The most obvious one is thermal compression: As the voice coil heats up, its resistance increases. With a voltage source, this means less current drawn and thus loud passages will be lowered in amplitude. A current source will not be affected by this type of distortion.janneman said:
Mr Evil said:
Huh? The voice coil moves as a result of ampere-windings through the magnetic field. If you have the same sound level, you have the same current, and the same heating, and the same compression. Whatever the driving source Zout.
You have another one?
Jan Didden
No, if the voice coil resistance increases then for a given input voltage the current will be less, and the SPL decreases with it. With current drive, the current will be independant of the resistance, thus sound output will be not be compressed.janneman said:
In the long past, Audio magazine did nice lodspeaker's tests, including compression.
I don't recall it as a significant problem with decent speakers.
I don't recall it as a significant problem with decent speakers.
Mr Evil said:
Evil,
You're mixing up two variables. With current drive, in your example, the current will be higher than with voltage drive, thus heating and compression are HIGHER. So, in your example, compression is higher with current drive, not lower. Current drive increases distortion here.
But also, the sound output level will be higher than with voltage drive. If you adjust the level for same sound level, you will see the same current, the same heating, the same compression.
Jan Didden
janneman said:
Sorry Jan, but Mr Evil is right here.
With voltage drive, the increase in VC resistence due to heathing will lower VC current, while with current drive the cureent will be constant, independent of VC temperature/resistence.
But, as I've posted before, I'm not aware of speaker drivers designed for current drive...
I didn't say it made a worthwhile difference, I was just pointing out that it is possible to have one of either current or voltage waveforms distorted without that distortion necessarily appearing at the output.Jorge said:
The change in voice coil resistance will be higher, yes, since the current doesn't decrease with temperature (which could be a problem), but the compression is lower because the output level is no longer affected by temperature. Certainly the sound output with current drive will be higher, because that with voltage drive is reduced below what it should be.janneman said:
You don't have to take my word for it, I'm just going by what I've read in various places, such as Hawksford's paper: 'DISTORTION REDUCTION IN MOVING-COIL LOUDSPEAKER SYSTEMS USING CURRENT-DRIVE'.
Jorge said:
Jorge,
I agree with what you say here, and that was also Evil's point. But higher vc current (due to current drive) means higher heating, doesn't that then increase compression??
Jan Didden
Jorge said:
Jorge, Evil,
come on.
Follow your own words to next step.
Current drive will keep the current constant, no matter which temperature and resistance the coil will have. Right.
But with increased temp and resistance a current drive will have to
provide more voltage. So current drive is compensating the speaker inherent thermal sound compression and sinking SPL simply by increasing the output voltage.
But the speaker itself remains unchanged. It's basic relations between voltage/current/cone movement are unchanged.
No matter which parameter you will regulate, the other two will
result. And of course every version will sound different...
Jan
First, I really don't follow your viewpoint that current drive WILL increase heathing. Could you pls clarify?
And no, even with more heathing there will not be compression (we are talking of current drive here, not low damping amps - some like Mauro's current output stage out of a NFB loop).
If the drive is current based, the heathing will increase resistence, but then the current source wil increase it's voltage so current remains the same, and there will not be compression.
First, I really don't follow your viewpoint that current drive WILL increase heathing. Could you pls clarify?
And no, even with more heathing there will not be compression (we are talking of current drive here, not low damping amps - some like Mauro's current output stage out of a NFB loop).
If the drive is current based, the heathing will increase resistence, but then the current source wil increase it's voltage so current remains the same, and there will not be compression.
It is not true to say there will be higher vc current (due to current drive). There will be no fall in vc current and there will be more heating and increased resistance leading to higher power dissipation and eventually to vc destruction but there will be no compression as the current is independent of vc resistance
Stuart
Stuart
...if you kep current constant and resistance is increasing...
P=R*I^2 will make more heat with stable current than with decreasing current.
P=R*I^2 will make more heat with stable current than with decreasing current.
Re: Jan
Yes, but the constant current means that as the vc heats up, its R increases, but the current keeps constant, so power = i^2*R so it even heats more, so the amp Vout goes up to keep the current constant so the increased R with the same current increased the heat even more so you have maybe a runaway condition.
Ahhhh! You're right! The force = B*i*L, B = magnet dependent = constant, L = # of windings * winding area and is constant, i is constant so force is constant. Shoot! I think I goofed! Blahhh.
Jan Didden
Jorge said:
Yes, but the constant current means that as the vc heats up, its R increases, but the current keeps constant, so power = i^2*R so it even heats more, so the amp Vout goes up to keep the current constant so the increased R with the same current increased the heat even more so you have maybe a runaway condition.
Ahhhh! You're right! The force = B*i*L, B = magnet dependent = constant, L = # of windings * winding area and is constant, i is constant so force is constant. Shoot! I think I goofed! Blahhh.
Jan Didden
Hi darkfenriz,
so what? the efficiency is 0.5% also for the useful signal, not only for the distorsion.
again, I don't see the point. If you have distorsion on the voltage you will have distorsion on velocity; if you have distorsion on the current, the force will suffer distorsion.
Either way, both velocity and force "concour" to form the acustic output thus both the distorsion of voltage and current will somehow "enter" in the sound output anyway. No matter how you drive the speaker.
From this perspective, what may (should) be discussed is wether it is preferable to have more distorsion on velocity or force (which is again the same as discussing wether it is better to have more distorsion on current or voltage...), i.e. how and how much that will impact the perceived sound quality, which after all is the only thing that really matters.
Even more interesting, is to see if we can minimize distorsion *on the sound*, i.e. to see if there is some "mid way" position where we'll have some distorsion on both the voltage and the current, buth the net result of the combination of the two after conversion will be less than that produced by either voltage or current drive.
Given the non-trivial relationships between current, voltage and sound, again I stress that possibly the only way to do this is by looking directly at the FFT of the actual sound output while gradually changing the output impedance of the source from ~0 to very high...
Mauro?! ...hai per caso un microfono a portata di mano? 😉
Hi Graham,
(I would replace the last 'or' with an 'and'... 😉 )
BTW, I guess you're right... could that be an explanation for the "harsh" sound which is typical of most (if not all) SS amplifiers? maybe...
Indeed, the harshness on trebles sounds a lot like a very fast ringing, to me.
(as a side: among those I had a chance to listen to, there were really a few SS amplifiers which did not sounded harsh and/or dull or anyway bad to me... maybe really only ONE: a ridiculously expensive top-of-the-line "Class A" monoblock monster from MBL... which costs like an apartment and draws up to 2.6KW -more than 5KW for a stereo setup!- from the mains... a really good true class A triode amplifier costs way much less and -amazingly?!- is both much smaller and requires way much less power than that monster... yet it sounds at least as good as it does!).
darkfenriz said:
so what? the efficiency is 0.5% also for the useful signal, not only for the distorsion.
again, I don't see the point. If you have distorsion on the voltage you will have distorsion on velocity; if you have distorsion on the current, the force will suffer distorsion.
Either way, both velocity and force "concour" to form the acustic output thus both the distorsion of voltage and current will somehow "enter" in the sound output anyway. No matter how you drive the speaker.
From this perspective, what may (should) be discussed is wether it is preferable to have more distorsion on velocity or force (which is again the same as discussing wether it is better to have more distorsion on current or voltage...), i.e. how and how much that will impact the perceived sound quality, which after all is the only thing that really matters.
Even more interesting, is to see if we can minimize distorsion *on the sound*, i.e. to see if there is some "mid way" position where we'll have some distorsion on both the voltage and the current, buth the net result of the combination of the two after conversion will be less than that produced by either voltage or current drive.
Given the non-trivial relationships between current, voltage and sound, again I stress that possibly the only way to do this is by looking directly at the FFT of the actual sound output while gradually changing the output impedance of the source from ~0 to very high...
Mauro?! ...hai per caso un microfono a portata di mano? 😉
Hi Graham,
Graham Maynard said:
(I would replace the last 'or' with an 'and'... 😉 )
BTW, I guess you're right... could that be an explanation for the "harsh" sound which is typical of most (if not all) SS amplifiers? maybe...
Indeed, the harshness on trebles sounds a lot like a very fast ringing, to me.
(as a side: among those I had a chance to listen to, there were really a few SS amplifiers which did not sounded harsh and/or dull or anyway bad to me... maybe really only ONE: a ridiculously expensive top-of-the-line "Class A" monoblock monster from MBL... which costs like an apartment and draws up to 2.6KW -more than 5KW for a stereo setup!- from the mains... a really good true class A triode amplifier costs way much less and -amazingly?!- is both much smaller and requires way much less power than that monster... yet it sounds at least as good as it does!).
Hello, Graham
The fact that some 99% of energy applied to a normal speaker is transformed in heath (mostly in the VC) wouldn't help here?
Graham Maynard said:
The fact that some 99% of energy applied to a normal speaker is transformed in heath (mostly in the VC) wouldn't help here?
Graham,quote:
Originally posted by Graham Maynard
With ultra low impedance solid state output back-EMF that is no longer directly related to on-going drive remains trapped to ring within the cable/loudspeaker system, until re-transduced or dissipated by components or drivers.
This is such a load of gibberisch I don't even want to try to comment on it. I must hand it to you, Graham, you invent physics as you go. Impressive!
Jan Didden
Originally posted by Graham Maynard
With ultra low impedance solid state output back-EMF that is no longer directly related to on-going drive remains trapped to ring within the cable/loudspeaker system, until re-transduced or dissipated by components or drivers.
This is such a load of gibberisch I don't even want to try to comment on it. I must hand it to you, Graham, you invent physics as you go. Impressive!
Jan Didden
Rodolfo,
In this post
You are declaring the following:
Though very late, but I would like to question Your reasoning.
I would start with the last statement. "Excursion dependent variations are an issue only at low frequencies". You repeat this claim later on as well. And this is simply true. In case of pure sine test signals. When was it last time that You been listening, with candlelight & glass of vine, to your favorite test tones CD?
All the other cases are a composit of bass & delicate midrange tones, TOGETHER, and your cone is NEVER at it's midpoint, that is, moving through it's FULL excursion while that delicate violin solo is on... and having this B*L here, then changing for B'*L' there..
Then. "I bet for cone velocity control"
As much as I have tried, I did not find any reference in Hawksford's paper, which would say that sound pressure =acceleration=force=current. Instead I found equation (1), which claims:
u= (Vo*B*l) / [Zm[Zs+(B*l)square/Zm] ]
That is, CONE VELOCITY, as function of all above factors in case of VOLTAGE DRIVE.
No acceleration was mentioned.
Then, in equation (2), he arrives to the expression of CONE VELOCITY again, as follows:
u=Io*B*l /Zm
Here, in the case of CURRENT DRIVE, he notes how much simpler it gets, eliminating a lot of dependencies, all included in the term:
[Zs + (B*l)square / Zm]
Then he explains that this means: (B*l)square eliminated, so the displacement dependency decreased [as You have noticed as well]
- a Zm term eliminated, so "a reduced dependence on compliance nonlinearity within Zm"
- a Zs term [inductance included] eliminated, so "Thus at the high frequency end of the drive unit's operating range, the elimination of performance dependence on coil inductance modulation and eddy current losses is seen to be a valuable asset"
-In Zs also comprised / eliminated the Re voice coil heating effect/ consequences.
So, contrary to Your claims, these effects are not exclusively connected to displacement. [but drive current intensity] So they are perfectly valid at higher frequencies. And eliminated by current drive.
So much so, that the distorsion MEASUREMENT executed by the Hawksford team shows a "small" !!!! 26 dB !!!!! difference at 3 kHz!
So, as a [provocative] conclusion, and as a reply also to Your post 80# in this thread, I would like to reinforce Hawksfords claims:
There are no different points of view [ acceleration/force or cone velocity] - both of you are talking about cone velocity. And Hawksford gives a series of clues why current drive is eliminating a serious part of the distortions.
Ciao, George
In this post
You are declaring the following:
Though very late, but I would like to question Your reasoning.
I would start with the last statement. "Excursion dependent variations are an issue only at low frequencies". You repeat this claim later on as well. And this is simply true. In case of pure sine test signals. When was it last time that You been listening, with candlelight & glass of vine, to your favorite test tones CD?
All the other cases are a composit of bass & delicate midrange tones, TOGETHER, and your cone is NEVER at it's midpoint, that is, moving through it's FULL excursion while that delicate violin solo is on... and having this B*L here, then changing for B'*L' there..
Then. "I bet for cone velocity control"
As much as I have tried, I did not find any reference in Hawksford's paper, which would say that sound pressure =acceleration=force=current. Instead I found equation (1), which claims:
u= (Vo*B*l) / [Zm[Zs+(B*l)square/Zm] ]
That is, CONE VELOCITY, as function of all above factors in case of VOLTAGE DRIVE.
No acceleration was mentioned.
Then, in equation (2), he arrives to the expression of CONE VELOCITY again, as follows:
u=Io*B*l /Zm
Here, in the case of CURRENT DRIVE, he notes how much simpler it gets, eliminating a lot of dependencies, all included in the term:
[Zs + (B*l)square / Zm]
Then he explains that this means: (B*l)square eliminated, so the displacement dependency decreased [as You have noticed as well]
- a Zm term eliminated, so "a reduced dependence on compliance nonlinearity within Zm"
- a Zs term [inductance included] eliminated, so "Thus at the high frequency end of the drive unit's operating range, the elimination of performance dependence on coil inductance modulation and eddy current losses is seen to be a valuable asset"
-In Zs also comprised / eliminated the Re voice coil heating effect/ consequences.
So, contrary to Your claims, these effects are not exclusively connected to displacement. [but drive current intensity] So they are perfectly valid at higher frequencies. And eliminated by current drive.
So much so, that the distorsion MEASUREMENT executed by the Hawksford team shows a "small" !!!! 26 dB !!!!! difference at 3 kHz!
So, as a [provocative] conclusion, and as a reply also to Your post 80# in this thread, I would like to reinforce Hawksfords claims:
There are no different points of view [ acceleration/force or cone velocity] - both of you are talking about cone velocity. And Hawksford gives a series of clues why current drive is eliminating a serious part of the distortions.
Ciao, George
Hi Paulo,
Yes, and that is why I suggest separate crossover/driver cables right back to the near short circuit of a solid state amplifier output so that back-EMF from bass and mid sections cannot energise the tweeter due to shared cable impedance.
Also try mimicking valve amplifier impedance by trying a series 'Q' busting 0.22 ohm resistor, and if necessary one for each circuit.
I don't believe that solid state amplifiers are less musical than tube. They are different beasts and so need to be used differently too.
Hi Jan,
Maybe you not wanting to comment means you are unable to discuss your reasons for making such an accusation.
Remember 'C's and 'L's store energy and delay voltage change/current flow. Energy does not just disappear once it fed to a loudspeaker, and a SS amplifier resists its return..
Cheers .......... Graham.
Yes, and that is why I suggest separate crossover/driver cables right back to the near short circuit of a solid state amplifier output so that back-EMF from bass and mid sections cannot energise the tweeter due to shared cable impedance.
Also try mimicking valve amplifier impedance by trying a series 'Q' busting 0.22 ohm resistor, and if necessary one for each circuit.
I don't believe that solid state amplifiers are less musical than tube. They are different beasts and so need to be used differently too.
Hi Jan,
Maybe you not wanting to comment means you are unable to discuss your reasons for making such an accusation.
Remember 'C's and 'L's store energy and delay voltage change/current flow. Energy does not just disappear once it fed to a loudspeaker, and a SS amplifier resists its return..
Cheers .......... Graham.
yupp...
current drive probably has just one single disadvantage:
The damping from the electromagnetic side is lost.
Damping is defined by the mechanics of the speaker & enclosure.
So you would need dedicated driver designs which have high mechanical losses and these losses must be trimmed to be linear...
Hm, difficult. Especially if your are running the driver through its
resonance frequency as it is common for the woofer.
But it would be interesting to hear such a thing.
current drive probably has just one single disadvantage:
The damping from the electromagnetic side is lost.
Damping is defined by the mechanics of the speaker & enclosure.
So you would need dedicated driver designs which have high mechanical losses and these losses must be trimmed to be linear...
Hm, difficult. Especially if your are running the driver through its
resonance frequency as it is common for the woofer.
But it would be interesting to hear such a thing.
- Status
- Not open for further replies.