It has been a long time since I have worked with electronics and I am finding I have forgotten more than I remember. Forgive me, but I don't know how to word this question in a forum search to find an answer.
How does "Le" relate to... let's say nominal driver impedance? By this I mean, does a 4ohm driver with an Le of 1mH compare to an 8 ohm driver with double the Le? or am I getting that backwards?
How does "Le" relate to... let's say nominal driver impedance? By this I mean, does a 4ohm driver with an Le of 1mH compare to an 8 ohm driver with double the Le? or am I getting that backwards?
If I remember correctly an 8 ohm driver usually has double the Le of a 4 ohm driver.
But in practice: Le impacts how much the impedance rises in high frequencies. With Le=0 then above the impedance hump the impedance will pretty much be flat. With Le>0 then impedance rises in high frequencies.
Then there is also the problem that usually, Le varies along the cone stroke. This means that when the cone is say +5mm Le might be 0.45, but at -5mm it might be 0.35. This is why Klippel tests for example plot Le as a function of the driver stroke. A high Le is more bothersome than a low Le but is not really a major issue. A high variation of Le on the stroke, however, is a problem and is often a major cause of distortion in the mid-high region. On average, better drivers (more expensive!) have less Le variations over the stroke.
Edit: Here are images to clarify from Erin's Audio Corner
Here is a plot of Le(x) where x is the driver stroke position:
.png)
But in practice: Le impacts how much the impedance rises in high frequencies. With Le=0 then above the impedance hump the impedance will pretty much be flat. With Le>0 then impedance rises in high frequencies.
Then there is also the problem that usually, Le varies along the cone stroke. This means that when the cone is say +5mm Le might be 0.45, but at -5mm it might be 0.35. This is why Klippel tests for example plot Le as a function of the driver stroke. A high Le is more bothersome than a low Le but is not really a major issue. A high variation of Le on the stroke, however, is a problem and is often a major cause of distortion in the mid-high region. On average, better drivers (more expensive!) have less Le variations over the stroke.
Edit: Here are images to clarify from Erin's Audio Corner
Here is a plot of Le(x) where x is the driver stroke position:
Last edited:
The voice coil inductance Le accounts only for a small portion of the net impedance with the major portion coming from the resistance Re, usually around 2.5-3.5 ohms (for 4ohm) and 5-6 ohms (for 8ohm).
For example, a very popular PA driver Faital Pro 15PR400 is available in both 4 ohm and 8 ohm versions, with Re/Le being 3.3ohm/0.6mH (4 ohms) and 5.1ohm/0.72mH (8ohms). Similarly, for the JBL 2226, these values are 2.5ohm/0.92mH (4ohm) and 5ohm/1.75mH (8ohm).
However, in the above examples, there appears to be no proportional relationship between nominal impedance and inductance of the voice coil. In fact, there are 8ohm drivers with very less inductance (e.g. JBL 2035H). The voice coil resistance Re may be safely assumed to be higher for higher impedance drivers, though.
I would really like to know what the "experts" have to say about this.
For example, a very popular PA driver Faital Pro 15PR400 is available in both 4 ohm and 8 ohm versions, with Re/Le being 3.3ohm/0.6mH (4 ohms) and 5.1ohm/0.72mH (8ohms). Similarly, for the JBL 2226, these values are 2.5ohm/0.92mH (4ohm) and 5ohm/1.75mH (8ohm).
However, in the above examples, there appears to be no proportional relationship between nominal impedance and inductance of the voice coil. In fact, there are 8ohm drivers with very less inductance (e.g. JBL 2035H). The voice coil resistance Re may be safely assumed to be higher for higher impedance drivers, though.
I would really like to know what the "experts" have to say about this.
Actually, the inductance should increase significantly for the higher nominal impedance version of the driver, assuming that the mass of wire in the voicecoil and the length of coil within and protruding the magnetic gap (for overhung coils) does not change. If you take a 4ohm coil, make the wire 2/3rds the cross-section and then increase the number of turns by 50% to fill the empty space made by using smaller wire, you have the same mass of copper on the coil, the resistance and inductance both increase to approximately double the original values therefore making an 8ohm driver that should perform equivalently. If they change the overall mass or length of voicecoil between impedance versions then you can get a non proportional change in inductance.
The 'nominal' impedance of a driver is entirely subjective. Usually an '8ohm nominal' driver will have a voicecoil resistance somewhere between 6 and 7 ohms, so once you factor in the impedance of a passive crossover, resistance of wiring, etc the minimum impedance the amplifier sees is close to 8ohm. That said, I have seen some drivers marketed as '8ohm nominal' having an Re closer to 5ohm and some which have Re almost exactly 8ohm. Le usually does not affect the minimum impedance. Take for example a 6.5" midwoofer, the minimum impedance is usually around 100-200Hz, and at these frequencies the inductance has almost no contribution.
Voicecoil inductance is an unwanted parasitic effect. The lower the inductance the better. It's difficult to design for a lot of excursion without also causing a lot of voicecoil inductance, therefore you tend to see subwoofers with large excursion, high inductance and therefore poor high frequency performance. Meanwhile dedicated midrange drivers have low excursion but low inductance and therefore stellar mid and high frequency performance.
The 'nominal' impedance of a driver is entirely subjective. Usually an '8ohm nominal' driver will have a voicecoil resistance somewhere between 6 and 7 ohms, so once you factor in the impedance of a passive crossover, resistance of wiring, etc the minimum impedance the amplifier sees is close to 8ohm. That said, I have seen some drivers marketed as '8ohm nominal' having an Re closer to 5ohm and some which have Re almost exactly 8ohm. Le usually does not affect the minimum impedance. Take for example a 6.5" midwoofer, the minimum impedance is usually around 100-200Hz, and at these frequencies the inductance has almost no contribution.
Voicecoil inductance is an unwanted parasitic effect. The lower the inductance the better. It's difficult to design for a lot of excursion without also causing a lot of voicecoil inductance, therefore you tend to see subwoofers with large excursion, high inductance and therefore poor high frequency performance. Meanwhile dedicated midrange drivers have low excursion but low inductance and therefore stellar mid and high frequency performance.
Last edited:
This could also happen if the 8-ohm version is made using two 4-ohm coils in series with one half of the turns wound in reverse (differential drive) to result in opposite coupling.
A significant factor in reactive characteristics of a speaker (or any electromagnetic motor) is the motor structure. A smaller gap (and lower excursion like TMM stated) is one part of it, but the overall gap geometry, motor strength and added features such as caps and sleeves determine the overall motor inductance.
'Impedance' is an equivalent method used to express the mechanical behaviour of the driver under applied signal. This is similar to, but not the same as a fixed inductor because the inductance (and impedance) changes as the coil moves through the gap. This is illustrated in #2, and in any driver impedance graph.
On the latter, the coil characteristics are dominant after resonance and the mechanical factors before resonance. The peak is where one hands off to the other. Some drivers have multiple impedance peaks due to their mechanical structure (some CDs come to mind).
Le is measured at a given frequency and voltage, and like every other inductor operated below saturation, changes with applied voltage and frequency.
Most of the changes between 4 and 8 ohm versions of a speaker are in the coil itself. It is not normal to see mechanical specifications change between versions except a slight change in mms due to the coil itself, and its resultant effects on all other dynamic parameters (such as Bl, all the Qs and Vas)
Another thing to note is that a (say) 4 ohm driver never has half the inductance of an 8 ohm driver. The relationship is analogous to twice the number of turns not being twice the inductane. It closer to 66% most of the time. Differences in Re are closer to 90%. Again, this is not always true. A designer may optimise 4 and 8 ohm drivers to keep Le differences smaller.
SB Acoustics SB17CRC35-4 | HiFiCompass
SB Acoustics SB17CRC35-8 | HiFiCompass
SB Acoustics SB20PFC30-4 | HiFiCompass
SB Acoustics SB20PFC30-8 | HiFiCompass
'Impedance' is an equivalent method used to express the mechanical behaviour of the driver under applied signal. This is similar to, but not the same as a fixed inductor because the inductance (and impedance) changes as the coil moves through the gap. This is illustrated in #2, and in any driver impedance graph.
On the latter, the coil characteristics are dominant after resonance and the mechanical factors before resonance. The peak is where one hands off to the other. Some drivers have multiple impedance peaks due to their mechanical structure (some CDs come to mind).
Le is measured at a given frequency and voltage, and like every other inductor operated below saturation, changes with applied voltage and frequency.
Most of the changes between 4 and 8 ohm versions of a speaker are in the coil itself. It is not normal to see mechanical specifications change between versions except a slight change in mms due to the coil itself, and its resultant effects on all other dynamic parameters (such as Bl, all the Qs and Vas)
Another thing to note is that a (say) 4 ohm driver never has half the inductance of an 8 ohm driver. The relationship is analogous to twice the number of turns not being twice the inductane. It closer to 66% most of the time. Differences in Re are closer to 90%. Again, this is not always true. A designer may optimise 4 and 8 ohm drivers to keep Le differences smaller.
SB Acoustics SB17CRC35-4 | HiFiCompass
SB Acoustics SB17CRC35-8 | HiFiCompass
SB Acoustics SB20PFC30-4 | HiFiCompass
SB Acoustics SB20PFC30-8 | HiFiCompass
Thanks for the info.
The inductance to DC resistance ratio in the examples you shared vary quite a lot between the 4 ohm and 8 ohm versions. I guess this must not affect the sound greatly since the manufacturers are pursuing other optimizations it would seem.
The inductance to DC resistance ratio in the examples you shared vary quite a lot between the 4 ohm and 8 ohm versions. I guess this must not affect the sound greatly since the manufacturers are pursuing other optimizations it would seem.
Exactly. Most manufacturers would want to keep frequency response of '4' and '8' ohm versions as close as possible, as well as things like enclosure dimensions. Hence the coil resistance and resulting inductance are usually effects of optimisation rather than design data input.
It may help understanding to see a full impedance plot, something very rarely shown for a driver, though not hard to measure. This is from an old General Radio Corp. manual for one of their bridges. Remember, capacitive reactance is negative, inductive reactance is positive.
Attachments
In that graph, should the line in the middle actually say impedance? And that would be showing the driver resonance to be around 352hz peaking at near 34 ohms? I'm hoping I understand what is going on here, LOL!
Yes, at resonance the reactances cancel out and the device is purely resistive. Thus, I think the x axis is correctly labeled as resistance. No idea what ancient driver they used, but I do think the graph is a bit idealistic, or maybe more points should have been measured.
Yes! In fact it is misleading to talk about a driver's "inductance" since that changes not only with cone position but also with frequency. The first derivation of this "semi-inductance" I recall was J.R. Wright: "The high-frequency electrical impedance characteristics of a moving-coil loudspeaker deviate significantly from those of the traditional series resistance-inductance model of the voice coil. The deviation is attributed to eddy current effects within the magnetic pole structure."
AES E-Library >> An Empirical Model for Loudspeaker Motor Impedance
looking for that I also stumbled on
https://www.klippel.de/fileadmin/klippel/Files/Know_How/Literature/Papers/Voice_Coil_%20Impedance_04.pdf
OK, gotcha... I understand now. It really makes a good argument for an active setup. All sorts of unwanted side effects get thrown into the mix when you start adding more reactance in series with that varying load.
Yes, the best sound I ever had was a full active 3-way with Marchand crossovers. Too much stuff in the living room and bad WAF so it's in the past now.
The ultimate question is, can it be managed. Some problems go away (like this one) and some can't be fixed.
One effective way for reducing the voice coil inductance is the use of differential drive (courtesy JBL). It would be nice to know of any other methods that achieve something similar.
Differential drive transducers
Differential drive transducers
While it doesn't reduce the inductance (but still reduces distortion in the same push-pull way): use dual drivers and invert one of them. This setup is even better since it can cancel even order distortion not just from the motor but from the suspension too. You can go even further by mounting them in a force opposing setup such that the mechanical vibration is cancelled.
Last edited:
Well JBL claims it does. Page 3, paragraph 3, point B says:
"The new voice coil arrangement will have less effective inductance than the standard one, since the reversely wound coils will have negative mutual inductance between them. This translates into a more uniform impedance curve and extended output at higher frequencies."
I think the idea is that the flux linking with the reverse wound coil is also opposite in direction, resulting in a force in the same direction as the other coil. The two coils may then be conveniently connected in series or parallel to operate the driver.
"The new voice coil arrangement will have less effective inductance than the standard one, since the reversely wound coils will have negative mutual inductance between them. This translates into a more uniform impedance curve and extended output at higher frequencies."
I think the idea is that the flux linking with the reverse wound coil is also opposite in direction, resulting in a force in the same direction as the other coil. The two coils may then be conveniently connected in series or parallel to operate the driver.
Last edited:
A more simple way of reducing the inductance is to just add shorting rings to the motor, which most manifacturers do nowadays on the more expensive drivers.
A more fancy solution though is 18Sound Active Impedance Control also uses multiple coils to control impedance.
A more fancy solution though is 18Sound Active Impedance Control also uses multiple coils to control impedance.