Here’s what I get from the Adire paper:
The response of a driver to an impulse was measured and the following three cases were compared:
1) Driver
2) Driver with mass added to the cone
3) Driver with inductor added in series with the coil.
Results:
1) Adding weight to a driver cone changed the HEIGHT of the high and low peaks, but not the time they occurred
2) Adding inductance to the driver changed WHEN the high and low peaks occurred.
My take-away:
1) More power can make a heavy cone respond just like a light cone.
2) There’s no work around for a high inductance driver to make it repond like a low inductance driver.
Comments? I don’t know what dynamics mean so please let’s discuss with respect to the impulse reponse graph.
All of us understand higher, lower, sooner, later, right?
The response of a driver to an impulse was measured and the following three cases were compared:
1) Driver
2) Driver with mass added to the cone
3) Driver with inductor added in series with the coil.
Results:
1) Adding weight to a driver cone changed the HEIGHT of the high and low peaks, but not the time they occurred
2) Adding inductance to the driver changed WHEN the high and low peaks occurred.
My take-away:
1) More power can make a heavy cone respond just like a light cone.
2) There’s no work around for a high inductance driver to make it repond like a low inductance driver.
Comments? I don’t know what dynamics mean so please let’s discuss with respect to the impulse reponse graph.
All of us understand higher, lower, sooner, later, right?
Let me continue the discussion with more impulse response concepts:
All else held constant,
1) If a driver’s BL was increased the HEIGHT of the peaks would increase.
2) If a drivers Mms was lowered the HEIGHT of the peaks would increase.
Do we agree that neither of these modifications changed the TIME at which the Peak and trough occur?
And back to the power, Do we agree that the lower BL driver’s response could be matched to the Higher BL driver’s response by adding power?
And also the Higher Mms driver’s reponse could be matched to the Higher Mms driver’s response by adding power?
My take-away: Buy the low inductance driver . Add more power if it wont play loud enough. High Mms and Low BL do not affect the time at which peaks occur, only the height of the peaks.
Again, please respond with respect to the impulse-response. Words like transients and dynamics confuse the situation.
All else held constant,
1) If a driver’s BL was increased the HEIGHT of the peaks would increase.
2) If a drivers Mms was lowered the HEIGHT of the peaks would increase.
Do we agree that neither of these modifications changed the TIME at which the Peak and trough occur?
And back to the power, Do we agree that the lower BL driver’s response could be matched to the Higher BL driver’s response by adding power?
And also the Higher Mms driver’s reponse could be matched to the Higher Mms driver’s response by adding power?
My take-away: Buy the low inductance driver . Add more power if it wont play loud enough. High Mms and Low BL do not affect the time at which peaks occur, only the height of the peaks.
Again, please respond with respect to the impulse-response. Words like transients and dynamics confuse the situation.
Explain again how the Bl product was changed? I don't believe they did this test. They were only measuring the effect of changing the cone mass. Adding inductance doesn't affect the Bl product. However, it does act like a low pass filter, limiting the uppermost frequency produced by the driver. This will most definitely affect the impulse response and the transient response. It doesn't say anything about Bl or Mms.
The impulse response and frequency response are duals of each other. If we change the cone mass, it most certainly changes the frequency response and it will most definitely change the impulse response as well.
If we keep the power constant, or current constant, the stronger motor should be able to stop a given cone before a weaker motor. The thing to look at here is not when the peaks and troughs occur but rather when the response decays to zero.
The impulse response and frequency response are duals of each other. If we change the cone mass, it most certainly changes the frequency response and it will most definitely change the impulse response as well.
If we keep the power constant, or current constant, the stronger motor should be able to stop a given cone before a weaker motor. The thing to look at here is not when the peaks and troughs occur but rather when the response decays to zero.
I'm not sure on #1. Changing BL means changing the inductance, no?
Driver inductance gives a phase change according to frequency, which would account for the time difference in when the spikes occur. (maybe)
Last edited:
If we keep the power constant, or current constant, the stronger motor should be able to stop a given cone before a weaker motor.
Currently, nobody tries to keep the power applied to drivers constant.
Keeping the current constant means no electrical damping.
The strength of a motor is not B*L but B*L* i.
For same L, B/2 and i *2, the same strength does not vary.
What changes is the efficiency and the electrical damping.
Apart in high power pro applications levels, the heating of the voice coil due to Joule's effect is not significant, even with low efficiency drivers.
Less electrical damping means higher Qe. The initial Qe can be retrieved using either a Transform, an amplifier with negative resistance output or a velocity feedback servo.
With adequate value of i, there is no change in the acceleration of the cone (acceleration = moving = starting and stopping) , the strength applied has not changed despite the lower B.
Many people, even some gurus, think that B*L/Mms and efficiency have a direct correlation with the transient response of drivers but never show it experimentaly.
Currently, nobody tries to keep the power applied to drivers constant.
Keeping the current constant means no electrical damping.
The strength of a motor is not B*L but B*L* i.
For same L, B/2 and i *2, the same strength does not vary.
What changes is the efficiency and the electrical damping.
Apart in high power pro applications levels, the heating of the voice coil due to Joule's effect is not significant, even with low efficiency drivers.
Less electrical damping means higher Qe. The initial Qe can be retrieved using either a Transform, an amplifier with negative resistance output or a velocity feedback servo.
With adequate value of i, there is no change in the acceleration of the cone (acceleration = moving = starting and stopping) , the strength applied has not changed despite the lower B.
Many people, even some gurus, think that B*L/Mms and efficiency have a direct correlation with the transient response of drivers but never show it experimentaly.
The quantity Qms, mechanical Q, is proportional to the mass of the cone. Bigger the mass, bigger the Q and the harder it is to damp it. In a spring mass system, if you increase the mass, you have lowered the resonant frequency. If you now displace the mass, the time it takes to sinusoidally decay to zero is more than that with a lighter mass, assuming the same spring and damping.
The damping force is the back EMF generated in the voice coil which is proportional to the velocity of the cone. But wait, it is also proportional to the flux density B, and the length of voice coil wire.
EMF = B.l.v
The damping provided by the electrical side, Qes, at resonance frequency depends on the flux density B and length of wire l, not v, because v will be the same for all drivers at the given resonant frequency.
Hence, the damping is most definitely proportional to the flux density and therefore, the time required to stop the cone from moving is also dependent on the magnet strength.
You don't have to believe me or anybody else. Just look at the equations and the fundamental principles.
The damping force is the back EMF generated in the voice coil which is proportional to the velocity of the cone. But wait, it is also proportional to the flux density B, and the length of voice coil wire.
EMF = B.l.v
The damping provided by the electrical side, Qes, at resonance frequency depends on the flux density B and length of wire l, not v, because v will be the same for all drivers at the given resonant frequency.
Hence, the damping is most definitely proportional to the flux density and therefore, the time required to stop the cone from moving is also dependent on the magnet strength.
You don't have to believe me or anybody else. Just look at the equations and the fundamental principles.
Last edited:
Hmm, transient [impulse] response is how well the driver can track the attack/decay of a signal, with the amp's slew rate being the ultimate limiting factor of how fast its ability to 'let go' of the signal, so for this then one needs a truly HIFI [ultra-wide BW] amp with the appropriate slew rate to cover the highest amplitude [‘fastest’] transients in recorded music. A driver can be transient perfect [0.5 Qts] in its rising response [acceleration] BW, yet have a very limited mass controlled [nominally flat] BW due to high inductance.
If a driver’s Qts then is <0.5, it’s intrinsically over-damped and will fall short of tracking the signal, leaving a time delayed ‘hole’ in the response if not increased via an appropriate box and/or filter loading; ergo a Qts >0.5 is under-damped and will overshoot the ‘let go’ point [‘ring’], delaying its ability to handle the next signal on time and/or overlapping [comb filtering] with it depending on a variety of factors.
In the driver’s mass controlled [falling response] BW, inductance sets its inherent transient response, as shown by the so-called ‘poor logic’ of Adire’s tech article, so the driver
GM
You're welcome!
Just barely, I'm where you were the other year with having to deal with major property long term 'wear n' tear'/storm damage repair plus my computer and back-up drive got ruined while doing a back-up in what appears to have been a motherboard meltdown, so with my memory and fairly large/diverse 'library' mostly MIA these days, I won't be posting much in the way of technical posts anymore if the latter can't be retrieved. The local data recovery service quote was high enough to make me wonder if it was worth it overall, but a forum member offered to try, saying he's had a bit of luck doing it before, so fingers X.
Otherwise, same ol' aces n' eights, with house repairs moving at a snail's pace due to my various infirmities, further slowed by some having to be done more than once since new storms would damage on-going repairs. Thankfully, I haven't suffered any of the [repeated] major property and/or bodily damage that Mother Nature has bestowed on my fellow metro Atlantans and many other locales around the country/world. At least all my roofing/leaking issues finally seem fixed, so have moved onto the more visible damage, i.e. siding, soffit, fascia, door/window framing, etc. repairs/replacements.
Anyway, for now, best to PM here as my yahoo mail still has 'issues' and haven't taken the time to get my bellsouth.net account functioning again.
GM
Just barely, I'm where you were the other year with having to deal with major property long term 'wear n' tear'/storm damage repair plus my computer and back-up drive got ruined while doing a back-up in what appears to have been a motherboard meltdown, so with my memory and fairly large/diverse 'library' mostly MIA these days, I won't be posting much in the way of technical posts anymore if the latter can't be retrieved. The local data recovery service quote was high enough to make me wonder if it was worth it overall, but a forum member offered to try, saying he's had a bit of luck doing it before, so fingers X.
Otherwise, same ol' aces n' eights, with house repairs moving at a snail's pace due to my various infirmities, further slowed by some having to be done more than once since new storms would damage on-going repairs. Thankfully, I haven't suffered any of the [repeated] major property and/or bodily damage that Mother Nature has bestowed on my fellow metro Atlantans and many other locales around the country/world. At least all my roofing/leaking issues finally seem fixed, so have moved onto the more visible damage, i.e. siding, soffit, fascia, door/window framing, etc. repairs/replacements.
Anyway, for now, best to PM here as my yahoo mail still has 'issues' and haven't taken the time to get my bellsouth.net account functioning again.
GM
Ummm.... no, not quite. An overdamped system is what you want. The impulse signal is a pulse of infinite amplitude and zero width. Once the pulse is applied and removed, the system decays to zero. But the system never stops right at the moment when the impulse is removed. An overdamped system will not leave a hole behind, instead it will overlap the least with the next incoming signal.
Anyway, I think the point GM is making is that B.l and Mms do play a role in the how one driver operates compared to another in the time domain. You can get two drivers to do the same thing ultimately, but taken just by themselves the drivers' time domain responses do depend on their magnet strengths and cone masses.
EDIT: wow! I hope things get back to normal soon, GM.
Anyway, I think the point GM is making is that B.l and Mms do play a role in the how one driver operates compared to another in the time domain. You can get two drivers to do the same thing ultimately, but taken just by themselves the drivers' time domain responses do depend on their magnet strengths and cone masses.
EDIT: wow! I hope things get back to normal soon, GM.
Hence, the damping is most definitely proportional to the flux density and therefore, the time required to stop the cone from moving is also dependent on the magnet strength.
[...]
You don't have to believe me or anybody else. Just look at the equations and the fundamental principles.
Have you wondered why there are no experimental proofs of what you claim ?
There are the fundamental principles and the reasoning about them, without forgetting that the force is controlled by one more variable than those you consider : the current.
[...]
You don't have to believe me or anybody else. Just look at the equations and the fundamental principles.
Have you wondered why there are no experimental proofs of what you claim ?
There are the fundamental principles and the reasoning about them, without forgetting that the force is controlled by one more variable than those you consider : the current.
But not solely by current. For a given driver, yes, it depends on current only. But when comparing one driver to another, we must also account for differences in magnet strength and mass of the cone.
Ra7,
Take a twin coil driver.
Run it with both coils in parallel and measure the parameters and response.
Then run it driving one coil only. Measure the new parameters. This halves the BL. Use a Liknwitz Transform to restore the initial Qe (hence Qt).
Drive it with twice the voltage to restore the initial level and measure the new response.
What differences did you expect before the experience ? Do they occur ?
Take a twin coil driver.
Run it with both coils in parallel and measure the parameters and response.
Then run it driving one coil only. Measure the new parameters. This halves the BL. Use a Liknwitz Transform to restore the initial Qe (hence Qt).
Drive it with twice the voltage to restore the initial level and measure the new response.
What differences did you expect before the experience ? Do they occur ?
Last edited:
forr, I don't own a dual coil driver. I wish I could try this. Can you tell me what happens?
Also, the Linkwitz transform works on the system Q, not the driver Q. It is basically EQ to get more bottom end out of a given enclosure at the expense of power and driver excursion. Nevertheless, I would be interested in the results.
I'm not convinced by what's been said so far. The equations definitely point towards the fact that a stronger magnet (independent of current) and lower cone mass result in higher efficiency. And all else being equal, which it rarely is, I'd take a higher efficiency driver anytime.
Also, the Linkwitz transform works on the system Q, not the driver Q. It is basically EQ to get more bottom end out of a given enclosure at the expense of power and driver excursion. Nevertheless, I would be interested in the results.
I'm not convinced by what's been said so far. The equations definitely point towards the fact that a stronger magnet (independent of current) and lower cone mass result in higher efficiency. And all else being equal, which it rarely is, I'd take a higher efficiency driver anytime.
Being a long time lurker and a self confessed complete noob I love to read these threads. They hold so much information but also so much tension.
The way I see it is you are all correct!! there I said it.
Each and every driver is a compromise of cost and target market. All we need to do is find the drivers that are aimed at us! Simple.
The drivers parameters mean nothing unless you have already determined How you want to mount them, how you are going to drive them and what they are going to being playing.
The probable reason there is no scientific data to back up your arguments is because there are NO constants. I could take any of your perfect response/transient/efficient drivers and make them sound slow and absolutely awful by mounting them in an inappropriate enclosure and under/overpower it playing awfully recorded music! am I wrong??
The way I see it is you are all correct!! there I said it.
Each and every driver is a compromise of cost and target market. All we need to do is find the drivers that are aimed at us! Simple.
The drivers parameters mean nothing unless you have already determined How you want to mount them, how you are going to drive them and what they are going to being playing.
The probable reason there is no scientific data to back up your arguments is because there are NO constants. I could take any of your perfect response/transient/efficient drivers and make them sound slow and absolutely awful by mounting them in an inappropriate enclosure and under/overpower it playing awfully recorded music! am I wrong??
Shall do Greg, will be in touch over the next day or so.
That's not under question. They do. And all other things being equal, high efficiency is preferable to lower efficiency.
Indeed, to the extent that they are constituant parts of what make up a driver's Qt. Also, as GM notes (clarifying my point above) within the driver's mass controlled BW, inductance sets the inherent transient response.
Last edited:
Hi,
I came to look at these matters in the following way, trying to sort things out (look here for a good reference) :
In a given situation for a driver, say mounted in a closed box, we can safely assume that whenever the cone movement (and acoustical output) is the same, current in the voice coil, as well and voltage, must also be same. It does not matter whether we inject energy more current-based (high-Z drive) or more voltage-based (amplifier Zout = 0, or even negative up to -Re). In fact we can record the current's impulse response under normal voltage drive and then feed the convolved signal into a current output amp, and the result (frequency response etc) will be the same... notably under linear, small signal conditions.
What is not the same, and it is the only thing, is how the induced voice coil voltage, the microphonic voltage, enters the scenario ("Back EMF" is not my preferred expression as it describes the effect in voltage drive, not its root cause).
With pure current drive, the microphonic voltage doesn't enter the picture at all. The driver's cone is simple presented, in a "fire-and-forget" fashion, a force B*l*i, to which it reacts by gaining or reducing velocity in true superposition with any motion already present. Any force on the cone causing VC movement, say from adjacent drivers, or cone breakup, or DC offset from nonlinearities, does not undergo any electrical damping.
With pure voltage drive (say Zout = -9/10 Re) things are radically different. Unless induced voltage does precisely track applied voltage, huge correction currents want to flow due to the "almost superconducting" VC, trying to establish an equilibrium. Any force trying to pull on the VC differently now creates a tremendously overdamped, almost clamped reaction to it. This is full local feedback, a control loop in action, while current drive is an "open-loop", steering-only approach.
Obviously this must have very different consequences in the way anything non-linear is handled, for these extreme cases at least. And there are tons of nonlinearities of many kinds.... IMHO some aspects of more or less perceived "dynamics" are buried here, not only in the plain transfer function.
Normal voltage drive happens to lie in between these drive extremes for good reasons, with moderate (and adequate) damping of basic system resonance(s) to a more or less aperiodic step response of cone displacement. At higher frequencies, less electrical damping can give better results (most drivers I examined did in fact). Fine-tuning of amplifier Zout so that it perfectly complements the driver's (and the user's) needs can have its merits, and IME a strong and linear motor is the best and most important thing to start with, right at the driver. Final target response is simply a matter of EQ.
- Klaus
I came to look at these matters in the following way, trying to sort things out (look here for a good reference) :
In a given situation for a driver, say mounted in a closed box, we can safely assume that whenever the cone movement (and acoustical output) is the same, current in the voice coil, as well and voltage, must also be same. It does not matter whether we inject energy more current-based (high-Z drive) or more voltage-based (amplifier Zout = 0, or even negative up to -Re). In fact we can record the current's impulse response under normal voltage drive and then feed the convolved signal into a current output amp, and the result (frequency response etc) will be the same... notably under linear, small signal conditions.
What is not the same, and it is the only thing, is how the induced voice coil voltage, the microphonic voltage, enters the scenario ("Back EMF" is not my preferred expression as it describes the effect in voltage drive, not its root cause).
With pure current drive, the microphonic voltage doesn't enter the picture at all. The driver's cone is simple presented, in a "fire-and-forget" fashion, a force B*l*i, to which it reacts by gaining or reducing velocity in true superposition with any motion already present. Any force on the cone causing VC movement, say from adjacent drivers, or cone breakup, or DC offset from nonlinearities, does not undergo any electrical damping.
With pure voltage drive (say Zout = -9/10 Re) things are radically different. Unless induced voltage does precisely track applied voltage, huge correction currents want to flow due to the "almost superconducting" VC, trying to establish an equilibrium. Any force trying to pull on the VC differently now creates a tremendously overdamped, almost clamped reaction to it. This is full local feedback, a control loop in action, while current drive is an "open-loop", steering-only approach.
Obviously this must have very different consequences in the way anything non-linear is handled, for these extreme cases at least. And there are tons of nonlinearities of many kinds.... IMHO some aspects of more or less perceived "dynamics" are buried here, not only in the plain transfer function.
Normal voltage drive happens to lie in between these drive extremes for good reasons, with moderate (and adequate) damping of basic system resonance(s) to a more or less aperiodic step response of cone displacement. At higher frequencies, less electrical damping can give better results (most drivers I examined did in fact). Fine-tuning of amplifier Zout so that it perfectly complements the driver's (and the user's) needs can have its merits, and IME a strong and linear motor is the best and most important thing to start with, right at the driver. Final target response is simply a matter of EQ.
- Klaus
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.