advertisement


How many amplifier watts am I using?

Nigel

Nigel

pfm Member
Is it simply a case of playing some music at the desired level, measuring the AC voltage across the amp's loudspeaker outputs, followed by this voltage squared, then dividing by the resistance of the loudspeakers (8 ohms)?
 
Yes. Most multimeters, whether digital or analogue, will not respond fast enough to show the peaks though.
It's also unlikely that your speakers will be exactly 8 Ohms and at any given frequency could be as high as say 12 or as low as say 4....
 
So in this video Nelson Pass seems to have hooked up an oscilloscope to his speaker which tells him how many watts the speaker is using, or have I completely misunderstood? The relevant bit is at about 7'10

 
Sorry can't be arsed to watch it but probably yes. I have an oscilloscope in my listening room that I can connect to any amplifier I'm using and spot clipping or instability. On the bench you usually use an oscilloscope to find the exact clipping point using a sine wave signal (it doesn't move about etc as music would!) and then measure it with an accurate AC voltmeter. Many "standard" multimeters are way off at anything much over 400Hz or so!
You can also take a reading straight from the oscilloscope screen of course.
 
Nigel said:
Is it simply a case of playing some music at the desired level, measuring the AC voltage across the amp's loudspeaker outputs, followed by this voltage squared, then dividing by the resistance of the loudspeakers (8 ohms)?
Click to expand...
It depends what you actually mean by 'amplifier Watts'. If you mean how much power the amp is generating then no this method will not be accurate. A speaker is a reactive load and unlike a resistor the current and voltage can go out of phase known as the speakers phase angle. This can be the destroyer of lesser build amps especially BJTs. As the phase angle between the voltage and current increases more power is dissipated by the output transistors as heat rather than driving the speaker load. This also varies with frequency. For example at 45 degrees 80% of the amplifier output is dissipated as heat in the output devices and only 20% actually drives the load.

So you have several moving parameters speaker impedance (changes with frequency and also with rising temperature of the voice coil) and phase angle (changes with frequency). Then has already been mentioned fast transients that need a lot of power but for a very small time duration. These give rise to attack and timbre.

Most power will be required at low and mid frequencies so for a given apparent loudness you'll need a lot more power there than at higher frequencies.

Lots of moving targets.

Cheers,

DV
 
I’ll give this a go at some point. I have a Fluke multimeter with a min/max function that hopefully might catch peaks, or at least close too.
 
I recall an article "explaining" that the reason transistor amplifiers don't sound as good as valve amplifiers is because they are often clipping, and valve amplifiers have a softer limit, and stating that a check with an oscilloscope will demonstrate clipping. But when I checked my amplifier (150watt/ch) typical output level (volts, and converting to power using above formula) into my Harbeth HL5s at normal listening levels, I found peak levels only rarely exceeded a few watts.
 
gustavm said:
I recall an article "explaining" that the reason transistor amplifiers don't sound as good as valve amplifiers is because they are often clipping, and valve amplifiers have a softer limit, and stating that a check with an oscilloscope will demonstrate clipping. But when I checked my amplifier (150watt/ch) typical output level (volts, and converting to power using above formula) into my Harbeth HL5s at normal listening levels, I found peak levels only rarely exceeded a few watts.
Click to expand...

We're gonna need a bigger kettle for all them fish!:)
 
I found this calculator online which will help folk like me who bunked off maths at school. I’ve just hooked my meter (a Fluke 87V) up to my Stereo 20 and used the ‘min/max’ function whilst playing my new Spendor S3/5Rs (see JR149 thread) at about 76-78db average on a George Duke album (quite weighty and dynamic) and it seems good and fast as you can hear it beep whenever it finds a new higher transient, and it certainly correlates to what I’m hearing. Max reading I got was 3.56V, the Leak is on its 8 Ohm tap, and the speakers are nominally 8 Ohm, so the result given is 1.58 Watts. Actually very good to know as I’ve always been worried as to whether I’m emptying the Leak into my speakers! It is after all only a ten Watt amp.

PS Just googled and the Fluke’s ‘min/max’ function can capture 250µs, which I guess should be plenty fast enough.
 
Last edited:
Could you sit 1 metre away from a speaker with a decibel meter phone app and adjust the volume until the meter reads the same as your speaker sensitivity rating. At that point you'll be using 1 Watt.

Or have I massively misunderstood something?
 
Don't know what those amps were measuring, but there was no way that was 500 watts average, it must have been peak hold or something.
 
deek said:
Could you sit 1 metre away from a speaker with a decibel meter phone app and adjust the volume until the meter reads the same as your speaker sensitivity rating. At that point you'll be using 1 Watt.

Or have I massively misunderstood something?
Click to expand...

Bairnstorm said:
I have a sound meter on my phone which shows the db. Don't know how accurate it is.
Click to expand...

No! Those methods are worse than useless.
 
Thorn said:
Is there a reason for wanting to find out?
Click to expand...

It can be very educating for those who one often reads posts from along the lines of "I have a 50WPC amp and want more control and drive to the sound. Should I get a 100WPC amp?" The short answer is no, power has virtually nothing to do with such things.
When you see that you are in fact usually only using say between 0.1 and 3W anyway then all becomes a little clearer.

So why do we have such powerful amplifiers often?

Ah...well, quite counter-intuitively to what I said above, the human ear is actually spectacularly low fi in many ways and perceives a doubling in volume from a tenfold increase in power!
So, whilst it can be a shock as to just how loud 1W actually is, and for many it can be louder than we can often get away with due to neighbours or family etc, if we had a few mates round and neighbours are out so we decide to play it twice as loud as normal this then takes 10W... and if a few too many beers go down and you want it twice as loud again you need 100W! Hence the common "speakers were blown up at party" scenario...

Obviously speaker efficiency and room size makes a massive difference as well! The volume Tony had from his 85dB/W Spendors with 1.58W could likely have been achieved with more like 0.2W with his much more efficient Tannoys....
 
You must log in or register to reply here.


advertisement


Back
Top