No worries, this is the first problem you think of when they hear about "drivers on strings" or the like. Drivers must be rigidly mounted not flapping around, right? Well it turns out that is not true, and decoupling them has several advantages.
In my 2022 article in AudioXpress I wrote a section about this very phenomenon and analyzed how much it influences the driver output. IMHO it is very informative. See below.
The frame of a decoupled, freely hanging driver will move in opposition to its cone movement under the effect of action-reaction. When the driver is hanging freely, the counter movement of the frame is completely linear and causes a reduction in the acoustic output generated by the cone. One concern that has been voiced to me is that the level of this reduction will be significant, so I will provide a brief analysis to address the issue. Newton’s Second Law states that acceleration depends on the force applied and the mass of the object (F=m*a) and Newton’s Third Law states that for every action there is an equal reaction, that is F1 = ─F2. The force F being applied is generated by the motor and this force pushes the cone against the other driver parts (frame, motor, etc). Since the driver parts are decoupled from the ground by the suspension system and are free to move, the motor force induces motion in them as well. Thus, we can write:
cone_mass * cone_acceleration = ─ driver_mass * driver_acceleration
Because the direction of the driver acceleration is opposite that of the cone acceleration, the negative sign can be dropped from the term on the right side and acceleration taken to mean the magnitude of the acceleration henceforth. Rearranging this we get
driver_acceleration = cone_mass * cone_acceleration / driver_mass
In this analysis, “driver” refers to everything except the cone, voice coil former, voice coil, and air load. Because the driver is accelerated away from the cone in the opposite direction, this reduces the net acceleration of the cone compared to when the driver is immobilized by a connection to the ground via a cabinet or baffle. The net cone acceleration will then be:
net_cone_acceleration = cone_acceleration - driver_acceleration
We define the loss_ratio as the ratio of the cone’s net acceleration to the acceleration without the loss from the driver acceleration:
loss_ratio = ( cone_acceleration - driver_acceleration ) / cone_acceleration
Substituting the expression for driver_acceleration and rearranging we get the result that:
loss_ratio = 1 ─ ( cone_mass / driver_mass )
Because the sound pressure in the far field is proportional to cone acceleration, the value calculated by the loss ratio also predicts the loss in far field sound pressure.
Let’s do the calculation for a lightweight neo magnet pro driver, the Eminence Deltalite II 2515: the driver weighs 2590 grams overall and the mass of the moving parts and the air load (Mms) is 79g resulting in a mass ratio of 36:1. Plugging in this ratio, we have loss_ratio = 1 - 1/36 = 0.972, which is approximately -0.25 dB. As an example of a home-audio 6” class driver, the SB Acoustics SB17MFC weighs 1550 grams overall and Mms is 12.3g. This represents a mass ratio of 126:1, which corresponds to -0.07 dB. These examples show that for both hi-fi and pro-audio drivers this source of sound pressure loss is typically much less than 1 dB and can be considered negligible. This analysis also highlights that the frame of the driver undergoes acceleration. It is useful to decouple it as much from other parts that can move, other drivers, or even the loudspeaker cabinet for boxed speakers since the induced vibrations will lead to cabinet panel resonances. For example, should the midrange and tweeter be affixed together but otherwise free to move, the tweeter’s position will undergo motion induced by the midrange’s reaction force, which would modulate the tweeter output and could cause doppler distortion. By decoupling drivers from one another within the wireframe scaffold this source of distortion can be eliminated.
To summarize, the frame/magnet is so much more heavy than the cone/moving parts that it works just fine.
Thanks for the in depth explanation Charlie!
Thanks for the in depth explanation Charlie!
I would think the Doppler distortion guys would be all over that -0.07db like a monkey on a football. One thing to consider is the wire frame, wire, loudspeaker system will vibrate harmonically with the music being played. This is probably trivial too but there will be some frequencies that are louder or softer due to the vibration. Tighten the wires and the frequency goes up. The midrange stand makes much more sense to me but there are advantages to baffle-less loudspeakers.
Doppler distortion is when a different frequency is modulating some other frequency or band of frequencies.
In the case of the hang-by-wire driver, the movement is completely the same as the cone movement and is linear. There is no doppler distortion generated.
In the case of the hang-by-wire driver, the movement is completely the same as the cone movement and is linear. There is no doppler distortion generated.
But surely some will argue there's loss of data, however small, and it's the finest details that will be lost. I don't know that's what happens; devil's advocate query. 😉
I think the most important aspect is that the vibrations are introduced to the enclosure structure and let the wooden panels swing.
I have made good experience to support the magnet of a big woofer in a 3-way from the back site. Just evaluated literally "by hand", the baffle resonances in the upper lows were much reduced, therefore the backside of the speaker got more swing. But overall result was cleaner sound, even had to revoice the xover a bit after that measure. Also made good experience to decouple the mid from the enclosure with viscoelastic sealing and hard rubber washers then for the mounting screws. Sound got a bit cleaner, and panel resonances were reduced. Typical panel resonances of wooden floorstanders are located somewhere the midrange. A critical location on the baffle is the thin part between midrange and tweeter cutout. Just crank up and feel with the fingers how it swings there without some stiffening or thick baffle. That's at least worth a small bracket inside.
I have made good experience to support the magnet of a big woofer in a 3-way from the back site. Just evaluated literally "by hand", the baffle resonances in the upper lows were much reduced, therefore the backside of the speaker got more swing. But overall result was cleaner sound, even had to revoice the xover a bit after that measure. Also made good experience to decouple the mid from the enclosure with viscoelastic sealing and hard rubber washers then for the mounting screws. Sound got a bit cleaner, and panel resonances were reduced. Typical panel resonances of wooden floorstanders are located somewhere the midrange. A critical location on the baffle is the thin part between midrange and tweeter cutout. Just crank up and feel with the fingers how it swings there without some stiffening or thick baffle. That's at least worth a small bracket inside.
If a sealed woofer was desired rather than OB, could one achieve the same (or similar) thing by hanging the enclosure from a frame? Also, do the wires want to be lossy like springs or taut and constraining?
The suspended speaker is going to move opposite of the membrane or cone but because it is suspended it will operate harmonically with the wire suspension and this will cause the woofer to move on its own. If the oscillations are big enough there will be 'Doppler' distortion because the woofer has inertia. Maybe a better example is when someone walks by the loudspeaker on a wooden floor the speaker will move, swing, or oscillate due to the floor moving and this will create distortion of some kind, what it's called is up for debate. You are talking about phase modulation distortion which is a kind of inter-modulation distortion that occurs due to the cone movement. With a speaker on strings you have to take into account the whole speaker moving not just the diaphragm. This might actually be Doppler distortion.
I understand that you are incredulous about my claims, however, you could simply provide your own analysis based on the physics of the problem, that calculates the doppler distortion and prove me wrong. Based on my own analysis I believe I am correct but I am more than willing to consider a counter-argument that has some rigorous math and physics behind it.
I am incredulous about your claims, but don't take it personally, I'm incredulous about everyone's claims. I'm a natural born skeptic. This is a problem of simple harmonic motion f=1/2pi sqrt(k/m), the suspended speaker on wires will oscillate or vibrate at it's natural frequency. The speaker doesn't have to be powered on to demonstrate this, just go give it a whack and it will ring like a bell. Using short tight wires you can get the frequency quite high so that it won't matter in the pass band.
The Doppler effect is the change in the frequency of a wave in relation to an observer who is moving relative to the source of the wave. What we hear in loudspeakers is a phase change as a higher frequency is modulated by a much lower frequency. You are the one that claims suspending speakers eliminates Doppler distortion, something that has not been proven to exist in loudspeakers https://sound-au.com/doppler.htm. So the first thing you have to do is prove Doppler exists for loudspeakers other than sirens on police cars. I believe the phenomenon is phase modulation distortion which is not Doppler effect. Reducing phase distortion by suspending a woofer is impossible in my opinion phase distortion will happen no matter the position or mounting of the loudspeaker, and hanging a loudspeaker may create more harmonic oscillations that are in and out of phase with the speakers output. So, back-up your claim of reduced distortion with some data.
Having said that, many baffle-less designers have their speakers on wires, the difference is they make no claims to reducing phase distortion but instead say it reduces diffraction up to f=1/4 the width of the speaker.
The Doppler effect is the change in the frequency of a wave in relation to an observer who is moving relative to the source of the wave. What we hear in loudspeakers is a phase change as a higher frequency is modulated by a much lower frequency. You are the one that claims suspending speakers eliminates Doppler distortion, something that has not been proven to exist in loudspeakers https://sound-au.com/doppler.htm. So the first thing you have to do is prove Doppler exists for loudspeakers other than sirens on police cars. I believe the phenomenon is phase modulation distortion which is not Doppler effect. Reducing phase distortion by suspending a woofer is impossible in my opinion phase distortion will happen no matter the position or mounting of the loudspeaker, and hanging a loudspeaker may create more harmonic oscillations that are in and out of phase with the speakers output. So, back-up your claim of reduced distortion with some data.
Having said that, many baffle-less designers have their speakers on wires, the difference is they make no claims to reducing phase distortion but instead say it reduces diffraction up to f=1/4 the width of the speaker.
@Arthur Jackson - Here is my opinion on what you wrote above:
You seem to be conflating two different types of motion. You posted a formula for "simple harmonic motion" but that is for a ball oscillating (up and down movement) on a SPRING - it's a completely different phenomenon and a type of motion that does not occur for a driver hung by a STRING (a pendulum, swinging front to back or side to side).
The natural resonant frequency of a pendulum is ONLY dependent on the length of the string. When you have for example 4 inches of string the natural resonant frequency of the pendulum is around 1.5 Hz. The longer the string the lower this frequency. So this will be far below, by an order of magnitude or two, the passband of the driver and this provides the decoupling effect of the motion of the driver. See:
https://www.omnicalculator.com/physics/simple-pendulum
I don't know where you got the idea that the string is tight, like in a piano. Even if you consider it under tension by the driver's own mass, this will not make it all that tight and the string or wire is just not going to behave like a guitar or piano string. So the string will not be a source of sound itself. What is causing the string to vibrate anyway? The only external force is from sound waves and they are extremely inefficient at inducing motion in a solid. There is no "whack" happening here...
As I said before, the movement of the entire driver is the same as the driver's own audio output. This can only add or subtract from the signal LEVEL, but not shift the frequency. There will be no more doppler distortion than there is in the rigidly mounted driver due to different tones that are present in its passband. From what you wrote I am not sure you understand what would cause doppler distortion and what would not.
You seem to be conflating two different types of motion. You posted a formula for "simple harmonic motion" but that is for a ball oscillating (up and down movement) on a SPRING - it's a completely different phenomenon and a type of motion that does not occur for a driver hung by a STRING (a pendulum, swinging front to back or side to side).
The natural resonant frequency of a pendulum is ONLY dependent on the length of the string. When you have for example 4 inches of string the natural resonant frequency of the pendulum is around 1.5 Hz. The longer the string the lower this frequency. So this will be far below, by an order of magnitude or two, the passband of the driver and this provides the decoupling effect of the motion of the driver. See:
https://www.omnicalculator.com/physics/simple-pendulum
I don't know where you got the idea that the string is tight, like in a piano. Even if you consider it under tension by the driver's own mass, this will not make it all that tight and the string or wire is just not going to behave like a guitar or piano string. So the string will not be a source of sound itself. What is causing the string to vibrate anyway? The only external force is from sound waves and they are extremely inefficient at inducing motion in a solid. There is no "whack" happening here...
As I said before, the movement of the entire driver is the same as the driver's own audio output. This can only add or subtract from the signal LEVEL, but not shift the frequency. There will be no more doppler distortion than there is in the rigidly mounted driver due to different tones that are present in its passband. From what you wrote I am not sure you understand what would cause doppler distortion and what would not.
A woofer moving +/-10mm@20Hz is going to move at up to 5km/h but slower most of the time and it will reverse. I have heard doppler on a motorcycle when gusts work around the fairing at speeds around and over 50km/h
Not that this proves anything, it might just be interesting.
Not that this proves anything, it might just be interesting.
But is it though?
Apart from the fact that they are both harmonic oscillators, you don't see up and down motion as different from front to back motion? Mass dependence versus none? Etc?
I have no idea what you are trying to say with your first sentence...
I have no idea what you are trying to say with your first sentence...
Gravity still acts as a compliance in the presence of mass and the energy increases in proportion to mass. Is a resonance just a resonance?
"Energy" is not really relevant - we are primarily interested in frequency of motion here, be it at resonance or otherwise. The period of a pendulum is independent of mass and there is no compliance type element in the system. The ball and spring are not and the spring is a compliant element. The fact that both undergo harmonic oscillation does not mean they are "the same" except in their behavior. None of this has anything to do with doppler distortion, which is what Mr Jackson seems "skeptical" about. I am trying to make him understand that the system in question does not have doppler distortion effects regardless of what he and you have observed about doppler effects with a motorcycle or police car. Those have nothing to do with this system.
You don't know what I think of doppler in speakers..
Gravity acts as a compliance in this case since energy in mass is converted to altitude, then back into motional energy where it again gains altitude in exchange.
Gravity acts as a compliance in this case since energy in mass is converted to altitude, then back into motional energy where it again gains altitude in exchange.
OK, I see, you are correct if/when "compliance" is the same as potential energy storage. Is it?
Don't really know or care what you think about doppler in speakers TBH. You did post something about a motorcycle so you seemed to be commenting on it? Not sure how that is related... because there is no doppler distortion in the driver hung by wires. That's what the discussion is about. Not police cars or motorcycles. The sound eminated from e.g. the horn of the police car (at one frequency) is made higher or lower by the DC effect of the movement of the car. This has nothing to do with the speaker in question...
I will give you a more pertinent example: a driver in a box. It's reproducing a pure tone (sine wave). You move the box back and forth (same axis of motion as the cone) at the same frequency. Any doppler? Can you show me with "maths"? Only consider the driver output and motion (e.g. not the effect of the cabinet moving).
Don't really know or care what you think about doppler in speakers TBH. You did post something about a motorcycle so you seemed to be commenting on it? Not sure how that is related... because there is no doppler distortion in the driver hung by wires. That's what the discussion is about. Not police cars or motorcycles. The sound eminated from e.g. the horn of the police car (at one frequency) is made higher or lower by the DC effect of the movement of the car. This has nothing to do with the speaker in question...
I will give you a more pertinent example: a driver in a box. It's reproducing a pure tone (sine wave). You move the box back and forth (same axis of motion as the cone) at the same frequency. Any doppler? Can you show me with "maths"? Only consider the driver output and motion (e.g. not the effect of the cabinet moving).
That was your original point, that separating the mid and tweeter reduced the potential for doppler due to the high frequencies not having to ride on different frequencies which are also lower and with more excursion.
Where is this going?
Where is this going?
Right, but Arthur Jackson seems to be promoting the idea the even for one driver (let's say just the midrange for example) that if that driver is freely hung from strings or wires so that it "moves" (because of the reaction force from the cone pushing against the air) when it is operating that will create doppler distortion. I am saying that it won't.
My little maths problem was an attempt to get the thinking process going about the physics of the problem. Let me explain how I envision it - as a superposition of two motions. You just add them up.
Model:
The driver is reproducing a pure sine wave. Then you move the cabinet at the exact same frequency and on the same axis as the cone is moving, following the same pure sine wave. The cone is pushing on the air to make sound. Let's ignore the cabinet itself in this "model".
Scenario 1:
If you moved the enclosure at exactly the same frequency and amplitude as the cone, but 180 degrees out of phase from it, the cone would become stationary in space from the point of view of a coordinate system based on something not moving, like the floor. Stationary cone WRT air = zero output.
Scenario 2:
Now change the cabinet movement to be in-phase with the cone and the same amplitude. The cone is now moving twice as far forwards and backwards against the air as seen from the coordinate system based on the non-moving floor. This will increase the amplitude and SPL produced by the driver. No new tones are introduced, and the frequency remains the same, so there is no doppler effect happening. The relative phase angle between the cone movement and the cabinet movement only influence how much SPL change is taking place.
Scenario 3:
Now instead of a pure sine wave the cone is playing a music type signal. At the same time the cabinet is moving with the same signal and some phase angle. I believe that the result will be the same, and only the SPL will be changed, and that no doppler effect will occur or at least no more than if the driver was rigidly mounted to the cabinet.
Now back to the real world driver on a string: in this case the driver frame and motor, etc. are moving out of phase from the cone but with much less amplitude. The movement of the frame and motor decrease the amount of motion the cone makes WRT to the room air and the SPL is reduced. This will not cause any increase in doppler distortion compared to the same driver mounted in a box. The movement is basically frictionless (friction with air is too low to be important) and because we are so far above the resonance frequency of the pendulum system, so I do not believe there will be any intertial component causing a relative phase difference other than 180 degrees.
If someone can show me, with maths, that the driver frame and motor move in some other way in this situation I would very much like to see that result.
I am sorry but I have been trying to make this point all along but was perhaps not very good at doing so.
My little maths problem was an attempt to get the thinking process going about the physics of the problem. Let me explain how I envision it - as a superposition of two motions. You just add them up.
Model:
The driver is reproducing a pure sine wave. Then you move the cabinet at the exact same frequency and on the same axis as the cone is moving, following the same pure sine wave. The cone is pushing on the air to make sound. Let's ignore the cabinet itself in this "model".
Scenario 1:
If you moved the enclosure at exactly the same frequency and amplitude as the cone, but 180 degrees out of phase from it, the cone would become stationary in space from the point of view of a coordinate system based on something not moving, like the floor. Stationary cone WRT air = zero output.
Scenario 2:
Now change the cabinet movement to be in-phase with the cone and the same amplitude. The cone is now moving twice as far forwards and backwards against the air as seen from the coordinate system based on the non-moving floor. This will increase the amplitude and SPL produced by the driver. No new tones are introduced, and the frequency remains the same, so there is no doppler effect happening. The relative phase angle between the cone movement and the cabinet movement only influence how much SPL change is taking place.
Scenario 3:
Now instead of a pure sine wave the cone is playing a music type signal. At the same time the cabinet is moving with the same signal and some phase angle. I believe that the result will be the same, and only the SPL will be changed, and that no doppler effect will occur or at least no more than if the driver was rigidly mounted to the cabinet.
Now back to the real world driver on a string: in this case the driver frame and motor, etc. are moving out of phase from the cone but with much less amplitude. The movement of the frame and motor decrease the amount of motion the cone makes WRT to the room air and the SPL is reduced. This will not cause any increase in doppler distortion compared to the same driver mounted in a box. The movement is basically frictionless (friction with air is too low to be important) and because we are so far above the resonance frequency of the pendulum system, so I do not believe there will be any intertial component causing a relative phase difference other than 180 degrees.
If someone can show me, with maths, that the driver frame and motor move in some other way in this situation I would very much like to see that result.
I am sorry but I have been trying to make this point all along but was perhaps not very good at doing so.
No more than if the driver was rigidly mounted to the cabinet, I would say (with the stated conditions)... This is the distinction and if we were unsure about this then there might be disagreement in other areas.