A nominally 8 ohm driver indicates its average impedance,
consequenquently DC resistance is in the 5 to 6 ohm range.
Perfect Zobel compensation requires a resistor equal in value
to the DC resistance to give a perfectly flat impedance.
I believe the requirements for speakers are minimum impedance
which should be resistive should not be less than 1/root2 of the
nominal impedance.
This sets the DC resistance at 5.6R, common as is 2.7R for 4 ohm drivers.
With voltage drive this is entirely untrue.
Driver current is unaffected by the Zobel network.
sreten.
consequenquently DC resistance is in the 5 to 6 ohm range.
Perfect Zobel compensation requires a resistor equal in value
to the DC resistance to give a perfectly flat impedance.
I believe the requirements for speakers are minimum impedance
which should be resistive should not be less than 1/root2 of the
nominal impedance.
This sets the DC resistance at 5.6R, common as is 2.7R for 4 ohm drivers.
With voltage drive this is entirely untrue.
Driver current is unaffected by the Zobel network.
sreten.
Theory's nice, but nothing beats experiment.
Theory says that the Zobel resistor should be the same as the DC resistance of the voice coil. In practice, the matter is complicated by the fact that voice coil inductance is not pure inductance, but a transformer winding loosely coupled to a short circuit (the pole pieces). The figures I have given were determined experimentally, and I wasted a lot of time starting with a 5R resistor...
Theory says that the Zobel resistor should be the same as the DC resistance of the voice coil. In practice, the matter is complicated by the fact that voice coil inductance is not pure inductance, but a transformer winding loosely coupled to a short circuit (the pole pieces). The figures I have given were determined experimentally, and I wasted a lot of time starting with a 5R resistor...
I may have misunderstood my theory, but I still can't see how
variations of the effective inductance affect the resistor value.
Resistances should be equal, seems to me its the capacitance
value you need to address for the problem you describe.
sreten.
edit :
according to Jordan DC resistance
is 4.5R and impedance is not given :
http://www.ejjordan.co.uk/JX92.html
By standard definitions this is a 6 ohm driver.
variations of the effective inductance affect the resistor value.
Resistances should be equal, seems to me its the capacitance
value you need to address for the problem you describe.
sreten.edit :
according to Jordan DC resistance
is 4.5R and impedance is not given :
http://www.ejjordan.co.uk/JX92.html
By standard definitions this is a 6 ohm driver.
Sreten, the impedance of the driver is a combination of voice coil inductance and resistance and the motional impedance. But it really doesn't matter what either of us think about theory. It was a combination of 8R2 and 4u7 that achieved the almost resistive 5R measured response above mechanical resonance.
.EC8010 said:
In this particular instance i''m not remotely inclined to argue,
I don't have the driver or the measuring equipment.
My only concern is other stuff I've come across that confidently
asserts DC resistance should be used rather than nominal
impedance for the R in a Zobel to allow crossover design that
can assume constant resistance values.
Apparently an R of Znominal is often used erroneously instead
of Rdc, you seem to be suggesting that both are wrong, and
that an entirely different Zobel R is required.
Not my problem ,
sreten.
EC8010 said:
May I jump in here. There is an alternative model to the series resistance-inductance model of the voice coil. It assumes that the impedance of the inductance goes as w^n * L rather than w * L, n=0.6..0.8 for most speakers. The phase of this impedance is *not* 90 degrees, so the impedance also has a frequency dependent resistive component.
Using this model, the resistive part of the impedance at higher frequencies will be higher than the DC resistance. Now, the perfect conjugate link for theis circuit cannot be made of a simple series resistor-capacitor network, but increasing the resistance of the zobel network a bit makes the impedance equalisation better than if Re is used.
In some sense sreten has a point, to make the perfect conjugate link for this impedance, the *capacitor* should be changed to behave "the opposite" of the coil. So, the capacitance should be made lossy in the same way as the inductance, ie it should contain a frequency-varying *resistance*. By approximating this varying resistance with a fixed one we can acheive an approximate parallel conjugate load for a certain frequency range. This is another way of saying the zobel network resistance should be a little bit higher than Re, the fixed series approximation in the capacitor, can be seen as part of the resistance.
Did that make *any* sense?
This voice coil inductance model is described in JAES #6 2002 p442, and fits real speaker data well.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.