Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
63.111.2.217
In Reply to: RE: terminology posted by pictureguy on March 25, 2014 at 18:42:53
The Electrical crossover(knee, half power point) along with the slope and their alignment (how the HP/LP filters interact) are (somewhat) easily addressed in theory and practice. In a passive crossover the components calculations are made a little more difficult because of the driver electrical characteristics. An active or passive line level crossover can be easier to implement. The acoustical crossover is a whole different thing and perhaps more important. No loudspeaker has a flat response - on or off axis. and the spacing between the drivers effects the acoustic crossover point. I prefer to do all my development with an active (DSP) crossover - which allows me the flexibility to assess on-axis and power response effects of a crossover design - this often results in asymmetrical electrical crossovers. A classic example is when you are using a two way system with no low pass filter - only a capacitor for the tweeter.
Three most important things in Audio reproduction: Keep the noise levels low, the power high and the room diffuse.
Follow Ups:
Your point is reasonable.
May I rephrase?
A speaker is a 'system' in that the crossover and the drivers act to produce sound. Changing one thing at a time would be preferred, at least in my way of thinking, to taking the same drivers and putting 'em in 2 different enclosures with 2 different crossovers…..How could you possibly compare those 2 systems? Or, if you were 'stuck' with the drivers, what would the path forward be towards whatever your sound goals are?
That's why I'm not going to MESS with the panel until I can reasonably duplicate the crossover as an active OR passive line-level item. Once a new baseline is established, frame mods can go forward.
Too much is never enough
To expound further, any raw loudspeaker driver has an acoustic bandwidth ( even without any crossover ). It's low-frequency *knee* ( e.g.: bass-extension on woofers ) has a highly documented "roll-off" ( see speaker building ).
The diaphragm/cone/dome has an effective moving mass ( Mms ). A magnetic driver has a suspension system to keep the voice-coil within the operational limits of the magnetic gap. Where a magneto-dynamic woofer cone has a spider, a planar-magnetic driver has diaphragm tension. Consequentially, this sprung mass has a natural resonance frequency ( think tuning fork ), "fs". This essentially defines the acoustic L.F. *knee* & its nature...
The *knee* is equivalent to a 2nd-order ( 12dB/octave ) high-pass filter. Even the driver's temporal "group-delay" perfectly fits the analogous electronic model's "phase", where the electronic "fc" is the driver "fs".
The magnetic driver is essentially an electric motor. In an efficient magneto-dynamic woofer this resonance makes the cone misbehave. Subsequently, the extraneous motion converts the driver into an electric generator. This is called back EMF & manifests itself as a measurable impedance spike. That spike quantifies Thiele-Small driver parameters "Qes", "Qms", & "Qts". The "Qts" equates to the "Qh" of the *knee". This is also the vary same "Q" corresponding to settling time of transient response where above one-half is under-damped & below one-half is over-damped.
Magnepan's planar-magnetic drivers are virtually resistive & constant. This means that there's no impedance spike to attenuate response at resonance. The math indicates that critical damping is unity ( rather one-half ). Thus, we can expect a S.P.L. spike at resonance before "roll-off".
Reverse engineering from Stereophile's plotted response graphs, I deduced the submitted newly broken-in Maggies have a "Q" of about three. ( I speculate that as Maggies age this might reduce to as much as two after a few years. ) The MG1.6's midtreble quasi-ribbon tweeter resembles a "fs" of 600hz.
H{s} = s² /( s² + ωₒ·s/Q + ωₒ² )where:
ωₒ = 2·π·fₒ
s = j·ω
The equivalent capacitance & inductance for response:
L = R /( 2×π×fs×Q )
C = Q /( 2×π×fs×R )
Apply like it were an active filter on the tweeter which is electronically buffered from actual cross-over on the tweeter... with complex multiplication of the acoustic output function & the electronic high-pass filter output ... and, then, use complex addition of the low-passed midbass panel output function.
... just my 2¢♪ moderate Mart ♫ ☺ Planar Asylum
where speakers are thin & music isn't
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: