How to shift the operating range of an amplifier while keeping the speaker volume constant?

[TurtlePaul]

...You didn't complain about noise. If you can't hear it, it's not a problem.

Yes, if you don’t hear noise at idle then you can ignore the whole left hand of the THD+N vs. power curve where the noise floor dominates.

 

[MaxwellsEq]

Thank you for the many inspiring answers. Meanwhile, I think I can answer the question myself. I will make exact measurements today. Here is only an example based on a guess:

View attachment 507824

To @UltraNearFieldJock : do NOT add resistors at the output of the power amplifier. It's a very bad idea in terms of fidelity.

1. Change the power amplifier gain. OR 2. Change the passive attenuator by putting resistors between the potentiometer and ground.

 

[UltraNearFieldJock]

do NOT add resistors at the output of the power amplifier. It's a very bad idea in terms of fidelity.

Could you specify this a little more? Which parameters of the signal will be distorted? What about passive loudspeakers—inside the crossovers there are a lot of resistors between the power amplifier and the drivers. I'am only worried about the damping factor: impulse‑response distortion (slower attack, longer ringing).

Even the distortion from the capacitors/inductors in the crossover will swamp any distortion improvements you can make in this and any good amplifier.

I don’t use any passive crossovers; I build and use only active systems. I use passive loudspeakers only for party music, not for "serious" listening. Active systems are technically better than passive ones exactly for this reason. Nevertheless, there are millions of happy passive loudspeaker listeners.

you cannot improve the SNR by modifiying your signal chain in any way

I’m entirely in agreement. Nothing could be not more nothing.

The most obvious question is why you don't use headphones?

Please look this: UNF Ultra Near Field. The answer is simple: UNF is much more enjoyable.

 

[MaxwellsEq]

Could you specify this a little more? Which parameters of the signal will be distorted

Modern power amplifier designers do everything they can to minimise the output impedance that the load (loudspeaker) sees.

Loudspeakers are not a pure resistance, if they were, what you are proposing would be "reasonable", if obviously wasteful in energy.

Instead loudspeaker drivers are motors and have inductive behaviour - i.e. the current lags the voltage. But they also have complex, non-linear relationships between voltages and currents depending on excursion, heat etc. These characteristics vary with frequency.

Because of this, even with an active setup where the driver is directly connected to the amplifier's output stage, your "resistor-divider" network idea is flawed. The voltage drop across the resistor will NOT be the same at all frequencies! You will have changed the frequency response of the amplifier+driver.

If you add in a passive crossover where designers have included capacitive as well as inductive components, your resistor-divider idea is even worse at changing the frequency response.

When engineers design the reactive elements of a passive loudspeaker, they assume that the output impedance of the power amplifier is effectively zero at all frequencies. It would be impossible for them to design for a wide range of output impedances.

If you instead create a resistor-divider network at line-level, everything is almost totally resistive in nature, the voltage across the resistor will be close to identical at all frequencies. This is the correct approach.

 

[Flaesh]

is what is the minimum current to make the diaphragms move? I suspect that most speaker drivers have “stiction” effects which will causes un-smooth motion, for example

Are there any objective data about this stiction or similar phenomena?

 

[Lambda]

A Class‑D Hypex Ncore NC400 power amplifier receives its input signal from a DAC through a passive volume control (potentiometer). The amplifier output drives a Visaton AL180 (8 Ω) loudspeaker directly. According to the specifications, the amplifier’s output impedance is only 3 mΩ. My ears are only a few centimeters away from the speaker, so I need extremely low listening levels—basically like using headphones.

Why don't you try using a headphone amplifier?
This seems to be the wrong amp for the job. I would not go with Class D for anything this low power.

You could and should calculate your max listening level, voltage and power wise, and then pick an appropriate amp, at least in the right order of magnitude.

For this you can set all your analog gains to a little bit more than normal max listening level for dynamic music.

Then with all the same gain settings play a sine wave at a known amplitude, negative gain basically, lets say -20 dB.

Measure the voltage at the amplifier output.
Since 20 dB is 10x voltage gain your amplifier Vrms at max listening level 0 dBFS will be 10x of your measured level.

Now choose an amplifier that can deliver this voltage into the impedance of your speaker.

To double check you should go the same route from your desired SPL volume level. Use your speaker efficiency rating to calculate the needed power or voltage to reach this level.
This is a good source if you need help with the calculations:

dB calculate - decibel calculation dB calculator voltage power ratio sound pressure level matching dBA SPL sound pressure intensity ratios converter thd percent % audio engineering impedance matching bridging - sengpielaudio Sengpiel Berlin

dB calculate dB calculator decibel calculator voltage power ratio reference voltage power level matching dBA dB SPL sound pressure intensity ratios converter thd percent % audio engineering impedance matching bridging - Eberhard Sengpiel sengpielaudio.com

sengpielaudio.com

dBu in Volt dBV dBu Pegel dBFS digital audio dB Dezibel Umrechnung umrechnen Daempfung Verstaerkung 0 dBFS Bezug Bezugswert Referenzwert Erklaerung absoluter Pegel Logarithmus Verhältnis Dezibel Schalldruck Druck Spannungspegel dBFS dBVU 0 dB Audio S

Umrechnung dB dBu dBFS digital audio dB Dezibel dBV Pegel in Volt zur Tontechnik Erklärung Dämpfung Verstärkung 0 dBFS Logarithmus Bezug Bezugswert Referenzwert absoluter Pegel dBFS SPL Verhältnis Dezibel Spannungspegel Spannung Leistung dBVU dBFS 0 dB Audio Schall mW Rechner pro audio consumer...

sengpielaudio.com

Sound power level SWL and sound pressure level SPL distance compare acoustic power sound source noise Conversion of sound pressure to sound intensity conversion sound level energy level strength directivity factor coefficient sound intensity SIL - se

Sound power SWL and sound pressure level SPL distance compare acoustic power sound source noiseConversion of sound pressure to sound intensity source conversion sound level energy intensity level strength directivity factor coefficient sound intensity SIL sound source - Sengpielaudio - Eberhard...

sengpielaudio.com

but honestly a modern LLM will have no trouble to do them for you and walk you throug:

 

[Lambda]

why would anyone want to listen to super high efficiency speaker near field?

"DIY construction with listening distance of 10-20 cm (depend on the Head position)"

So apparently he wants to...

If we do this back-of-the-envelope calculations.

400W in a 88dB/W speaker at 0.1M would give hypothetical 134dB (ignoring near field effects)
If the Amp has 120dB S/N at rated power https://www.audiosciencereview.com/...and-measurements-of-hypex-nc400-diy-amp.5907/
(and if we make the generous assumption it's all noise)

So hypothetical in the order of ~14dB SPL noise at the ear.

If we go from the quoted output noise voltage datasheet 25uV (ASR 8ohm test? 35uV?)
We still only get ~10dB SPL.

"Not great not terrible"

I would try to use an amp with 1/100 of the power and see if this is sufficient power

 

[Waveform Fidelity]

Are there any objective data about this stiction or similar phenomena?

I am sure someone, somewhere has done research, maybe the work of Klippel. However is should be practically obvious that there is a minimum current to make the cone move, otherwise one could simply breathe on a woofer cone and watch it move!

 

[Waveform Fidelity]

Could you specify this a little more? Which parameters of the signal will be distorted? What about passive loudspeakers—inside the crossovers there are a lot of resistors between the power amplifier and the drivers. I'am only worried about the damping factor: impulse‑response distortion (slower attack, longer ringing).


I don’t use any passive crossovers; I build and use only active systems. I use passive loudspeakers only for party music, not for "serious" listening. Active systems are technically better than passive ones exactly for this reason. Nevertheless, there are millions of happy passive loudspeaker listeners.


I’m entirely in agreement. Nothing could be not more nothing.


Please look this: UNF Ultra Near Field. The answer is simple: UNF is much more enjoyable.

I guess you do realize that most of the measurements FR, polar response, diffraction, phase response, distortion etc SPL are based upon being in the far field? That’s > 6 times the diameter of each driver. So when you are in the near field, I am not surprised it sounds very different. If its more enjoyable to you, good for you.

Ps by the way your active crossovers most likely have higher THD that the Hypex amp!

 

[DVDdoug]

I am sure someone, somewhere has done research, maybe the work of Klippel.

If it was a problem, or an audible problem, I'm sure it would be commonly reported on. The thing about low-level distortion is that it's harder to hear... If you can barely hear the signal you aren't going to notice 10% distortion because the distortion is below audibility.

A speaker is more like a spring so it should have little or no friction or stiction.

However is should be practically obvious that there is a minimum current to make the cone move, otherwise one could simply breathe on a woofer cone and watch it move!

You can hear sound before a speaker moves enough to see the movement. I don't know about "breathing" but you can use a speaker as a microphone and it will convert sound in the room into electrical signals. (Dynamic microphones are very similar to speakers with a voice coil and diaphragm, and a dynamic mic will put out sound if you feed-in a signal.) A speaker won't make an efficient or good-quality mic, but it will work. Sometimes a speaker is used as a mic for a kick drum.

 

[antcollinet]

Are there any objective data about this stiction or similar phenomena?

No, it is not a thing. Otherwise quiet parts of music would suffer from it. There are no sliding contact surfaces in a speaker, so no stiction.

 

[antcollinet]

Could you specify this a little more?

I'm going to add to this advice clear instruction. DO NOT TRY TO ADD RESISTORS TO THE OUTPUT OF THE AMP.

1 - You mess up the frequency response - the high output resistance will interact with the impedance curve of the speaker and mess it. up.

2 - You are effectively trying to divert power from the speakers to the resistors. You will need high power resistors and they will get very hot.

3 - DONT DO IT. It is not safe, you risk damaging your amp or speakers or both.


What you should be doing is using addtional attenuation on the input either by (as has been pointed out above)


Selecting the lowest gain on your amp.
An input attenuator.
I resistor in your volume control
Turning down the output of your DAC if it can do this.

As has already been pointed out - if you are not hearing noise from your speakers when music is not playing, you don't need to worry about noise.

 

[antcollinet]

However is should be practically obvious that there is a minimum current to make the cone move,

No, it is not practically obvious.

Resistance to speaker movement comes from the spider. With no current the spider is holding the cone in it's central position. Force forwards and backwards is equal. (Like an object held between two stretched springs pulling in opposite directions.

As soon as there is current (any current at any magnitude) that will create a force in the magnetic field that will work with one of the springs and against the other. There will be movement. It will be tiny movement if the current is tiny, but no matter how tiny, there will be movement proportional to the current.

 

[Waveform Fidelity]

No, it is not a thing. Otherwise quiet parts of music would suffer from it. There are no sliding contact surfaces in a speaker, so no stiction.

Surely you don’t think that a rubber roll surround is linear?

 

[Waveform Fidelity]

No, it is not practically obvious.

Resistance to speaker movement comes from the spider. With no current the spider is holding the cone in it's central position. Force forwards and backwards is equal. (Like an object held between two stretched springs pulling in opposite directions.

As soon as there is current (any current at any magnitude) that will create a force in the magnetic field that will work with one of the springs and against the other. There will be movement. It will be tiny movement if the current is tiny, but no matter how tiny, there will be movement proportional to the current.

Well, except non of this stuff is linear. Time to go off and research and prove me wrong then lets start with Klippel and see if we can educate ourselves further.

The reason, IMHO, that this is not widely researched is that the LOUD speaker is designed to produce significant SPL. They are not designed with the use-case of ultra near-field use

 

[Waveform Fidelity]

If it was a problem, or an audible problem, I'm sure it would be commonly reported on. The thing about low-level distortion is that it's harder to hear... If you can barely hear the signal you aren't going to notice 10% distortion because the distortion is below audibility.

A speaker is more like a spring so it should have little or no friction or stiction.


You can hear sound before a speaker moves enough to see the movement. I don't know about "breathing" but you can use a speaker as a microphone and it will convert sound in the room into electrical signals. (Dynamic microphones are very similar to speakers with a voice coil and diaphragm, and a dynamic mic will put out sound if you feed-in a signal.) A speaker won't make an efficient or good-quality mic, but it will work. Sometimes a speaker is used as a mic for a kick drum.

So what makes it a bad quality mic then? Also, you cede that its should have little or no friction, but we are only discussing very small amounts of movements here.

 

[antcollinet]

Well, except non of this stuff is linear. Time to go off and research and prove me wrong then lets start with Klippel and see if we can educate ourselves further.

The reason, IMHO, that this is not widely researched is that the LOUD speaker is designed to produce significant SPL. They are not designed with the use-case of ultra near-field use

Non linearity is not the same as discontinuity. Stiction is a discontinuity. It happens with two surfaces sliding together. There are no such surfaces in a speaker.

And actually - in the region we are discussing - tiny movement: A speaker is effectively linear.

 

[Lambda]

Surely you don’t think that a rubber roll surround is linear?

The suspension stiffens is usually the lowest in the rest position.
https://www.klippel.de/know-how/mea...rs/transducer-nonlinearities-curve-shape.html

JL Audio 12W7AE-3

Klippel LSI and Distortion measurements on the JL Audio 12W7AE-3 subwoofer.

resonixsoundsolutions.com

otherwise one could simply breathe on a woofer cone and watch it move!

And who says its not moving if you breathe on it? its just not moving a lot.

 

[MaxwellsEq]

Well, except non of this stuff is linear. Time to go off and research and prove me wrong then lets start with Klippel and see if we can educate ourselves further

You are the one making EXTRAORDINARY CLAIMS. Extraordinary claims require extraordinary evidence

So it's now YOUR task to provide US with this extraordinary evidence.

 

[Waveform Fidelity]

Non linearity is not the same as discontinuity. Stiction is a discontinuity. It happens with two surfaces sliding together. There are no such surfaces in a speaker.

And actually - in the region we are discussing - tiny movement: A speaker is effectively linear.

The suspension stiffens is usually the lowest in the rest position.
https://www.klippel.de/know-how/mea...rs/transducer-nonlinearities-curve-shape.html

JL Audio 12W7AE-3

Klippel LSI and Distortion measurements on the JL Audio 12W7AE-3 subwoofer.

resonixsoundsolutions.com

And who says its not moving if you breathe on it? its just not moving a lot.

the point is that the spider and surround is lossy. It takes a minimum current to make the cone overcome frictional effects and start to move. You can very easily show that by pressing your finger on the diaphragm and feeling the resistance to motion. Physically, its not possible for it to be frictionless on earth and thus there is a minimum current that is needed to the voice coil to make it move. This is basic physics and I don’t understand why there is any argument about this. I used the extreme example to make this point, that if there were no losses as others seem think, it would freely move, with a faint force like from your breath. Frictional losses within the materials are the main reason why a microphone has a minimum SPL sensitivity and why there is a minimum SPL floor in a loudspeaker driver.