tri-amping and active xover - TOTAL SEPARATES?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Daveis:

I reread my post and after thinking it sounds like from the tone I'm trying to insult you or to be a jerk. I'm not, I just get passionate when people dismiss ideas or technologies because they are either inexpensive or aren't what traditionally falls into the catagory of "audiophile". Don't take it the wrong way, I didn't post it to insult you.
 
As a casual observer of this thread it appears that there are several different issues being all mixed together and considered all at once. It may not be an all or nothing situation here.

For example the original idea was digital, active, line level crossovers with multiple power amps driving individual drivers. Now for someone into T-amps and computers this might be a very reasonable approach but the guy sitting there surrounded by SET amps, well, I think not. He might go for tube based passive-with-gain line level crossovers and drivers with naturally smooth response that do not require notch filtering. He would tend to use acoustic measures to deal with any remaining response shaping. Now the guy with a huge SS amp might feel better with power level crossovers.

So the issues would seem to be.

- Line v.s. power level crossover.
- Active v.s. passive (passive could be line or power level).
- Digital v.s. analog
- Cheap v.s. golden ear.
- Electronic v.s. acoustic.

Combination approaches might well be the best approach. My personal bent would be toward a single class A amp ( preferably SE tube ) per channel driving a wide range with low order passive power level crossover on a helper tweeter. Line level crossover (would consider either passive or active) driving a seperate sub woofer or two using a non-switching amp. I would however have no real objection to a seperate tweeter amp using line level analog crossover.

mike
 
m0tion said:
...But you don't need super "audiophile" crossover components to do it and there is absolutely no argument to be made what-so-ever that passive crossovers are TECHNICALLY superior to active crossovers. You can argue all day long that they subjectively sound better, but don't make the mistake that active crossovers don't have measurable technical superiority if designed correctly. Just ask Siegfried Linkwitz...

When I refer to passive crossovers, I mean power level, I don't consider passive line level crossovers viable - See Linkwitz for many reasons.

I think a point often missed by many active crossover experimenters is that the comparison must be made apples to apples - the transfer functions should be the same for the active and passive crossovers for a comparison to be valid.

Simply dialing in an active LR4 at the desired frequency is only slightly more valid than using the textbook formula to design a passive crossover (the active xo won't be affected by the drivers' impedance changes making it "slightly more valid").

A competently designed passive crossover can sound about as good as a competently designed active crossover, at the expense of a system efficiency and (slight) loss of amp control over the drivers. However, some systems make passive crossovers impractical, such as Linkwitz' Orion.

An advantage of the active approach is fairly easy incorporation of notches, peaking and shelving equalization as well as phase correction. Rare is the accurate driver that does not need some or all of these.

Of course this all can be accomplished passively, at the cost of further power loss in the crossover. Better crossover designers can incorporate shelving types of EQ (such as baffle step compensation) into the basic crossover without increasing parts count. See Zaph's ZD5 project for an example (although the phase alignment in the tweeter section adds a few parts)

Don't forget that arriving at your passive crossover with $300 in parts will probably have many parts leftover. If you're a serial builder, like many here, fine, you'll probably end up using them in another project. I have a hard time with "carrying inventory" although I have built up quite a stock of crossover parts over the years.

For prototyping I now have several adjustable analog active crossover boards (MOX) that can be daisy chained to achieve a variety of slopes. I'm also building modular EQs in their prototyping area. Pretty much nothing will get thrown away in development.

On the hiss front, I've often seen it said that the Behringer and dBx boxes need to be run pretty much at full signal to maximize their S/N, and attenuated afterwards. The only way to do this and have each section track accurately is with a stepped attenuator on each output. This way you'll attenuate any hiss along with the signal. I haven't felt a compelling need to do this with my analog crossovers yet.
 
Before we too fixated on the hiss issue, let me try and calibrate how loud the hiss is in my system and what I have done to improve the situation.

At first I had the dBx inut sitting right on top of my amps and the noise level was significant, audible at my listening position when no music was playing. Since then I have done the following :

1. Moved the dBx out away form the other equipment.
2. Improved the "internal" speaker wiring and connections.

Both of these changes reduced the noise level. The hiss right now is not audible at my listening position and is low when I get very close to the speaker drivers. The speakers are extremely efficient. By low I mean less than the background noise on an average record (remember black disks with grooves) and less than the noise I once heard in a high end dealer's show room when he was demoing Chello (spelling?) electronics and B&W speakers. I would say it is on the same level as most reasonably priced commercial SS amps or maybe just slightly above. It is definitely lower then what I have heard from some tube amps.

I consider this good enough, not a limiting problem. I have not heard any other weird buzzing or other noises. There are other bigger fish to fry in terms of making my system perform better.
 
m0tion said:
Daveis:

DCX2496 != a dozen cheap opamps. Thats a Behringer CX3400.

DCX2496 = about 2 cheap ADCs, 6 cheap DACS, and a cheap Sharc DSP.


I take it you've listened to the DCX2496 in stock or semi-modified form and think that all the opamps it uses after the DAC actually sound good? They dont.

Over the years, the the one common thread I've discovered is the fewer opamps in the signal path the better. The only exception I have ever had to that was an M3 headphone amp using OPA627.
That design, BTW, is quite nice.

Another thing I've heard consistently, is that good passive components (eg. transformer, caps) often sound better than active circuits. examples, I/V conversion in a DAC.

The gain by going active versus passive is a poor engineering choice. Better to put your dollars into better speaker drivers, etc.

Since I cant hear much difference between passive or active(and I've tried both), why should I pay more? So my distortion analyzer or Praxis can tell me my speaker is .5% better? I call that poor engineering.

I hope that my expensive experiments might save someone else from going down a road that's only going to cost them money and not gain them much better sound.

Linkwitz may be a genious, but to say that his design choice is the only good engineering choice is silly. The best speakers I've heard all use passive crossovers.

If you are using chipamps then by all means try it. But I wouldnt go into a totally active speaker project thinking you need alot of high quality amps. You'll break your bank account and be unhappy in the end.

One nice benefit of active is that using expensive amps doesnt really help. I've almost found that a bunch of cheap Panasonic receivers sounds as good as Class-A SS amps for much less money.

The crossover for me is WHERE the mojo happens. It has to be first rate. The only thing more important is your selection of drivers and cabinet size/porting. If you dont get the crossover right you will hate your speakers.
 
Engineering is all about trade-offs to reach a given set of goals. Everyone has different goals and values the trade-offs differently. I enjoy building amps, so have plenty. The cost of the amplifiers is not an issue for me. For those who must buy quality amps, cost is obviously an issue.

Apparently the great unwashed masses cannot hear the difference between a low bitrate MP3s and SACDs, so for them we're not going anywhere near the aspirations of the majority of us here. It's a matter of how far out we are willing to go on the cost/benefit curve and how we define the point of diminishing returns.

I will gladly eliminate opamps when possible, but if one of the design priorities happens to be board space, discrete buffers, gain stages and line drivers may not be a viable option. Certainly there are opamps that don't sound good, but there are also good sounding opamps and opamp implementations. Passive XO components, inductors in particular, are not without their issues, but seem to be a better option for you.

I think that we have pretty well established that to my ears the benefits of active crossovers outweigh the cost and to your ears they do not. You dislike equipment that others find acceptable. Fine, lets agree to disagree.

That's a big part of the fun of this hobby for me - finding out what makes an audible difference to me and what doesn't. I really didn't expect to hear a difference between my Leach amps and my A75s, but I did. Hmm, lets try something else now and see if I like it better. :cool:

BTW, for me crossover design is where the MOJO is, and is MORE important than box tuning and driver selection (assuming a basic quality level has been met).
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I consider this good enough, not a limiting problem. I have not heard any other weird buzzing or other noises. There are other bigger fish to fry in terms of making my system perform better.

I urge you to see about averaged power response at listening position. That is a mighty fish to check out and its nothing to fix since you have the dbx already. What measurement system you use?
 
The finest speaker system I have ever heard is a 4 way active system built by lenard audio. lenardaudio.com.au

Lenard only makes 4 way active systems since 1977 and horn loading is his favorite. He uses industrial style of components ie PA but his results are true audiofile:angel: . The Opal series are heaven and the Cinema System is the finest sounding cinema I have experienced (the Chauvel in Paddington, Sydney. Australia.)

Lenard has told me that active can and is the ideal, but to beware of poor quality active crossovers.
 
I am finding it interesting the several references to this piece of equipment as non-audiophile, while many of the reviews (including the AudioXpress "mod" recently published, refer to it as a production/profressional quality piece of equipment. What is it about DCX2496 that would make any/some/all of you consider it not up to "audiophile" quality/standards? Thanks,

Terry



BobEllis said:


When I refer to passive crossovers, I mean power level, I don't consider passive line level crossovers viable - See Linkwitz for many reasons.

I think a point often missed by many active crossover experimenters is that the comparison must be made apples to apples - the transfer functions should be the same for the active and passive crossovers for a comparison to be valid.

Simply dialing in an active LR4 at the desired frequency is only slightly more valid than using the textbook formula to design a passive crossover (the active xo won't be affected by the drivers' impedance changes making it "slightly more valid").

A competently designed passive crossover can sound about as good as a competently designed active crossover, at the expense of a system efficiency and (slight) loss of amp control over the drivers. However, some systems make passive crossovers impractical, such as Linkwitz' Orion.

An advantage of the active approach is fairly easy incorporation of notches, peaking and shelving equalization as well as phase correction. Rare is the accurate driver that does not need some or all of these.

Of course this all can be accomplished passively, at the cost of further power loss in the crossover. Better crossover designers can incorporate shelving types of EQ (such as baffle step compensation) into the basic crossover without increasing parts count. See Zaph's ZD5 project for an example (although the phase alignment in the tweeter section adds a few parts)

Don't forget that arriving at your passive crossover with $300 in parts will probably have many parts leftover. If you're a serial builder, like many here, fine, you'll probably end up using them in another project. I have a hard time with "carrying inventory" although I have built up quite a stock of crossover parts over the years.

For prototyping I now have several adjustable analog active crossover boards (MOX) that can be daisy chained to achieve a variety of slopes. I'm also building modular EQs in their prototyping area. Pretty much nothing will get thrown away in development.

On the hiss front, I've often seen it said that the Behringer and dBx boxes need to be run pretty much at full signal to maximize their S/N, and attenuated afterwards. The only way to do this and have each section track accurately is with a stepped attenuator on each output. This way you'll attenuate any hiss along with the signal. I haven't felt a compelling need to do this with my analog crossovers yet.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
To Cameron:

Best professional system I ever heard was based on the same active drive concept you refer and horns, but it was Funktion One and XTA with MC2 amps and a Sony SACD top loader (best and first one) as source.

This guy you told us about was right for paying special attention to the digital processor and that was my point in this thread.
 
tlparker,

I used a DCX2496 for about a year. The under $300 I paid for it was a bargain. One night I decided to hook the DCX2496 upto my M3 headphone amp and use it as a dedicated DAC with the crossover functions turned off. It sounded far worse than my MAudio SuperDAC which I only paid $200 for. Now, the interesting thing is that the SuperDAC has similar AKM DAC's in it.

My DCX2496 is currently using an opamp board mod by bdipoala. It's a little better than stock as it removes many opamps and pro-level balanced circuitry and gives your remote volume control.
It, too, isnt as good as the MAudio SuperDAC in sound quality.

There's an active Yahoo forum for the DCX2496. You'll find the hardcore DCX modders hanging out there(including the AudioXPress article author you mention). They are ALL extremely happy with their modding efforts. Some claim their modded SRC2496 sounds better than a Benchmark DAC1. That's a serious DIY accomplishment.

The DCX2496 is very modular with ribbon cables connecting various boards. Some have built X-BOSOZ outputs for it as well.

The conflicting reports about the relative value of the DCX2496 is coming from people who are using the unit stock versus heavily modified. Since the DCX2496 has fixed analog outputs you have to work through how you are going to control volume of that many channels. Two solutions might be the Twisted Pear Joshua Tree Attenuator or a DACT CT2 6-channel.
 
Well, I'll jump into the fray and share my opinions and experiences.

I run a tri-amped configuration. This is something I've evolved into for fun and in pursuit of better sounding music. since this isn't an amp discussion I'll stay out of those details. I do however use a DBX 234XL active crossover which I'm not happy with and will probably soon replace. The crossover is loaded with features, but I don't like the low-freq crossover. It isn't as clean as the Paradigm X-30 I previously used. Will likely buy an Ashly or build something.

I've separated my speakers into three types of enclosures. This allows them to be placed in the listening area in a way that best suits the characteristics of each. It also allows more flexibility in cabinet design. The sub x-over point is 80 hz and the HF x-over is 10K hz. This leaves a lot of bandwidth in the middle. No single driver will properly handle it. The three drivers in my JBL L100 do a wonderful job.

A good active crossover allows you to control the gain of each channel as well as contour the crossover frequencies. I wouldn't go any other way.
 

Attachments

  • stack1.jpg
    stack1.jpg
    82.3 KB · Views: 485
Ah ha! This was one of the points I was first making (it's even in the title), yet it's been (up to now) totally ignored. The ability to put each driver into its own "perfect" enclosure would seem to me a tremendous benefit with this approach. So much of speaker design (and I've read a LOT about it and built a few of my own) seems to deal with cancelling the interactions between drivers in the same enclosure that I thought this would be an obvious advantage to go "TOTALLY SEPARATES", yet you're the first one to bring it up. Interesting.


HipoFutura said:
Well, I'll jump into the fray and share my opinions and experiences.

I run a tri-amped configuration. This is something I've evolved into for fun and in pursuit of better sounding music. since this isn't an amp discussion I'll stay out of those details. I do however use a DBX 234XL active crossover which I'm not happy with and will probably soon replace. The crossover is loaded with features, but I don't like the low-freq crossover. It isn't as clean as the Paradigm X-30 I previously used. Will likely buy an Ashly or build something.

I've separated my speakers into three types of enclosures. This allows them to be placed in the listening area in a way that best suits the characteristics of each. It also allows more flexibility in cabinet design. The sub x-over point is 80 hz and the HF x-over is 10K hz. This leaves a lot of bandwidth in the middle. No single driver will properly handle it. The three drivers in my JBL L100 do a wonderful job.

A good active crossover allows you to control the gain of each channel as well as contour the crossover frequencies. I wouldn't go any other way.
 
hipoofutura's design

Hippo, I'm curious, after looking at your pic. Do you connect to the JBL's and treat them as a single drive in your active crossover, which means you're also using the passive crossover inthe JBL's. Or, Are you bypassing the JBL xover and direct connecting to the drivers?
 
Hi, I use active crossover for 3 years now and will never go back to passive. I had 2 Rane 4 way mono crossover since my speakers are 4 way, one for each speaker, but was not happy at all with the sound. Since I knew what freq. was best for my speaker, I decided to build the one from the ELLIOT SOUND PRODUCT website but built it 4 way instead of 3 way. I use NE5532 for the op amp , metal film resistors and all caps are polystyrene except for the 10uF polypropylene at the output. What a difference , better bass, better mids and better high, nothing to compare with the Rane. What I like about active crossover is no CD are recorded the same so it's easy to adjust only the tweeter or the mid or the bass to get the sound you're looking for. Sure this set up take more space because you must have 4 power amp + preamp ect but I think the sound you get from this kind of set up worth it.
 
The L100 is fed by the "middle" band on the active crossover. This means it gets everything between 80 Hz and 10K Hz. Since this is such a wide band all three drivers in the L100 are used and controlled by the JBL passive crossover in the cabinet.

I know this isn't specifically what you were asking for, but I'll throw it in any way. Tri-amp allows you so select specific amps that work best within their respective frequencies. As far as I'm concerned this is as important as having separate speakers. I'm stunned by those who feel separating the amps isn't that important. I use a pair of tube mono-blocks to drive the L100s. Handling the mids is what tubes do best. The voicing and distinction between different instruments is outstanding. I use SS amps for the subs and UHF. The subs are driven by a 200 WPC @ 8 ohm amp and the ring radiators have a 50 WPC amp. As an experiment I built a pair of 2 WPC chip amps and put them in the UHF cabinets. They had more than enough power! The JBL 2405 UHF driver are so efficient I was using less than 2 watts. And to be honest I couldn't tell the difference in sound quality between the chip amps and my hi-powered SS amp. Once you get over 10K Hz it gets quite difficult to hear subtle differences. However, the quality of the UHF driver makes all the difference in the world! I feel the JBL 2405 is the best of the best. I have several other UHF drivers and none compare.
 
Passive filter have a few advantages worth mentioning. In some frequency ranges the driver distortion becomes less if the driving impedance is high. With a careful crossover design, the driving impedance can be made higher in those ranges resulting in a system with less distortion.

Also, it seems that passive filters are more commonly custom made for a given driver combination, active filters seem to use standard cookbook transfer functions, not considering the responses of the drivers at all.

There are of course exceptions both ways (cookbook passive filters, and digital crossovers with various compensations). However, the decreased distortion that results from higher driving impedance can rarely be used with active filters.
 
HipoFutura said:

Tri-amp allows you so select specific amps that work best within their respective frequencies. As far as I'm concerned this is as important as having separate speakers. I'm stunned by those who feel separating the amps isn't that important.

Once you get over 10K Hz it gets quite difficult to hear subtle

Ear sensitivity over 10khz and under 100hz is certainly less than in the vocal range. When the crossover points were under 100hz and above 10khz I found that matching amps timbre wasnt very critical. When crossing over around 2500 to 4kz, I found the best results when the amps were the same model. I suspect that if you can keep the range from 300hz through 4Khz on the same amp you will have the most success.

Now, here's the interesting thing I found. My 100-watt Class-A amp mates better with a cheap Panasonic receiver than it does an Aleph30. I speculate that the overly warm sounding Aleph is standing out from other two amps in ways that dont sound natural.

As for putting your speakers in seperate boxes....

That might make it easy to experiment. You dont have to rebuild your entire box every time you change drivers. But I suspect that you could create an oversized rigid box with 3 separate chambers. This could be similar to the way PartsExpress has boxes with removeable front baffles. It's got to be easier to route out holes for new drivers in a single front baffle than to build a complete box???

Also, there may be advantages to having the drivers vertically aligned and as close as possible together. Certainly this is true for line arrays. I dont think I'd bother putting 2 bass drivers in a separate box. I'd prefer to have them close together with some bracing close by.
 
The amp component and overall config

Actually, I did want to keep it out of the discussion, only because it's such a huge area that I didn't want to get too wide on the topic.

However, the point you make re matching amps and the freq ranges is certainly one I'm considering. Here is the configuration I was considering:

Best two-way design I can find that will work with home theater and music (I know I may have to make a cormpromise one way or the other) and use these for both fronts, both sides (5.1 rear, 7.1 side) and for the rears. In addition, I would use a hgih quality bi-amp center (I'm pretty much in love with my paradigm 100 center, so I think I'll keep her - in fact, these mounted vertically might make the basis for all of the 6 "two ways"). Then I'd keep my Polk 300W sub, possibly add another (it's a big, 25 by 60 room), both for .1 channel. This would comprise my basic "home theater mode" setup. I've got a BUNCH of different Carver amps I'd play around with for driving these.

Then I would have separate standalone tweeets (something like the LCY-100K, though they're pricey) and woofs (homebuilt -- basicallyl a high-range sub-woofer like you'd create for a bass guitar speaker) alongside the two fronts and two rears (tweets at ear level, of course) and switch them in for music listening and/or really "cranking up" the Dolby Digital 7.1 simulation (adds in the rears) that my Denon will do. For the woof amps, I was considering the Carver monobloks I have, and for the tweets, probably a low-watt chip-based (LM3875 or similar) amp. My Denon would primarily be resigned to pre-amp and signal processing status (which it does very well, particularly for hooking up the SACD analog channels to the multi-channel input cluster).

Anyways, that's the current thinking, what I'm really struggling with at this point is what programmable active xover to go with. I REALLY liked the DCX's programmability and flexibility (chaining, etc.), but am concerned re the quality from some responses I'm getting (then others who've had em for years love them). I wouldn't mind pay a bit more for quality, but I don't want to give up the flexibility (filter type selection, EQ flexibility, PC control, etc.). In the end I see this setup also as my "reference" point for future speaker designs/creations (I do some custom home installs/creations -- I don't advertise here though -- but if you ever see two 500W sub's hidden "inside" a river rock fireplace, that's me :). I was even thinking about using the DCX as a platform for designing passive xovers for implementations where active might not make sense (or just to save money).

Although I'm not a big "tube nut", your mention of using it for mid-range got me thinking about using one in this config for vinyl playing (which I consider the best combo for tube amps). Now this is getting REALLY interesting...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.