frequencies that the driver/speaker can actually do very well?
I'm just curious to know if an equation exists. I am guessing that the power requirement increases exponentially as the frequency goes down? And, does this involve using the Fletcher-Munson curve also? Then, it becomes a subjective and not an objective calculation...arrrgggghhh.
One important example that comes to mind is that my mids/tweets in my vehicle are actively sent only 160Hz and above (-12dB/oct cutoff). The amplifier is 150Wrms/channel and each mid/tweet "speaker" is 125Wrms/channel. So, by not sending anything much under 160Hz to those "speakers," am I maybe overdriving them in the 160Hz to 20,000Hz range (grin)?
Thanks!
My usage would be to determine, for example, how much amplifier wattage I would be saving if my active crossover design did not give a driver/speaker full range but rather just 200Hz and above. Or, any range in between certain frequencies for active bandpass crossovers. Or, instead of feeding 0Hz to xxHz, I cut that to 40Hz to xxHz sent to a subwoofer.
The big question and Why?: What is the wattage "saved" that can then be used to provide power to the frequencies that the driver/speaker can actually do very well?
I'm just curious to know if an equation exists. I am guessing that the power requirement increases exponentially as the frequency goes down? And, does this involve using the Fletcher-Munson curve also? Then, it becomes a subjective and not an objective calculation...arrrgggghhh.
One important example that comes to mind is that my mids/tweets in my vehicle are actively sent only 160Hz and above (-12dB/oct cutoff). The amplifier is 150Wrms/channel and each mid/tweet "speaker" is 125Wrms/channel. So, by not sending anything much under 160Hz to those "speakers," am I maybe overdriving them in the 160Hz to 20,000Hz range (grin)?
Thanks!
It is not an simple answer.
This actually depends on what type of music you listed to.
Rock ? Classical ? Techno ? Other ?
Each type generally has a power vs frequency spectrum that is not the same.
I've seen Bi-amp / Tri-amp setup that call for 10W tweet, 20W Mid and 40W sub. (But that was a AF PWR IC mfg trying to sell their line)
What you save is not having to pay for "full power" at all frequencies when you may not need it.
Take a look at Rod Elliott's page:
ESP - Frequency, Amplitude and dB
This actually depends on what type of music you listed to.
Rock ? Classical ? Techno ? Other ?
Each type generally has a power vs frequency spectrum that is not the same.
I've seen Bi-amp / Tri-amp setup that call for 10W tweet, 20W Mid and 40W sub. (But that was a AF PWR IC mfg trying to sell their line)
What you save is not having to pay for "full power" at all frequencies when you may not need it.
Take a look at Rod Elliott's page:
ESP - Frequency, Amplitude and dB
see the power distribution chart here: Why Do Tweeters Blow When Amplifiers Distort?
the catch with this idea is that in an ideal world, the HF amp need to be able put out the same peak power as the LF amp, in the real world you can get away with a smalle one
the catch with this idea is that in an ideal world, the HF amp need to be able put out the same peak power as the LF amp, in the real world you can get away with a smalle one
Thanks DUG and Pete. I'll check out the links and learn all that I can.
As for music styles, mine runs the entire gamut.
I have found extremely low end music by a band named Front 242 where I can actually see the woofer cone moving in and out at around 3Hz! But, that MUST have been a mixing or mastering error. It is fun to watch the woofer cone move like that (and it isn't full excursion at loud volume either, just a slow fluctuation). The exact track is called "Soul Manager" from the 1991 disc Tyranny For You. Disc markings include EK 46998 and DIDP 073112 by Sony Music Entertainment. Pretty hard core music, I admit.
Then, after listening to something like that, I'll go to classical with a composition that includes cannon shots. Then, something that is acoustical guitar like Abba. Whatever comes to mind.
But, I do understand that it does depend on the type of music. I was hoping for an equation that took driver params into account and determined the wattage required to produce equal sound levels at each frequency. Oh well. Nothing in the art and engineering of sound systems works that simply.
or infrasonics...i have quite a few DVDs that have silly levels of sub 10hz rubbish for 'effect', all it does is steal power from the amp, and unload the woofers in a similar way to turntable rumble. I think i have a couple of cds which do the same. Bad mastering or accidental recording of seismic events? Who knows? It shouldnt be there on a cd in my opinion. A lot of synth based music has the same problem, analogue noise generator synth sections used for 'snare' sounds being the worst. Ive seen more excursion from these synth snares than the kick drum sounds. Clearly the electronic music camp has yet to discover rumble filters.
The power requirements are not only different for different music, but for different cabinet frequency response.
For instance, a sealed cabinet with a response 10 dB down at 30 Hz could be equalized flat using 10 times the power at 30 Hz than at 60 Hz, while another design that is flat to 30 and the same sensitivity at 60 Hz would use 1/10 the power at 30 Hz.
If you look at a musical spectrum that is 10 dB more bass heavy (say the difference between folk and hip-hop) and different cabinet designs, the difference could be 100 times, the folk music with an efficient cabinet would use 1 watt while the hip-hop with a sealed cabinet uses 100 watts.
Hi,
Its a long running debate regarding average power to the frequency
ranges and the crest factor of each range. Certainly bass gets a lot
more average power than the mid, the mid more than the treble.
Your really stuck with rules of thumb. Its around 50:50 for a 3 way
with active bass to mid and passive mid to treble, active mid to treble
around 35:15 to 30:20, 2 ways around 80:20 though often less.
Active EQ and different driver efficiencies can really skew the numbers,
and of course they vary with the x/o points used, no simple answer.
Also note going active increases peak handling compared to average
handling, an active 3 way with a total of 100W (say 50+30+20)
can produce wideband peaks equivalent to a 270W amplifier.
rgds, sreten.
Its a long running debate regarding average power to the frequency
ranges and the crest factor of each range. Certainly bass gets a lot
more average power than the mid, the mid more than the treble.
Your really stuck with rules of thumb. Its around 50:50 for a 3 way
with active bass to mid and passive mid to treble, active mid to treble
around 35:15 to 30:20, 2 ways around 80:20 though often less.
Active EQ and different driver efficiencies can really skew the numbers,
and of course they vary with the x/o points used, no simple answer.
Also note going active increases peak handling compared to average
handling, an active 3 way with a total of 100W (say 50+30+20)
can produce wideband peaks equivalent to a 270W amplifier.
rgds, sreten.
In active multi-way you can save power on midrange and tweeter due to their higher sensitivity.
Rod Elliot has a good example on his website:
"Assume we have a loudspeaker rated at 90dB/W/m (a softspeaker?) versus another rated at a much more respectable 100dB/W/m. With one Watt of electrical energy applied, one will have 10 times (10dB) the SPL of the other. While this is insignificant if we are happy with 90dB SPL, if we try to obtain 110dB SPL at one metre, the efficient driver will do this with only 10W, while the inefficient driver needs 100W. Another 10dB makes that 100W vs. 1000W - anyone want to guess which speaker will last longer before the voicecoil melts?"
Power Handling Vs. Efficiency
Rod Elliot has a good example on his website:
"Assume we have a loudspeaker rated at 90dB/W/m (a softspeaker?) versus another rated at a much more respectable 100dB/W/m. With one Watt of electrical energy applied, one will have 10 times (10dB) the SPL of the other. While this is insignificant if we are happy with 90dB SPL, if we try to obtain 110dB SPL at one metre, the efficient driver will do this with only 10W, while the inefficient driver needs 100W. Another 10dB makes that 100W vs. 1000W - anyone want to guess which speaker will last longer before the voicecoil melts?"
Power Handling Vs. Efficiency
I don't agree with either of your first two statements.
That probably puts be at odds with 95% of the Members.
I claim that all the drivers must be capable of similar maximum SPL and that the power of each of the frequency band amplifiers should be chosen to achieve that same maximum SPL in their respective frequency bands.
If you start with that as a hypothesis, you may end up with a balanced sounding active speaker system that does not grossly clip on any particular frequency band because you have selected a "too small" amplifier based on "average" power levels across the frequency bands.
But my suggested starting point will not appeal to many.
This is a complex subject and not so good to generalize too much about, but some of your starting assumptions may be a little off.
First off, power requirements doesn't necessarily go up as frequency goes down. Within the range of a driver you can have flat response and not need any more power at, say 50 Hz as you need at 500, for the same output. If you want to go below the corner frequency of a woofer (below the system resonance) then sensitivity will roll off at a considerable rate and you would need lots more power to counteract the rolloff. If instead you wanted to extend the woofer response lower (to always be above the corner frequency) then you will have to drop efficiency considerably (such as heavy mass loading the woofer), but still, if you have flat response in the woofers range then power required will not change for frequencies in that range.
If you are talking about the power spectrum of music then search around and you will find a number of studies and curves. For the long term average the power requirements of music peaks in the midrange (around 200 Hz) and falls off either side. That means that you need to generate more acoustical power in the mids than at higher or lower frequencies. Classical music tends to have a more noticeable peak and jazz and pop less of a falloff away from the midrange, but they all have greater output requirements in the midrange. These are the long term averages, and the short term peaks (peak spectrum rather than long term average spectru) curves tend to be a little flatter, but still don't rise for the bass frequencies.
Now if you are talking natural driver sensitivity then midrange and tweeter sensitivities can be significantly higher than LF sensitivities. So, yes, you may need bigger amplifiers for the bass section of a triamped speaker than for midrange and treble. Again, this is a tradeoff between woofer sensitivity and extension. It is very easy to design a sensitive woofer but the bass extension will be inadequate. Achieving low bass extension requires low efficiency or large system volume, so you may see woofer sensitivities around 85 to 88 db (1m with 2.83 V input) while tweeters may more frequently be around 88 to 92.
Some of the above comments about true power (I x V(cos theta)) are correct but we tend to talk of power loosely as V squared over the nominal impedance. This may be referred to as nominal power or apparent power. If true power is considered efficiency will actually peak at resonance since the impedance curve goes high and current through the driver drops.
David S,
First off, power requirements doesn't necessarily go up as frequency goes down. Within the range of a driver you can have flat response and not need any more power at, say 50 Hz as you need at 500, for the same output. If you want to go below the corner frequency of a woofer (below the system resonance) then sensitivity will roll off at a considerable rate and you would need lots more power to counteract the rolloff. If instead you wanted to extend the woofer response lower (to always be above the corner frequency) then you will have to drop efficiency considerably (such as heavy mass loading the woofer), but still, if you have flat response in the woofers range then power required will not change for frequencies in that range.
If you are talking about the power spectrum of music then search around and you will find a number of studies and curves. For the long term average the power requirements of music peaks in the midrange (around 200 Hz) and falls off either side. That means that you need to generate more acoustical power in the mids than at higher or lower frequencies. Classical music tends to have a more noticeable peak and jazz and pop less of a falloff away from the midrange, but they all have greater output requirements in the midrange. These are the long term averages, and the short term peaks (peak spectrum rather than long term average spectru) curves tend to be a little flatter, but still don't rise for the bass frequencies.
Now if you are talking natural driver sensitivity then midrange and tweeter sensitivities can be significantly higher than LF sensitivities. So, yes, you may need bigger amplifiers for the bass section of a triamped speaker than for midrange and treble. Again, this is a tradeoff between woofer sensitivity and extension. It is very easy to design a sensitive woofer but the bass extension will be inadequate. Achieving low bass extension requires low efficiency or large system volume, so you may see woofer sensitivities around 85 to 88 db (1m with 2.83 V input) while tweeters may more frequently be around 88 to 92.
Some of the above comments about true power (I x V(cos theta)) are correct but we tend to talk of power loosely as V squared over the nominal impedance. This may be referred to as nominal power or apparent power. If true power is considered efficiency will actually peak at resonance since the impedance curve goes high and current through the driver drops.
David S,
Hi,
That is because you are wrong and always have been. The idea that
max SPL at a single frequency is the issue irrespective of the actual
bandwidth the driver is handling is just very, very wrong.
You suggest that a 3 way x/o at 3KHz and 10KHz needs the same
amplifiers as one crossed over at 300Hz and 3KHz, that is wrong.
You keep reiterating a standpoint / opinion in various threads you
cannot justify just because its a (wrong) assumption you once
made and used (or still use) to build something (or things).
We have been round the houses before with this, and you keep
refusing to accept the accepted theory (and measured practice)
about the subject, that is clearly right.
"I claim that all the drivers must be capable of similar maximum SPL and
that the power of each of the frequency band amplifiers should be chosen
to achieve that same maximum SPL in their respective frequency bands."
Similar maximum SPL in each band at multiple frequencies requires
amplifier power to be proportional to each bands bandwidth, QED.
rgds, sreten.
proportional to each bands bandwith
Yes, exactly what David said in post #13. The midrange often receives a great share of the musical power, but would take no more to reproduce than the bass, given the same speaker sensitivity (and BW, as sreten points out).
The problem being that bass drivers and cabinets are often less efficient than midranges and tweeters.
The problem being that bass drivers and cabinets are often less efficient than midranges and tweeters.
Its just not that simple.
Lets say I have a 2 way sytem with a passive crossover and a 100 Watt amplifier. I take long term measurements and find that the average power distribution is about 80% to the woofer and 20% to the tweeter. If I biamp with an 80 Watt woofer amp and a 20 Watt tweeter amp have I achieved equivalence (lets assume similar driver sensitivity and little attenuation in the previous passive crosover)?
Unfortunately not. With my passive system there might have been instances where full peaks fell in the tweeters range. The full 100 Watts would be available for these peaks.
When going from passive to active we can calculate the average power distribution and also the peak power distribution. They will give different answers. The only way the active system is guaranteed to surpass the passive system, for all input signals, is when each section has the full power the original passive system had.
David S
I've often thought about this too, but in looking around online and the answers generally presented seem to dance around the topic without ever really being solid.
Let's say you generate a noise signal that's equal in level across all frequencies from 20 cycles to 20,000 cycles. Let's also assume all the drivers have perfectly even impedances and sensitivities. You look at your RTA and you've got a perfectly straight line going from 20-20,000. You put a voltmeter on your subwoofer driver and increase the level until it's getting 1 volt.
Now, move the meter to the tweeter driver. Does it also read 1 volt?
That seems like what the question really is once you strip away the variables. If the answer is yes, then there is no chart.
Let's say you generate a noise signal that's equal in level across all frequencies from 20 cycles to 20,000 cycles. Let's also assume all the drivers have perfectly even impedances and sensitivities. You look at your RTA and you've got a perfectly straight line going from 20-20,000. You put a voltmeter on your subwoofer driver and increase the level until it's getting 1 volt.
Now, move the meter to the tweeter driver. Does it also read 1 volt?
That seems like what the question really is once you strip away the variables. If the answer is yes, then there is no chart.
Hi,
I never said its that simple, but that the simple principle cannot be ignored.
The above is overengineering and ignores a basic principle of active speakers,
they can handle wideband transients far better than passive speakers, at
the cost of lower narrowband peak levels, for the same total amplifier power.
e.g. A 50W+50W FAST two way crossed over at around 300Hz will on
most quality material equal the dynamics of a 200W amplifier, though
clearly in each band it can't beat a 100W amplifier.
For an active 3 way with properly chosen amplifiers, say each band
is equivalent to a 100W amplifier, wideband transients go up to an
equivalent of a 900W amplifier, but narrowband it is still 100W.
One could argue you'd be better off with a 300W amplifier and passive x/o.
Its a self defeating argument regarding 300W passive versus 3x100W active.
The statistical nature of music signals indicates the active system would
be superior most of the time and nearly all the time across an LP side.
Intelligent choice of x/o points and amplifier power is needed. That is
why I often advocated a stereo amplifier with active bass/mid around
300Hz, active EQ and a simple passive mid/treble x/o for active.
If you have a say 35W per channel valve amplifier for mid/treble the
argument does not go you need a 30W per channel amplifier for the
bass, it does go 70W to 100W for the bass will be a better choice.
(Might as well have the extra bass handling for bass heavy stuff).
Its not simple regarding the most cost effective way of doing things.
For a typical active 3 way 170W/90W/40W will better 3x100W.
rgds, sreten.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.