• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

How much amplifier power is required?

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,521
Likes
3,086
Location
Palatinate, Germany
Okay. So I control the complete digital chain, and I can see digital signal levels before they leave for the DAC. Since I use Topping D50, I also know the volume there, and we know that it outputs 2V RMS at full scale.

My test track is Westernhagen - Mit 18:

Dynamic range meter for foobar2000 quotes DR8, here's iZotope RX:


Waveform:


And here's TotalMix, showing a level of ~ -5dB at peaks:


I have the Topping D50 set up at -7dB for the satellites, so in total we're down -13dB already. If my calculations are correct, that means the D50 is only outputting ~ 0.44v into the Benchmark AHB2. We know the gain in the high setting is 23dB and it would produce 100W into 8 ohms at 2V. But since we only have 0.44v and this is a linear relationship (right?) I would assume max power is actually only 22W, with the amp outputting 6.21V. Is that correct?

My SPL meter set to slow and C weighting at the listening position is showing 107dB max hold. Mind you, the subs have their own amp and I cross over at 100hz. Also, I'm not using KEF LS50 as you would guess from the TotalMix screenshot but ER18DXT at the moment.

e: The calculator gives 80dB sensitivity for 22W at the distance when placed in a corner/near a wall. It fits. So there you have it. All I need is 22W :eek:
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Okay. So I control the complete digital chain, and I can see digital signal levels before they leave for the DAC. Since I use Topping D50, I also know the volume there, and we know that it outputs 2V RMS at full scale.

My test track is Westernhagen - Mit 18:

Dynamic range meter for foobar2000 quotes DR8, here's iZotope RX:


Waveform:


And here's TotalMix, showing a level of ~ -5dB at peaks:


I have the Topping D50 set up at -7dB for the satellites, so in total we're down -13dB already. If my calculations are correct, that means the D50 is only outputting ~ 0.44v into the Benchmark AHB2. We know the gain in the high setting is 23dB and it would produce 100W into 8 ohms at 2V. But since we only have 0.44v and this is a linear relationship (right?) I would assume max power is actually only 22W, with the amp outputting 6.21V. Is that correct?

My SPL meter set to slow and C weighting at the listening position is showing 107dB max hold. Mind you, the subs have their own amp and I cross over at 100hz. Also, I'm not using KEF LS50 as you would guess from the TotalMix screenshot but ER18DXT at the moment.

e: The calculator gives 80dB sensitivity for 22W at the distance when placed in a corner/near a wall. It fits. So there you have it. All I need is 22W :eek:
I see one error. In power 3 db doubles or halves the power. So you actually are a touch shy of 5 watts with 6.2 volts out of the amp into 8 ohms.
6.21 volts / 8 ohms=.77625 amps so .77625 amps x 6.21 volts = 4.82 watts.
 

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,521
Likes
3,086
Location
Palatinate, Germany
Holy shit that can't be. I mean. I had an amp with VU meters before. I rarely pushed above 10W when pushing it, so that could be right.

e: I connected a smart power plug. Power consumption with the amp on but no music is ~ 24W, power consumption at full volume is 28W....
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
IME most people use far less average power than they think, but have much higher peaks than they think.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Holy shit that can't be. I mean. I had an amp with VU meters before. I rarely pushed above 10W when pushing it, so that could be right.
I'd suggest turning off the sub and trying REW again to see peak. It probably will be much less than 107 db.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,586
Likes
38,284
Location
Gold Coast, Queensland, Australia
IME most people use far less average power than they think, but have much higher peaks than they think.

Maybe if those people who think they only need a few watts, attached a 'scope to the speaker terminals, let it free run and watched their peaks being chopped off flat on transients, while playing moderately loud dynamic music, they would throw out all their carefully calculated numbers...
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
How much amplifier power is required?

Depends on whether you want to turn it up sometimes, or not.

100W is adequate. Power requirements go up sharply if you want to be "sure".

Estimation for a speaker of 8 ohms and 87dB sensitivity at 1W (2.83V).

Add 3 to 6dB SPL for stereo.

1578686680469.png


My observation is that about 20dB headroom is necessary for peaks above average listening levels with most recordings, so if you listen at 80dB average, you'll see 100dB peaks, and the associated power demands.

My amps are rated for 350W at 8 ohms, and I've seen (briefly) 116.9dBspl at the sofa, during expermentation with a drum solo.

Typical listening loudly loud peaks are 105 to 110dB.

Moderate peaks 90 to 95dB.

Quiet listening peaks maybe 80 to 85dB.

Talk Radio is playing right now, through the JBLs, 70.9 average SPL over 47 seconds, 89.4 peak SPL. Talk is maybe a little more "dynamic" than tunes, at least, it registers at the top of my observed peak to average range.

1578687090295.png


So, operating in this range, more or less.

1578687567502.png
 
Last edited:

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
I have the Topping D50 set up at -7dB for the satellites, so in total we're down -13dB already. If my calculations are correct, that means the D50 is only outputting ~ 0.44v into the Benchmark AHB2. We know the gain in the high setting is 23dB and it would produce 100W into 8 ohms at 2V. But since we only have 0.44v and this is a linear relationship (right?) I would assume max power is actually only 22W, with the amp outputting 6.21V. Is that correct?

I'm trying to do a similar calculation for my Auralic Vega DAC and Bryston 3B-SST amp. Stereophile measured the max output of the Vega at 4.37 V (RMS, I assume). Bryston specifies 2.6 Vin for the rated output (250W @ 4 Ohm) at the lower gain setting, again I assume RMS. For serious listening to a good digital recording of symphonic music, I usually have the Vega set at no more than -15 dBFS. Let's say I go a bit nuts, and I'm at -10 dBFS. Program peaks can be -3 dBFS, but let's assume they get to 0 dBFS (ignoring intersample overs). So output voltage is

10**(-10/20) * 4.37 Vrms = 1.38 Vrms

which would produce

1.38/2.6 * 250W = 133W into 4 Ohms
 

HemiRick

Active Member
Joined
Jan 2, 2020
Messages
133
Likes
150
I think how much amp power you need/want is directly related to your need/desire for good bass. The more/better bass you want the more amp power you want.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,586
Likes
38,284
Location
Gold Coast, Queensland, Australia
People also tend to forget that all stages in the path have both peak and RMS values.

Your D/A may output 4.37V RMS @ 0dBFS, but the top of the waveform is actually +6.18V.

Your amplifier may output 250W@4R which corresponds to 31.6V RMS, but it needs to able to swing +/-44.7V to do that cleanly.

So in the example above, @Daverz , -10dBFS from 6.18V is 1.95V. The amplifier takes that 1.95V peak and amplifies it by its gain, which is 12.16 times (21.7dB) ,according to your example above.

Your amplifier doesn't care if the input voltage is a peak or somewhere else on the waveform, it treats all signals the same. It needs to produce an accurate RMS voltage, times its gain, for any input voltage, at any point on the waveform. So we get 23.71V RMS at the speaker terminals on your peak. That is 140.5W@4R. Also, bear in mind that to do that, the amplifier needs to be able to swing another 1.414 times that voltage internally to produce that...



10**(-10/20) * 4.37 Vrms = 1.38 Vrms

which would produce

1.38/2.6 * 250W = 133W into 4 Ohms

Your figures above are not correct.
Your amplifier would have a voltage gain of 12.16 times.

12.16 times 2.6V is 31.62V, and over 4R that equates to 250W
12.16 times 1.38V is 16.78V, and over 4R that equates to 70.4W.

Remember:

1578692715926.png
 
Last edited:

Putter

Senior Member
Forum Donor
Joined
Sep 23, 2019
Messages
493
Likes
771
Location
Albany, NY USA
All right the consensus seems to be 200 wpc from a power amp is required, but 2 things that don't seem to be covered.

One is almost everyone agrees that to get decent bass you need a subwoofer, preferably multiple subs. This means that your 50 wpc amp or under performing 100 wpc home theater receiver isn't required to produce below say about 80 hz. How much does that reduce the amp requirements on the primary amp or for that matter what is the minimum wattage needed for the sub amp?

Two let's assume that your 50 wpc amp is actually clipping briefly on transient peaks. What is the actual audibility or for that matter the frequency of those peaks esp in regard to blind testing?

I'll admit that it's easy to ask these questions without providing any possible answers, but it does apply to most of my systems where a sub supplements a relatively low wattage amp (40 wpc and 80 wpc respectively) and I don't hear obvious distortion related to clipping.
 
D

Deleted member 9286

Guest
This is an interesting topic and is something I have been pondering myself recently.

So I am driving a pair of Fyne Audio F500 bookshelf speakers directs from a Chord TT2.

The speakers sensitivity is rated as 89db and I have them setup in a room that is 5m x 3m with approx 3m ceilings.

Now the TT2 can only muster 7w running single ended, but we are told by Rob Watts that due to the 6 large super capacitors inside, it can easily handle peaks without clipping.

It can certainly fill the room referenced above without distorting. Is it at ear bleeding levels? Definitely not, but 7 watts seems to deliver much more volume than I would have expected in my setup.
 
OP
W

Willem

Major Contributor
Joined
Jan 8, 2019
Messages
3,658
Likes
5,276
@Putter Amplifier power in an active sub is obviously also part of the story, even if the very low frequencies that the sub(s) will be capable of add extra demands. The effect of sub power will be greater if the sub(s) are combined with some high pass filtering of the main speakers and amplifier.
@BenniMac Your case is a good example of the non linearity of the issue. Your room is small, and your speakers are efficient. Even so, I think 7 watts is anaemic.
 

murraycamp

Senior Member
Forum Donor
Joined
Jul 10, 2019
Messages
421
Likes
647

anmpr1

Major Contributor
Forum Donor
Joined
Oct 11, 2018
Messages
3,722
Likes
6,406
So all systems witout subs have deficient bass performance?
This, of course, depends upon what you are attempting to reproduce. For chamber music, Chinese opera, solo Bach cello suites, etc., it is not likely necessary. If your musical tastes run along the lines of Death Star explosions and Godzilla...
 

RichB

Major Contributor
Forum Donor
Joined
May 24, 2019
Messages
1,946
Likes
2,611
Location
Massachusetts
People also tend to forget that all stages in the path have both peak and RMS values.

Your D/A may output 4.37V RMS @ 0dBFS, but the top of the waveform is actually +6.18V.

Your amplifier may output 250W@4R which corresponds to 31.6V RMS, but it needs to able to swing +/-44.7V to do that cleanly.

So in the example above, @Daverz , -10dBFS from 6.18V is 1.95V. The amplifier takes that 1.95V peak and amplifies it by its gain, which is 12.16 times (21.7dB) ,according to your example above.

Your amplifier doesn't care if the input voltage is a peak or somewhere else on the waveform, it treats all signals the same. It needs to produce an accurate RMS voltage, times its gain, for any input voltage, at any point on the waveform. So we get 23.71V RMS at the speaker terminals on your peak. That is 140.5W@4R. Also, bear in mind that to do that, the amplifier needs to be able to swing another 1.414 times that voltage internally to produce that...





Your figures above are not correct.
Your amplifier would have a voltage gain of 12.16 times.

12.16 times 2.6V is 31.62V, and over 4R that equates to 250W
12.16 times 1.38V is 16.78V, and over 4R that equates to 70.4W.

Remember:

View attachment 45349
Isn't it fair to say that any amp capable of the RMS voltage has sufficient headroom to produce the peak of the signal?

If you ignore that that listening levels include room gain and two speakers. At least in my case, those factors combine to allow the Salon2's to deliver the 86dB at 2 watts (stereo) at my listening position.

Assuming digital sources, a full scale sine wave at the listening position represents the maximum signal, except for digital clipping. It's a shame that such recordings are made but 3dB for short term peaks seems to be a practical limit.

- Rich
 
Last edited:

RichB

Major Contributor
Forum Donor
Joined
May 24, 2019
Messages
1,946
Likes
2,611
Location
Massachusetts
Maybe if those people who think they only need a few watts, attached a 'scope to the speaker terminals, let it free run and watched their peaks being chopped off flat on transients, while playing moderately loud dynamic music, they would throw out all their carefully calculated numbers...

In my room 2 watt s is 86 DB average which is loud which could have 20 dB peaks.
The problem is if you ignore rooms and average levels, like the crown calculator, you will compute power requirements close to 1 kw.
The difference between 200 to 300 watts is about 1 dB.

- Rich
 
Last edited:

Not Insane

Member
Joined
Jan 8, 2020
Messages
31
Likes
35
He refers to the Audio Precision analyzer used to make the measurements like Amir uses here. Very expensive BTW
555B-ADIO-PDM-DSIO-BT.jpg
Heh. I used to work as a sales engineer for compuware (mostly mainframe software). Our software was much simpler than something like Microsoft Word or any PC operating system. But it cost millions per copy with a hefty annual fee (sometimes seven figures). The reason: We didn't sell millions of copies. When your R&D goes into only a few hundred or thousand customers, the price per copy to the end user can be rather high. Then there is the critical aspect of relability.

i.e. I'll bet these things would be cheaper if everybody needed one. :D
 

Xulonn

Major Contributor
Forum Donor
Joined
Jun 27, 2018
Messages
1,828
Likes
6,311
Location
Boquete, Chiriqui, Panama
Top Bottom