How can we do this? How can we make a well integrated sub?
If the two way is a ported design with BSC, how can we make the subwoofer to click/integrate well with the midwoofer? Will the midwoofer be filtered electronically? Or only working out the subwoofer filter? Will you need an RTA to tell you how to set the subwoofer?
What if the 2-way is a transmission line?
I can make a 3-way, but I have never had a success adding a subwoofer to a working 2-way. And I don't like to redesign the already working 2-way into 3-way. So these are questions I'm looking the answer for...
Hi,
First, by watching the various behaviours of the Subwoofer, so that we can match group-delay and transient behaviour etc. with the mains. Means we need what is often called a "fast" subwoofer, meaning low group-delay, minimal transient overswing, an amplifier with crossovers that have the appropriate slope and so on.
Few if any commercial Subwoofers and few any so-called "subwoofer" drivers need apply.
Probably not at all.
BSC causes a 6dB rise in LF in the power-response (in room response) over a neutral response...
This lifted "mid-bass" can subjectively compensate for a lack of really low bass, but if we add a matched and integrated sub this fallacy becomes painfully obvious with a completely unbalanced, overcooked, overbearing bass response. If we turn the subwoofer down enough it will sound as if it is not there or at least does very little.
Probably what you call a "working 2-way" was not actually really "working" in the literal, technical sense of the word, even though it probably sounded fine, subjectively.
Try using the midwoofer in a maximally flat sealed enclosure without BSC, use a sealed woofer with a 15 - 18" Pro Driver equalised for flat LF and make sure to use a 2nd order LPF...
Ciao T
First, by watching the various behaviours of the Subwoofer, so that we can match group-delay and transient behaviour etc. with the mains. Means we need what is often called a "fast" subwoofer, meaning low group-delay, minimal transient overswing, an amplifier with crossovers that have the appropriate slope and so on.
Few if any commercial Subwoofers and few any so-called "subwoofer" drivers need apply.
Probably not at all.
BSC causes a 6dB rise in LF in the power-response (in room response) over a neutral response...
This lifted "mid-bass" can subjectively compensate for a lack of really low bass, but if we add a matched and integrated sub this fallacy becomes painfully obvious with a completely unbalanced, overcooked, overbearing bass response. If we turn the subwoofer down enough it will sound as if it is not there or at least does very little.
Probably what you call a "working 2-way" was not actually really "working" in the literal, technical sense of the word, even though it probably sounded fine, subjectively.
Try using the midwoofer in a maximally flat sealed enclosure without BSC, use a sealed woofer with a 15 - 18" Pro Driver equalised for flat LF and make sure to use a 2nd order LPF...
Ciao T
That's what I thought. It means $$$ 😛
It means that the main speaker will have "no" use without the subwoofer? 😕 Then it means a redesigned 3-way, with passive and active combination of filters. Well okay I will try this, but...
If the midwoofer is as fast as a Seas Excel, what is the minimum (cheapest cheapest cheapest) subwoofer that may apply?
Hi,
Well, quality is never cheap.
Depends. You could wean yourself off the over-blown mid-bass you gotten used to... Maybe replace the BSC with an EQ to flatten out the lower midrange loss from narrow baffles without excessively boosting the mid-bass?
Maybe just use wide baffles, maybe use speakers systems with better controlled directivity at all frequencies (dipoles, unipoles) than what you currently use?
There are many ways you can solve the various issues...
Well, try a decent 18" Pro Driver, fairly long throw, middeling efficiency and lowish Fs, if the Seas goes low enough to integrate with it.
Which model Excel is it?
Ciao T
Well, quality is never cheap.
Depends. You could wean yourself off the over-blown mid-bass you gotten used to... Maybe replace the BSC with an EQ to flatten out the lower midrange loss from narrow baffles without excessively boosting the mid-bass?
Maybe just use wide baffles, maybe use speakers systems with better controlled directivity at all frequencies (dipoles, unipoles) than what you currently use?
There are many ways you can solve the various issues...
Well, try a decent 18" Pro Driver, fairly long throw, middeling efficiency and lowish Fs, if the Seas goes low enough to integrate with it.
Which model Excel is it?
Ciao T
Some suggestions with respect to crossover points and subwoofer integration:
This is what I have in mind. I already prepared the walnut enclosure. 15" wide. But I haven't calculated what cut-off frequency is in effect by such baffle. Just an experiment.
W15CY. 18" is too big. Overseas purchase can be too expensive. Local drivers are cheap (and with no measurement data).
Now I'm starting to give high value on TL design....
This is exactly what I experience with music while listening to my smallish 2 way with a sub. I just cannot enjoy it with an actively crossed sub around 80hz. 60hz is ok if the sub is turned down. However, I have to adjust the sub almost every time I change CDs. Depending on how the music was mixed. It has made me nearly swear off subs.
Not sure I totally understand why what you say happens. But I know it happens. Thanks for confirming what I've been hearing.
Having said that. For movie watching, the sub is fine. Which makes sense because it's mostly a dedicated LFE channel, and even the LF content from the rest of the speakers usually doesn't require the same fidelity the way music does.
Thank you thank you thank you
Also wanted to throw in my 
Given my experiences above, I have been wanting to try a FAST actively crossed using a minidsp or possibly my receiver. I'd like to cross around 150 to 200hz. Not sure that is the correct XO frequency. But something I plan on. My current 2300hz XO doesn't bother me. Seemless to my ears. I think I'm actually most sensitive to upper bass lower midrange, but still my plan.
Given my experiences above, I have been wanting to try a FAST actively crossed using a minidsp or possibly my receiver. I'd like to cross around 150 to 200hz. Not sure that is the correct XO frequency. But something I plan on. My current 2300hz XO doesn't bother me. Seemless to my ears. I think I'm actually most sensitive to upper bass lower midrange, but still my plan.
Hi,
If your baffle is 15" wide and no more than around 1.5 to 2 times the width from the rear room wall you need not worry much.
With a W15 I would not remotely attempt to push it below around 200Hz, it is just too much of a toy driver. It is okay as midrange for 3-Way, but I would not attempt to even let it handle upper bass...
18" is not too big. You can sucessfully do a 2-Way with around 1KHz crossover using a good 18"...
But given your main driver I think 3Way is the only viable option. And at the minimum you will need 3rd order (acoustic) slopes on the Midrange.
You may find my current speakers at home slightly interesting.
They use a 15cm Metal Cone fullrange driver as midrange, circular ribbon tweeter and a 28cm Metal Cone bass.
The Bass is set to 28Hz in a MLTL tuned quite low, more for reducing LF excursion than for extra SPL.
The crossover is electrical first order series with NO BSC, crossover points at 300Hz and 3KHz (forced by the drivers). As a result the acoustic slopes on both midrange and tweeter highpass are 3rd order, while the woofer and midrange lowpass slopes are 1st order, so the acoustic crossover is asymmetric (works very well in practice, actually)...
There is no magic in TL's and they do not offer any actual improvement on suitably designed reflex box. And they cannot make a small driver big, or make it go low and loud, either solution...
Ciao T
If your baffle is 15" wide and no more than around 1.5 to 2 times the width from the rear room wall you need not worry much.
With a W15 I would not remotely attempt to push it below around 200Hz, it is just too much of a toy driver. It is okay as midrange for 3-Way, but I would not attempt to even let it handle upper bass...
18" is not too big. You can sucessfully do a 2-Way with around 1KHz crossover using a good 18"...
But given your main driver I think 3Way is the only viable option. And at the minimum you will need 3rd order (acoustic) slopes on the Midrange.
You may find my current speakers at home slightly interesting.
They use a 15cm Metal Cone fullrange driver as midrange, circular ribbon tweeter and a 28cm Metal Cone bass.
The Bass is set to 28Hz in a MLTL tuned quite low, more for reducing LF excursion than for extra SPL.
The crossover is electrical first order series with NO BSC, crossover points at 300Hz and 3KHz (forced by the drivers). As a result the acoustic slopes on both midrange and tweeter highpass are 3rd order, while the woofer and midrange lowpass slopes are 1st order, so the acoustic crossover is asymmetric (works very well in practice, actually)...
Now I'm starting to give high value on TL design....
There is no magic in TL's and they do not offer any actual improvement on suitably designed reflex box. And they cannot make a small driver big, or make it go low and loud, either solution...
Ciao T
Yes I understand about the bass reinforcement. And that a wide baffle (half space?) may give a "good" sound. What bothers me is whether the speaker will be able to "disappear". This is my priority in speaker design. If I cannot make it disappear, I'll forget it.
Yes, that MLTL. Long time ago I wondered, why my small thin vented speaker (designed by ears) can reproduce such a deep bass. Everybody thought it was the effect of port turbulence 😛
Cannot help it but I'm against series crossover design. I always thought that it was a bypass method for those who do not want to try to do the best.
Knowing that a ported system can employ a quarter wave technique, I think you are right. But for long pipes, they aren't reflex boxes.
And I cannot explain what I have achieved with a TL design to make it sound big and loud, like a real concert! Sure I can explain a bit, but better not as I don't have theory to back it up.
Jay,
I believe Thorsten was referring to an assymetrical crossover which can be done with either a parallel or a series crossover.
(Correction: Thorsten did mention he used a series crossover as well.)
I suspect that your aversion to series crossovers is because you've never been able to make one work.
It actually takes a bit of work to really get it dialed in correctly and can be a lot more work than most L/R 4th orders.
However, you can rest assured that I did not use a series crossover (it's also assymetrical, btw), because it was easier.
Best Regards,
TerryO
Last edited:
Is this in a domestic listening room, and if so, how is this relevant?
I used to think so too, then I found some people like them. Personally, I don't believe that the voice coil is an appropriate location to marry drivers.
Nope, it is because I can make it "work" with eyes closed 😉 What I mean is, it is so easy to make it work, but the quality is... let me make an analogy...
In specific area (can I call it dynamic?), SERIES crossover to PARALLEL crossover is like PASSIVE crossover to ACTIVE crossover.
I avoid using active crossover, not because I think it is worse (true that many implementation is bad), but I cannot make it as good as I can make a passive one, not yet.
Of course, many people will like the image it creates, though so unreal to my ears (I also don't like headphone imaging). Especially if they compare with parallel crossovers with bad imaging capability. Personally, I have built enough passive crossovers, and I trust what I heard.
Suggestion
@ audiothings
Hi, further to my response to you in - http://www.diyaudio.com/forums/multi-way/187778-20-hz-20-khz-2-way.html?highlight=audiothings
As most likely you will you using drive units made from different materials, where you xover will sound differently. Apart from personal choice, differences in peoples hearing etc, there are various things to consider.
As desirable as it is to xover lower, it may not be practicle. Apart from anything else, depending on, especially the efficiency/power rating of the mid/high driver, it "might" not be able to keep up with the efficiency/power of the woofer. So you would need to xover higher than you might have wished. This may or may not make much/any difference to the final outcome. Only by actually testing will you know.
I would suggest initially using an active xover etc to determine where the most desirable/practicle xover point is, then design a passive xover to suit. Or just go with the active set up solution. If you choose a suitable active xover, you can easily/quickly select the slopes etc you require 🙂
@ audiothings
Hi, further to my response to you in - http://www.diyaudio.com/forums/multi-way/187778-20-hz-20-khz-2-way.html?highlight=audiothings
As most likely you will you using drive units made from different materials, where you xover will sound differently. Apart from personal choice, differences in peoples hearing etc, there are various things to consider.
As desirable as it is to xover lower, it may not be practicle. Apart from anything else, depending on, especially the efficiency/power rating of the mid/high driver, it "might" not be able to keep up with the efficiency/power of the woofer. So you would need to xover higher than you might have wished. This may or may not make much/any difference to the final outcome. Only by actually testing will you know.
I would suggest initially using an active xover etc to determine where the most desirable/practicle xover point is, then design a passive xover to suit. Or just go with the active set up solution. If you choose a suitable active xover, you can easily/quickly select the slopes etc you require 🙂
Last edited:
Hi,
Baffle width as such has nothing to do with "disappearing" speakers. In fact, I find narrow baffles make the problems worse, not better.
This is completely irrational. Series crossovers are one of the possible applications. They have different sets of strength.
A parallel crossover midrange bandpass, first order, for a cone driver or dome midrange with their strong resonance and voice voil inductance will need complex impedance linearisation and even then the "lookback" impedance for the driver at frequencies above and below the passband will be much higher.
Using instead a series X-over means the choke and capacitor are in parallel with the driver and the lookback impedance for the driver is very low at all frequencies, not just in the passband.
I have used about everything under the sun, parallel crossover, series crossovers, mixed mode ones, active ones, hybrid ones, line level LCR and so on.
I would not put any methode specifically before any other, it depends on design goals and to a degeree preferences as to which one is best in a given application.
The specific one I had (3-Way, passive X-Over, cone midrange in sealed enclosure) suggested quite strongly that a 1st order series crossover with high quality parts was the right choice.
FWIW, a friend who has the whole Lyngdorf setup (dipole mains, sealed wallplaced woofers, with digital amp including room correction and active crossover between woofer and mains etc.) heard that speaker driven by a simple 25W Tube Amp (no looped feedback) and has been badgering me since to build him a set, which I have been refusing (let him buy an AMR system at friendship discount, I don't have time to build another set).
Oh yes, they are, at least all the comparably short lines. They are "unconventional" reflex boxes, that I grant, but they are... Actually, look at the bass driver impedance, it tells all you need to know...
I can explain how you can achieve that using a reflex box of more traditional dimension and design, because I understand how it works what some call "TL" (which aren't).
Well, that cool, kudos for you. I find getting any 3-Way crossover to work (or 2-way for that) is quite a bit of work, a truely minimalist 3-Way series crossover is harder than most, certainly next to that integrating any subwoofer with a 2-Way speaker is trivial by comparison.
I guess given that you are that good, you better answer the questions here...
Ciao T
Baffle width as such has nothing to do with "disappearing" speakers. In fact, I find narrow baffles make the problems worse, not better.
This is completely irrational. Series crossovers are one of the possible applications. They have different sets of strength.
A parallel crossover midrange bandpass, first order, for a cone driver or dome midrange with their strong resonance and voice voil inductance will need complex impedance linearisation and even then the "lookback" impedance for the driver at frequencies above and below the passband will be much higher.
Using instead a series X-over means the choke and capacitor are in parallel with the driver and the lookback impedance for the driver is very low at all frequencies, not just in the passband.
I have used about everything under the sun, parallel crossover, series crossovers, mixed mode ones, active ones, hybrid ones, line level LCR and so on.
I would not put any methode specifically before any other, it depends on design goals and to a degeree preferences as to which one is best in a given application.
The specific one I had (3-Way, passive X-Over, cone midrange in sealed enclosure) suggested quite strongly that a 1st order series crossover with high quality parts was the right choice.
FWIW, a friend who has the whole Lyngdorf setup (dipole mains, sealed wallplaced woofers, with digital amp including room correction and active crossover between woofer and mains etc.) heard that speaker driven by a simple 25W Tube Amp (no looped feedback) and has been badgering me since to build him a set, which I have been refusing (let him buy an AMR system at friendship discount, I don't have time to build another set).
Oh yes, they are, at least all the comparably short lines. They are "unconventional" reflex boxes, that I grant, but they are... Actually, look at the bass driver impedance, it tells all you need to know...
I can explain how you can achieve that using a reflex box of more traditional dimension and design, because I understand how it works what some call "TL" (which aren't).
Well, that cool, kudos for you. I find getting any 3-Way crossover to work (or 2-way for that) is quite a bit of work, a truely minimalist 3-Way series crossover is harder than most, certainly next to that integrating any subwoofer with a 2-Way speaker is trivial by comparison.
I guess given that you are that good, you better answer the questions here...
Ciao T
Hi,
I was refering to the fact that in my setup at least the acoustic slopes for the highpass on the midrange and tweeter are 3rd order acoustic, as they combine the first order electrical slope with a 2nd order acoustic slope.
Meanwhile the lowpasses on the Woofer and Midrange are pretty much textbook 1st order (at least up to where the drivers linear operating range reaches, two octaves+ or so).
So the acoustic crossover (which is after all the one that matters) is asymmetric. As a result my 1st order series crossover is not really textbook either, if it was I'd get a 3-humped response and not a flat one.
Actually, making this kind of crossover work is pretty darn hard, because I decided to limit myself strictly to two caps and two inductors with only an RC to equalise and level match the tweeter (a circular magnetostat of somewhat difficult to employ frequency response, but exceptionally good sounding if it works). It would have been much easier with the addition of extra impedance equaliser circuits and such.
But this way I was able to use tinfoil Cap's for the tweeter and midrange, copperfoil chokes all around and filmcaps for the bass.
Enclosure BTW is 20mm stone lined 18mm MDF, avalonish style... Midrange has "hard ceramic non-regular spheroid" rear enclosure. Well, okay. It is a ceramic flowerpot (one has to be inventitive in one-offs/diy).
The colourations from these are not quite as low as the commercial all Alu speakers I designed a while back, but they do not give up a lot either...
Ciao T
I was refering to the fact that in my setup at least the acoustic slopes for the highpass on the midrange and tweeter are 3rd order acoustic, as they combine the first order electrical slope with a 2nd order acoustic slope.
Meanwhile the lowpasses on the Woofer and Midrange are pretty much textbook 1st order (at least up to where the drivers linear operating range reaches, two octaves+ or so).
So the acoustic crossover (which is after all the one that matters) is asymmetric. As a result my 1st order series crossover is not really textbook either, if it was I'd get a 3-humped response and not a flat one.
Actually, making this kind of crossover work is pretty darn hard, because I decided to limit myself strictly to two caps and two inductors with only an RC to equalise and level match the tweeter (a circular magnetostat of somewhat difficult to employ frequency response, but exceptionally good sounding if it works). It would have been much easier with the addition of extra impedance equaliser circuits and such.
But this way I was able to use tinfoil Cap's for the tweeter and midrange, copperfoil chokes all around and filmcaps for the bass.
Enclosure BTW is 20mm stone lined 18mm MDF, avalonish style... Midrange has "hard ceramic non-regular spheroid" rear enclosure. Well, okay. It is a ceramic flowerpot (one has to be inventitive in one-offs/diy).
The colourations from these are not quite as low as the commercial all Alu speakers I designed a while back, but they do not give up a lot either...
Ciao T
Attachments
Last edited:
For this you should reduce diffraction. You should also work at reducing early reflections.
Thorsten mentions narrow baffles being a problem here and I tend to agree, for specific reasons. For one, diffraction effects are likely to be closer in time to the direct sound. Secondly, there may be (depending on the design) more off axis sound to contribute to the earlier reflections.
I found that crossover+driver is the most critical, second is the box, third is room (many will surprise why I put this on the last).
I just found that the baffle and box is something that I can not easily control. First, I don't know the underlying theory. Second, the effort is too high.
But I don't believe in wide baffle for the tweeter (such as the Orion), unless the waveguide is made perfect (but I don't know the underlying theory either to make waveguide that will 100% work)
Sure. I built more series than parallel in the past. I found that series has one perfect disadvantage. Now for the advantage, can't I achieve the same advantage with parallel? The answer has always been yes.
If you have found that in one of your implementation, a parallel (with the complexity of the compensation circuitry) cannot be used to solve a problem/objective better than a series, of course you know better the situation.
I agree. I just want the best speaker, especially when the drivers are expensive. The real test is always the final result. Look at available series design on the net, compare it with parallel design using the same driver, aren't ears great when you can build and compare multiple options?
I believe you. Those drivers are not easy to work with. First order series with high quality capacitor is acceptable to me, no problem at all, is preferable actually, as long as the driver is smooth enough not to show distortion/peak (my ears are sensitive to peaks and distortion)
As long as theory goes, no doubt, you are an expert, I'm learning from you.
This I also believe you. It is just beyond my knowledge and energy to reproduce the reflex box by trial and error 😀
Now I see/remember why other poster also think that getting series xo to work is more difficult:
1) I have assumed that drivers that have not smooth response need not apply for series design.
2) Effect of compensation circuitry is worst in series xo than in parallel xo (so going series with complex circuitry is never my solution)
Sure it is. I'm not trying to say that you cannot do complex things by choosing a series xo. Your response to this subwoofer (and other) issue showed your deep knowledge of the subject, and I thank you for that.
Yes, I understand. I mean, I have read what the experts say. I just don't get my hands dirty with the subject/topics (because it is too expensive for me). You know what I mean 😀
- Status
- Not open for further replies.