Infinite Slope Crossover


2001-02-19 10:38 pm
Joseph Audio has a patent on this crossover.
Joseph Audio

I haven't done a patent search but I did find an
explantion on how it works on the madisound forum.

It works using the mutual inductance of windings inside a transformer.

Picture a transformer, with one primary winding, two secondary windings, the turns ratio being 1:1:1.

Connect the amp to the primary winding. Connect a driver one of the secondary windings. Now connect
the other secondary winding to the driver, except inverted and with a capacitor in series.

At low freqs, the cap disconnects the second coil and the output is largely unaffected.

At high freqs, the cap provides an AC short, which then opposes the formation of the mutual inductance
which would be necessary to induce current in either secondary winding. Output current drops toward zero
at high freqs because the mutual inductance between the primary and secondary drop toward zero at the
affected freqs.

The reflected impedance will begin to increase as the capacitor starts to provide the short. The amp
side sees the resulting cancellation of the mutual inductance as an open circuit on the secondary.
Also as the capacitor provides the short, the driver starts to see only the paralled DCR of the two
windings (and the cap) as the magnetic field necessary to create the inductance can't form (enough) a
round the closely coupled bifillar windings when one is inverted.

So... the question is.. is it possible to make an active crossover using this "patent" idea using a transfomer and capacitor and an op-amp to get a 120db/octave crossover ?

Sounds too simple to be true - lol

I don't know how the x-over works, but I have a pair of Joe Audio RM22si floor standing speakers, and they sound awesome. They were the most musical speaker of all the models I auditioned under $2500. I don't know how much is attributable to the crossover, but I think it must be a large part of the sound of this speaker.
I just read through the explanation quickly so if I missed the point I appologised, but wouldn't you need some series resistors in there too?

At high frequencies there would be a looped current in effect through the two secondary windings which would only be limited by the output resistance of the secondaries in series?

Just a point that came into my head - if the coil resistance of the secondaries is adequate to restrict current in the "high stop" operation of the circuit ignore me, lol

:apathic: :( :apathic:
<i>US Patent 4,771,466 Inventors: Modafferi; Richard September 13, 1988
<a href="'4771466'.WKU.&OS=PN/4771466&RS=PN/4771466">Multidriver loudspeaker apparatus with improved crossover filter circuits</a>
"A loudspeaker system includes at least two loudspeaker drivers, together with an electrical crossover network having filter circuits with at least two separate mutually exclusive frequency passbands. The filter circuits comprising the crossover network each possess brick-wall amplitude responses, i.e., passbands with very high band-edge amplitude vs. frequency response slops, on the order of 100 dB/octave in the better embodiments. The high passband band-edge slopes, which are realized by the inclusion of transmission zeros in the separate crossover filter transfer functions, takes together with further appropriate crossover filter transfer function synthesis causes the separate loudspeaker drivers comprising the loudspeaker system to function independently of one another in their contribution to total system acoustic output. It is shown that the loudspeaker system permits an accurate approximation to the ideal delay function in acoustic space, while minimizing acoustic wave interference among drivers operating in adjacent frequency band, and also reducing overall system nonlinear distortion."</i>



  • pat pic.gif
    pat pic.gif
    31.1 KB · Views: 1,126
good timing

I too loved these speakers when I heard them some time ago. Just today I started getting curious about this particular crossover, and a I found this thread via the search. I like how you guys are one step ahead of me :)

my question is how does this crossover affect the phase relationship of the signals? I do know know enough to figure it out from the schematic.
totally frelled

One thing I did notice about the speakers was that there was an enormous "sweet spot" and the imaging was dead on thoughout.

maybe I need even more about phase relationships though... ok so I understand how you can get interference if the tweet and woof are trying to drive the same signal but out of phase over a certain wide frequency range. But we've sufficiently narrowed that frequency range so all thats left is how your ears interpret it. What I'm wondering is does the phase relationship affect the way you hear the music? I'm inclined to think yes because there are people who assert that you can hear absolute phase (although I'm very curious as to what it sounds like, and am slightly skeptical) but this seems to fall along those lines...

I think its in my nature to look for hidden drawbacks. If phase isn't a problem, then what else is? Why isn't anyone else using this type of crossover? Is it only because nobody else has figured out how to get around the patent, and nobody else thought of this circuit till now? Dave, you already pointed out that a variation is in a textbook, so how could they patent it anyhow? Although I guess that the difference in active and passive crossovers is enough to call it new technology...


Re: totally frelled

totally frelled

Ref: Farscape. An Oz SciFi show

Tradeoffs are always being made. The trade-off here is a small region of phase anomallies traded for almost no interaction between the drivers. From most of the comments i have heard, Joeseph makes this work well.

The patent (and the cost) probably keeps others from using it.

As with all these circuits I'm interested in, I decided to model this circuit in Orcad / PSpice. The LP section works fine, but the HP section claims there is a "node x is floating" problem between C1 and C10...Any ideas?! I have traced it to being a problem with C1 as bypassing it corrects the problem but gives me about a 6dB rolloff! Thing is, at 2am I am having problems seeing let alone troubleshooting!


P.S. Sorry about the quality...Had to drop it right down to get under limit!


  • xover.jpg
    51.2 KB · Views: 492
I'm gonna dredge up this ancient thread, because steep crossover slopes sound significantly better. They work the best when done by a PC, because then you have no phase shift a long with no cone break up, lobing, or any of the other issues caused by crossovers.

I see some companies doing active speakers bu they still have somewhat shallow slopes. I'm not sure why they do this but probably the DSP they use can't process a very steep slope. Using equalizer APO you can do brick wall slopes with no complicated circuits. Direct connection between amp and driver. Problem solved.
Last edited: