[chadh]

Following a couple of threads in the previous incarnation of this circle, I've started dreaming about adding a pair of subs to supplement my single-driver speakers.  I'm not sure about the best plan to adopt though, so I was hoping for some advice.  My over-arching goal is not necessarily to attain the ultimate bass response or even to plumb the depths of the lowest octaves, but instead to secure a more satisfying bass experience and relieve the wide-band drivers from working as hard on the lower octaves.

I envisioned using a stereo pair of small, sealed subs (maybe 8", maybe 10", maybe 12" drivers), crossed over at around 100 Hz. 

I received the strong recommendation to ensure that the low pass filter slope was very steep (24dB/octave).  I also decided I wanted to crossover at the line level rather than at the speaker level.  As I understood that a passive line-level solution wouldn't really let me achieve the required low-pass filter slope, I initially planned on adopting an active low-pass filter.  The high pass filter wouldn't need such a steep slope, so I imagined that a passive filter arrangement would be satisfactory there (something like a 6 dB/octave slope).  The problem with this approach is the expense:  the active crossover, plus a pair of passive subs plus a stereo amp starts looking expensive.

The alternative would be to use a passive high pass filter to the main speakers, and then feed a line level signal to a pair of active subwoofers with low pass filters of 24 dB/octave built-in.

My first concern is over the disadvantages I face if I adopt the latter plan.  Are the crossovers in these active subs at the speaker level?  And if so, does this matter very much (given that the subs are really only employed for the range < 100Hz)?  Is the possibility of getting some EQ control in the active sub a substantial benefit that might outweigh other shortcomings of using the built-in crossover?  Or am I really better off with a well designed and well built active crossover?

My second concern is about using a passive high-pass filter.  One of the goals of adding bass augmentation is to relieve the wide-band drivers from playing the very low octaves.  In that case, does it make sense to use a steeper slope in the high-pass filter than just 6 dB/octave?  Might it be better to adopt an active high pass filter as well?

Thanks for any and all insights.

Chad

[Scott F.]

Chad,

The low pass crossover slope for your wide range driver depends on the frequency response of that driver at a given crossover point. As an example, if your driver mounted in your baffle doesn't have a strong output at say 150Hz, a shallower slope will work just fine.

To further define, my Lowther PM2As sound pretty good crossed over a few different ways; I can use a 6db slope and cross the 15" Altecs (vented cabinet) over 125Hz and the Lowthers at 250Hz. The gentle slope and pushing apart the XO point gives a reasonably flat FR in that range. This XO is the Pioneer SF-850.

On the other hand I've got a different crossover that when I set the XO frequency at 150Hz on both drivers and use a 12db slope and it sounds great too (but slightly different). On the Pioneer, I can also set the XO for 125Hz on both drivers and use an 18db slope and it works well too, again it sounds slightly different.

I like the sound of all of them though I haven't measured them. My ear tells me that the 6db slope probably has a slight dip at the XO point but it is quite slight. So far I like my current XO which crosses at 150 at 12db. Unfortunately its not available as its a prototype.

I would suggest looking for a Dalquist DQ10 active XO. It is fairly inexpensive and can be updated (caps) for not much money. There are a couple of guys here locally that are using that XO. John Hillig at Musical Design has modded the units. I'm pretty sure the DQXO uses a 12db slope. It is line level. Avoid passive XOs if at all possible as they suck the life out of music (you'll soon learn this once you hear a decent line level XO).

If you were to run your wide range drivers full range and try to roll a sub under them, then I would suggest a steep XO slope. If you are going line level or even passive on your WR drivers, then I would say a gentler slope will work just fine based on all the active XOs I own or have owned and played over the years (about eight or ten to date).

[chadh]

Scott,

Thanks for your thoughts.

The idea of using the steep slope on the low pass filter was to ensure that the subs wouldn't be producing anything within the vocal range at all (a danger, given that they'd be crossed over as high as around 100Hz).

My preference is to avoid speaker level crossovers altogether.  But using passive filters at the line level seems to have some advantages too.  That's what I'm considering using for a high pass filter.  I'm still not sure whether the standard active subwoofer utilizes a speaker level or line level crossover, though.  I suspect that the filter built into these is at the speaker level, though.  This sounds less than ideal.  But active subs also give you some EQ ability.  So maybe there are some advantages of using this kind of approach.

I'll look into the Dalquist unit.  My original thoughts about employing an active crossover involved having Roger Modjeski build a tube-based active crossover into a pre-amp for me.  He seemed pretty enthusiastic about the idea too.

Chad

[konut]

My experience has been that any active XOs impose a sonic signature which I find unappealing. This is especially true of the high pass, as the signal occupies the most sensitive area of hearing. One can get away with an active on the low pass as only a small percentage of the sound spectrum is being affected, and that part of the frequency spectrum is much less perceptibly critical. Currently I use a pair of passive Marchand passive high pass at 70hz, 24db per octave with my Omega A8s. Caution must be used with passives, as Scott F has found, because they will 'suck the life out' if they have too much of an insertion loss. This is especially true in my situation as I use a passive preamp as well. The Marchands have only 1db insertipon loss. I get around this by using very high gain, at 36db, amps. My low pass is a Onix RDES digital EQ that allows me 5 bands of parametric EQ. 1 band is used as a low pass which allows me to 'feather' the XO to the Marchand. IOW, I can tailor the slope as needed for the most pleasing effect. This allows me to bypass the low pass in my Rocket UFW-12 powered sub.

[chadh]

I just spent some time investigating the active crossover that was used with the Dahlquist DQ-10.  The crossover was labeled the DQ-LP1.  It provides an active low pass filter (with continuously variable cross-over point between 40 and 400 Hz), as well as a passive line-level high pass filter.  This is essentially the same kind of set-up I'd discussed with Roger Modjeski.  Except the crossover Roger would have built would have used tubes in the active circuit, would have a fixed crossover point, and would have had a pre-amp built into the same enclosure.

I couldn't find any DQ-LP1s for sale anywhere.  Does anyone have any idea how much they go for when they show up on ebay?

As far as the line level passive high-pass filter is concerned, if one limits the slope of this filter to 6 dB/octave, I understand it requires only two passive components in the signal path.  The parts count increases if you wish to increase the slope.  So if I  limited myself to a 6dB slope on the high pass filter, I should minimize any insertion loss.

Chad

[konut]

You are going to have a hard time finding something readily available with a lower parts count and lower insertion loss, much less low distortion and high S/N ratio, than this
http://www.marchandelec.com/xm46.html
as well as having the advantage of the steep slope. Keep in mind that the lower the frequency of high pass the lower the chance of phase anomalies intruding, on the ears' most sensitive area of hearing, with any type of filter. It all depends on the response of the main driver. The generally accepted rule is to double the -3db point of the main driver. This also assumes the response of the bass augmentation is capable up to the XO point.

[chadh]

You are going to have a hard time finding something readily available with a lower parts count and lower insertion loss, much less low distortion and high S/N ratio, than this
http://www.marchandelec.com/xm46.html
as well as having the advantage of the steep slope. Keep in mind that the lower the frequency of high pass the lower the chance of phase anomalies intruding, on the ears' most sensitive area of hearing, with any type of filter. It all depends on the response of the main driver. The generally accepted rule is to double the -3db point of the main driver. This also assumes the response of the bass augmentation is capable up to the XO point.

That's awesome.  Thanks.

I have two important questions about these passive Marchand units which may be trivial.  But remember I'm a bit of an idiot with these things, so these may be trivial.  What's more, they may actually be the same question just asked twice...

First, when it comes to a high pass filter for my application, is there really a big advantage to having a steep slope?  I presume that a steep slope more effectively would relieve the main drivers from the difficulties of producing the low octaves.  But that would come at the cost of a more complex circuit which is likely more detrimental to the sound.  Is it clear that the steep slope is a better option?

The second question relates to the low pass filter.  When I discussed the issue with Roger, he seemed to feel that a 24dB slope on the low pass filter could only be accomplished adequately with an active circuit.  Would this simply be a result of the parts count and subsequent attenuation of the signal with a passive circuit of this complexity?  Do you feel that the 24dB slope on the Marchand passive units is accomplished with little detriment to the sound?

Chad 

[Duke]

Chadh, you sent me a PM some time ago and I never responded.  I've tried to several times but for some reason I can only rarely send PMs; almost all the time I get an error message.

Anyway, if your question is still applicable, shoot me an e-mail at audiokinesis at yahoo dot com.

Duke

[DaveC113]

Is it clear that the steep slope is a better option?

Chad

IMO... no for the hi-pass, yes for the low pass.

I'd start simple and build a 6 dB 1st order into a set of interconnects using high quality parts before trying more complex and expensive alternatives that are likely to sound worse. I will try this myself one of these days, but I'm pretty happy with running my 4.5" Omegas full range and bringing in the sub around 50 Hz.

Also, I'd look for a dedicated sub amp(s) w/ a built-in 24 dB xo. Or, Omega and Horn Shoppe build powered subs for use w/ single drivers that should work well if your looking for mid 20 Hz range extension. 

[konut]

There is an important reason why a 24db per octave slope is used, and that is because it maintains the same phase as the incoming signal. Only 6db and 24db per octave do this. 12 and 18db per octave filters do not.  The reasons why 24db is preferable to 6db are:
1. Less chance of an overlapping hump.
2. Less chance of out of band resonances being excited.
3. Better amp control of the driver.
4. More power available to the driver in its primary area of reproduction.
5. Less chance of harmonic distortion by the amp.

Of course every choice has its pluses and minuses. In my experience its better to have the passive in the high pass, rather than an active, because, as you rightly point out, the fewer parts count and absence of power supply impose less noise, even though one pays a price of insertion loss. As long as your amp, or preamp, has enough gain its not that much of a penalty. The reason why I find the active more acceptable in the low pass is that it allows me to have parametric control over that part of the frequency spectrum where it needs it most, and where the ear is least sensitive to noise. 

The Marchand passive is the most transparent that I have heard. Admittedly, I haven't heard them all. I find its simplicity and low parts count appealing.

[chadh]

The reason why I find the active more acceptable in the low pass is that it allows me to have parametric control over that part of the frequency spectrum where it needs it most, and where the ear is least sensitive to noise. 

The Marchand passive is the most transparent that I have heard. Admittedly, I haven't heard them all. I find its simplicity and low parts count appealing.

So, if I was prepared to live without parametric EQ, do you think the 24dB passive Marchand filter would remain a good option for a low pass filter?

Chad

[konut]

As a filter, no question. The result might not be totally satisfactory though depending on many variables not limited to, but including, the design, and alignment of the sub, the size of the sub and sub amp, the acoustics and modes of your room, and the placement of the sub in the room. There are so many variables when it comes to tuning in a sub, or subs,  that it can become a truly frustrating experience even if you've got the right gear. It helps if you've got a sound pressure level meter, a sub sweep CD, or tone generator software, plenty of patience, and a good friend to help you out.   

[Telstar]

My experience has been that any active XOs impose a sonic signature which I find unappealing.

I agree.
And i also have had a hard time to find a passive XO that did not put a blanket on the music.

What's left? Line-leve PLLXO and digital XO (before D/A conversion). If you have an analog source, forget the latter.

[chadh]

My experience has been that any active XOs impose a sonic signature which I find unappealing.

I agree.
And i also have had a hard time to find a passive XO that did not put a blanket on the music.

What's left? Line-leve PLLXO and digital XO (before D/A conversion). If you have an analog source, forget the latter.

The passive crossovers we've been mentioning (like the Marchand, or something simply built into a pair of interconnects) are line level crossovers (PLLXOs), right?

Chad

[Telstar]

The passive crossovers we've been mentioning (like the Marchand, or something simply built into a pair of interconnects) are line level crossovers (PLLXOs), right?

Chad

Yes, if they are put before the amps

[chadh]

Chadh, you sent me a PM some time ago and I never responded.  I've tried to several times but for some reason I can only rarely send PMs; almost all the time I get an error message.

Anyway, if your question is still applicable, shoot me an e-mail at audiokinesis at yahoo dot com.

Duke

Duke,

Thanks for your message.  I just sent you an email.

Cheers.

Chad