Direct Coupled Drivers Vs Crossover

[Kingrex]

I was wondering about the concept of altering a signal vs altering the output from a amp. I figured the signal is very fragile and easy to degrade. It seems that is the case.

I understand how difficult it is to make a crossover. On the surface that is. I was wondering in a general way if a company with resources and tallent like Magico wouldn't be able to make a higher performance product with direct coupled amps. Or are the trade off of crossover to signal manipulation a wash.

Customers do loose the option to shape sound with an amp. Or do they. A good amp is not supposed to have a voice.

And yes, I do have an itch to try. Mabe I can get a few Allo Volt Plus D and a Danville crossover.

 

[Audiohertz2]

Are you saying use an active analog on the mid/highs and a full active on the woofers?

Yes passive on mid/highs , full active on bass ..!

 

[Audiohertz2]

Most commercial loudspeakers have compromises that require tweeks in the crossover to work well. I have talked to Greg Timbers about the JBL DD67000 Everest, the last statement loudspeaker he designed at JBL before being forced out by the new owners. He said he was frustrated because the product design guys kept changing his design for their aesthetic goal. He therefore needed to design a crossover that is more complicated than necessary to compensate. He told my friend who owns a pair of these to remove the supertweeters and reposition them optimally to simplify the crossover, amongst other tweeks, to undo the problems introduced by the aesthetics of the design.
Therefore, it is not wise to try and modify commercial speakers for active crossover unless this is recommended by the manufacturer.
I build my speaker setup from the ground up with active crossovers in mind. I use an Accuphase F25 analog crossover, running 4th order L-R slope. I find this gives a cleaner result compared to 12dB slope. I have experimented with different amps before settling on a low powered class A transistor module for the Acapella plasma tweeters, 300B push-pull amps for the field coil midrange horns and Parasound transistor amp for the field coil woofers in ported enclosures. The field coils allow optimisation by adjusting the supply current to change the drive behaviour. I also physically time aligned the drivers. There are lots of variables but I am pretty settled now after experimenting over a span of a decade. The gain in clarity and dynamics by directly coupling the drivers with the amps are well worth the effort.

Well you will have less outer band distortion, sweet, but that 24db sucks the life out of everything thou ..!


Regards

 

[Audiohertz2]

In theory, direct connection from a power amp to a driver is superior. Damping factor is optimal, resulting in greater dynamics and faster transients. As you mention, power requirement drops significantly. A passive crossover has to drop the level of each driver to match the least sensitive driver. For example, say your tweeter is 92dB/W/m, mid is 89dB, bass driver 86dB. You would lose 6dB to resistors, and instead of a 25W amp, you need 100W just to equalize driver sensitivity. Then you'll need baffle step correction, generally 4-6dB. Now you need ~300W for the same level as your crossover-less 25W system.

In the real world however, an active system is no panacea. I have tried to go active with a Marchand XM44 electronic crossover, Marchand XM46 passive line-level crossover and FMod modules. None were transparent, all compromised the sound of my amps. I then tried a single high quality cap on the input of my power amp to roll off the bottom end by 6dB/octave at 80Hz, relieving the monitors of the bass range to cross to a sub. The cap did what it was supposed to, it relieved the mid-woofers of deep bass, but caused grit at high frequencies, a bad trade. You are going to hear anything inserted between a preamp and power amp. The crossover had better be at least as good as your pre and power amplifiers.

You need to know something about speaker design to implement an active crossover. Simply ripping out a crossover and substituting an active crossover is almost guaranteed to fail. Mid and bass drivers often need a notch filter to deal with resonance above their operating range. Even with an active system, speakers need baffle step correction. Then there's time alignment, a tricky problem, compounded by the acoustic slope (natural roll-off) of the drivers and the electronic slope of the crossover. Speaker design is complicated, whether active or passive.

Amps are not cheap, nor are cables, and IME they need to extremely well matched for anything above the subwoofer range. Some people seem happy with different amps on their mid, tweeter and bass drivers. That was just awful to me. It was easy to get better bass and treble with non-identical amps, but coherence suffered; it was hi-fi sounding, not musical. I gave up on active crossovers. I run my main speakers full range, with subs connected in parallel and DSP corrections on the subs only.

A DSP solution with convolution filters, auto-eq software like REW or Accuphase and a USB microphone may be an ideal solution for someone with a fully digital system. I looked into it but haven't gone much further. It's a steep learning curve, and my system sounds so good to me right now that incentive is low for another audio system battle.

Its not simple by any means , active or passive , but you can do proper compensations with the passive unlike the active , this shows much in the critical midband /Highs section so not surprised you had to insert a 6db passive roll off, to start ...!

The biggest fly in the ointment is the grungy active Xover signal and tone, hi quality pre signal , getting each amp to be phase correct to each other, on and on ..!


Regards

 

[adrianywu]

Well you will have less outer band distortion, sweet, but that 24db sucks the life out of everything thou ..!


Regards

Why should that be ?

 

[audiobomber]

Sage advice.
Whats Baffle Step Correction? Is that another way to say time align?

Baffle step correction is a way of compensating for the way a sound wave bounces off of, and is reinforced by, the shape of the cabinet surrounding the driver. At lower frequencies, a loudspeaker cabinet is too narrow to reinforce the sound waves. Higher frequencies OTOH, launch off the cabinet, and are therefore louder.

A driver mounted in a wall does not need BSC. Traditional rectangular speaker cabinets have an easily identifiable baffle step frequency, and can be compensated using a simple calculation. f3= 4560/Wb where Wb is the baffle step in inches. A 12" wide cabinet for example will have a -3dB point at 380Hz. Non-rectangular cabinets are more difficult to correct.

"At high frequencies, a loudspeaker radiates sound directly forward in half space. At low frequencies, the sound is not direectional and radiates into full space. This results in a gradual shift in the frequency response of -6dB from the highs to the lows."

Loudspeaker Diffraction Loss and Baffle Step Compensation Circuits

Loudspeaker Diffraction Loss and a Baffle Step Compensation (BSC) Circuit Calculator for Speakers.

diyaudioprojects.com

 

[Audiohertz2]

Wide at the woofer narrow at the tweeter unless some smearing and a lack of imaging is not a big deal ..!


Regards

 

[Kingrex]

Yes passive on mid/highs , full active on bass ..!

Passive to me means caps, inductors and resistors after the amplifier.

 

[audiobomber]

I understand how difficult it is to make a crossover. On the surface that is. I was wondering in a general way if a company with resources and tallent like Magico wouldn't be able to make a higher performance product with direct coupled amps. Or are the trade off of crossover to signal manipulation a wash.

I expect it's mostly about marketing, not engineering. I am very confident that Magico could make an outstanding active speaker system, as many other firms do. But audiophiles seem pretty resistant to active systems so far.

Customers do loose the option to shape sound with an amp.

Definitely one reason.

Or do they. A good amp is not supposed to have a voice.

In theory, but is there an amp with no voice? Every amp I've heard has had a signature.

And yes, I do have an itch to try. Mabe I can get a few Allo Volt Plus D and a Danville crossover.

You would modify your existing speakers, not considering an active loudspeaker system, correct?

I looked at the miniDSP Flex system, but I was concerned that the DACs weren't up to the level of my existing DAC. I would have the same reservation about any digital solution in a box.

Danville says they provide semi-custom solutions. I believe you would still need to take on the speaker designer role, to implement the crossover slopes and frequency, BSC, notch corrections, and time alignment. If the Danville box allows integration with Accuphase, REW or other Auto-EQ capable software like the miniDSP SHD or minDSP Flex products do, that would simplify things.

 

[Maril555]

In theory, direct connection from a power amp to a driver is superior. Damping factor is optimal, resulting in greater dynamics and faster transients. As you mention, power requirement drops significantly. A passive crossover has to drop the level of each driver to match the least sensitive driver. For example, say your tweeter is 92dB/W/m, mid is 89dB, bass driver 86dB. You would lose 6dB to resistors, and instead of a 25W amp, you need 100W just to equalize driver sensitivity. Then you'll need baffle step correction, generally 4-6dB. Now you need ~300W for the same level as your crossover-less 25W system.

In the real world however, an active system is no panacea. I have tried to go active with a Marchand XM44 electronic crossover, Marchand XM46 passive line-level crossover and FMod modules. None were transparent, all compromised the sound of my amps. I then tried a single high quality cap on the input of my power amp to roll off the bottom end by 6dB/octave at 80Hz, relieving the monitors of the bass range to cross to a sub. The cap did what it was supposed to, it relieved the mid-woofers of deep bass, but caused grit at high frequencies, a bad trade. You are going to hear anything inserted between a preamp and power amp. The crossover had better be at least as good as your pre and power amplifiers.

You need to know something about speaker design to implement an active crossover. Simply ripping out a crossover and substituting an active crossover is almost guaranteed to fail. Mid and bass drivers often need a notch filter to deal with resonance above their operating range. Even with an active system, speakers need baffle step correction. Then there's time alignment, a tricky problem, compounded by the acoustic slope (natural roll-off) of the drivers and the electronic slope of the crossover. Speaker design is complicated, whether active or passive.

Amps are not cheap, nor are cables, and IME they need to extremely well matched for anything above the subwoofer range. Some people seem happy with different amps on their mid, tweeter and bass drivers. That was just awful to me. It was easy to get better bass and treble with non-identical amps, but coherence suffered; it was hi-fi sounding, not musical. I gave up on active crossovers. I run my main speakers full range, with subs connected in parallel and DSP corrections on the subs only.

A DSP solution with convolution filters, auto-eq software like REW or Accuphase and a USB microphone may be an ideal solution for someone with a fully digital system. I looked into it but haven't gone much further. It's a steep learning curve, and my system sounds so good to me right now that incentive is low for another audio system battle.

I was trying to say the same early on, apparently not as eloquently. But I couldn't have agreed more.
Yet again, the solution points towards high efficiency, preferably 2-way speakers and low power tube amplification, like it was done 60 years ago.
Ain't that a full circle. Taken me 20 years of playing around to realize that.

 

[Kingrex]

I expect it's mostly about marketing, not engineering. I am very confident that Magico could make an outstanding active speaker system, as many other firms do. But audiophiles seem pretty resistant to active systems so far.


Definitely one reason.

In theory, but is there an amp with no voice? Every amp I've heard has had a signature.

You would modify your existing speakers, not considering an active loudspeaker system, correct?

I looked at the miniDSP Flex system, but I was concerned that the DACs weren't up to the level of my existing DAC. I would have the same reservation about any digital solution in a box.

Danville says they provide semi-custom solutions. I believe you would still need to take on the speaker designer role, to implement the crossover slopes and frequency, BSC, notch corrections, and time alignment. If the Danville box allows integration with Accuphase, REW or other Auto-EQ capable software like the miniDSP SHD or minDSP Flex products do, that would simplify things.

This is my crossover. I can only imagine how much is being sucked up and turned into heat. And how much I am hearing all the parts.

I was on a chat group with a very DSP friendly following. Too friendly and I felt the need to check out. There were too many there that were adamant a Mini DSP SHD was so transparent you would be stupid to not run your analog through it as the gains by adjusting phase, freq and impulse would far outweigh the loss of anything the listener thinks they hear with analog. And maybe they are right. I have no idea. But I'm not going to listen to people tell me my stereo will never sound any better than matginal if I don't embrace DSP.

That still does not mean I don't wonder what might be done with an active system. An active system and adding DSP to a passive crossover are not the same thing.

It has been informative to hear the signal is sensitive and can easily pick up distortions from an active crossover. And that its really hard, even with good equipment to dial in the speakers. Heck, Adrian said the spent 10 years on his.

My gut also tells me the perceived results will be impacted by what level am I at now. And how much time and investment am I willing to put forth to make it right. A Allo Volt+D is far from right. Paying Scott to make me a pair of 3 channel amps might end up outstanding.

I did try biamping with the Audion 845 and it was such a disaster I unplugged it all after 10 minutes.

Scott had told me to get air core chokes. And Carlos talks about ditching the resistors and going with custom transformers to adjust the gain. I can get my head around a choke. But I fear you change out a resistor and you impacting more than gain. Can of worms.

 

[DasguteOhr]

you're talking about this speaker right?ob coax10"with 2× 15" emminence woofer.

 

[Kingrex]

you're talking about this speaker right?ob coax10"with 2× 15" emminence woofer.
View attachment 129366

Exactly

 

[DasguteOhr]

Exactly

Disconnect the woofer cable from the passive xover. Get yourself a longer Klotz speaker cable (original cable) . Then buy a Yamaha studio amp with internal subwoofer xover e.g model p3500s (~500$ used).


Connect the woofer directly to the amp then you have a stable active bass. Then buy a handy app audio tool and a handy microphone and measure the frequency response. On the amplifier, adjust the xover frequency and the volume to the passive coax. The partially active loudspeaker is ready. Greetings from the partially active openbaffle listener.
I forget you need t adapter and rca to xlr cable for simultaneous operation of both power amplifiers on one pre-amplifier.

P.S i drive passive my coax 12db filter tweeter, midrange/bass only a notch filter better rolloff with set amps

 

[SCAudiophile]

lf you have sensitive drivers, why would you use Class D?

Possibly to keep cost down versus the generally more costly Class A or Class AB alternatives. Also, if building internally, there may be space and weight considerations (and cost) that make Class D a better option.

 

[SCAudiophile]

I'm curious what the delta would be if you had say a Wilson speaker with a Dagostino amp vs the same Wilson and no crossover with say Legacy class D amps driving each speaker and a factory tuned analog active crossover driving it.

@SCAudiophile
Mark, is Legacy all direct coupled Amps? The Wavelet is the crossover? Does Legacy offer a active analog crossover for vinyl enthusiasts? Is the Wavelet analog or digital?

@Kingrex Rex, There's the older Wavelet and newer Wavelet II; the latter is what I would definitely recommend to use now as it has an much more capable DAC and 64-bit DSP versus the prior version 56-bit.

Either Wavelet is primarily a digital device. It does have 2 XLR analog inputs and 2 RCA analog inputs along with multiple digital inputs. Anything coming in to the XLR or RCA analog inputs goes through a very good ADC stage; after that, like the digital inputs, all core functionality is done in digital domain. The 24/192 DACs are then used to output up to an 8-way output (4 channels per speaker or 3 per speaker plus 2 for subs), etc. It's highly programmable at Legacy and at least other manufacturers OEM it for their own speakers.

There is no pure analog version of Wavelet as what is does for Room Correction I believe is impossible purely in the analog domain and without the 64-bit DSP engine's power.

The outputs are XLR analog out (1 set of 8) or RCA single-ended analog out (1 set of 8). Note you can turn room correction on or off on the web/mobile control panel software. You can also select "Bypass" all together to have what comes digital have minimal processing through
the DACs and then straight to the outputs or anything that came in from an analog input sent straight through as I understand it direct to the
analog outputs (but obeying the output map that's programmed in).

The room correction algos all operate in the 64-bit digital domain of the excellent DSP engine the Wavelet II has.

Whether the outputs are direct-to-amp after the split-out by Wavelet, in the purest sense depends on the speaker in question,
the output map needed, etc... For example in Valor, some channels go direct in to the speaker drivers from either Wavelet or external
amps. In at least two inputs to the speaker there is a minimal passive crossover internally when there is are different drivers in the speaker
to cover that band. The same for Aeris,..you can actually see the internal passive crossover portion through the grille cloth of the top portion
for certain frequency.

Any more detail would be an even longer post and I don't think that's what you are looking for. Perhaps Ed from Legacy who appears to be here on this forum now could give more details that are being my knowledge/capabilities.

***For vinyl, tape, other enthusiasts: The Wavelet comes with multiple analog inputs. A vinyl lover with a TT, phono pre and preamp could easily run from their preamp straight into Wavelet II's analog inputs. Same for RTR tape enthusiasts or anyone with anything (including disc or server-streamer sources) feeding a SOTA external preamp (or a DAC with a great volume control for its analog outputs. I do exactly this with a full Esoteric Grandioso + Cybershaft digital stack feeding an Esoteric preamp which then feeds Wavelet.

 

[Kingrex]

Disconnect the woofer cable from the passive xover. Get yourself a longer Klotz speaker cable (original cable) . Then buy a Yamaha studio amp with internal subwoofer xover e.g model p3500s (~500$ used).
View attachment 129381View attachment 129382

Connect the woofer directly to the amp then you have a stable active bass. Then buy a handy app audio tool and a handy microphone and measure the frequency response. On the amplifier, adjust the xover frequency and the volume to the passive coax. The partially active loudspeaker is ready. Greetings from the partially active openbaffle listener.
I forget you need t adapter and rca to xlr cable for simultaneous operation of both power amplifiers on one pre-amplifier. View attachment 129383

P.S i drive passive my coax 12db filter tweeter, midrange/bass only a notch filter better rolloff with set amps

Interesting. Do you find the need for a insolation transformer to split the signal from the preamp to 2 amps. I was not sure of part of the reason my biamp sucked so bad was the pre not driving 2 sets of amps well.

 

[Kingrex]

Disconnect the woofer cable from the passive xover. Get yourself a longer Klotz speaker cable (original cable) . Then buy a Yamaha studio amp with internal subwoofer xover e.g model p3500s (~500$ used).
View attachment 129381View attachment 129382

Connect the woofer directly to the amp then you have a stable active bass. Then buy a handy app audio tool and a handy microphone and measure the frequency response. On the amplifier, adjust the xover frequency and the volume to the passive coax. The partially active loudspeaker is ready. Greetings from the partially active openbaffle listener.
I forget you need t adapter and rca to xlr cable for simultaneous operation of both power amplifiers on one pre-amplifier. View attachment 129383

P.S i drive passive my coax 12db filter tweeter, midrange/bass only a notch filter better rolloff with set amps

Have you tried this with a PAP Trio 15 coax. I think the woofers cross in at 1200 hertz. The Yamaha only goes to 150 hertz. Am I reading the spec on the Yamaha wrong?

 

[Kingrex]

@Kingrex Rex, There's the older Wavelet and newer Wavelet II; the latter is what I would definitely recommend to use now as it has an much more capable DAC and 64-bit DSP versus the prior version 56-bit.

Either Wavelet is primarily a digital device. It does have 2 XLR analog inputs and 2 RCA analog inputs along with multiple digital inputs. Anything coming in to the XLR or RCA analog inputs goes through a very good ADC stage; after that, like the digital inputs, all core functionality is done in digital domain. The 24/192 DACs are then used to output up to an 8-way output (4 channels per speaker or 3 per speaker plus 2 for subs), etc. It's highly programmable at Legacy and at least other manufacturers OEM it for their own speakers.

There is no pure analog version of Wavelet as what is does for Room Correction I believe is impossible purely in the analog domain and without the 64-bit DSP engine's power.

The outputs are XLR analog out (1 set of 8) or RCA single-ended analog out (1 set of 8). Note you can turn room correction on or off on the web/mobile control panel software. You can also select "Bypass" all together to have what comes digital have minimal processing through
the DACs and then straight to the outputs or anything that came in from an analog input sent straight through as I understand it direct to the
analog outputs (but obeying the output map that's programmed in).

The room correction algos all operate in the 64-bit digital domain of the excellent DSP engine the Wavelet II has.

Whether the outputs are direct-to-amp after the split-out by Wavelet, in the purest sense depends on the speaker in question,
the output map needed, etc... For example in Valor, some channels go direct in to the speaker drivers from either Wavelet or external
amps. In at least two inputs to the speaker there is a minimal passive crossover internally when there is are different drivers in the speaker
to cover that band. The same for Aeris,..you can actually see the internal passive crossover portion through the grille cloth of the top portion
for certain frequency.

Any more detail would be an even longer post and I don't think that's what you are looking for. Perhaps Ed from Legacy who appears to be here on this forum now could give more details that are being my knowledge/capabilities.

***For vinyl, tape, other enthusiasts: The Wavelet comes with multiple analog inputs. A vinyl lover with a TT, phono pre and preamp could easily run from their preamp straight into Wavelet II's analog inputs. Same for RTR tape enthusiasts or anyone with anything (including disc or server-streamer sources) feeding a SOTA external preamp (or a DAC with a great volume control for its analog outputs. I do exactly this with a full Esoteric Grandioso + Cybershaft digital stack feeding an Esoteric preamp which then feeds Wavelet.

Thanks Mark. I have to digest this all. It looks a bit expensive. I need the Wavelet and the IV2

 

[DasguteOhr]

Interesting. Do you find the need for a insolation transformer to split the signal from the preamp to 2 amps. I was not sure of part of the reason my biamp sucked so bad was the pre not driving 2 sets of amps well.

What is output impedance from your preamp? I think no need for isolation transformer the cold signal path is bridged to ground.
The tweeter is crossed at 2 khz and 10" cone is lower then 200 hz
Exact values will only bring measurement, disconnect the speaker cable from the woofer and then only measure coax frequency response.
Normally the open baffle or ripol woofers are limited to 150hz with a blocking circuit
Can you make a photo from backside of coax, i think its a modified bms coax driver