[bhobba]

Hi All

I am looking at putting together a system around the DEQX digital crossover and am currently looking at a suitable speaker system.  It occurred to me that the tweakabilty of the AKSONIC may be the perfect match for the DEQX.  What do others think?  Would the kit be available without the crossovers?

Thanks
Bill

[EchiDna]

I'm with you there bhobba - the AKSonic would be the all but perfect solution when using a DEQX, but this is a call for Hugh to make
The kit pretty much only exists of the crossovers (the real IP in the design) and plans for the cabinetry, so you would be wanting a small portion of the package.

[AKSA]

Hi Bhobba,

Thanks for your post about the AKSonic.

I'm committed with the designer for the sale of the AKSonic;  it must be sold as a complete kit.  Separating the crossovers is sadly not an option.

That said, I would suggest a lot of very smart people have tried to implement a digital filter (if indeed the DEQX is such) and failed.  This is a very tricky exercise, where the passive crossovers of conventional speakers are a fully mature technology.  A good passive crossover will also incorporate driver impedance correction.  This is very important, and should not be forgotten with active systems.

Cheers,

Hugh

[andyr]

Hi Bhobba, ...

That said, I would suggest a lot of very smart people have tried to implement a digital filter (if indeed the DEQX is such) and failed.  This is a very tricky exercise, where the passive crossovers of conventional speakers are a fully mature technology.  A good passive crossover will also incorporate incorporate driver impedance correction. This is very important, and should not be forgotten with active systems ...

Yes, Hugh, ze Aussie-designed DEQX is indeed a diggital filter ... I understand the guy who designed the Fairlight Synthesiser back in the 70s?? 80s?? is behind it.  

From wot I have read, it has some very cool features - up to 96dB slopes (for minimum driver overlap) AND phase correction, so you can avoid the phase changes which are intrinsic to "conventional" analogue active/passive filters.  

BTW, can U hexplain wot you mean by "A good passive crossover will also incorporate incorporate driver impedance correction. This is very important, and should not be forgotten with active systems."??  

Do you mean that a good passive crossover typically incorporates resistors (an L-pad?) to balance the output levels of each driver?  (An active crossover generally has gain controls for each frequency band.)

Or do you mean that because driver impedance typically varies all over the place with frequency, a good passive crossover will incorporate some correction for this?  (Isuggest it is important to do this in a passive crossover bcoz the crossover cap and coil values are dependent on driver impedance ... but I would've thought, in an active setup, the amp is attached directly to the driver and couldn't care less what its impedance is.)    

Regards,

Andy

[andyr]

Hi All

I am looking at putting together a system around the DEQX digital crossover and am currently looking at a suitable speaker system.  It occurred to me that the tweakability of the AKSONIC may be the perfect match for the DEQX.  What do others think?  Would the kit be available without the crossovers?

Thanks
Bill

Hi Bill,

Lucky fellow ... I'd love to get my hands on a DEQX!!

Seeing as how the AKSonic ain't an option for you, I suggest you have several possibilities ... depending on how much of a "handyman" you are!

With a DEQX, you could make up a world-beating speaker system by building Siegfried Linkwitz's "Orion" system.  Or there's Lynn Olsen's "Arial" speaker.

But you've gotta be a good woodworker for these - partickerly for the "Arial".

Or, can you find any 2nd-hand Magneplanar speakers in Brisbane or elsewhere in Oz?  You could always try the Oz distributor (Bill McLean in Long Jetty, NSW, 02-4333-3545) to see whether he happens to have any trade-ins in stock.  The reason I mention these is that they are known to be superb speakers, compromised by being built to a price in terms of their crossovers.  If you rip out the stock crossovers and use a DEQX instead, wow! you have something!  

I have old Maggie IIIAs which I run 3-way (analogue) active with my AKSAs.

Regards,

Andy

[AKSA]

Hi Andy,

Matching driver sensitivities in multiple driver speakers is de rigeur.  It has to be done to get a smooth frequency response.

However, as it happens, amps care very much what load they are connected to.  The impedance of a speaker is a mighty complex load, and the myriad phase shifts at different frequencies can lead to some very strange effects on amps with global negative feedback.  If the phase shift at the speaker exceeds the phase tolerance of the amp, it will oscillate and may destroy itself - and the drivers.  Of course with purely resistive speakers like the Magneplanar, this is not at issue.

A typical dynamic driver has a highly inductive voice coil, and numerous capacitors introduced into the crossovers.  Unless impedance is corrected for the voice coil, the load presented to the amplifier could well cause problems.  In cases where it does not cause problems, there is inevitably some disturbance of the phase relationships in the music - and this seriously affects the imaging.  Thus, I would argue that Zobel correction, a capacitor and a resistor in series across the voice coil, should be mandatory EVEN ON ACTIVE SYSTEMS!!

And of course, once you put in a Zobel, you might as well shove in a few extra caps and inductors and make a crossover of it........

Passive crossovers are not the wooden cross they are made out to be....

Cheers,

Hugh

[John Ashman]

I'd only like to say that speakers that are designed around passive crossovers are typically not as appropriate for DEQX as they could or should be which is why the DIY and custom market is diving it.  An ideal (though obviously people are exploring variations)  DEQX speaker is

1.  Monopole
2.  3-way
3.  Rigid, preferable metal or ceramic, crivers
4.  Minimum baffle
5.  Low diffraction
6.  Proportional drivers (as in a 5" going with an 8" or 10")
7.  Has one midrange driver

Anyway, that's ideal - something like the Salk HT3 which can be ordered "digital-ready" (finally, a good use for that term!) or one of the Selah models or a 3-way Thiel.  You can use just about any speaker you want, but what gets you great results in the passive world won't always translate into the digital world.

[bhobba]

Hi Bhobba,

Thanks for your post about the AKSonic.

I'm committed with the designer for the sale of the AKSonic;  it must be sold as a complete kit.  Separating the crossovers is sadly not an option.

That said, I would suggest a lot of very smart people have tried to implement a digital filter (if indeed the DEQX is such) and failed.  This is a very tricky exercise, where the passive crossovers of conventional speakers are a fully mature technology.  A good passive crossover will also incorporate dr ...

Hi Hugh

Thnaks for your post.  I see your point about the AKSonic .  What I am planning to do is build it in stages.  One option is to get a passive speaker - get it working properly then move over to the DEQX.  So getting an AKSonic is still not out of the question.  And it would be very interesting to compare the two.

Yes indeed the DEQX is a digital crossover.  It corrects for room defects, and frequency and phase response.  Although I have not purchased one yet all information I have about it is positive - some of it provided by people like Vance Dickinson.  It look as though it is done right.


Thanks
Bill

[bhobba]

Passive crossovers are not the wooden cross they are made out to be....

I just want to mention I agree with Hugh on this.  Indeed if you wish to use a ribbon tweeter you need to put a capacitor in series with it because they usually short at DC.  The value of a device like the DEQX IMHO is not only does it make crossover design a doodle it does frequency, phase, and room correction - the last one being - IMHO - the weakest link.  Which is one reason I am attracted to the AKSonics - matching cabinet volume, port size, the room and source material is very important IMHO.  I think the real value of the DEQX as far as crossover design is concerned is taming difficult drivers like the resonance of the EXCEL magnesium cones that would normally be a challenge for the crossover designer.  Another issue with the DEQX is I have spoken to someone who works for a well known Australian loudspeaker designer (who I will not name) and he told me that he used it on a number of his designs and it produced an impressive but different sound - so different that even though it was totally phase and frequency corrected with the drivers perfectly time aligned many people did not like it.  

Thanks
Bill

[andyr]

I just want to mention I agree with Hugh on this.  Indeed if you wish to use a ribbon tweeter you need to put a capacitor in series with it because they usually short at DC. ...

Not so, Bill.  

My MGIIIA (true) ribbon has been driven quite happily by my AKSA 25 for several years.  What you DO need to be careful of in an all-active setup (partickerly those fragile ribbons!) is that the amp driving the tweeter does not exhibit a power-on "thump" ... which might blow them to bits.

Regards,

Andy

[andyr]

Hi Andy,

... A typical dynamic driver has a highly inductive voice coil, and ... Unless impedance is corrected for the voice coil, the load presented to the amplifier could well cause problems. In cases where it does not cause problems, there is inevitably some disturbance of the phase relationships in the music - and this seriously affects the imaging. Thus, I would argue that Zobel correction, a capacitor and a resistor in series across the voice coil, should be mandatory EVEN ON ACTIVE SYSTEMS!!...

Hi Hugh,

Now you've REALLY got me confoosed!!  

As you say, "a typical dynamic driver has a highly inductive voice coil".  To aid this problem, I can understand that Zobel correction is a good thing ... and yes, I can see this is true whether it's an active or passive crossover setup.  (Now I'm even more rapt in my Maggies!    )

HOWEVER, first ... am I correct in thinking the amount of voice coil inductance varies with frequency - and the cap/res values chosen for the Zobel correction are selected by taking the inductance at one particular frequency?

Which means at all other frequencies, they are the wrong values ... so your dynamic driver will be causing problems for the amp or affecting the imaging (due to disturbance of the phase relationships in the music) at all frequencies except the one used for the calculation?

At least in an active setup, the changing driver reactance just affects (or may be affecting) the amp ... in a passive setup, the varying driver reactance is screwing up the calculations on which the crossover component values were based - changing the crossover frequency!  

Second, series inductors (for LP slopes) and series capacitors (for HP slopes) add their own sonic colourations to the signal.  A Zobel cap is across the voice coil and so is not colouring the signal nearly as much ... ie. there's a heap of difference, sonically speaking, between just using a Zobel correction across the voice coil and adding the large series components to make up 1st or 3rd order slopes.

Regards,

Andy

[Tinker]

HOWEVER, first ... am I correct in thinking the amount of voice coil inductance varies with frequency - and the cap/res values chosen for the Zobel correction are selected by taking the inductance at one particular frequency?

Which means at all other frequencies, they are the wrong values ... so your dynamic driver will be causing problems for the amp or affecting the imaging (due to disturbance of the phase relationships in the music) at all frequencies except the one used for the calculation?

Hi Andy,
 the primary use of the Zobel is usually to tame the inducative reactance component that increases smoothly with frequency. This is usually well-enough behaved that a good approximation of flat can be obtained with this simple correction. The  peak associated with resonance is a whole different thing and is very hard to deal with.

Going back a post or two... having built and fiddled with variations in Siegfried Linkwitz's Phoenix, there are some nasty resonances and other anomalies that really benefit from the correction digital crossovers/DSPs can bring. I've experimented with the Behringer and great results can be had, but driver impedence correction is still mooted.

T.

[rabbitz]

Impedance correction tends to work well with parallel crossovers but not so hot with series crossovers from my experience. I use series on all my designs and I've tested several with a CR network and they had a detrimental effect on the dynamics and slam. I've tried values of the CR network by text book, software and manufacturer's recommendations and all were so so.... but other builders have had great success using them... go figure.

There are probably heaps of reasons why but a series xo is more forgiving with a reactive load. There is a benefit if the tweeter and woofer impedance at the crossover point are close to being equal. Since I don't use L-Pads but a single series padding resistor, this works to my advantage as the resistance values on each leg are almost the same due to the woofer's rising impedance. Another other reason is that a series xo uses the driver as a path to the other driver and using a CR increases the resistance in that part of the network. These observations only apply to series crossovers and really nothing to do with parallel xo's or active.

One thing this does give me is a higher impedance for the amp further up the frequency range and the corresponding reduction in power, but AFAIK you need a lot less power at 3K onward than you do at say 50Hz. Anyway, my AKSA 55 takes it all in it's stride and and copes with a 3R6 load at about 240Hz without a hiccup.

Just a different point of view on impedance correction for a different type of passive crossover.... and yeah... passive xo's have a lot of life in them yet.

[bhobba]

My MGIIIA (true) ribbon has been driven quite happily by my AKSA 25 for several years.  What you DO need to be careful of in an all-active setup (partickerly those fragile ribbons!) is that the amp driving the tweeter does not exhibit a power-on "thump" ... which might blow them to bits.

I bow to your superior experience in such matters.  My comment was about the ribbons I know; namely the Fountek, Arum Cantus and Raven.  Fountek for example carry the following warning: 'Ribbon tweeters show a dead short to your amplifier!  You must use a capacitor on them, when testing them or when using them with an electronic crossover!'  Considering they have a DC resistance of .02 ohms I would take such a warning very seriously and not risk it - the slightest offset in the amp and bang goes your tweeter.  One option I am considering is to use a pair of AKSA amps into Ambience ribbons fed via the DEQX.  I contacted Tony Moore at Ambience and he refused to sell me any without such a blocking capacitor.

Thanks
Bill

[AKSA]

Bill,

I hadn't realised that ribbons had such a low DC resistance, though I had wondered a few times as I gazed at that tiny strip of aluminium foil......  A blocking cap is indeed de rigeur.

I know AndyR's system well, but had not realised he was driving the tweeter ribbons with a direct coupled AKSA.  Since even the low power AKSA has a turn on thump, admittedly measured in milliseconds and down to less than 6V, I'm surprised he's reported no problems with it.  It sounds like manna from heaven, certainly, Andy, care to comment?

Cheers,

Hugh

[andyr]

I bow to your superior experience in such matters.  My comment was about the ribbons I know; namely the Fountek, Arum Cantus and Raven.  Fountek for example carry the following warning: 'Ribbon tweeters show a dead short to your amplifier!  You must use a capacitor on them, when testing them or when using them with an electronic crossover!'  Considering they have a DC resistance of .02 ohms I would take such a warning very seriously and not risk it - the slightest offset in the amp and bang goes your tweete ...

Hi Bill,

Well I guess as the saying goes "ribbons ain't ribbons"!  

Maybe the ones you listed are much shorter?  The 5' long, big-Maggie ribbons have a resistance of 2 ohms or 3 ohms, depending on which model you have.  But yes, 0.02 ohms would be another thing entirely!  

Still, some people regard putting electronic crossovers in front of any tweeter is "polluting" the sound so, if the ribbon can take it (in terms of its low frequency extension), a series cap giving a 6dB rolloff may not be a bad way to go?     All the other drivers can be actively filtered.

Regards,

Andy

[bhobba]

I hadn't realised that ribbons had such a low DC resistance, though I had wondered a few times as I gazed at that tiny strip of aluminium foil......  A blocking cap is indeed de rigeur.

Please remember the drivers I mentioned contain a transformer driving the ribbons.  I suspect that is the reason it is so low - not the properties of the ribbon itself.

If I recall correctly Tony's Ambience Ribbons do not have a transformer so the resistance is unlikely to be that low - still I suspect he would not strongly suggest a series cap without it being an issue.

Thanks
Bill

[andyr]

Please remember the drivers I mentioned contain a transformer driving the ribbons.  I suspect that is the reason it is so low - not the properties of the ribbon itself.

If I recall correctly Tony's Ambience Ribbons do not have a transformer so the resistance is unlikely to be that low - still I su ...

Hi Bill,

If you want to get ribbon magic with your DEQX then Maggie is the way to go!

Regards,

Andy