Center speaker, The "Birdcage" (former CO design, phase question)

[anders]

Hi all

This is my first post in here, I have been around for a while reading tons of good stuff in here and I am pretty shure that many of you guys (or girls) have a nice answer for my question.

Recently I started to build a speaker similar to the Spassvogel project, identical drivers but my speaker is (hopefully) going to replace my center speaker in a 5.1 system.

I have just started to learn Speaker Workshop and today I did some simulations just for learning purpose. I came up with a crossover design that looks quit OK (at least for me) but I have some doubts.... (I have not measured anything yet, just using curve tracking files from the manufacturers datasheet for now)

I have read a lot but I cannot really figure out what the goal is if we are looking at the phase behavior of the crossover. I understand what phase "is" and why it can cause problems if two drivers are playing 180 degrees apart (at same frequency) but what "goal" should I have?

In "my world" it would be OK if two drivers are playing with almost the same phase at the crossover point so my goal tonight was to keep the phase of the two drivers as close as possible while trying to get the frequency respons as flat as possible. Is this a proper approach or am I doing something very wrong here?

A maybe related question is that I get 4 ohm impedance at 180 degree phase, will this kill my amplifier?

Ill post both frequency/phase diagram and crossover design...

Hopefully someone here have the time to explain this very interesting but maybe also complicated question.

Thanks!

/Anders

[rc white]

You are on the right track Anders.
In order to get the ideal L-R characteristic the high and low pass sections of the combined electrical and acoustic characteristic must sum to all pass, this is usually put as..

(Lp + Hp)/d = 1

The ideal occurs if the drivers are always in phase and the frequency response always sums to unity.
The result you simulated is close enough in practice, now all you have to do is replicate it in a real device.

[anders]

Thanks for answering!

...In order to get the ideal L-R characteristic the high and low pass sections of the combined electrical and acoustic characteristic must sum to all pass, this is usually put as..

(Lp + Hp)/d = 1
...

I dont really follow you. Whats the "d"?, Are the Highpass and Lowpass expressed in dB in the formula?

... now all you have to do is replicate it in a real device.

Yea, I belive this will be a completely different story compared to simulations. 8O

/Anders

[servicetech]

The distance between drivers and room acoustics play a factor. You may try wiring one tweeter reversed and one in phase, test both a/b and pick the one that sounds the best.

While the science behind speaker design has improved dramatically in the past few years there is still an "art" element to design.

[rc white]

A first order filter has low pass and high pass functions of

1/(s + 1) and s/(s + 1) if you add them you get

(1 + s)/(1 + s) = 1 for all frequencies, this is known as an all pass characteristic.

This still has lobe steering because there is a 90degree phase shift between filers, we can fix this by making it second order..

(1 + s)^2/(1 + s)^2 = 1

Now one is 180 degrees phase shifted and can be made in phase by reverse connection.
This has the desirable L-R characteristics of flat amplitude summing and zero lobe steering and describes the second order L-R crossover filter.

In order to have this characteristic both low and high pass sections need to have the same denominator and sum to zero at all frequencies, the "d" symbol
is the desired common denominator and Hp, Lp are the polynomials of the filter.

The difficulty arises because the Hp and Lp characteristics are only partially the filter itself and the driver characteristics can make it very hard to impossible to achieve the ideal with ought a large degree of complication.
This means that most practical systems are in fact pseudo L-R and the debate is generally about how close to the ideal is close enough, and there are many firmly held opinions on this matter.

[anders]

A first order filter has low pass and high pass functions of

1/(s + 1) and s/(s + 1) if you add them you get

(1 + s)/(1 + s) = 1 for all frequencies, this is known as an all pass characteristic.

This still has lobe steering because there is a 90degree phase shift between filers, we can fix this by making it second order..

(1 + s)^2/(1 + s)^2 = 1

Now one is 180 degrees phase shifted and can be made in phase by reverse connection.
This has the desirable L-R characteristics of flat amplitude summing and zero lobe steering and describes the second order L-R crossover filter.

In order to have this characteristic both low and high pass sections need to have the same denominator and sum to zero at all frequencies, the "d" symbol
is the desired common denominator and Hp, Lp are the polynomials of the filter.

The difficulty arises because the Hp and Lp characteristics are only partially the filter itself and the driver characteristics can make it very hard to impossible to achieve the ideal with ought a large degree of complication.
This means that most practical systems are in fact pseudo L-R and the debate is generally about how close to the ideal is close enough, and there are many firmly held opinions on this matter.

This is probably a great explanation but I think I have to dig a little bit deeper into the theori part If I am going fully understand it.
For now I am quit satisified with you saying that I am on the right track

Actually I checked my old books from the university but I couldnt find any good examples of passive filtering and Laplace transforming (which I believe this is). I have a lot to learn and I will keep digging...

Thanks!

/Anders

[anders]

...Now one is 180 degrees phase shifted and can be made in phase by reverse connection.
This has the desirable L-R characteristics of flat amplitude summing and zero lobe steering and describes the second order L-R crossover filter....

My interpretation of this is that the "loobing problem" is minimized (or not existing) if the two drivers plays with identical phase, is this correct? (at frequencys around the crossover point).

I guess loobing could cause a lot of problems since I am planning to build a center speaker with only ONE bass/midrange driver (the speaker have almost horizontal aligned drivers). Is this why almost all commercially designed center speakers have two bass/midrange drivers (to minimize loobing effects)?

/Anders

[Curt C]

I have read a lot but I cannot really figure out what the goal is if we are looking at the phase behavior of the crossover. I understand what phase "is" and why it can cause problems if two drivers are playing 180 degrees apart (at same frequency) but what "goal" should I have?

In "my world" it would be OK if two drivers are playing with almost the same phase at the crossover point so my goal tonight was to keep the phase of the two drivers as close as possible while trying to get the frequency respons as flat as possible. Is this a proper approach or am I doing something very wrong here?

A maybe related question is that I get 4 ohm impedance at 180 degree phase, will this kill my amplifier?

Anders, This is a fairly involved subject. I’ll try to hit the high points.

In a perfect world (or crossover) with equal order slopes, crossovers with even order acoustic slopes will be –6 dB down at the crossover frequency and will (mostly) sum flat. The two driver phase responses will ‘track’ each other (one superimposed on the other) for a significant portion of the frequency band either side of the crossover frequency. Odd order acoustic slopes will be –3dB down at the crossover frequency and sum reasonably flat. The differential phase in this case will be 90 degrees apart for the ideal situation. Unfortunately, the real world is rarely ideal.

One issue that wreaks havoc with our perfect world is the difference in acoustic centers of the drivers. While the exact position of the acoustic center of a given driver is very difficult to measure, it is very easy to measure the difference in acoustic centers between two drivers. Generally the tweeter AC will be quite close to the driver flange, while a woofer will be several cm deeper into the frame. On a flat baffle measured perpendicular to the baffle, this difference in AC means the time of flight from the driver AC of the woofer is longer than the tweeter, so the woofer signal arrives later. This delay causes additional phase rotation of the woofer, and the our perfect 2nd order LR crossover will no longer sum flat or show good phase tracking.

One solution often seen for this is to slant the baffle back so the acoustic centers of the drivers align on the measurement axis. Another is to adjust the slope of the crossover transfer functions to delay the tweeter. This will result in a mixed order slope, usually with the woofer slope being a lower order than the tweeter. Mixed order slopes can sum flat and exhibit good phase tracking, and is generally the prevalent method to compensate for AC differences.

On your crossover, you’ve simulated a 4th order acoustic transfer function on the woofer, and a 3rd order function on the tweeter. This may be due to the fact the two drivers were not measured at the exact same distance, or that differing amounts of excess phase (time of flight of the acoustic waveform from the driver to the measurement mic) was removed. Outside of that caveat, it was well done. Congrats! :T

For more accurate sims, the best thing to do here is to perform a Hilbert transform on the response of each driver response, then add your best guess as to the difference in acoustic centers of the drivers. The Hilbert transform derives the minimum phase response of the driver, as if you measured exactly at its acoustic center. Good approximations of the acoustic center of a driver is the front plate of the magnet assy. Subtract the tweeter depth from the woofer depth and enter this difference in mm into SW in the woofer ‘Z’ position.

I'll recommend Jeff Bagby's Response Modeler program to modify the manufacturers anechoic driver files to add diffraction effects (Another major topic) and derive minimum phase. You might also find his Passive Crossover Designer more user friendly than SW.

Impedance minima of 4 ohms at the crossover frequency should not be an issue with most well designed amplifiers. -Even those spec’d for 8 ohm impedances only.

I hope this rambling dissertation was of some assistance.

C

[Curt C]

My interpretation of this is that the "loobing problem" is minimized (or not existing) if the two drivers plays with identical phase, is this correct? (at frequencys around the crossover point).

I guess loobing could cause a lot of problems since I am planning to build a center speaker with only ONE bass/midrange driver (the speaker have almost horizontal aligned drivers). Is this why almost all commercially designed center speakers have two bass/midrange drivers (to minimize loobing effects)?

/Anders

The orientation of the main lobe is dependent on both driver orientation and spacing, and the crossover topology utilized.

For example: Assuming equal acoustic centers, a 1st order BW 2 way crossover will have the main lobe tilted down 15 degrees in the vertical axis. The differences in acoustic centers of a typical tweeter/woofer will move it even further down from perpendicular to the baffle. Reversing the tweeter polarity will position the lobe 15 degrees above the design axis. Since the difference in acoustic centers with a normally orientated baffle (tweeter on top) now subtracts from the 15 degrees, and may be a better choice in this situation.

Even order crossovers: will have the main lobe aligned along the Zero Delay Plane. -This is the plane where the acoustic centers of the drivers are aligned. Odd order crossover will be alternately tipped up or down 15 degrees from the ZDP.

Mixed order crossovers: Well, it’s a mixed bag. Its possible to align the main lobe perpendicular to the baffle in spite of the ZDP being not being on the design axis, but careful planning and design is required to accomplish this.

In general: The higher the order, the narrower the main lobe, and the higher the number of side lobes.

Yes, an MTM (or an Wt/mW) forces the main lobe to be centered on the design axis. A center with only one woofer (A Wt/mW with one W missing) will suffer non symmetrical horizontal off axis issues, but its audibility can be minimized by making the woofer/mid crossover frequency as low as the mid will tolerate.

Keep in mind that all these effects only occur at those frequencies a couple of octaves either side of the crossover frequency Higher order crossovers will have a narrower frequency band, while a 1st order effects may cover 3 octaves either side of the crossover frequency.

C

[anders]

...

I hope this rambling dissertation was of some assistance.

C

Definitely!

All the bits and pieces are slowly coming together...

There is sooo much information out there and the hard part is to separate pure fact from "not so pure" fact and with this in mind I think this site is no1, many people with loads of experiance, thats great!.

...
...I'll recommend Jeff Bagby's Response Modeler program to modify the manufacturers anechoic driver files to add diffraction effects (Another major topic) and derive minimum phase. You might also find his Passive Crossover Designer more user friendly than SW.
C

I think I have this software on my PC at home. If I remember correct it was a couple of excel files that didnt work properly on my computer. Anyway, Ill give it another try later on tonight.

Actually, I am almost ready to start doing proper measurments (as soon as I get my microphone back from calibration), wouldnt this be the best thing to start with?

Hilbert transform, do I find this in the Jeff Bagby's Response Modeler program ?

/Anders

[Dennis H]

I think I have this software on my PC at home. If I remember correct it was a couple of excel files that didnt work properly on my computer.

You need the Analysis Toolpack and VBA for Excel installed. They aren't by default so you may need to get out the Excel disk.

[anders]

You need the Analysis Toolpack and VBA for Excel installed. They aren't by default so you may need to get out the Excel disk.

Thank you, now its working

...A center with only one woofer (A Wt/mW with one W missing) will suffer non symmetrical horizontal off axis issues, but its audibility can be minimized by making the woofer/mid crossover frequency as low as the mid will tolerate.

Keep in mind that all these effects only occur at those frequencies a couple of octaves either side of the crossover frequency Higher order crossovers will have a narrower frequency band, while a 1st order effects may cover 3 octaves either side of the crossover frequency.

OK, in other words......I'm in trouble

How low is "as low as the mid will tolerate"?
I am using the SS18W-8531 combined with SS6600 and planned crossover point is around 2000Hz or slightly lower.

I have allready built the box but I am ready to push the "emergency stop button" if you convince me that I am doing the wrong thing. First priority is to build a good speaker and it doesnt really matter if I have to redesign the box and buy another driver. A couple of MDF boards and 35 kg of concrete is wasted but I can live with that.

OK, square one....
Ill give you some background since this "project" is bigger than what it seems.

My plan is (in step 1) to replace a center speaker that I dont like. This center speaker is nowhere near the quality of the other speakers in my system, simply due to that it didnt exist when I bought my front speakers. Step 2 is to actually replace my front speakers with (very) similar design as the new center. Last step will be the surround speakers, also with the same design. It is important that all the speakers match each other as this is one of my goals, to have identical sound from every speaker.

So, why dont I just go out and buy home cinema kit for 990 dollars and live happily ever after?

Well, thing is......my front speakers today is more or less high end and I havn't found anything (commersial) on this side of 10000 dollar that is substantially better. This is one of the reasons, I believe that in almost all comersially designed speakers there is a huge pile of economic and cost reduction involved, almost regardless of the pricetag.

Another reason is that we are rebuilding our house and I have 30 square meters devoted to a home cinema. I would like something new and fresh in there and it would be very nice if I had built them myself 8)

So (I might have to change the subject of this thread ) there it is.....

In the light of all this I pretty much realize that I should actually press that button, start over and plan for another design.

/Anders

[Dennis H]

Consider what you've done so far good practice.

Actually, I am almost ready to start doing proper measurments (as soon as I get my microphone back from calibration), wouldnt this be the best thing to start with?

Yup that's the first step. I think you'll be surprised how much different your real measurements in a real box are. The factory Scanspeak curves are done in an anechoic chamber with the speaker mounted into a wall (infinite baffle) so at a minimum you will see a several dB drop in the midbass (baffle step) compared to what they measured.

As far as the design goes, 5 (nearly) identical speakers are ideal, especially if you have room to leave the center MT upright. Then you won't have any lobing problems with the center as long as you go for a nominal 2nd or 4th order crossover and get a good null when you reverse the tweeter. 3rd order is a little more complicated as you want them 90 degrees out of phase and you can't check it by reversing the wires on the tweeter and looking for a null.

[Curt C]

Thank you, now its working

OK, in other words......I'm in trouble

Well maybe not…

Sorry I misunderstood your post. A 2 way will be fine for a center as long as the tweeter is above the midwoofer, not beside it. Assuming you have the real estate in your cinema to accommodate the vertical orientation, you are golden. If, on the other hand, you were going to lay it on it’s side, it will not provide the off-axis performance you desire.

I agree with you that a well turned out DIY design can easily outperform commercial designs, both sonically and economically. Commercial designs aren’t trying to be moneygrubbers; they just have a lot more mouths to feed than a DIY guy.

Since you are ultimately thinking of replacing your system, you may want to look at the Statements thread here:

https://www.htguide.com/forum/showthread.php?t=28728

And the MiniStatments/Center thread:

https://www.htguide.com/forum/showthread.php?t=28728

I suggest this, as while it doesn’t utilize your drivers (Good choices, by the way.) it may give you some ideas, as there is a whole family of designs around the same set of drivers, including a forthcoming Statement monitor, a smaller 3 way. The center channel topology with the vertical tweeter and mid flanked by a pair of woofers is probably the best topology for a center speaker, although it does take up more vertical space than most would prefer.

C

[anders]

Consider what you've done so far good practice.



Yup that's the first step. I think you'll be surprised how much different your real measurements in a real box are. The factory Scanspeak curves are done in an anechoic chamber with the speaker mounted into a wall (infinite baffle) so at a minimum you will see a several dB drop in the midbass (baffle step) compared to what they measured.

As far as the design goes, 5 (nearly) identical speakers are ideal, especially if you have room to leave the center MT upright. Then you won't have any lobing problems with the center as long as you go for a nominal 2nd or 4th order crossover and get a good null when you reverse the tweeter. 3rd order is a little more complicated as you want them 90 degrees out of phase and you can't check it by reversing the wires on the tweeter and looking for a null.

Yep, it was practice...agreed

Yes identical speakers is my kind of tune, I have been listening to quite cheap systems that works surprisingly well when watching movies. I can never get that solid sound with my system even though each speaker on its own is far better compared to a "home cinema kit" speaker. This is actually quit annoying, I think you know what I mean...

Well maybe not…

Sorry I misunderstood your post. A 2 way will be fine for a center as long as the tweeter is above the midwoofer, not beside it. Assuming you have the real estate in your cinema to accommodate the vertical orientation, you are golden. If, on the other hand, you were going to lay it on it’s side, it will not provide the off-axis performance you desire.

I agree with you that a well turned out DIY design can easily outperform commercial designs, both sonically and economically. Commercial designs aren’t trying to be moneygrubbers; they just have a lot more mouths to feed than a DIY guy.

Since you are ultimately thinking of replacing your system, you may want to look at the Statements thread here:

https://www.htguide.com/forum/showthread.php?t=28728

And the MiniStatments/Center thread:

https://www.htguide.com/forum/showthread.php?t=28728

I suggest this, as while it doesn’t utilize your drivers (Good choices, by the way.) it may give you some ideas, as there is a whole family of designs around the same set of drivers, including a forthcoming Statement monitor, a smaller 3 way. The center channel topology with the vertical tweeter and mid flanked by a pair of woofers is probably the best topology for a center speaker, although it does take up more vertical space than most would prefer.

C

I checked the MiniStatments/Center thread and I really liked that design...
You guys have me all fired up here....almost planning to buy a couple of extra drivers and make it a 3-way design
Similar design to the B&W HTM3S

Well as you understand I have really decided to start from scratch. I will order another SS18W and start all over using a "normal" center design with two woofers and one tweeter. Unfortunately it doesnt work with a vertical center since the home cinema room not will be ready until earliest next summer. Until then I have to squeze everything into a normal living room. (but when the time comes I might consider redesigning the center box )

I really appreciate that you "stopped" me before buying a lot of expensive crossover components and I will probably be back in a while with loads of more strange questions.

Thanks!

/Anders

[Curt C]

The minimum center to center spacing of a horizontal center channel MTM utilizing the 18W's and the 6600 would suggest an 'optimal' crossover frequency of 1200 Hz. This is what compromises most large MTM center channels.

I've not used the 6600, but looking at the HD plots, you should be able to cross it below 2K, although the energy storage increases somewhat. I'd suggest trying something like 1500 Hz 4th order acoustic and see how low you can take it before it starts complaining. A 2K crossover will work, but the horizontal nulls either side of the main lobe will be closer to on axis than a lower crossover. -Just another compromise you will need to consider, but not a show stopper. Keep the drivers as close together as possible on the baffle.

C

[anders]

The minimum center to center spacing of a horizontal center channel MTM utilizing the 18W's and the 6600 would suggest an 'optimal' crossover frequency of 1200 Hz. This is what compromises most large MTM center channels.

I've not used the 6600, but looking at the HD plots, you should be able to cross it below 2K, although the energy storage increases somewhat. I'd suggest trying something like 1500 Hz 4th order acoustic and see how low you can take it before it starts complaining. A 2K crossover will work, but the horizontal nulls either side of the main lobe will be closer to on axis than a lower crossover. -Just another compromise you will need to consider, but not a show stopper. Keep the drivers as close together as possible on the baffle.

C

OK!
Well 1200 Hz seems low, resonanse frequency of this driver is about 500Hz, low but not enough I think.
I have read about several designes where the crossover frequency was down to 1600 Hz so I believe I will aim towards that area.

Drivers close together.
Yes, since the height of the speaker makes it impossible for a vertical alignement I was thinking of a middle way. Would i be an advantage to have the drivers in a triangel? Tweeter on top and bas/midrange on each side (I have attached a small sketch).

Even very high end center speakers have a pure horizontal alignement (except the B&W). Are we magnifying this problem?

Looking at the JMLab Utopia center





They have a center speaker with pure horizontal alignement on midrange and tweeter. I believe they have put a lot of money into the crossover network on this one so maybe they have solved it this way...

/Anders

[Curt C]

Drivers close together.
Yes, since the height of the speaker makes it impossible for a vertical alignement I was thinking of a middle way. Would i be an advantage to have the drivers in a triangel? Tweeter on top and bas/midrange on each side

In my opinion, this is the best compromise in an MTM center channel design. The first woofer null will be far to the sides, and with a reasonably low crossover frequency will not be a factor. Vertical polar response will of course be non symmetrical, but this is not generally an issue as the listeners stay within a small vertical angle from sitting to standing.

Even very high end center speakers have a pure horizontal alignement (except the B&W). Are we magnifying this problem?

Looking at the JMLab Utopia center





They have a center speaker with pure horizontal alignement on midrange and tweeter. I believe they have put a lot of money into the crossover network on this one so maybe they have solved it this way...

/Anders

I can guess as to why many of them choose compromised center channel designs:
-Many have an MTM they designed for vertical use, and simply lay it on its side to save engineering costs, etc.
-Others choose to maximize the aesthetics, at the expense of performance.
-I suspect some even assume the buying public is simply too ignorant to know the difference.

-IMO, the majority of the current generation on young adults have been conditioned by the marketers to accept mediocre performance as normal. When exposed to true fidelity, they don't recognize it, as they have never experienced it bofore.

A little theory:
For two coincident point sources, measured at some angle off axis, and at a given frequency multiple, the wave front from the more distant driver will arrive with sufficient delay to cause an additional 180 degrees of phase rotation compared to the nearer driver. When this occurs, the two wave fronts will combine destructively and the result will be a null.

The further apart the two point sources are, the lower the frequency is that this first occurs. This is simply physics. There is no way to correct for it in the crossover. Angling the drivers will help only if one of the driver’s off axis response at that frequency and angle is significantly reduced. This will result in a broad reduction in acoustic energy off-axis instead of a series of nulls. In my opinion, neither solution provides optimal off-axis performance.

C

[Dennis H]

Even very high end center speakers have a pure horizontal alignement (except the B&W).

Many 'high-end' speaker companies build WM/TW centers, if they build a center at all. Revel and Aerial come to mind off the top of my head.

[anders]

In my opinion, this is the best compromise in an MTM center channel design. The first woofer null will be far to the sides, and with a reasonably low crossover frequency will not be a factor. Vertical polar response will of course be non symmetrical, but this is not generally an issue as the listeners stay within a small vertical angle from sitting to standing....
C

OK, nice that you agree, for me this means that I have learned something. Attached picture is a small sketch of my idea. I managed to get the drivers fairly tight together by moving the tweeter outside the box. The box will be more complicated to build but it really doesn't matter, think I can handle this.
The horizontal spacing between woofer and tweeter (center to center) is about 105 mm and if I've done the calculations correct this would correspond to 1600 Hz crossover point.

One question regarding this design is the volume needed behind the tweeter, any recommendations?

...I can guess as to why many of them choose compromised center channel designs:
-Many have an MTM they designed for vertical use, and simply lay it on its side to save engineering costs, etc.
-Others choose to maximize the aesthetics, at the expense of performance.
-I suspect some even assume the buying public is simply too ignorant to know the difference.

-IMO, the majority of the current generation on young adults have been conditioned by the marketers to accept mediocre performance as normal. When exposed to true fidelity, they don't recognize it, as they have never experienced it bofore.

A little theory:
For two coincident point sources, measured at some angle off axis, and at a given frequency multiple, the wave front from the more distant driver will arrive with sufficient delay to cause an additional 180 degrees of phase rotation compared to the nearer driver. When this occurs, the two wave fronts will combine destructively and the result will be a null.

The further apart the two point sources are, the lower the frequency is that this first occurs. This is simply physics. There is no way to correct for it in the crossover. Angling the drivers will help only if one of the driver’s off axis response at that frequency and angle is significantly reduced. This will result in a broad reduction in acoustic energy off-axis instead of a series of nulls. In my opinion, neither solution provides optimal off-axis performance.

C

You are probably correct in every word.
JMLab DO have a speaker that is identical to the center speaker in my link. Only difference is that it is standing vertically

It's just that I have always considered the JMLab design to be state of the art. Since I am starting to learn about the theories I also realize that even this company is "cutting corners"

...-IMO, the majority of the current generation on young adults have been conditioned by the marketers to accept mediocre performance as normal. When exposed to true fidelity, they don't recognize it, as they have never experienced it bofore...

Bullseye!..........most people today have 128 kbit mp3 files as the only daily music source.

Many 'high-end' speaker companies build WM/TW centers, if they build a center at all. Revel and Aerial come to mind off the top of my head.

Actually, I had never seen those before. Pretty nice trick to cut the drivers to really get them together



Very well, today I bought some new MDF boards and will probably start building this weekend

Another question:
According to the designer of the SS18W driver the sweet spot should be 22 liters box volume. Many simulation programs doesn't agree, instead the best fit looks to be 30 liters or even more. I've read about many well known DIY people whose design is over 30 liters with this driver but at least one of them say that the midrange is affected if the box is to big. I am planning to go for the designers recommendations, but how do I approach this with two drivers? Just double the volume?
Should i separate the box into two (internally) so that each driver have 22 liters?


/Anders

[rc white]

The size of the box depends upon the bass extension you are after.
You can make the box smaller if you want a satellite system that has a lower frequency cut off in the 80-100Hz. region for use with a subwoofer.
If you want a full range system the drivers you are using do not have much useful output bellow 40Hz.

[Curt C]

One question regarding this design is the volume needed behind the tweeter, any recommendations?

The tweeter has a sealed rear chamber. No enclosure is required. I would suggest however, that the semi-circular portion of the baffle will result in a tweeter diffraction spike at approximately 10 kHz. Utilizing a 20mm roundover bit on the edge will mitigate the issue somewhat. Better would be to bring out the sides of the baffle next to the tweeter, resulting in something like an inverted V shape for the top edge of the baffle rather than the semi-circle. See my crude drawing below...

Another question:
According to the designer of the SS18W driver the sweet spot should be 22 liters box volume. Many simulation programs doesn't agree, instead the best fit looks to be 30 liters or even more. I've read about many well known DIY people whose design is over 30 liters with this driver but at least one of them say that the midrange is affected if the box is to big. I am planning to go for the designers recommendations, but how do I approach this with two drivers? Just double the volume?
Should i separate the box into two (internally) so that each driver have 22 liters?

If I were using the 18W, I'd probably use 30 liters per driver and tune it to 25Hz. I like this tuning as it emulates a sealed rolloff down to f10, and has an f3 of 40 Hz. 60 liters is a big center channel however, so 22 liters/driver tuned to 22 Hz will provide a similar rolloff. The f3 is raised consequently raised to 47 Hz.

There is no reason to add the complexity of 2 separate enclosures. I'd suggest mounting the vent on the rear centered between the drivers.

C

[anders]

The tweeter has a sealed rear chamber. No enclosure is required. I would suggest however, that the semi-circular portion of the baffle will result in a tweeter diffraction spike at approximately 10 kHz. Utilizing a 20mm roundover bit on the edge will mitigate the issue somewhat. Better would be to bring out the sides of the baffle next to the tweeter, resulting in something like an inverted V shape for the top edge of the baffle rather than the semi-circle. See my crude drawing below.....

OK, no volume behind the tweeter will make the woodwork much easier, but.....will it not get hot in there?

The semi circular baffle...
Of course, I should have known this, after all the reading I've done ops:

Your suggestion not only looks better but this will also make the woodwork easier, perfect!

...If I were using the 18W, I'd probably use 30 liters per driver and tune it to 25Hz. I like this tuning as it emulates a sealed rolloff down to f10, and has an f3 of 40 Hz. 60 liters is a big center channel however, so 22 liters/driver tuned to 22 Hz will provide a similar rolloff. The f3 is raised consequently raised to 47 Hz.

There is no reason to add the complexity of 2 separate enclosures. I'd suggest mounting the vent on the rear centered between the drivers.

C

Yes, the size have almost become a problem here and as you say, 60 liters will be a hugh center speaker. The "tweeter trick" helped a lot though, makes tha box look smaller than it is.

I am quit happy if I reach 50Hz with this speaker, dont really see the need to push it any lower. If the center speaker is activated in my system it means that the subwoofer is as well and I can have the sub to take care of everything under 50 Hz.

There is a rule of thumb that I cannot recall right now.....something about not designing a box with lower tuning frequency than fs of the woofer, or is it the f3 point that should be higher than fs (due to compromizing of the power handling)?

Anyway, I think you know this way better than me so I will follow your recommendations.

Thanks again!

/Anders

[rc white]

You can use two of the ss18w in a 30 litre sealed box. This in conjunction with a high pass filter at 80Hz. typical of THX type center channel outputs, gives a fourth order characteristic at 80Hz. whilst keeping within the linear cone excursion for 100Watts input.

[JonW]

Hi Anders,

This looks like a nice project. With those drivers and the right effort here and there, it should turnout sounding very good. I’ve been using the same 6600/18W combo and can tell you that they sound great together. And that’s in a flimsy test cabinet with only a preliminary crossover.

Your first posted crossover looks excellent, really super for a first shot. It sounds like you’ve got a good hang of things.

Two things I will mention, one of which has already come up. First, you need to get your own measurements in your own cabinets. You’ve taken the manufacturer's data as far as you can. It’s the biggest pain about this whole process but it really is necessary.

Second, you have to do *a lot* of listening. Change around every component in your preliminary crossover. And also try lots of different woofer tunings (volumes and/or port lengths). Don’t just use a given volume or tuning because someone else did or Unibox/WinISD says to. I found that lower tunings yielded more bass but at the expense of midrange clarity. So you need to find the middle ground there that works for your ears.

I need to update the Spassvogel thread with the latest minor progress. Work has been really busy so I haven’t been able to spend much time here on the boards.

Your project looks excellent. Keep at it. :T

-Jon

[Curt C]

That rule of thumb probably made sense back before we had software to see the actual performance of the drivers. There are actually several advantages in smaller volumes/lower tuning compared to your typical QB3 alignment. The one most germane to our situation is in the 22 liter/driver enclosure; the cone will not unload until below 18 Hz. All of the volumes/tunings we have discussed will make at least 106 dB/driver SPL without exceeding Xmax. Certainly limiting low-end extension can improve the harmonic distortion levels in the lower midrange, but this is very driver specific as to where and what level these effects will be.

I suspect the tweeter will have no thermal issues…

C

[anders]

Hi Anders,

This looks like a nice project. With those drivers and the right effort here and there, it should turnout sounding very good. I’ve been using the same 6600/18W combo and can tell you that they sound great together. And that’s in a flimsy test cabinet with only a preliminary crossover.

Your first posted crossover looks excellent, really super for a first shot. It sounds like you’ve got a good hang of things.

Two things I will mention, one of which has already come up. First, you need to get your own measurements in your own cabinets. You’ve taken the manufacturer's data as far as you can. It’s the biggest pain about this whole process but it really is necessary.

Second, you have to do *a lot* of listening. Change around every component in your preliminary crossover. And also try lots of different woofer tunings (volumes and/or port lengths). Don’t just use a given volume or tuning because someone else did or Unibox/WinISD says to. I found that lower tunings yielded more bass but at the expense of midrange clarity. So you need to find the middle ground there that works for your ears.

I need to update the Spassvogel thread with the latest minor progress. Work has been really busy so I haven’t been able to spend much time here on the boards.

Your project looks excellent. Keep at it. :T

-Jon

Thanks for the kind words!

I actually found your thread the same day as my drivers were delivered. When I did a search on Google your project popped up and actually this was how I discovered this site, many thanks!

Yes I belive it doesnt hurt to try a lot of combinations of box volumes and port tuning. Problem for me right now is that this center speaker will probably end up in Guiness book of record if I increase the volume any further. I am a big fan of "sturdy" speaker cabinets and I hate flimsy boxes. This cabinet have a wall thickness of 4.5 cm, sandwich construction made of 19 mm MDF + 13 mm gypsym and finally 10 mm MDF. In between the layers I have a vibration reducing "silicon glue". This of course increase the outer dimension quit much even if the internal is quit moderate. I have pushed it to about 48 liters and that is as far as it goes. I still have the port tuning to play around with.

I have read about the midrange beeing affected by larger volumes (both in your thread and other places) and I think I am on the correct approach in keeping the volume down. Midrange performance is very important to me and if I am going to replace my front speakers later on with a similar design this range must be 100%.

I will follow your thread carefully, I've allready found lot of useful information in there and I guess there is more to come.

That rule of thumb probably made sense back before we had software to see the actual performance of the drivers. There are actually several advantages in smaller volumes/lower tuning compared to your typical QB3 alignment. The one most germane to our situation is in the 22 liter/driver enclosure; the cone will not unload until below 18 Hz. All of the volumes/tunings we have discussed will make at least 106 dB/driver SPL without exceeding Xmax. Certainly limiting low-end extension can improve the harmonic distortion levels in the lower midrange, but this is very driver specific as to where and what level these effects will be.

I suspect the tweeter will have no thermal issues…

C

OK!
I will definitaly try the lower port tuning. When I did some simulations tonight I understand what you mean by "it emulates a sealed rolloff ". Maybe it is a must to have a low tuning if I am to avoid a peak in the 100-150 Hz region that comes with a smaller size...



I did some "design shootout" and the one attached was the best shot tonight, tell me what you guys think of it. I will probably change it alot before I descide permanently so I am open to suggestions 8)

/Anders

[anders]

You can use two of the ss18w in a 30 litre sealed box. This in conjunction with a high pass filter at 80Hz. typical of THX type center channel outputs, gives a fourth order characteristic at 80Hz. whilst keeping within the linear cone excursion for 100Watts input.

This is a good option and I will test this as well (when the time comes) by closing the vent and reduce the volume (by putting something big inside the box).

Without any real experiance I might dare to say that I believe this option is the best if we consider midrange performance.

/Anders

[JonW]

Hi Anders,

That sounds like one very thick, very heavy and pretty large enclosure. I approve whole heartedly. :T

OK, I see the constraints upon your volume. So then, in this case, I wouldn’t worry about it. I’ve seen people recommend various volumes for this woofer- from as low as 18 L to as high as 32 L. In my opinion, if you’re making such a high end speaker, I say figure out the best tuning yourself, rather than taking other people’s word on it. But your 48 L for 2 woofers sounds very good to me. I forget what the optimal volume I settled on was for one woofer. Maybe 20 L? Details are in the Spassvogel thread (and my notes at home). But 48 L is definitely in the “excellent” range. It will provide a small difference, but I might recommend sticking with the 48 L and then trying a few different port lengths to see what you like best. If you don’t want to be bothered, just go with the tuning I had and all will go just fine.

For the proposed drawing you attached... I think it looks good. Make sure the black area is flush with (not sticking out in front of) the brown area. Otherwise you’d get some poor baffle diffraction effects. The “unusual” shape of your baffle might be good for minimizing baffle diffraction. You have all your drivers close to each other which is good. I found some real benefits to offsetting the tweeter from the center line, so you might want to consider that. But then it won’t look so good, I know.

Glad to hear my thread has been helpful to you. I’ve been trying to add in as much detail as possible, in case it could be of benefit to anyone else. But really it’s the helpful people on this board that have made the project possible. Lots of greta folks here. Yes, more details to follow as they happen. I’ve been taking lots of photos and detailed notes on how I built the cabinets. If anyone was crazy enough to want to make the exact same design, I could burn a few hours to write up a detailed cabinet construction process.

[Curt C]

I will definitaly try the lower port tuning. When I did some simulations tonight I understand what you mean by "it emulates a sealed rolloff ". Maybe it is a must to have a low tuning if I am to avoid a peak in the 100-150 Hz region that comes with a smaller size...

I did some "design shootout" and the one attached was the best shot tonight, tell me what you guys think of it. I will probably change it alot before I descide permanently so I am open to suggestions 8)

/Anders

I like the looks, Anders. With your constrained layer construction, your enclosure will certainly will limit panel flex. -I'm just glad I don't have to be the one to lift it! :E

If it were my project, I'd use the 10mm MDF on the inside rather than the 19mm. I think that, along with staggered bracing to break up the large panel areas will be sufficient, and will help reduce the enclosure size somewhat.

I also suggest lining the enclosure with something like black hole 5, sonic barrier, wispermat, or equivelant. This will further attenuate the unwanted acoustic radiation, and give you yet another constrained layer.

C

[Curt C]

Make sure the black area is flush with (not sticking out in front of) the brown area. Otherwise you’d get some poor baffle diffraction effects. The “unusual” shape of your baffle might be good for minimizing baffle diffraction.

I modeled Anders raised baffle shape in a diffraction simulation program. For the tweeter, it showed a smoother diffraction curve than a flush one. Diffraction effects from the raised panel will have less impact the 18W due to the larger piston diameter and limited pass band. Of course the frequencies where baffle step occurs will be low enough that the raised baffle is immaterial.

It appears that raising the hexagonal panel from the enclosure certainly won't hurt and quite possibly may help. Irregular, non-symmetrical shapes almost always exhibit better diffraction effects.

C

[JonW]

I modeled Anders raised baffle shape in a diffraction simulation program. For the tweeter, it showed a smoother diffraction curve than a flush one. Diffraction effects from the raised panel will have less impact the 18W due to the larger piston diameter and limited pass band. Of course the frequencies where baffle step occurs will be low enough that the raised baffle is immaterial.

It appears that raising the hexagonal panel from the enclosure certainly won't hurt and quite possibly may help. Irregular, non-symmetrical shapes almost always exhibit better diffraction effects.

C

Sounds good. I recommend that Anders listens to the results of your simulations rather than my broad, sweeping, and rather useless generalizations. :B

[anders]

Hi Anders,

That sounds like one very thick, very heavy and pretty large enclosure. I approve whole heartedly. :T

OK, I see the constraints upon your volume. So then, in this case, I wouldn’t worry about it. I’ve seen people recommend various volumes for this woofer- from as low as 18 L to as high as 32 L. In my opinion, if you’re making such a high end speaker, I say figure out the best tuning yourself, rather than taking other people’s word on it. But your 48 L for 2 woofers sounds very good to me. I forget what the optimal volume I settled on was for one woofer. Maybe 20 L? Details are in the Spassvogel thread (and my notes at home). But 48 L is definitely in the “excellent” range. It will provide a small difference, but I might recommend sticking with the 48 L and then trying a few different port lengths to see what you like best. If you don’t want to be bothered, just go with the tuning I had and all will go just fine.

For the proposed drawing you attached... I think it looks good. Make sure the black area is flush with (not sticking out in front of) the brown area. Otherwise you’d get some poor baffle diffraction effects. The “unusual” shape of your baffle might be good for minimizing baffle diffraction. You have all your drivers close to each other which is good. I found some real benefits to offsetting the tweeter from the center line, so you might want to consider that. But then it won’t look so good, I know.

Glad to hear my thread has been helpful to you. I’ve been trying to add in as much detail as possible, in case it could be of benefit to anyone else. But really it’s the helpful people on this board that have made the project possible. Lots of greta folks here. Yes, more details to follow as they happen. I’ve been taking lots of photos and detailed notes on how I built the cabinets. If anyone was crazy enough to want to make the exact same design, I could burn a few hours to write up a detailed cabinet construction process.

Yes, the negative aspect of this wall design is definately that the box size is growing towards unacceptable measures.

Yesterday I made a paper model just to actually "see" the box in my living room and ouch :E My first box (the one I scrapped) was big....this one was....hmmm, how should I put it.....enourmous!!
.......something got to be done.

I dont want to lower the volume below 48 liters and I dont want to give up on the sandwich design, sounds lik mission impossible but after a few tricks I am getting quite close.

Curts idea of using 10 mm MDF on each side of the gypsum was excellent. The wall will still be about 35 mm thick and this is OK.

I removed the bracing panels and instead I will use "MDF sticks" that is 19x20 mm. Five rows and 3 columns of these sticks will substitute the bracing panels.

Backside of enclosure will be a "normal" 19 mm MDF board.

Inside MDF boards on top/bottom and sides will have round holes with 4 cm diameter, totally about 65 circular holes in the inner MDF. This not only gave me further headroom but it will hopefully prevent/reduce standing waves.

All of this combined gives me a box of the original (like the first one) size but with the volume of 48 l and I think that I managed this without compromizing my original design ideas to much.



I will try to save as much detailed info as possible along the way and post it here, maybe the information will help someone some day . Afterall, as you say, it is thanks to all the helpful and experianced people here that a newbie like me actually dare to enter a project like this.

Ill post the box drawings once the design is 100% finished.

/Anders

[anders]

I modeled Anders raised baffle shape in a diffraction simulation program. For the tweeter, it showed a smoother diffraction curve than a flush one. Diffraction effects from the raised panel will have less impact the 18W due to the larger piston diameter and limited pass band. Of course the frequencies where baffle step occurs will be low enough that the raised baffle is immaterial.

It appears that raising the hexagonal panel from the enclosure certainly won't hurt and quite possibly may help. Irregular, non-symmetrical shapes almost always exhibit better diffraction effects.

C

Curt

Thanks for helping me out!

What software are you using for this kind of simulations?

Just to confirm......Yes, the black area is raised above the lower baffel, one MDF board thick (19mm).

I was thinking of not doing this with the tweeter and in this way get a better time alignement but it makes the woodwork harder since the drivers are so close together (at least hard to get a nice design). Would the difraction curve be worse off if I lower the tweeter to the cabinet "base level"?

/Anders

[rc white]

You can kill two birds with one stone with a waveguide on the tweeter. It allows you to compensate for acoustic offset and it reduces the illumination of the enclosure boundaries by the tweeter, lessening the total amount of diffraction.
Another useful technique is to make the distances from the driver centers to the enclosure boundaries in the 1:1.2, 1.6 golden ratios, this has the effect of smoothing out the diffraction field, (try this in a simulation program and you will see that it works).

[Curt C]

Curt

Thanks for helping me out!

What software are you using for this kind of simulations?

Just to confirm......Yes, the black area is raised above the lower baffel, one MDF board thick (19mm).

I was thinking of not doing this with the tweeter and in this way get a better time alignement but it makes the woodwork harder since the drivers are so close together (at least hard to get a nice design). Would the difraction curve be worse off if I lower the tweeter to the cabinet "base level"?

/Anders

There are a couple of freeware apps out there. There is 'The Edge' from Tolvan data:

http://www.tolvan.com/edge/help.htm

Quite versatile in enclosure shape, but doesn't show the effects of edge treatments.

Another good one is the Baffle Diffraction Simulator:

http://www.pvconsultants.com/audio/frdgroup.htm

Although limited to quadrilateral baffles, it models the effects of roundovers and chamfers. Both allow you to export the diffraction effects as .frd files.

Trying to time align the acoustic centers of the drivers when they are this close together will prove quite challenging. The abrupt transition in baffle height between the mids and the tweeter in a stepped baffle will cause some significant tweeter response abberrations due to the diffraction from the step, and there is no room to slope the baffle in between them. As RC indicated, you could use a waveguide, but again, due to the spacing constraints, there is not room for a waveguide. Instead, you'd end up with a horn with a diameter no larger than the tweeter flange.
Since it is fairly simple to compensate for the differences in acoustic centers in the crossover, I'd suggest that might be the most elegant solution for that issue.

C

[JonW]

Yes, the negative aspect of this wall design is definately that the box size is growing towards unacceptable measures.

Yesterday I made a paper model just to actually "see" the box in my living room and ouch :E My first box (the one I scrapped) was big....this one was....hmmm, how should I put it.....enourmous!!
.......something got to be done.

I dont want to lower the volume below 48 liters and I dont want to give up on the sandwich design, sounds lik mission impossible but after a few tricks I am getting quite close.

Curts idea of using 10 mm MDF on each side of the gypsum was excellent. The wall will still be about 35 mm thick and this is OK.

I removed the bracing panels and instead I will use "MDF sticks" that is 19x20 mm. Five rows and 3 columns of these sticks will substitute the bracing panels.

Backside of enclosure will be a "normal" 19 mm MDF board.

Inside MDF boards on top/bottom and sides will have round holes with 4 cm diameter, totally about 65 circular holes in the inner MDF. This not only gave me further headroom but it will hopefully prevent/reduce standing waves.

All of this combined gives me a box of the original (like the first one) size but with the volume of 48 l and I think that I managed this without compromizing my original design ideas to much.



I will try to save as much detailed info as possible along the way and post it here, maybe the information will help someone some day . Afterall, as you say, it is thanks to all the helpful and experianced people here that a newbie like me actually dare to enter a project like this.

Ill post the box drawings once the design is 100% finished.

/Anders

That all sounds good. You're on the right track. I'd guess you don't really need walls that are a mile thick. With my Spassvogels I'm trying single layer thickness (except 3 layers on the front) for the cabinet but with a few shelf braces added. We'll see how that works out but so far it looks OK. You could probably cut down on your wall thicknesses and still be very good. But keep your front baffle 2 layers thick. For volume, your 48 liters sounds good. If you did 40, with the 2 woofers, you're probably OK too. So if your mockup looks to be too large for the room, smaller volume and thinner walls might be OK. However if you don't mind the size and weight, the thicker walls and larger volume are good things.

[anders]

The "Edge" was a nice program.
Amazing what small changes in the baffle design can lead to quit noticable changes in output.
I will use this program later on when the final design of the baffle is to be done.

This weekend I started to build the "inner" box out of 10mm MDF. 132 holes were made with a circular saw (!), my drilling machine almost died before I was finished. 8O

The "bracing sticks" that is used to replace the bracing panels seems to work out OK, the box is quit stiff even though most of the box materal is removed (I'll post some pictures later on today) The inner box almost looks like a birdcage, maybe that should be tha name of my project, "The Birdcage" 8)

I have one problem that I havn't sorted out yet. The port was ment to be located on the backside right between the two woofers according to Curts recommendation. However, I didn't think of this when positioning the bracing and now there is a row of bracing sticks right where the port was suposed to be. So, now I am planning to have the port in either the right or the left corner instead, wouldnt this be better compared to have the port just behind one of the woofers? Maybe it doesnt matter, it just feels strange to have the port exactly behind one of the woofers.

...Trying to time align the acoustic centers of the drivers when they are this close together will prove quite challenging. The abrupt transition in baffle height between the mids and the tweeter in a stepped baffle will cause some significant tweeter response abberrations due to the diffraction from the step, and there is no room to slope the baffle in between them. As RC indicated, you could use a waveguide, but again, due to the spacing constraints, there is not room for a waveguide. Instead, you'd end up with a horn with a diameter no larger than the tweeter flange.
Since it is fairly simple to compensate for the differences in acoustic centers in the crossover, I'd suggest that might be the most elegant solution for that issue.
C

Yes, I made som small pieces of MDF just to see how this could be solved and it will be virtually impossible when the drivers are so close together as they are in this case.

I will go for the flat baffle....

/Anders

[JonW]

I have one problem that I havn't sorted out yet. The port was ment to be located on the backside right between the two woofers according to Curts recommendation. However, I didn't think of this when positioning the bracing and now there is a row of bracing sticks right where the port was suposed to be. So, now I am planning to have the port in either the right or the left corner instead, wouldnt this be better compared to have the port just behind one of the woofers? Maybe it doesnt matter, it just feels strange to have the port exactly behind one of the woofers.

I think (so check on this) that you do not want the port directly behind a woofer. Otherwise it's not really acting as a port, just letting air out when the woofer moves. Can you put it behind the tweeter? That's pretty standard, although not necessarily easy with your design. I guess a corner would be OK if it's not directly behind a woofer. Make sure the port is at least one diameter from each cabinet wall.

[Curt C]

I have one problem that I havn't sorted out yet. The port was ment to be located on the backside right between the two woofers according to Curts recommendation. However, I didn't think of this when positioning the bracing and now there is a row of bracing sticks right where the port was suposed to be. So, now I am planning to have the port in either the right or the left corner instead, wouldnt this be better compared to have the port just behind one of the woofers? Maybe it doesnt matter, it just feels strange to have the port exactly behind one of the woofers.

To equally load the midwoofers, a single port should be symmetrically placed in between them. Offsetting the port will cause the acoustic impedance of the midwoofers to be unbalanced, and probably not the best choice for a quality design. There are options, however:

1) Remove the center row of braces and replace them with an offset row of braces, or two rows of smaller braces either side of the port. In general braces should be offset to make the resultant smaller panels (between the braces) different sizes. Making the spacing between them all the same results in many small panels that all resonate at the same frequency. In you case, I get the impression that all these panels will resonate above the pass band of the midwoofers, so I throw this out as 'sauce for the goose'.

2) Use 2 ports instead of one. Ideally these ports should not be aimed directly at the drivers, so perhaps moving them to the rear corners, as you suggest, may be the best solution. 2 ports each with 1/2 of the area of the single port should result in the same port length.

3) Actually a subset of the dual ports, dual slot ports fit easily into designs like this and are easy to maximize the port area to the available enclosure depth. The link is a drawing showing dual front slot porting, but obviously the same technique could be used on the rear of the enclosure.

http://www.geocities.com/cc00541/Eric_D_MTM_CC_DWG.jpg

C

[anders]

...2) Use 2 ports instead of one. Ideally these ports should not be aimed directly at the drivers, so perhaps moving them to the rear corners, as you suggest, may be the best solution. 2 ports each with 1/2 of the area of the single port should result in the same port length.
C

I definitely prefer this one....

Otherwise it's not really acting as a port, just letting air out when the woofer moves.
C

This was my feeling as well.......


I will upload some photos when I get home today, I feel that I am beeing mysterious, talking about sticks and birdcages :B

/Anders

[Curt C]

I will upload some photos when I get home today, I feel that I am beeing mysterious, talking about sticks and birdcages :B

Well, it does paint an interesting mental picture... :P

C

[anders]

Well, it does paint an interesting mental picture... :P

C

OK, let me ruin your fantasy... 8)

/Anders

ps. How do I attach pictures that are visible directly in the thread? (without clicking on the thumbnail)

[Hdale85]

Your just going to do an MTM instead of a WTMW?

[anders]

Your just going to do an MTM instead of a WTMW?

Yes, this is my first attempt and I dont want to complicate things more than necessary. A two way design seemed like a nice starting point for me.

/Anders