Ah, thank you.
RMS Watts?
- Thread starter Millennium
- Start date
-
- Tags
- help
irb
pfm Member
A few thoughts:
(a) How confident are you that your speakers produce 92dB from 1 watt at 1 metre? Manufacturers claims about sensitivity are often optimistic, and 92 dB/w is very high. Few speakers can actually achieve this, and the ones that can (without sacrificing bass extension) will be large. It's possible that the spec is wrong, therefore. Let's assume, for the sake of argument, that your speakers only actually produce 88 dB from 1 watt at 1 metre. That's still above average, as someone else has pointed out.
(b) The impedance spec you quote is odd - a 'nominal' figure is usually a single number, not a range. Leaving that aside, it's safe to assume that the 92 dB claimed by the manufacturer (which may only be 88dB in reality) is for an "8 ohm" watt i.e. a voltage of 2.83 V RMS. If your speakers present a 3 ohm load at some frequencies, as the spec indicates, then the amplifier will have to supply something like 2.7 watts in order to generate that output.
(c) If 3 to 8 ohms is the 'nominal impedance', the minimum impedance could be less than 3 ohms, and the power required could be greater again.
Also, speakers often present a reactive load, which means they draw more current than you'd expect, given the impedance curve. Taking both these points into account, it's possible that your speakers draw something like 5 watts in practice, in order to output 88 dB at 1 metre.
(d) Incidentally, are you really listening from 1 metre? That's very short - real near-field listening. If so, you will need less power than most.
(e) Crest factor is a big issue. As others have said, some music might have peaks 20 dB above the mean level. However rock/pop may have a mean level only 3 or 4 dB below the max peak level. Assuming you want to listen to a range of music, you certainly need to allow something like 10 dB. Your 5 watt amplifier requirement goes up to 50 watts. You might get away with a bit less, I suppose. You're presumably listening in stereo, not mono, for one thing, and (depending on your amp) you might be able to tolerate a little clipping without much audible loss. Will 15 watts be enough? Hard to say, but I doubt it. (Having said that, if your amp is solid state, and is specified at 15 watts into 8 ohm, and has a stiff enough power supply that it can cope with sub 3 ohm loads, then it will probably be OK.)
However, if you want to be able to reach peak levels around 110 dB, you can add another 10 dB (or more) - so your amplifier requirement goes up to 500 watts.
(And if you decide, one day, to take the system into a bigger room where you listen from 3 metres away, you might need over 1000 watts to get those 110 dB peaks. Your voice coils might release some 'magic smoke' before then, though.)
Last edited:
John Phillips
pfm Member
On the latter point: yes, I think so.
I have only suspicions and no hard evidence, but I would be wary of any amplifier that on test looks like it measures worse at 1W power output than at 10W power output or more. I have seen that at times in reviews with technical measurement and asked why. Class A and high-bias Class AB should not, I think.
However, I still take the engineer's view and dimension my system so that my anticipated SPL peaks in my room are handled without clipping when (on occasions) I am listening loud to very dynamic music. And that involves the more difficult matter of loudspeakers having enough SPL capability too. But low amplifier power output is undoubtedly where most listening takes place.
Normally the measurement is THD+noise. In low distortion designs (with very low or non-existant crossover distortion) it is often the noise element that dominates at low power levels. This increases as a percentage as you drop between 10W and 1W and gives the impression of worse performance.
John Phillips
pfm Member
Indeed it was and @martin clark's expansion of the initialism represents the fundamental truth.
I recall seeing the progression: RMS power output -> music power output -> peak music power output. As the marketing department deviated from the relative reality of the starting point, as far they could get away with, in the chase to convince unwary customers.
Last edited:
I tend to think that if the trade had been more self disciplined a metric such as the PMPO to RMS output could have been a useful concept for replay of music with a given DR.
Getting that 25dB extra impact from a drum, guitar or piano must be a design challenge.
I don't think i get it from my Crown amp but the Brystons sound sound as if they can.
Getting that 25dB extra impact from a drum, guitar or piano must be a design challenge.
I don't think i get it from my Crown amp but the Brystons sound sound as if they can.
irb
pfm Member
I'm not so sure. There's a confusion that's popped its head up a few times in this thread: RMS means root mean square, which is a way of describing the magnitude of a waveform - even a steady tone. After all, a wave is (by definition) constantly changing, so how do you measure its amplitude? At the highest point, the peak? But almost all of the wave has a lower amplitude. The average? Strictly speaking, that's zero. RMS is the agreed practical measure, which represents, to all intents and purposes, the average value of the waveform's amplitude.
Now all of that applies to all signals, even to constant tones, such as a 1kHz sine wave.
The difference between the peak and average levels of a music signal is something else entirely. 'Crest factor' is one commonly used term for this - the difference between the highest peak in a piece of music and the average level.
MPO and PMPO were nothing to do with crest factor. They were simply a marketing strategy - changing the way you quote figures to benefit your own product, in comparison with other manufacturers. (A different version of this slipped into the market with T-amps, the makers of which often quoted power output into 4 ohms, and 10% thd, allowing them to claim 15 watts output when they'd otherwise only have been able to claim 6 watts.)
Is there a way to put a number on an amplifier's 'dynamic' ability rather than its 'static' output - apart from quoting output capability in watts RMS? Probably not.
Darth Vader
From the Dark Side
The interaction between an amp and speaker is way more complex than people understand and in most case people don't have the knowledge nor understanding hence all this confusion. I and loads of others have tried many times to help over the years but ultimately we failed to get the message across so we attempt simpler ways but yhey all have their drawbacks.
To answer your question about measuring true RMS Voltage of a nonsinusoidal waveforms of course this can be measured. I had until I recently sold it a lovely HP RMS Voltmeter 3400A. It had a frequency range of 10Hz to 10MHz and a Voltage range from 100mV to 300V. Although I sold the meter I did keep the manual to drool over. Page after page after page of circuit diagrams. This is a serious nit of kit!
Sine waves are not just waves. They are mathematically derived from a rotating circle. Once someone has fully grasped that significance it automatically follows that our electricity follows a Sine curve because part of the dynamo rotates in a circle. Next any waveform can be broken down into its composite Sine components - the maths gets a bit complicated!
Cheers,
DV
To answer your question about measuring true RMS Voltage of a nonsinusoidal waveforms of course this can be measured. I had until I recently sold it a lovely HP RMS Voltmeter 3400A. It had a frequency range of 10Hz to 10MHz and a Voltage range from 100mV to 300V. Although I sold the meter I did keep the manual to drool over. Page after page after page of circuit diagrams. This is a serious nit of kit!
Sine waves are not just waves. They are mathematically derived from a rotating circle. Once someone has fully grasped that significance it automatically follows that our electricity follows a Sine curve because part of the dynamo rotates in a circle. Next any waveform can be broken down into its composite Sine components - the maths gets a bit complicated!
Cheers,
DV
This isn't a special quality of sine, but it has the benefit that it makes other calculations easier because of Euler's equation. I know, a nitpick.
irb
pfm Member
Apologies if I'm missing something obvious, but who are you replying to?
Tony L
Administrator
I did much the same trying to figure out if I was anywhere close to emptying my 10 Watt Leak Stereo 10 into the JR149s:
I came to the conclusion I was using a little more than I thought I was on transient peaks of very dynamic jazz recordings (e.g. ECM), but that I had sufficient headroom for my requirements. Basically 6-8 Watt peaks were louder than I’d ever want. The Fluke has a peak hold so is useful for this. It certainly captures higher numbers than I can see.
PS The idea of using Avometers as VU meters is excellent. I shall certainly try that!
gavreid
Pretty Words...
Those Hameg scopes were very nice too - not a tektronix but not to be sniffed at either. I used a few in the lab.
martin clark
pinko bodger
It's what I use the lesser of the two I bought from you for: cleaned up it fits a treat. Totally passive, high-Z input, visible in comfort. As I said - rarely approaches 2vac. I also know that getting to end of scale at 3vrms -way louder than I ever want - is about the point the ESL989s are producing music audible to (tho well short of annoying) to neighbours above and below on quiet evenings - we spent sometime happily testing this together, they're a great bunch - so that's a nice sense check too.
hurried crappy pic:
Tony L
Administrator
That’s the one MyHermes threw right over the back wall into the concrete yard despite my being in and listening for the doorbell! Thankfully the seller packed it very well indeed so it survived without even a scratch. It’s quite a late one for a CRT scope, so fairly light, which no doubt helped here!
gavreid
Pretty Words...
I remember them being light certainly
Arkless Electronics
Trade: Amp design and repairs.
I recall back in the 90's there being powered speakers for PC's with a 9V 0.5A wall wart powering them and claiming things like 300W PMPO!! A total pack of lies obviously as the PSU can only provide 4.5W. They were usually a genuine 0.5 - 1W.
Peak power or dynamic headroom is a double edged sword. The stiffer the supply and the greater the ability of the amplifier to deliver current the less will be any extra peak power available. With an amplifier that uses fully regulated supplies, such as Naim NAP 250 or Quad 303 etc, there is no extra peak power available.
It's mainly a thing with cheap amplifiers with weedy power supplies. A small mains transformer charges up the smoothing caps to say +55 and -55 Volts. This is enough for about 120W into 8R but remember we're talking about a small under rated transformer here and if you actually measure the RMS continuous output of this amp then it's likely the transformer can't sustain this voltage and drops to say +40 and -40V under load... enough for about 70W output.
Such an amp would perform badly in the 4R max RMS power stakes of course and may not give much more into 4R than it does into 8R.
Now of course the max peak power of such an amp will be given when the amp is at it's lowest load ie if you are only averaging 1W than you may be able to get 120W for a single kick drum beat. That would require a hellish dynamic range recording and crest factor which will never happen in practice.
Peak power or dynamic headroom is a double edged sword. The stiffer the supply and the greater the ability of the amplifier to deliver current the less will be any extra peak power available. With an amplifier that uses fully regulated supplies, such as Naim NAP 250 or Quad 303 etc, there is no extra peak power available.
It's mainly a thing with cheap amplifiers with weedy power supplies. A small mains transformer charges up the smoothing caps to say +55 and -55 Volts. This is enough for about 120W into 8R but remember we're talking about a small under rated transformer here and if you actually measure the RMS continuous output of this amp then it's likely the transformer can't sustain this voltage and drops to say +40 and -40V under load... enough for about 70W output.
Such an amp would perform badly in the 4R max RMS power stakes of course and may not give much more into 4R than it does into 8R.
Now of course the max peak power of such an amp will be given when the amp is at it's lowest load ie if you are only averaging 1W than you may be able to get 120W for a single kick drum beat. That would require a hellish dynamic range recording and crest factor which will never happen in practice.
Power, a great and fascinating subject.
But we must include impedance changes and phase changes in the load and air pressure, furniture even the bodies in the room will all effect the power require for the amp to control and drive the load, Room, Air, Bodies, Vertical Motor (speaker) etc.
I did some work for Audium a few years ago and the French came back with the new chip design and on average driving a rubbish speaker it only need a max of 100mW but the peaks went to 15W, this was designed into a Class D/T audio chip, now used in lots of phones, it sounded crap.
A complex concept to apply to a mathematically solution without the perfect transducer. Even on XL from Microflop. Maths Cad can get close but no there yet.
Oh and not forgetting the impedance of the amps output or damping factor ratio will for best control need to be very high, transformer coupled amps cant do this very well, unless Anode Current and voltage is high and then the transformer ratio will be better.
So use a TH250 at 10,000V at 100mA mmm nice.
But we must include impedance changes and phase changes in the load and air pressure, furniture even the bodies in the room will all effect the power require for the amp to control and drive the load, Room, Air, Bodies, Vertical Motor (speaker) etc.
I did some work for Audium a few years ago and the French came back with the new chip design and on average driving a rubbish speaker it only need a max of 100mW but the peaks went to 15W, this was designed into a Class D/T audio chip, now used in lots of phones, it sounded crap.
A complex concept to apply to a mathematically solution without the perfect transducer. Even on XL from Microflop. Maths Cad can get close but no there yet.
Oh and not forgetting the impedance of the amps output or damping factor ratio will for best control need to be very high, transformer coupled amps cant do this very well, unless Anode Current and voltage is high and then the transformer ratio will be better.
So use a TH250 at 10,000V at 100mA mmm nice.
Jim Audiomisc
pfm Member
The problem is there are various ways to specify it, and that the peak/mean behaviour of music also varies from case to case. So no single number can fully represent it.
That said, I'd say that a peak level capability rating with some specific details might well help to clarify some of the lack of awareness at least.
In some designs noting the max rms power *one channel driven* and seeing if that is higher that both channels being driven can be a pointer. Not ideal, but at least finds out when the max is clamped by the PSU or some similar factor so the real peak/rms for sinewave is just x2, no higher. But if it is more than x2 then it probably won't repesent the 'musical peaks' capability potentially being bigger.
Sadly, when no-one measures it, the designer may tend not the bother.