🚀 go-pugleaf

In article <3922537c.1132967@news.one.net.au>,
Johnny <johnny_c@NOSPAM.one.net.au> wrote:
>On Wed, 10 May 2000 12:30:58 GMT, DPierce@world.std.com (Richard D
>Pierce) wrote:
>
>>The argument is advanced that the series inductors in a
>>crossover ruin the amplifier's high damping factor. That's wrong
>>on two fronts:
>>
>>1. As has been stated here and elsewhere numerous times,
>>   "damping factor" is meaningless and misleading single-
>>   number advertising spec that has no significance in
>>   loudspeaker performance in all but the most pathological
>>   of cases. It is the resistance of the voice coil that is
>>   THE single greatest factor in determining the damping of
>>   the system, and all other electrical losses fall FAR behind
>>   that in significance.
>>
>>2. The competent engineer will have already compensated for
>>   these and other losses in the design of the INTEGRATED
>>   system, and removing these coils can, in fact, undo the
>>   system tuning that leads to a speakers good performance
>
>Are point 1 & 2 not contradictory?  

No, they are not.

>Surely if the total AC impedance
>of the inductor was not an issue as you claim in point 1 it would not
>need to be considered as part of the design as you mention in point 2.

It's not the AC impedance of the inductor that's the issue,
because it is not the inductive part of the impedance that
provides the damping, rather it is the equivalent series losses.

>My experience with 3-way designs is that a large inductor of (say 6mH
>and DCR = 0R8) has a very significant effect on the bass tuning.
>Which agrees with point 2, but not point 1.

Well, let's see. Assume a driver with a DC resistance of 6.5
ohms, a Qmc of 4.0, a Qec (with zero source resistance) of 0.86
and a system Qtc of 0.707 (Butterworth, maximally flat
response). Your claim is that 0.8 ohms series resistance will
have a "significant effect" on bass tuning.

Well, the increase in series resistance will be (according to
Small):

                 Re + Rg
   Qec'  =  Qec ---------
                    Re

where Re is the DC resistance of the voice coil, 6.5 ohms and Rg
is the source resistance, 0.8 ohms. Thus:

                  6.5 + 0.8
   Qec'  =  0.707 ---------  = 1.12
                     6.5

Fine, how does that effect the system Qtc?

            Qmc * Qec
   Qtc  =  -----------
            Qmc + Qec

so

            4.0 * 1.12
   Qtc  =  ------------  = 0.875
            4.0 + 1.12

Okay, now how does that effect the response of the system? Let's
look at the maximum deviation from flat response that results:

                                     4
                                  Qtc
   G(max)dB  =  20 log sqrt ( ----------- )
                                 2
                              Qtc  - 0.25

which is about a half a dB peak at system resonance. I would say
that in a room, this might constidute a marginal change in the
effective system tuning.

Great, so now if the designer of the system were to compensate
for this marginal tuning change by choosing a driver with a
lower Qes, and you were to remove the series resistance, then,
under your criteria yes, indeed, it would be "significantly
detuned as a result.

Point 1 and point 2 that you quoted are not at all
contradictory. In fact, your assertion supports #2 but does not
contradict #1 at all.

There are other effects that are not included in 1 anbd 1 above.
For example, there is the frequency-dependent attenuation due to
the series resistance that will change if that resistance
changes. In the example above, lets say the impedance over the
bandwidth of the woofer varied between 7.5 ohms around 250 Hz to
as high as 50 ohms at resonance. The resultant frequency
dependent attenuation due to a series 0.8 ohm resistance would
vary from a high of 0.88 dB at 250 Hz to a low of 0.137 dB at
resonance. That's a difference of nearly 3/4 of a dB,
significantly eclipsing the peak due to an increase in the Qtc
of the system. Thus, while we might quible over whether the
system is or is not "significantly" detuned, it's clear that in
non-"tuning" aspects, the attenuative effects exceed that and
must be compensated for.

Competent speaker engineers will have compensated for this
effect by, say, using a driver with a lower Qt or a slightly
larger enclosure. They will have done so by knowing what the
properties of the integrated system will be when done. A third
party does not have the benefit of this design process and thus
could never hope to get anything as good using two subsystem
that were designed without knowledge of the properties of each
other, advertising hooey and magazine rags claims
notwithstanding.

And this is the point: You CANNOT simply rip out a competently
designed-for-the-purpose passive crossover, replace it with a an
equally competently designed BUT GENERIC active crossover and
assume the result will be better. It's simply NOT a matter of
adjusting crossover frequencies, selecting a slope, choosing the
frequency, and you're done. Speaker design is SYSTEM design, and
there is NO gaurantee that taking any arbitrary collection of
arbitrarily high-quality parts and throwing them together is
going to give you anything other than a random pile of
non-integrated high-quality parts that probably will not sound
as good. In fact, it's likely to be quite unsatisfactory,
despite the high expenditure.

-- 
|            Dick Pierce            |
|   Professional Audio Development  |
|   1-781/826-4953  Voice and FAX   |
|       DPierce@world.std.com       |