convert 2-way bookshelf into 3-way tower?

[Guiria]

As the title suggests I am interested in converting my two way bookshelves (see sig) into a three way tower and obviously completely rebuilding the enclosures. Come May 11th (school is out) I plan on cranking out a speaker project before I start summer school in July. Initial design/constraints I see are:

Matching sensitivity to the current design.
Finding the ideal crossover point between midrange and bass driver(s).
Ideally keeping the system nominally between 6-8 ohms.

http://www.drmarksays.com/wp-content/uploads/2008/09/polk-rt-3-mod_crossover.pdf

My current crossover is the above design.

That's a start at least. Hopefully I can find some drivers in the Dayton RS line that would work but I am not set on using those drivers. My biggest problem is that I don't have the software, design crossover skills, etc to run scenarios on this project so I'm banking on the good people here on the audioholics DIY to help out. Thanks in advance.

 

[WmAx]

As the title suggests I am interested in converting my two way bookshelves (see sig) into a three way tower and obviously completely rebuilding the enclosures. Come May 11th (school is out) I plan on cranking out a speaker project before I start summer school in July. Initial design/constraints I see are:

Matching sensitivity to the current design.
Finding the ideal crossover point between midrange and bass driver(s).
Ideally keeping the system nominally between 6-8 ohms.

http://www.drmarksays.com/wp-content/uploads/2008/09/polk-rt-3-mod_crossover.pdf

My current crossover is the above design.

That's a start at least. Hopefully I can find some drivers in the Dayton RS line that would work but I am not set on using those drivers. My biggest problem is that I don't have the software, design crossover skills, etc to run scenarios on this project so I'm banking on the good people here on the audioholics DIY to help out. Thanks in advance.

TLS(Dr. Mark) is going to say the same thing I'm about to: It is not good to cross at the low frequency range where those bookshelves will optimally integrate to the woofers(100-150Hz band). You should use an active crossover on bands this low, as the crossover parts to do it passively are expensive, very expensive, due to the huge values and filter complication to do it correctly.

The Behringer CX2310 is an inexpensive but very high quality active crossover suitable for the job.

-Chris

 

[TLS Guy]

As the title suggests I am interested in converting my two way bookshelves (see sig) into a three way tower and obviously completely rebuilding the enclosures. Come May 11th (school is out) I plan on cranking out a speaker project before I start summer school in July. Initial design/constraints I see are:

Matching sensitivity to the current design.
Finding the ideal crossover point between midrange and bass driver(s).
Ideally keeping the system nominally between 6-8 ohms.

http://www.drmarksays.com/wp-content/uploads/2008/09/polk-rt-3-mod_crossover.pdf

My current crossover is the above design.

That's a start at least. Hopefully I can find some drivers in the Dayton RS line that would work but I am not set on using those drivers. My biggest problem is that I don't have the software, design crossover skills, etc to run scenarios on this project so I'm banking on the good people here on the audioholics DIY to help out. Thanks in advance.

We need a budget.

You have a couple of options.

Use a sub type driver with a crossover around 120 to 150 Hz.

If you want to stay passive, I have a 10" Peerless driver in mind we could crossover at 350 to 400 Hz.

 

[Guiria]

We need a budget.

You have a couple of options.

Use a sub type driver with a crossover around 120 to 150 Hz.

If you want to stay passive, I have a 10" Peerless driver in mind we could crossover at 350 to 400 Hz.

Planned budget is around 200 dollars for drivers and crossover parts. Given the budget, I think an electronic crossover + amplification + driver parts is out of the question...if that's the only way to do it then this project will be thrown out for the time being.

I was thinking 8" drivers but really for no specific reason, that's just what came to mind. A 10" driver would make for a pretty wide cabinet...I'd still consider it though.

 

[Swerd]

I'm not sure what driver TLS is thinking of, but I've seen a Peerless SLS 830668 10" woofer work nicely when added to another small 2-way.

It was crossed at 450 Hz, 2nd order Linkwitz-Riley, to a GR Research M-130 used as a midrange. It's a Dennis Murphy DIY design, called the MBOW1 3-way.

The 10" woofer is in a separate 2 ft³ sealed cabinet that sits below a smaller cabinet housing the midrange and tweeter. The bass cabinet is 24.25" H, 12" W, 17.5" D (external) and made with ¾" MDF. The woofer is mounted close to the top so that its center is 12" below the midrange center.

 

[TLS Guy]

I'm not sure what driver TLS is thinking of, but I've seen a Peerless SLS 830668 10" woofer work nicely when added to another small 2-way.

It was crossed at 450 Hz, 2nd order Linkwitz-Riley, to a GR Research M-130 used as a midrange. It's a Dennis Murphy DIY design, called the MBOW1 3-way.

The 10" woofer is in a separate 2 ft³ sealed cabinet that sits below a smaller cabinet housing the midrange and tweeter. The bass cabinet is 24.25" H, 12" W, 17.5" D (external) and made with ¾" MDF. The woofer is mounted close to the top so that its center is 12" below the midrange center.

You have to one Swerd! I will try and take time to modify his crossover for his driver over the weekend if I get time.

For his drivers that unit is the most cost effective for his purpose.

 

[Guiria]

I'm not sure what driver TLS is thinking of, but I've seen a Peerless SLS 830668 10" woofer work nicely when added to another small 2-way.

It was crossed at 450 Hz, 2nd order Linkwitz-Riley, to a GR Research M-130 used as a midrange. It's a Dennis Murphy DIY design, called the MBOW1 3-way.

The 10" woofer is in a separate 2 ft³ sealed cabinet that sits below a smaller cabinet housing the midrange and tweeter. The bass cabinet is 24.25" H, 12" W, 17.5" D (external) and made with ¾" MDF. The woofer is mounted close to the top so that its center is 12" below the midrange center.

Something like that is exactly what I'm after. I will check out the MBOW_3-way you mentioned. I built the large Dayton RS WMTW center with the Dennis Murphy xover last year...he does good work.

 

[Guiria]

You have to one Swerd! I will try and take time to modify his crossover for his driver over the weekend if I get time.

For his drivers that unit is the most cost effective for his purpose.

TLS,

If you ever make it out to Boise, ID let me know...I'll take you out to dinner

Your help is very appreciated.

 

[Guiria]

...The woofer is mounted close to the top so that its center is 12" below the midrange center.

What is the significance of having the distance be 12" center to center? Does it correlate to the size of the wavelength at the xover frequency?

 

[Swerd]

What is the significance of having the distance be 12" center to center? Does it correlate to the size of the wavelength at the xover frequency?

Yes, a general rule of thumb is try to keep the center-to-center distance no greater than the crossover wavelength. That's easy for the woofer-to-mid distance, but for a mid-to-tweeter we usually settle for as close as possible .

 

[TLS Guy]

Here you go my friend.

This is your design.

Right away you will see the trouble with three ways, as some of the cap and one of the inductor values is high.

I will be happy to address any questions.

 

[highfigh]

Here you go my friend.

This is your design.

Right away you will see the trouble with three ways, as some of the cap and one of the inductor values is high.

I will be happy to address any questions.

You wrote "I suspect this is much more a loudspeaker correction program than room correction in a lot of cases."- how would a program determine which it was correcting, unless the speaker's anechoic response had been loaded in as a default curve? If flat & smooth is the desired response curve and it's measuring the in-room response, how could it correct for one without affecting the other? Speakers have enough issues and most rooms have even more, making good speakers sound much worse than they normally would. If it catches the problems and fixes most of them (realizing that it's not professional acoustical analysis and correction), is it such a bad thing? For those who don't know or care how to correct the problems, I have to say that the results are better than connecting a 12-band graphic equalizer and letting some friend who's an 'expert' do their magic, complete with 6-10dB of gain, the noise and distortion that go along with that and blown speakers because of the massive boost at the wrong frequencies.

When you looked for the corrections after running Audyssey, was it the Audyssey Flat setting? With my Denon AVR-990, Audyssey Flat is definitely too bright and there's a big lift in the HF response that, according to RTA, isn't needed (since RTA shows that the response is flatter than the correction).

 

[Guiria]

Here you go my friend.

This is your design.

Right away you will see the trouble with three ways, as some of the cap and one of the inductor values is high.

I will be happy to address any questions.

I will hopefully get to spend the rest of my peaceful Sunday afternoon working on a build design but we'll see. I already had in mind a cabinet with a wider base and narrower top. Now a couple of design questions.

What is the optimal driver spacing between midrange and bass woofer when crossed over at 350 Hz? Is it 12"? Is there a calculator or something out there where I can figure that out for any frequency?

I have not calculated how big a 4.XX cubic foot tower will be but if it is too large will the Peerless work well in a smaller sealed alignment?

Off to the drawing board...

 

[TLS Guy]

I will hopefully get to spend the rest of my peaceful Sunday afternoon working on a build design but we'll see. I already had in mind a cabinet with a wider base and narrower top. Now a couple of design questions.

What is the optimal driver spacing between midrange and bass woofer when crossed over at 350 Hz? Is it 12"? Is there a calculator or something out there where I can figure that out for any frequency?

I have not calculated how big a 4.XX cubic foot tower will be but if it is too large will the Peerless work well in a smaller sealed alignment?

Off to the drawing board...

I can do a sealed alignment later. Also I think you could do a 3 cu.ft, ported, but I suspect there would be some ripple.

I personally would locate the 10 Woofer, where it fits preferably at least with the center 10" above the floor if you can. Try and get it so it is in the narrowest part of the cabinet where it will fit. At 350 Hz driver spacing will not be highly critical. I would concentrate on diffraction, which is why I suggested the cabinet shape I did.
Best Regards, Mark.

 

[TLS Guy]

You wrote "I suspect this is much more a loudspeaker correction program than room correction in a lot of cases."- how would a program determine which it was correcting, unless the speaker's anechoic response had been loaded in as a default curve? If flat & smooth is the desired response curve and it's measuring the in-room response, how could it correct for one without affecting the other? Speakers have enough issues and most rooms have even more, making good speakers sound much worse than they normally would. If it catches the problems and fixes most of them (realizing that it's not professional acoustical analysis and correction), is it such a bad thing? For those who don't know or care how to correct the problems, I have to say that the results are better than connecting a 12-band graphic equalizer and letting some friend who's an 'expert' do their magic, complete with 6-10dB of gain, the noise and distortion that go along with that and blown speakers because of the massive boost at the wrong frequencies.

When you looked for the corrections after running Audyssey, was it the Audyssey Flat setting? With my Denon AVR-990, Audyssey Flat is definitely too bright and there's a big lift in the HF response that, according to RTA, isn't needed (since RTA shows that the response is flatter than the correction).

I agree with everything you have to say. I'm just surprised that Audyssey worked as well as it did. It did catch the response error I introduced on purpose.

As far as the HF correction was concerned, what Audyssey did was not correct. It boosted the HF to all speakers except the rear surrounds. I think it is caused by the distance from the mic. In three out of six testing positions the mic was fairly close to the rear backs.

My guess is that Audyssey id trying to correct the HF roll off with distance from the speaker and I don't think it should.

I have these speakers all pretty well balanced out prior to running Audyssey. The sound was pretty uniform throughout the room, but the alterations that Audyssey made to the time delays, which were significantly different from measured I felt did result in even more uniform response throughout the room.

How applicable the system is to a broad range of situations, I have no idea. But it was far from a dead loss in my evaluation, and I have to admit I was highly sceptical of it before letting it loose.

The thing that amazed me the most is that it did not change the speaker set up, except for the delays. I would have bet big money that it would have derailed my best laid plans.

 

[Guiria]

I can do a sealed alignment later. Also I think you could do a 3 cu.ft, ported, but I suspect there would be some ripple.

I personally would locate the 10 Woofer, where it fits preferably at least with the center 10" above the floor if you can. Try and get it so it is in the narrowest part of the cabinet where it will fit. At 350 Hz driver spacing will not be highly critical. I would concentrate on diffraction, which is why I suggested the cabinet shape I did.
Best Regards, Mark.

After doing some modeling the ported tower would be gigantic, relative to what I was originally after. If the sealed alignment models well I will most likely go for a sealed system. I will always be running these with a sub during HT and most of the time with music. Occasionally I may switch my receiver to direct mode but not too often. A couple of reasons behind my desire to convert from the 2-way to 3-way...

-The 2-way was my first real speaker build and I've learned a lot since then. There are some significant cosmetic errors that only rebuilding can take care of.
-I will most likely be moving in 2012 into a much smaller form of residence (house/apartment) and I want something floorstanding with protected xovers (currently xovers are outside of my speaker) to define "speaker space" and not worry about stands that my kids will bump anyways and knock off my expensive DIY bookshelves
-In my current HT I've found that having additional bass drivers in the mains location smooths out some serious peaks and valleys between 65-200 Hz. It also adds more impact to the HT experience.

I think a triangular prism design would be totally awesome and plan on using that for my plans...I just hope I can build it a little smaller than 4.4 cubic feet (internal dimensions). I modeled the Peerless 10" in WinISD pro and it gave me 2.77 cubic feet to get to a .704 Qtc...seems reasonable.

 

[TLS Guy]

After doing some modeling the ported tower would be gigantic, relative to what I was originally after. If the sealed alignment models well I will most likely go for a sealed system. I will always be running these with a sub during HT and most of the time with music. Occasionally I may switch my receiver to direct mode but not too often. A couple of reasons behind my desire to convert from the 2-way to 3-way...

-The 2-way was my first real speaker build and I've learned a lot since then. There are some significant cosmetic errors that only rebuilding can take care of.
-I will most likely be moving in 2012 into a much smaller form of residence (house/apartment) and I want something floorstanding with protected xovers (currently xovers are outside of my speaker) to define "speaker space" and not worry about stands that my kids will bump anyways and knock off my expensive DIY bookshelves
-In my current HT I've found that having additional bass drivers in the mains location smooths out some serious peaks and valleys between 65-200 Hz. It also adds more impact to the HT experience.

I think a triangular prism design would be totally awesome and plan on using that for my plans...I just hope I can build it a little smaller than 4.4 cubic feet (internal dimensions). I modeled the Peerless 10" in WinISD pro and it gave me 2.77 cubic feet to get to a .704 Qtc...seems reasonable.

Your 2.77 cu. ft. sealed enclosure seems very reasonable for your needs.

I make Qtc arounf 0.5, which would give a clean tight bass I prefer.

Name: SLS 830668
Type: Standard one-way driver
Company: Peerless (Denmark)
No. of Drivers = 1
Fs = 33.3 Hz
Qms = 4.85
Vas = 69.3 liters
Cms = 0.45 mm/N
Mms = 51.1 g
Rms = 2.19 kg/s
Xmax = 8 mm
Xmech = 12 mm
P-Dia = 207 mm
Sd = 335 sq.cm
P-Vd = 0.263 liters
Qes = 0.57
Re = 5.6 ohms
Le = 3.3 mH
Z = 8 ohms
BL = 10.2 Tm
Pe = 100 watts
Qts = 0.51
no = 0.433 %
1-W SPL = 88.51 dB
2.83-V SPL = 88.7 dB
-----------------------------------------
Box Properties
Name:
Type: Closed Box
Shape: Prism, square
Vb = 2.77 cu.ft
Qtc = 0.575
QL = 19.58
F3 = 53.23 Hz
Fill = heavy

To get a Qtc of 0.7 that a lot, if not most seem to prefer then the box volume is just over a cubic foot.

Driver Properties
Name: SLS 830668
Type: Standard one-way driver
Company: Peerless (Denmark)
No. of Drivers = 1
Fs = 33.3 Hz
Qms = 4.85
Vas = 69.3 liters
Cms = 0.45 mm/N
Mms = 51.1 g
Rms = 2.19 kg/s
Xmax = 8 mm
Xmech = 12 mm
P-Dia = 207 mm
Sd = 335 sq.cm
P-Vd = 0.263 liters
Qes = 0.57
Re = 5.6 ohms
Le = 3.3 mH
Z = 8 ohms
BL = 10.2 Tm
Pe = 100 watts
Qts = 0.51
no = 0.433 %
1-W SPL = 88.51 dB
2.83-V SPL = 88.7 dB
-----------------------------------------
Box Properties
Name:
Type: Closed Box
Shape: Prism, square
Vb = 1.108 cu.ft
Qtc = 0.707
QL = 19.58
F3 = 55.99 Hz
Fill = heavy

However the 2.77 cu. ft. box has severe problems with power handling in the bass, because of exceeding xmax.

Overall the smaller enclosure is your best bet. Xmax is not exceeded within the power limits of the driver at any frequency.

So I would build the smaller box. This will likely fit the circumstances you describe better.

 

[Guiria]

I am planning on a sealed box design with the base about 14 inches wide and 40 inches tall with the top of the speaker being about 6 inches wide.

When building a speaker box with reduced parallel sides is bracing still critical? I planned on doing some shelf bracing because I will have some pretty long lengths of wood without anything on them.

I also thought of making the back of the speaker sloping as well as the shelf on the inside that will separate my mid enclosure from the bass enclosure. I would end up with a box with zero parallel sides. Sounds cool in theory. Probably a pain to build

 

[TLS Guy]

I am planning on a sealed box design with the base about 14 inches wide and 40 inches tall with the top of the speaker being about 6 inches wide.

When building a speaker box with reduced parallel sides is bracing still critical? I planned on doing some shelf bracing because I will have some pretty long lengths of wood without anything on them.

I also thought of making the back of the speaker sloping as well as the shelf on the inside that will separate my mid enclosure from the bass enclosure. I would end up with a box with zero parallel sides. Sounds cool in theory. Probably a pain to build

You will still have to brace it. It will be a pain to build, but worth it.

 

[Guiria]

I also thought of making the back of the speaker sloping as well as the shelf on the inside that will separate my mid enclosure from the bass enclosure. I would end up with a box with zero parallel sides. Sounds cool in theory. Probably a pain to build

To clarify...the base, top and front baffle would be perpendicular. The sides would be wide at the base and slope up towards the top, like a triangle. The back would also be wide at the base and slope towards the top. So the base would be something like 14x12 and the top would be something like 6x6 (external dims)

When building such a box how in the world do you calculate volume on something like this...it's not a pyramid, front face is perpendicular to base. It's not a triangular prism because you have flat sections on top and an extra sloping side. When only two sides were slanted it was easy. Now with the addition of a third side sloping in I am lost on how to calculate volume. Any pointers? Do I need to find a friend that knows CAD to draw it up and have a program calculate volume?