Behaviour of amplifiers, power & transparency

[hwlr]

Hello everyone,

I've discovered this superb website a few months ago and I will start by saying thank you for what I've already learnt, how to equalize my headphones being the most important thing that comes to my mind. Now, the time has come for my first post...

Please correct me if I am wrong: in its purest principle, the role of an amplifier is to take a voltage at its input, to increase it, and to throw this increased voltage at its output. From this statement can be deduced that:
1/ Listening volume has to do with voltage, not with current or power.
2/ The power (or current as both are linked in the P = R I² formula) is here only to ensure the good voltage at the output of the amplifier. I mean that any load (speaker or headphone in our case) will demand current according to Ohm's law ( U = R I ), so that for a voltage U to be maintained across a speaker of impedance R, the amplifier needs to supply a current I.
3/ If the amplifier is not powerful enough, that is to say if it can't provide enough current, the voltage is capped and clipping occurs.

Am I still correct? If so, is it true to say that as long as there is no clipping during use, any additional usable power from your amplifier is useless? In other words, the transparency of an amplifier does not depend on its power rating at all as long as there is enough of it to prevent clipping, does it? The transparency only depends on figures like SINAD for example?

PS: apologies for the imperfect and maybe more complicated than needed language

 

[voodooless]

Your spot on. Why buy a Porsche if you only ever drive in the city.. shure it looks great, but does not really do a better job than a Toyota.

 

[hwlr]

Great, thank you very much for your fast answer ! I will completely ignore power then, I never heard clipping (at least, due to power limitation) despite using my headphones from my DAP, so I really don't need a lot of power.

 

[faheem]

I hear clipping on the Ifi Xcan slightly before the O2 Amplifier ( Hifiman Ananda and Empire Ears Valkyrie 3ohm 96db/mw rated iem). Is it safe to say that means I am current limited. As the Xcan is rated at 5.7V / 1,000 mW (@ 32 Ohm) ?

 

[hwlr]

What do you mean by "slightly before the amplifier"?

 

[faheem]

What do you mean by "slightly before the amplifier"?

ah. sorry. meant volume of the Objective 2 is a little higher before it starts clipping / distorting, compared to the Ifi Xcan.

 

[hwlr]

I think you are correct then, if my understanding of how an amplifier works is correct.

 

[faheem]

I think you are correct then, if my understanding of how an amplifier works is correct.

Thank you.

 

[charleski]

The most comprehensive guide to this is NwAvGuy's post from 10 years ago: More Power?
As long as an amp can produce the power you need without clipping (which depends on the sensitivity of your speakers/'phones) you only need to look at the SINAD and output impedance (which may vary with frequency).

 

[faheem]

The most comprehensive guide to this is NwAvGuy's post from 10 years ago: More Power?
As long as an amp can produce the power you need without clipping (which depends on the sensitivity of your speakers/'phones) you only need to look at the SINAD and output impedance (which may vary with frequency).

Thank you, that was an excellent read. My curiosity was based on the specification provided by ifi and jds labs respectively and the science behind it, volume levels are more than adequate.

 

[boXem]

3/ If the amplifier is not powerful enough, that is to say if it can't provide enough current, the voltage is capped and clipping occurs.

Sorry for the late reply, just saw this thread.
To get power you need both voltage and current. If the load impedance is low, the current will be the limiting factor exactly as you described. But when load impedance is high, the limiting factor will be the maximum voltage usable.
This is why Amir measures headphone amps in both 32 (current) and 600 Ohm (voltage).

 

[Willem]

So you do not need more power than you need, but you may need more than you think, particularly for the lower frequencies. You must realize that for dynamic peaks you will need a lot of power, often in the hundreds of watts.The reason is that the relation between power and sound level is not linear but logarithmic. So you need a lot more power for a little bit more sound. The good news is that these days power is cheap.

 

[voodooless]

Crest factor in music is about 10 to 20 dB. So for an average music listening level of 90 dB, you'd have peeks of 100 to 110 dB. So that means that you need 10 to 100 times the power for those peeks compared to the average.

However, note that to our ears, 10dB more only sounds about twice as loud, but need 10x the power. The higher the amp power, the lower the actual returns will be because of this.

 

[hwlr]

Thank you for all your answers!

 

[txbdan]

I disagree with #1. Power is all that matters. A speaker is a transducer that converts electrical power to acoustic power. More power, more volume. Of course electrical power is the product of voltage and current.

A voice coil is motivated by current via Lorentz Force equations, nothing else. See: https://www.linearmotiontips.com/voice-coil-actuator-basics/

Voltage is simply the potential that motivates current to flow through an impedance. You only need enough voltage to supply the necessary current into the load.

 

[KSTR]

A "laboratory grade" audio amplifier would have two different power-related limits:
- voltage clipping (amp cannot go beyond its supply rails)
- current clipping (amp must limit the current to not overload or even damage)
Those would be completely independent from each other. For a class-A/B this puts extreme demands on the circuit design and sheer brute power reserves. In class-D this can be done much better.
(Long-term thermal shutdown shall be neglected here)

A speaker is a complex load and therefore the above limits should be established to best effort.
In affordable consumer units, corners must be cut, the limits are less well defined and influence each other. Still, any competent design will not have any "real vs. rated" power issues with the majority of speakers out there.

Bottom line:
Once you run into either form of clipping, the amp is not powerful enough, obviously. Note that the clipping character can be made quite benign so it doesn't sound completely broken when sligthly overdriven occasionally. Think good tube amps.
But unless you have proper clipping indicators on the amp (which are seldom), you never know if you clipped already, let alone know how much headroom is left before clipping sets in. And some amps may start to sonically slowly change, actually degrade at way lower levels. High efficiency speakers are good way to lower the power demands ;-)