Le inductance Relevency in Subwoofers

[MacManNM]

I'm sorry Chris, but inductance is VERY important to woofers. Read this article.

http://www.adireaudio.com/Files/TechPapers/WooferSpeed.pdf

Paul

I was looking for that paper. Glad you found it.

 

[WmAx]

The fact is that inductance is proportional to reactance. The greater the reactance the greater the delta in impedance, also the phase variance is greater (not taking into account the dc resistance of the coil because it is always there). When placed in a room, peaks and nulls are going to be greater with a higher inductance driver. Thus a difference can be measured and heard.

If you are referring to a tube amplifier with an output transformer with several ohms output impedance, then yes, it will cause additional frequency response irregularities that can be of audible magnitude(s). If you are referring to the other 99.9% of amplifiers used in the world, then the output impedance is a small fraction of an ohm, and the frequency response will not audibly deviate, unless of course you use an absurd example of a speaker that has non-typical variations of impedance swings. But that would not be represenative of real-world typical products.

-Chris

 

[MacManNM]

If you are referring to a tube amplifier with an output transformer with several ohms output impedance, then yes, it will cause additional frequency response irregularities that can be of audible magnitude(s). If you are referring to the other 99.9% of amplifiers used in the world, then the output impedance is a small fraction of an ohm, and the frequency response will not audibly deviate, unless of course you use an absurd example of a speaker that has non-typical variations of impedance swings. But that would not be represenative of real-world typical products.

-Chris

The impedance swing is because of the inductance.
A speaker with a 2mH VC @60Hz will add 0.756 ohms to the total Z
A speaker with a 1.5mH VC @60 Hz will add 0.56 Ohms.

For a driver with a 2 ohm dc resistance that is a 38% change for the 2mH, and only an 18% change for the 1.5mH coil, this is almost a factor of 2! That is one example. I'm not even including any phase data in this, which is affected just as much.

 

[WmAx]

The impedance swing is because of the inductance.
A speaker with a 2mH VC @60Hz will add 0.756 ohms to the total Z
A speaker with a 1.5mH VC @60 Hz will add 0.56 Ohms.

For a driver with a 2 ohm dc resistance that is a 38% change for the 2mH, and only an 18% change for the 1.5mH coil, this is almost a factor of 2! That is one example. I'm not even including any phase data in this, which is affected just as much.

Examing only DCR does not tell one very much. You must analyse the impedance. 2 ohm DCR? Only a car audio SPL subwoofer is going to have DCR this low, as it's going to be about a 3 ohm nominal impedance, if 2 ohms is the DCR. In a typical 6-8 ohm nominal home application woofer/subwoofer, the effect of 0.25 ohms is going to result in about 0.25dB fluctuation, give or take a fraction of a dB, in amplitude response. This is not audibly relevant at these low frequencies, even on headphones(with no room interactions), much less when analysed in a room where room response masking of +/-3dB would be considered superb within this bandwidth.

I will ammend my first statment where I say that inductance has no relevance on woofer quality. I was wrong to say this as a blanket statement. In very large differences of inductance, it will have a considerable frequency response effect(but I was assuming this discussion would remain focused on typical examples, not extreme ones). Also, lower inductance allows for greater linearity in long-excursion woofers as used in SPL applications, since higher inductance makes it more difficult to design a linear long-stroke dynamic motor, due to the unavoidable issue of dealing with the significant eddy currents in such an application. But most woofers that are high quality, by default, have a faraday loop(shorting ring) installed to reduce these effects -- therefor very high inductance variation(of a magnitude that would cause audible frequency response irregularities with a standard amplifier) between good woofers is not common. But poor quality woofers skimp on a lot more than just a shorting ring. So, it's not easy to compare such things.

-Chris

 

[RJB]

Gotta love Engineers....

 

[MacManNM]

Examing only DCR does not tell one very much. You must analyse the impedance. 2 ohm DCR? Only a car audio SPL subwoofer is going to have DCR this low, as it's going to be about a 3 ohm nominal impedance, if 2 ohms is the DCR. In a typical 6-8 ohm nominal home application woofer/subwoofer, the effect of 0.25 ohms is going to result in about 0.25dB fluctuation, give or take a fraction of a dB, in amplitude response. This is not audibly relevant at these low frequencies, even on headphones(with no room interactions), much less when analysed in a room where room response masking of +/-3dB would be considered superb within this bandwidth.

I think you should check your numbers. The high end of DCR is around 4 ohms in home apps( for most decent subwoofers). 2nd what numbers did you use to get 0.25ohms? That means the woofer would have to be around 0.6mH. Not realistic. My numbers show its more like 3 times that, and those are low. Most of your everyday ok subs are 2.6 mH. A good (not Great) parts express sub has 2.2 mH, that turns into 0.83 Ohms with a 3.1 ohm DCR. Thats almost a 3db change, compared to one with a 1.5mH driver with the same stats. Change that freq, and its even larger. You still are ignoring the phase issue.

I will ammend my first statment where I say that inductance has no relevance on woofer quality. I was wrong to say this as a blanket statement. In very large differences of inductance, it will have a considerable frequency response effect. Also, lower inductance allows for greater linearity in long-excursion woofers as used in SPL applications, since higher inductance makes it more difficult to design a linear long-stroke dynamic motor, due to the unavoidable issue of dealing with the significant eddy currents in such an application. But most woofers that are high quality, by default, have a faraday loop(shorting ring) installed to reduce these effects -- therefor very high inductance variation(of a magnitude that would cause audible frequency response irregularities with a standard amplifier) between good woofers is not common.

-Chris

I agree, this is my point.

 

[WmAx]

I think you should check your numbers. The high end of DCR is around 4 ohms in home apps( for most decent subwoofers). 2nd what numbers did you use to get 0.25ohms? That means the woofer would have to be around 0.6mH. Not realistic. My numbers show its more like 3 times that, and those are low. Most of your everyday ok subs are 2.6 mH. A good (not Great) parts express sub has 2.2 mH, that turns into 0.83 Ohms with a 3.1 ohm DCR. Thats almost a 3db change, compared to one with a 1.5mH driver with the same stats. Change that freq, and its even larger. You still are ignoring the phase issue.

My estimate was based on memory. But, here is a specific example, that I prepared: An average quality(nothing special) Dayton 12" Series II woofer. It is 8 ohms nominal impedance. 2.3 mH inductance. Using impedance/phase plot data, here is the difference in amplitude of x(driver) and x+0.25(0.25 ohms placed in series between source). The flat response line(x) is not indicative of the subs actual acoustic response, but is a flat response line that I substitued in place of the acoustic SPL measurement. The reason for this is so that only the difference caused by adding the isolated variable(0.25 ohms) remains. In this case, it's actually a little over 0.3 dB difference, as opposed to the 0.25 dB difference that I estimated earlier. Even 1 dB fluctuation would not be readily audible within the bass range.

-Chris

 

[MacManNM]

My estimate was based on memory. But, here is a specific example, that I prepared: An average quality(nothing special) Dayton 12" Series II woofer. It is 8 ohms nominal impedance. 2.3 mH inductance. Using impedance/phase plot data, here is the difference in amplitude of x(driver) and x+0.25(0.25 ohms placed in series between source). The flat response line(x) is not indicative of the subs actual acoustic response, but is a flat response line that I substitued in place of the acoustic SPL measurement. The reason for this is so that only the difference caused by adding the isolated variable(0.25 ohms) remains. In this case, it's actually a little over 0.3 dB difference, as opposed to the 0.25 dB difference that I estimated earlier. Even 1 dB fluctuation would not be readily audible within the bass range.

-Chris

This does not make sense. Why would you use 0.25 Ohms? Impedance varies.The Impedance of a 2.0mH coil goes from 0.2 ohms @ 15hz to 1.13 Ohms @ 90 Hz. Your graph (allbiet nice) does not take that into account. Can you plot it with the correct varying impedance? Just by the delta in resistance, there is well over 6 db in amplitude driven linearly.

 

[WmAx]

This does not make sense. Why would you use 0.25 Ohms? Impedance varies.The Impedance of a 2.0mH coil goes from 0.2 ohms @ 15hz to 1.13 Ohms @ 90 Hz. Your graph (allbiet nice) does not take that into account. Can you plot it with the correct varying impedance? Just by the delta in resistance, there is well over 6 db in amplitude driven linearly.

This is the impedance plot of X that was used in the prior graph:

You stated in a prior post that 0.25 ohms as a variable applied over total Z of a driver @ 60Hz was significant. The prior graph shows the difference of this isolated variable that actually occurs, with a common quality woofer, with 0.25 ohms of resistance added across the total Z.

The fact is that every woofer adheres to it's calculated freqency response(derived from T&S parameters) within a specific alignment, within usually a dB of this predicted response, when it is used with a typical low output impedance amplifier, regardless of whether the woofer's Le is 1.5mH or 2.0mH.

I don't understand your point in reference to exactly where the supposed large fluctuations in amplitude response supposedly occur. As of this moment, I am wondering if you mean theoretically changing the Le of a driver after it was installed into an enclosure with a specific alignment. Since change in Le would change the Theile/Small parameters, you must compensate for the change in electrical/mechanical Q with the enclosure alignment after making such a modification, in order to regain a linear response. Or, I wonder if you mean to add inductance to a pre-existing driver by placing an inductor in series with said driver. This is unrealistic, as each driver is specifically designed with the VC parameters in consideration, and adding 1 mH of inductance to a driver with 2 mH inductance, with a series inductor, for example, would not be the same as comparing two drivers, one with 2 mH and one with 3 mH inductance.

-Chris

 

[mulester7]

.....did all this have anything to do with cornbread?.....

 

[MacManNM]

This is the impedance plot of X that was used in the prior graph:

You stated in a prior post that 0.25 ohms as a variable applied over total Z of a driver @ 60Hz was significant. The prior graph shows the difference of this isolated variable that actually occurs, with a common quality woofer, with 0.25 ohms of resistance added across the total Z.

This is not what I said. 0.25 ohms is added to the DCR @ 60Hz only. The total impedance changes with frequency, so adding 0.25 ohms across the board does nothing.

The fact is that every woofer adheres to it's calculated freqency response(derived from T&S parameters) within a specific alignment, within usually a dB of this predicted response, when it is used with a typical low output impedance amplifier, regardless of whether the woofer's Le is 1.5mH or 2.0mH.

I don't understand your point in reference to exactly where the supposed large fluctuations in amplitude response supposedly occur. As of this moment, I am wondering if you mean theoretically changing the Le of a driver after it was installed into an enclosure with a specific alignment. Since change in Le would change the Theile/Small parameters, you must compensate for the change in electrical/mechanical Q with the enclosure alignment after making such a modification, in order to regain a linear response. Or, I wonder if you mean to add inductance to a pre-existing driver by placing an inductor in series with said driver. This is unrealistic, as each driver is specifically designed with the VC parameters in consideration, and adding 1 mH of inductance to a driver with 2 mH inductance, with a series inductor, for example, would not be the same as comparing two drivers, one with 2 mH and one with 3 mH inductance.

-Chris

Here is a comparison of two drivers that have virtually the exact same driver parameters, in the exact same infinite baffle setup. The only difference is that one has a 1.5mH coil, and the other has a 2.2mH coil. This is a true representation of the affect of inductance on freq response.

 

[WmAx]

Here is a comparison of two drivers that have virtually the exact same driver parameters, in the exact same infinite baffle setup. The only difference is that one has a 1.5mH coil, and the other has a 2.2mH coil. This is a true representation of the affect of inductance on freq response.

If the drivers had the same T&S parameters, the responses would be identical. Besides, the program that you are using to plot that simulation calculates based on the T&S parameters that you manually enter or import into it's database. So, this now leaves the question: Which parameters are you talking about?

This has now come full circle, becuase in my last reply, I stated: As of this moment, I am wondering if you mean theoretically changing the Le of a driver after it was installed into an enclosure with a specific alignment. Since change in Le would change the Theile/Small parameters, you must compensate for the change in electrical/mechanical Q with the enclosure alignment after making such a modification, in order to regain a linear response.

It would now appear that this basicly what you meant. Is this correct?

-Chris

 

[MacManNM]

If the drivers had the same T&S parameters, the responses would be identical. Besides, the program that you are using to plot that simulation calculates based on the T&S parameters that you manually enter or import into it's database. So, this now leaves the question: Which parameters are you talking about?

This has now come full circle, becuase in my last reply, I stated: As of this moment, I am wondering if you mean theoretically changing the Le of a driver after it was installed into an enclosure with a specific alignment. Since change in Le would change the Theile/Small parameters, you must compensate for the change in electrical/mechanical Q with the enclosure alignment after making such a modification, in order to regain a linear response.

It would now appear that this basicly what you meant. Is this correct?

-Chris

All of the T S parameters are very close; the only one that is substantially different is the inductance. This is raw driver data, no enclosure. Enclosures tend to hide how bad the driver really is. It is not a theoretical case. This is code run with actual numbers.

 

[MacManNM]

I hope you realize that the higher SPL # is the one with the higher Le. The lower, linear response is of the lower Le. The phase on the lower Le driver is much flatter as well.

 

[WmAx]

All of the T S parameters are very close; the only one that is substantially different is the inductance. This is raw driver data, no enclosure. Enclosures tend to hide how bad the driver really is. It is not a theoretical case. This is code run with actual numbers.

First, I doubt the simularity of the T&S parameters used in your example, because the LF response of a driver is directly related to the T&S parameters. Theile/Small parameters will always predict the true response within very close tolerances. The T&S parameters are, in part, calculated based upon the actual impedance plot of the driver, so that it's specific reactivity is known and predictable using standarized mathematical formulas. It is a proofed method, extensively verified/tested in audio engineering, decades ago. In fact, WinISD uses the T&S parameters to calculate the LF response curve(s). The Le parameter in WinISD appears to do one thing: calculate what resembles a low pass filter based on the BL(force-current ratio) parameter, or if you skew the proportionate(BL vs. Le) values enough, it appears to act as a bandpass filter. But what this has to do with anything in application, I don't know, since in the real world, if you change the driver motor design to alter Le and Bl, you'll also affect other parameters. But a new set of T&S measurements after this physical modification would reflect the changes. The T&S parameters would not remain exactly the same after such physical motor modification(s). This is not something that this program provides for; WinISD is not a transducer engineering program. Second, what on Earth are you bickering about? Making a mountain out of a mole hill, are we? The overall Q(which is the only thing really discerned in that frequency response graph that you provided) of the driver's free air response is not relevant to driver quality/fideltiy; it merely dictates the ideal application/alignment for the specific driver. Let's consider the low Le example you refer to as not being linear -- if you were using a critically damped sealed enclosure or infinite baffle alignment, the non-linear one is far more suited. The linear example would perform horribly in such an application.

Let's examine a driver compared under two conditions, with the variable being Le, using the same software(WinISD) that you used:

This is a standard 8 ohm driver, with the parameter of Le manually changed and compared. Both examples in the same oversized volume(critically damped). Both examples having exactly the same T&S parameters. The effect on frequency response that WinISD shows is one caused directly by the low pass filtering of the inductance in interaction with force-current ratio.

-Chris

 

[WmAx]

This is not what I said. 0.25 ohms is added to the DCR @ 60Hz only.

How can you add DCR @ 60Hz only? The frequency specification implies that it is A.C. current. D.C. Resistance can not be selectively applied. It is D.C.. Another typo, I know, but I think one of the core issues in this discussion is that you are not putting much effort into clear communication. The result is this(essentially) meaningless debate that is now occuring. To be fair: I don't have to reply and thus perpetuate this discussion, so I have to take some blame too.

The total impedance changes with frequency, so adding 0.25 ohms across the board does nothing.

A low output impedance amplilfier(99.9% of amplifiers) will not have trouble feeding a very constant voltage(and at very low distortion) to delta resistance. Therefor I fail to see the point of caring about the insignificant impedance dynamics of typical dynamic speakers, unless the speakers are used with poor quality amplifiers that have high output impedance.

-Chris

 

[Tomorrow]

The stuff you two are spraying into the wind is splashing out here.

 

[MacManNM]

The stuff you two are spraying into the wind is splashing out here.

your point?

 

[MacManNM]

How can you add DCR @ 60Hz only? The frequency specification implies that it is A.C. current. D.C. Resistance can not be selectively applied. It is D.C.. Another typo, I know, but I think one of the core issues in this discussion is that you are not putting much effort into clear communication. The result is this(essentially) meaningless debate that is now occuring. To be fair: I don't have to reply and thus perpetuate this discussion, so I have to take some blame too.

The DCR is applied across the board, the impedance of the VC is then added on top of that. Because it is impedance, it CHANGES with the frequency. The fact is that the DCR is the lowest resistance the amp will ever see. The Impedance was calculated at 60Hz, therefore it is valid for that data point only. A plot of Z vs. Freq must be applied to make any argument valid.

A low output impedance amplilfier(99.9% of amplifiers) will not have trouble feeding a very constant voltage(and at very low distortion) to delta resistance. Therefor I fail to see the point of caring about the insignificant impedance dynamics of typical dynamic speakers, unless the speakers are used with poor quality amplifiers that have high output impedance.

-Chris

It is not the ability of an amplifier to drive the load that is in question. It is the ability of the field of the coil to move the mass of the speaker. I did make a mistake with the code. I haven't been around much to really put a lot into this discussion. I will hunt it down and display the difference the Le makes ASAP.

 

[WmAx]

It is not the ability of an amplifier to drive the load that is in question. It is the ability of the field of the coil to move the mass of the speaker. I did make a mistake with the code. I haven't been around much to really put a lot into this discussion. I will hunt it down and display the difference the Le makes ASAP.

You might find it easier take a look(at a real world example of a driver which can have it's inductive field changed) at this link:

http://www.caraudiomag.com/testreports/0311cae_infinity/

This is perfect example of what I was talking about. This woofer has interchangable inserts to fit into the motor assembly. By changing these parts, you change the magnetic flux(along with directly affecting the inductive field(eddy currents) and Bl). The net effect is the ability to change driver Q. The result: a driver that can be optimized to work in any type of application by varying it's overall Q. The Theile & Small Q parameters are very different for each variance of the magnetic field(as expected). But no mode is better than another, except in reference to a specific application. Example: The high Q mode is optimal for infinite baffle, but not for other applications.

-Chris