Damping Factor

[Walor]

Hello!

When comparing transistor and tube amplifiers, the significantly higher damping factor of the transistor amplifier is presented as a major advantage. I began to doubt this for some time. I designed and tested an audio current amplifier. Here is the functional description:

https://www.elektormagazine.de/labs/versatile-audio-amplifier-with-the-current-output-signal

This amplifier drives a guitar speaker or a full-range speaker flawlessly. Meanwhile, its damping factor will be practically zero. So I started thinking, why is there something wrong with the theory? Finally, simulation with LTspice gave me the answer (at least I think so).

I found a model for the guitar speaker here: http://tamivox.org/dave/speaker/graphs.html

The test setup looks like this:

V1 is the ideal voltage amplifier with the infinitely large damping factor. Therefore, its output voltage does not depend on the load and can be loaded as desired. The current through the Rref can show us an ideal current pattern. Now the first measurement:

Surprisingly, the after-oscillations from the current amplifier have a much smaller amplitude than the after-oscillations from the voltage amplifier... Why?

Because the output voltage of the current amplifier counteracts the voltage induced in the loudspeaker coil and prevents the first huge swing as is the case with the voltage amplifier.

Furthermore, this exactly matches my listening experience. This sound is cleaner than the sound of my (standard) hi-fi stereo system.

 

[fpitas]

Guitar and fullrangers aside, a modern hi-fi speaker is designed to run from a voltage source:

Audio Myth - "Damping Factor Isn't Much of a Factor"

Myth - "Damping Factor Isn't Much of a Factor" Myth - "A Damping Factor of 10 is High Enough" Myth - "All Amplifiers Have a High-Enough Damping Factor" Where did these Myths Originate? These myths seem to trace back to a well-known paper (PDF) written by Dick Pierce. His analysis shows that a...

benchmarkmedia.com

 

[DVDdoug]

Speakers are "voltage operated". A speaker should have flat frequency response with respect to the voltage, regardless of how the speaker's impedance varies over the frequency range. Same with headphones.

 

[DonH56]

I am not sure what you are showing. The first plot only shows currents, and the second shows the voltage from an ideal voltage source and voltage at the output of the amplifier, plus the current through the load (again). I am not sure what Rref is doing; it is shunting ideal voltage source so affecting its current, but why is it there?

The difference in current waveforms may be due to the biasing and bandwidth of the amplifier limiting its performance relative to the ideal voltage source.

You can measure the output impedance of the amplifier in the simulation and use that to calculate the damping factor. The amplifier has feedback and finite output impedance so I doubt it is zero.

The advantage of low output impedance from the amplifier depends upon the speaker's impedance and design goals. As stated above, most speakers are designed with a voltage source as the driving source. There are books on designing current-mode outputs but my limited research years ago led me to conclude that there was no real advantage, and some significant disadvantages, driving real speakers.

Here is an old post about the impact of output impedance (related to 1/damping factor) for some various amplifier and speaker models: https://www.audiosciencereview.com/...amping-factor-and-speakers.23968/#post-807327

 

[Walor]

Speakers are "voltage operated"

The force that drives the diaphragm is the Lorentz force: F = I*L*B. L and B are constant, so there is a direct dependence on the current.

A speaker should have flat frequency response with respect to the voltage

You write correctly: "should have". But not for natural physical reasons, but because of artificial, unnatural adaptation to manufactured voltage amplifiers.

 

[fpitas]

For better or worse, the standard that speakers are designed to is voltage drive.

 

[Walor]

I am not sure what Rref is doing; it is shunting ideal voltage source so affecting its current, but why is it there?

This is the reference. This is how the current would flow through an idealized speaker.

so I doubt it is zero

I measured the output impedance of real current amplifiers, it is in the range of 200 ohms. 6 ohms / 200 ohms = 0.03, so almost zero.

 

[voodooless]

From Wikipedia:

Speaker diaphragms have mass, and their compliant suspension components have stiffness. Together, these form a resonantsystem, and the mechanical cone resonance may be excited by electrical signals (for example, pulses) at audio frequencies. But a driver with a voice coil is also a current generator, since it has a coil attached to the cone and suspension, and that coil is immersed in a magnetic field. For every motion the coil makes, it will generate a current that will be seen by any electrically attached equipment, such as an amplifier. In fact, the output circuitry of the amplifier will be the main electrical load on the "voice coil current generator". If that load has low resistance, the current will be larger, and the voice coil will be more strongly forced to decelerate. A high damping factor (which requires low output impedance at the amplifier output) very rapidly damps unwanted cone movements induced by the mechanical resonance of the speaker, acting as the equivalent of a "brake" on the voice coil motion (just as a short circuit across the terminals of a rotary electrical generator will make it very hard to turn). It is generally (though not universally) thought that tighter control of voice coil motion is desirable, as it is believed to contribute to better-quality sound

 

[Cbdb2]

As said, speakers are designed to be voltage driven, a current amp will not produce a flat freq. response. Can you show the voltage at each speaker?
That current amp is not a great design, no global feedback. Can you measure (in ltspice) freq. response output impedance and the distortion?

 

[popej]

Guitar speakers aside, this amplifier looks like really bad design. You probably have to tune R3 and R4 depending on used parts and then it still can be thermally unstable. Better to not use it for any real task.

 

[fpitas]

Guitar speakers aside, this amplifier looks like really bad design. You probably have to tune R4 depending on used parts and then it still can be thermally unstable. Better to not use it for any real task.

Maybe it's meant to distort?

 

[Cbdb2]

View attachment 315388

Furthermore, this exactly matches my listening experience. This sound is cleaner than the sound of my (standard) hi-fi stereo system.

Your voltage drive speaker will be reproducing the Viout. Thats what you will hear. Lots of ringing, exaggerating certain freqs.

 

[solderdude]

Now try this amp with a hifi speaker with a wildly varying impedance and see if the sound is still 'cleaner' and better...
Then you'll know why hifi amplifiers are voltage sources.

That said, some full range drivers do sound 'better' from a current source. Just connect such a speaker through a 100ohm resistor (turn up the volume a little).
Chances are that guitar amp speaker does sound 'better'.

 

[fpitas]

That said, some full range drivers do sound 'better' from a current source. Just connect such a speaker through a 100ohm resistor (turn up the volume a little).
Chances are that guitar amp speaker does sound 'better'.

I think that's a big part of the confusion here.

 

[Walor]

For better or worse, the standard that speakers are designed to is voltage drive.

Current Source Amps and Sensitive/Full-Range Drivers

In this fascinating article, published in audioXpress, October 2004, Nelson Pass investigates loudspeaker performance when driven by a current source amp. This extensive research shows what Pass and Kent English found with a variety of full-range sensitive loudspeaker drivers when they drove...

audioxpress.com

 

[voodooless]

It’s a very elaborate way to get some bass boost…

 

[solderdude]

Not only that, but that bass boost is solely determined that is by the resonance frequency of a speaker (width and peaking) and treble boost that is determined by the voice coil.
Bass will get weird when that speaker is ported as well.
It could work out well with some full range speakers in a small cabinet that need some boost in the lows and highs.

 

[Walor]

Can you measure (in ltspice)

I even have some real measurements

 

[Walor]

thermally unstable

This is Class A, so the amplifier needs a decent heatsink anyway.

 

[Walor]

Maybe it's meant to distort?