Watts; the worst measurement unit ever for amp output?

[acmebass]

As an engineer with experience in the field myself, I'd put much much more stock in what Agedhorse has to say about it.

As for back-emf, yes it can be looked at as an impedance from the point of view of the amp outputs but the source impedance of the amplifier driving the speaker makes a difference in how much effect the back-emf has on the overall behaviour of the speaker. Higher source impedances ( i.e. low damping factor) can allow the back-emf to have sonic effects as the voltage created by that back-emf at the amplifier's terminals is higher if the amp's output impedance is higher. The higher the voltage is, the higher the impedance seen by the amp is at a particular frequency. This actually is a good example that highlights how we can't look at any of the Ohm's law factors in a vacuum when it comes to amp-speaker behaviour.

Assuming I have that correct, of course. Andy?

I agree on your comment about agaedhorse, who appears to be the "real deal."

There are obviously some other very smart, and very funny people here, as well.

In fact, the voltage known as Back EMF is simply a BLv voltage. It is in proportion to, and in phase with voice coil velocity in the gap. While the motion itself can be affected by damping factor, in turn affecting that BLv voltage, the voltage is actually independent of amplifier impedance, and proportional only to velocity, as with any other conductor immersed in a field.

Andy

 

[acmebass]

Missed the back emf comments... Back emf is in fact energy that can be reflected or absorbed. The combination of source impedance (inverse if damping factor) an any distributed loads affect this. Back emf superimposes (sums) on the original signal (subject to reflection and absorption) and the result gets fed back along with the original signal. Generally it's a small contribution but it could be significant in some instances.

Where it can become a really big problem is in VI limiting circuits with substantial phase shift between V & I. In this case, the feedback loop is opened under limiting conditions and interrupting or clamping the drive signal (voltage) can result in a large inductive spike being generated due L x dI/dT. This can destroy an amp if mechanisms are not in place to absorb these spikes.

Back EMF is a BLv voltage which is, as I said, proportional to, and in phase with, voice coil velocity, as with any other conductor immersed in a field.

It is helpful to look at the example of a DC motor, in which the motor's resistance is equal to the resistance of the windings, plus the Back EMF, which as a voltage proportional to velocity/current, can be expressed in ohms (R = E/I). This DC example presents a purely resistive Back EMF to the voltage source, which can't tell the difference between winding resistance and Back EMF resistance. An ohm is an ohm.

The same is true in a moving coil driver, but it's an AC motor, not a DC one. The "slip" between the input voltage and the resulting cone velocity, due to the inertia of the moving mass, results in a phase shift between the input voltage and the impedance, known as a "reactance." In the case of a complex impedance, an ohm is still an ohm, phase shifted or not.

There are three modes of Back EMF generation around resonance, because of the increasing "shift" as the frequency is increased. Below resonance, Back EMF is inductive. At resonance, it is purely resistive, because the "slip" is actually 180 degrees, as I recall (it's been a few years since I worked out all these sinusoidal relationships for an article which is well over 20 pages, and incomplete as I waste my time here... ). Above resonance, the Back EMF becomes capacitive.

But an ohm is still an ohm, albeit a vector quantity in the case of the reactive Back EMF generated in the gap of a moving coil loudspeaker.

But it is not "energy," is a voltage, which manifests itself as a reactance. Around resonance, where actual VC inductive reactance is negligible, a driver's impedance is equal to resistance, plus Back EMF. That's it. Just like in a DC motor, only AC.

This could be a different way of saying what you say, that is "superimposes itself on the original signal," but that's not how I would put it.

Thanks,
Andy

 

[agedhorse]

I agree on your comment about agaedhorse, who appears to be the "real deal."

There are obviously some other very smart, and very funny people here, as well.

In fact, the voltage known as Back EMF is simply a BLv voltage. It is in proportion to, and in phase with voice coil velocity in the gap. While the motion itself can be affected by damping factor, in turn affecting that BLv voltage, the voltage is actually independent of amplifier impedance, and proportional only to velocity, as with any other conductor immersed in a field.

Andy

Sort of...

Because there is a resistive component to the voice coil impedance, this is in series with the reactive component, and factors into the voltage that is seen at the output terminals ofthe amplifier. It's a "simple" voltage divider, with the resulting voltage being multiiplied by the ratio of output impedance to VC series resistance. With a typical output impedance of 0.01 ohms, the DCR is a very big number and say .01/5 = .002 so the resulting voltage would be Vinductive x .002.

Where it becomes a problem is under a clipping event where the global feedback is effectively removed and the output impedance goes up to say 0.1 or 0.5 ohms, the contributing factor is maybe now roughly 10 to 50 times higher. Now begins the task of managing clipping recovery and the resulting artifacts.

 

[acmebass]

Sort of...

Because there is a resistive component to the voice coil impedance, this is in series with the reactive component, and factors into the voltage that is seen at the output terminals ofthe amplifier. It's a "simple" voltage divider, with the resulting voltage being multiiplied by the ratio of output impedance to VC series resistance. With a typical output impedance of 0.01 ohms, the DCR is a very big number and say .01/5 = .002 so the resulting voltage would be Vinductive x .002.

Where it becomes a problem is under a clipping event where the global feedback is effectively removed and the output impedance goes up to say 0.1 or 0.5 ohms, the contributing factor is maybe now roughly 10 to 50 times higher. Now begins the task of managing clipping recovery and the resulting artifacts.

Yeah, I'd certainly go along with that, although I must plead ignorance when it comes to the fine points of how an amplifier reacts (no pun intended) to this complex, widely varying impedance.

A

 

[RockNRollAl]

A DI schematic design and OT is running into resistive load whereas a tube amp design and his OT is running into reactive load, that is the difference.
The so called "vierquadrantenbetrieb" of an power amplifier.
Try to translate for yourself. I failed to find out a proper translation on the web.

I read it as "four quadrant operation". It makes sense to me as it takes into account both the resistive and reactive components of a load.

 

[fdeck]

How speakers work, including back emf. I'm working on an improved version of this.

Invalid Link Removed

 

[acmebass]

How speakers work, including back emf. I'm working on an improved version of this.

Invalid Link Removed

Wow! Hats off to fdeck!

I think you could probably expand that to about 200 pages, if you wanted to really go in depth. I'm not sure about the efficacy of your conversions- it's a little like using a Spanish-English dictionary.

You don't go into Back EMF deeply, but what you have looks spot-on to me, if I understand.

Thanks,
Andy

 

[fdeck]

Wow! Hats off to fdeck!

I think you could probably expand that to about 200 pages, if you wanted to really go in depth. I'm not sure about the efficacy of your conversions- it's a little like using a Spanish-English dictionary.

You don't go into Back EMF deeply, but what you have looks spot-on to me, if I understand.

Thanks,
Andy

Thanks. I plan to fill in more background on some things such as back emf, rather than just stating them. All of this emerged from my interest in understanding speakers myself. Thiele-Small is "speaker theory for filter jocks" which I am not, resulting in why I prefer to make a direct appeal to the physics. It arrives at the same result, but offers an alternative view of how to get there.

 

[P Town]

Mr. Lewis,

I am trying to make sense of this statement:

"Above resonance, the Back EMF becomes capacitive."

Is this because the voltage of the back EMF is 180 degrees out of phase of the input signal, (which is causing the motion, and therefore opposes the input signal voltage)?

This question is coming from a person who last laid eyes on the formulas for inductive, and capacitive reactance in 1976, and who got lost rather quickly in Mr. Decks treatise. So I may in over my head in even asking.

Thanks.

 

[agedhorse]

The reactance of an impedance plot can be represented graphically by slope. Flat slope is resistive, rising slope is inductive, falling slope is capacitive.

Thinking of a good analogy for back-EMF, I came up with the following (and there are several mechanisms):

1. Driving a speaker VC with a current that resides inside a magnetic field, when the current is removed the speaker returns to it's neutral bias and in doing so represents a generator with a moving coil inside a static magnetic field. It's the conversion of stored electro-mechanical energy.

2. Driving a fixed coil (of any type) with a current within or outside of a magnetic field causes electro-magnetic energy to be stored in the field and when the current is interrupted, a voltage will be impressed on the coil's terminals as the field collapses. The voltage = L x dI/dT, so the faster the switchoff of the current, and the larger the inductance, the larger impressed voltage.

Now this is going back to theory I haven't really thought about in many years, so pardon the rusty approach.

 

[B-string]

Since it is not an absence of voltage, but a controlled collapse the back EMF presents out of phase to the signal voltage yes?

 

[P Town]

I initially assumed the back emf voltage would be 180 degrees out of phase with the input voltage, but I have a feeling it is not that simple.

 

[agedhorse]

I initially assumed the back emf voltage would be 180 degrees out of phase with the input voltage, but I have a feeling it is not that simple.

Correct, while a purely inductive back EMF component would be inverted, all the other elements come into play in real world circuits, looking from one direction can be quite different than looking from the opposite direction. Think about an RC filter... one way the R is in series with the C in shunt but the other way it looks almost purely capacitive.

 

[Major Softie]

This thread has totally geeked out.

Not that there's anything wrong with that!

 

[ThisBass]

In fact, the voltage known as Back EMF is simply a BLv voltage. It is in proportion to, and in phase with voice coil velocity in the gap.

with v=dx/dt this is BL*dx/dt

BL = costant, dx/dt varies depending on a given signal waveform (the moving of the voice coil in the field).


DC => Zero
Square Wave => Dirac Pulse (periodically +/- spikes)
Triangle => Square Wave with pi/2 phase shift
Sawtooth => Dirac Pulse + DC (rough worded)
Sine Wave => Sine Wave with pi/2 phase shift (which is cosinus)

I have just this moment no idea how the back emf waveform looks like for a bass guitar signal but it seems to be a very complex one.

Considering the standard waveforms like the sawtooth it is the sawtooth that equals most of all to an audio signal.
Like every audio signal the sawtooth has got a asymmetrical waveform and contains even and odd Harmonics.
Other waveforms like triangle and square are symmetrical and contain only odd Harmonics.

 

[Mark Reccord]

I agree on your comment about agaedhorse, who appears to be the "real deal."

Andy

And his name is Andy too.

I think we're all on the same page re back-emf, with slight differences in perspective.

How speakers work, including back emf. I'm working on an improved version of this.

Invalid Link Removed

I've said it before and I'll say it again: This is an excellent resource for anyone who wants to really get into the nuts and bolts of speaker behaviour.

 

[acmebass]

Mr. Lewis,

I am trying to make sense of this statement:

"Above resonance, the Back EMF becomes capacitive."

Is this because the voltage of the back EMF is 180 degrees out of phase of the input signal, (which is causing the motion, and therefore opposes the input signal voltage)?

This question is coming from a person who last laid eyes on the formulas for inductive, and capacitive reactance in 1976, and who got lost rather quickly in Mr. Decks treatise. So I may in over my head in even asking.

Thanks.

I think agedhorse fielded that one. It's all about degree of phase shift between the input voltage and the current. Above resonance the voltage lags behind the current, which is the definition of capacitive.

BTW, upon further reflection, and without digging out my notes, i think the "slip" at resonance is 90 degrees. But don't hold me to it. I hope to get that stuff out before long and finish my work on it. Believe it or not, the two articles I did have published were before my business got busy, and two that I was working on at that time, including this, were shelved at the time, and have never seen the light of day, which really bums me out. I have a lot of work in that stuff.

Andy

 

[acmebass]

The reactance of an impedance plot can be represented graphically by slope. Flat slope is resistive, rising slope is inductive, falling slope is capacitive.

Thinking of a good analogy for back-EMF, I came up with the following (and there are several mechanisms):

1. Driving a speaker VC with a current that resides inside a magnetic field, when the current is removed the speaker returns to it's neutral bias and in doing so represents a generator with a moving coil inside a static magnetic field. It's the conversion of stored electro-mechanical energy.

2. Driving a fixed coil (of any type) with a current within or outside of a magnetic field causes electro-magnetic energy to be stored in the field and when the current is interrupted, a voltage will be impressed on the coil's terminals as the field collapses. The voltage = L x dI/dT, so the faster the switchoff of the current, and the larger the inductance, the larger impressed voltage.

Now this is going back to theory I haven't really thought about in many years, so pardon the rusty approach.

I think you nailed it.

A

 

[acmebass]

This thread has totally geeked out.

Not that there's anything wrong with that!

A deep understanding of the nature of Back EMF is utterly unnecessary when designing a loudspeaker.

Or playing music, for that matter.

I suppose that would be the definition of geeked out.

A

 

[acmebass]

with v=dx/dt this is BL*dx/dt

I have just this moment no idea how the back emf waveform looks like for a bass guitar signal but it seems to be a very complex one.

The waveform of the Back EMF is an exact reflection of instantaneous VC velocity, dx/dt, if you will.

Designers of servo-corrected systems will sometimes drop an extra turn or two into the gap, to provide an exact real-time picture of what the voice coil is doing, such that the servo circuits can can compare it to the input, and correct.

This signal would be EMF. The adjective "back" refers to the phase of the EMF generated by the voice coil, so to get geeky, one is Back EMF, and the other is just regular ol' EMF.

Andy