How much amplifier power is required?

[arisholm]

Those are "Hypex Watts", so it equates to about 100W/ch in Accuphase Watts.

I kid, I kid....

You kid, but I have wondered about the difference between "continuous" vs "peak" in those things and whether it is the former or the latter (or something in between) being reported in those specs...

 

[tomtoo]

Watt's are like money.

It's always good to have more than you actually need.

 

[HansHolland]

Under normal listining situations the output of my amplifier stays under 1 Volt. I measured that by putting the volume control a little above my average listining position and run a -40 dB 1kHz tone via foobar2000, the output of my amplifier showed 0.01 Volt, and then multiply by 100 (= +40 dB). This measurement simulates the peak levels in music, the sound cannot go louder than 0 dB.

Sometimes I listen at 20 dB higher levels. That is still below 10 Volt at the loudspeaker outputs of my amplifier.

So, how much power do I need?
I have multiple loudspeakers. All with a efficiency around 90 dB/2.83 Volt. And some with a 4 Ohm impedance.
10 Volt * 10 Volt / 4 Ohm = 25 Watt

25 CLEAN Watts!
Look at the different amplifier measurements. With some amplifiers the distortion starts rising just a little before clipping. While for other amplifiers the distortion already starts rising at 100 times lower powers than the clipping power.

Why do I need not more? Because I have a desktop setup! I have my (loud)speakers within 1 meter distance.

 

[veeceem]

Under normal listining situations the output of my amplifier stays under 1 Volt. I measured that by putting the volume control a little above my average listining position and run a -40 dB 1kHz tone via foobar2000, the output of my amplifier showed 0.01 Volt, and then multiply by 100 (= +40 dB). This measurement simulates the peak levels in music, the sound cannot go louder than 0 dB.

Sometimes I listen at 20 dB higher levels. That is still below 10 Volt at the loudspeaker outputs of my amplifier.

So, how much power do I need?
I have multiple loudspeakers. All with a efficiency around 90 dB/2.83 Volt. And some with a 4 Ohm impedance.
10 Volt * 10 Volt / 4 Ohm = 25 Watt

25 CLEAN Watts!
Look at the different amplifier measurements. With some amplifiers the distortion starts rising just a little before clipping. While for other amplifiers the distortion already starts rising at 100 times lower powers than the clipping power.

Why do I need not more? Because I have a desktop setup! I have my (loud)speakers within 1 meter distance.

If that's the case, Will Pass Labs XA25 drive Revel Salon2 with absolute authority at 3 meters I wonder? :| iirc it runs 25w in class A and approximately 50w in class AB at 8 Ohm.
From a glace, XA25 seems to not as good at controlling bass as my NAD M22 v2 (300w @ 8 Ohm)

 

[HansHolland]

At 3 meters I need 3 * 3 = 9 times more power, so let's say 10 times more. That is 250 Watt at 4 Ohm (or 125 Watt at 8 Ohm). And that is with loudspeakers with the same efficiency.
But, when I run at 25 Watt, that is very loud, I do it only for a short time. But I want to be able to do it, so I need that power!

 

[RichB]

At 3 meters I need 3 * 3 = 9 times more power, so let's say 10 times more. That is 250 Watt at 4 Ohm (or 125 Watt at 8 Ohm). And that is with loudspeakers with the same efficiency.
But, when I run at 25 Watt, that is very loud, I do it only for a short time. But I want to be able to do it, so I need that power!

Room gain and two speakers should also be considered. With those included, I suspect it is half that amount.

Here are some posts with maximum (0 dBFS signals) power estimates based on measurements at my listening position (10 feet) into 4 ohms:
https://www.audiosciencereview.com/...mplifier-power-is-required.10781/#post-300381

Here is a power calculator that takes into account speaker placement, distance, and dual speakers:
https://www.audiosciencereview.com/...mplifier-power-is-required.10781/#post-300476

- Rich

 

[Koeitje]

How accurate is measuring power draw at the wall? My class AB is pulling 80W peak from the wall when playing loud but at a comfortable level, does that mean it is putting 2x25W-ish into my speakers? I don't think it does right? Because of the caps providing power?

 

[DonH56]

How accurate is measuring power draw at the wall? My class AB is pulling 80W peak from the wall when playing loud but at a comfortable level, does that mean it is putting 2x25W-ish into my speakers? I don't think it does right? Because of the caps providing power?

How accurate is your measuring device? And how fast does it respond? "Peak" power can be tricky to measure (and record).

A typical class AB amp is probably somewhere around 60% efficient and that may be high depending upon what else is in the component. At 60%, you need about 1/0.6 = 1.7x the input power so 80 W yields about 48 W or 24 W/ch output so yes that seems in the ballpark. The SPL depends upon your speakers, your distance from them, room effects, etc. A very rough ballpark can be gleaned from an online calculator such as this one: Peak SPL Calculator . A pair of speakers with 90 dB/W/m sensitivity at 8' away will produce about 100 dB SPL driven by 24 W each. That is pretty loud to me but I've no idea what you consider "loud but comfortable". And don't forget most of the power is probably in the lower frequencies where it is easy to play at much higher SPL before it sounds "loud" ( https://en.wikipedia.org/wiki/Equal-loudness_contour ).

HTH - Don

 

[Koeitje]

How accurate is your measuring device? And how fast does it respond? "Peak" power can be tricky to measure (and record).

A typical class AB amp is probably somewhere around 60% efficient and that may be high depending upon what else is in the component. At 60%, you need about 1/0.6 = 1.7x the input power so 80 W yields about 48 W or 24 W/ch output so yes that seems in the ballpark. The SPL depends upon your speakers, your distance from them, room effects, etc. A very rough ballpark can be gleaned from an online calculator such as this one: Peak SPL Calculator . A pair of speakers with 90 dB/W/m sensitivity at 8' away will produce about 100 dB SPL driven by 24 W each. That is pretty loud to me but I've no idea what you consider "loud but comfortable". And don't forget most of the power is probably in the lower frequencies where it is easy to play at much higher SPL before it sounds "loud" ( https://en.wikipedia.org/wiki/Equal-loudness_contour ).

HTH - Don

Probably not very accurate, its just a cheap thing but I'm guessing within 10%. It updates several times a second and saves the highest wattage it encounters.

I'll have to grab my UMIK to get some SPL measurements , but my speakers are probably more like 86 dB/W/m. I'm guessing I'm at like 85dB.

 

[HifiLove]

First post here, learning and still learning .., pls excuse me if this was discussed earlier..

I'd like to revisit and get some clarification on "Recommended amplifier power" , that frequently appear in Speaker specs. When it says '40W-250W' , what does the high value 250W really mean ?

Is that average/continous or peak power? Is this at the nominal impedence of the speaker or minimal impedence or it does not matter?

Is that an absolute max power safely supported by the speaker , beyond which it can damage the speaker? Or more of an upper soft threshold or guidance that can be breached occasionally without any issue ?

Its been repeatedly stated in this thread , that its better to have more power than less(ie headroom for peaks) . , but its not clear how high is too high?

I was looking at an amp with spec "Output Power: 300 RMS watts/channel into 8 Ohms, 450 watts/channel RMS into 4 Ohms ". Can I safely use this amp to drive speaker rated 250W " max recommended amplifier power" . This amp outputs 450W RMS when impedence of speaker drops to 4 Ohms(which the speaker does...into 2.9Ohms in this case) and that far exceeds 250W.


Thanks

 

[HifiLove]

Just to add, I'm familiar estimating the power requirements , for a particular speaker sensitivity,listening distance,speaker placement etc from online calculators.

[For eg, I get 0.7W power requirement for reference level(85db) listening and 64w for 20db peak(105db). This is for an 8 ohm nominal impedence and double this (1.4W at avg spl, 128W at peak spl ) when it dips into 4 ohms ].

But my earlier question is a generic one, as I want to try to buy an amp , that'll last a long time , possibly across different speakers , including those whose nominal impedences are rated 4 Ohms(like many of the MartinLogans are). Lets say I stick to speakers with ~90db senstivities.

Any ideas on my previous question on the ceiling of "Recommended amplifier power"?

 

[NTK]

.... Any ideas on my previous question on the ceiling of "Recommended amplifier power"?

The ceiling of the "recommended power" is basically meaningless. It is highly frequency dependent. The tweeter of a speaker usually has a much lower power handling capacity than its woofer. What is enough to fry the tweeter may just be enough to give you sufficient bass. Too little power may arguably be worse, as clipping (which generates large amounts of harmonic distortions) will send large amounts of high frequency junk to the tweeter, and may damage it as a result.

The general rule of thumb is to err on the high side -- get as much as you can afford. For a very detailed explanation of how to estimate the amount of power you need, take a look at part 3 of Dr. Toole's "How to design a home theater".
https://routledgetextbooks.com/textbooks/9781138921368/home-theatre.php

 

[TomB19]

It's true, a lot of power isn't needed for almost everyone, including audiophiles.

The reason larger amps sound better, or at least are more solid on the bottom end, is because they have more control authority over the reactive load.

Demonstrating an amplifier responds with fantastic linearity into an 8 ohm static load means something but not much.

Amplifiers are open loop. I've long thought they should be closed loop. An infinitesimal amount of feedback would go a long way toward making all amplifiers sound the same. When that day comes, I will be happy with 10 Watts per channel.

I look forward to the day when servos aren't just for subs, anymore.

 

[RichB]

I am in the process of upgrading and renovating the game room system. The Revel M20's have been replaced by Revel M126Be's.
I measured the Oppo UDP-205 at 2.829 volts so basically 1 watt at 8 ohms and played some music.
Even at 2 watts, it is amazing how loud this is and emphasizes the importance of the 1'st watt.

I encourage you to determine your desired listening level and estimate your power requirements.

- Rich

 

[peanuts]

the myth of "too little power" blowing tweeters
https://sound-au.com/tweeters.htm

 

[sergeauckland]

the myth of "too little power" blowing tweeters
https://sound-au.com/tweeters.htm

It may be a myth that too little power blows tweeters, but as the article shows very well, it's excessive level that blows tweeters (and woofers). This comes earlier on a low powered amplifier as the amp clips and the average level goes up.

Unfortunately 'loud' when measured subjectively depends greatly on the distortion present so using a higher powered amplifier that doesn't clip i.e distort as easily means that the volume can go up higher, damaging the drivers more easily.

That's why on my system I have peak reading output level meters (not VU which are Virtually Useless) and on the crossover HF peak limiters to avoid overdriving the tweeters.

S

 

[TomB19]

VU meters do a great job of creating a nice aesthetic.

 

[DonH56]

It's true, a lot of power isn't needed for almost everyone, including audiophiles.

The reason larger amps sound better, or at least are more solid on the bottom end, is because they have more control authority over the reactive load.

Demonstrating an amplifier responds with fantastic linearity into an 8 ohm static load means something but not much.

Amplifiers are open loop. I've long thought they should be closed loop. An infinitesimal amount of feedback would go a long way toward making all amplifiers sound the same. When that day comes, I will be happy with 10 Watts per channel.

I look forward to the day when servos aren't just for subs, anymore.

My comment for years has been that most folk overestimate their average power needs and underestimate their peak power needs. Often by an order of magnitude or more either way. I agree output impedance explains some of the difference among amps, but there is SNR (noise floor, actually, since a little hiss is often easier to hear than a lot of THD) and having enough power to preclude clipping (20 - 30 dB over average power for music or movies is typical, though if an explosion in a movie clips I doubt anyone would know or care).

There have been several amplifiers that added a feedback line from the speaker terminals starting in the 1970's if not before (@restorer-john has shown examples; I only remember a Kenwood unit and couple of one-offs at the moment, senility). By and large the scheme never caught on. It costs money for the manufacturer and consumer (extra wires) and can lead to stability problems (especially when someone forgets to hook up the feedback wire; I always used a high-value resistor to ensure the loop was never open even if the wire was gone, and I think commercial units did the same). It is much easier to simply use large wire gauge and these days amplifiers with sub-100 m-Ohm and lower Zout are almost universal.

Another feedback loop, even one that realizes 0-ohm driving impedance at the terminals, will not change the passive crossover that generally sets the impedance the drivers see nor will it change the need for enough power to deliver your desired SPL at all frequencies. I suspect that, even with a perfect amplifier and perfect servo circuit, you'd still want more than 10 W unless the speaker was extremely sensitive and/or you sit very, very close...

IME/IMO - Don

 

[Frank Dernie]

My 109dB/watt horns are driven by a 175 W/channel Job amp.
I was worried noise may be a problem but if I set the level at my symphony listening level I have to put my head into the horn to hear hiss but I listen 16' from the speakers, so no problem at all.

 

[restorer-john]

but if I set the level at my symphony listening level I have to put my head into the horn to hear hiss

I can see it now. Frank with his head down the horn, symphony level volume, listening for hiss and the cat jumps on the remote...