Speaker Design Proposal



Speaker Design Proposal

Dan Bluhm

FA 4740

2/19/10

Introduction.

“Our goal in designing loudspeakers I to create systems that accurately reproduce the full range of frequencies that are audible to the human ear. Our loudspeakers are clearly the single most critical link in the complete audio signal chain.”[1] This accurate reproduction is what allows us as engineers to hear how our mixes will translate to other systems. It also allows us to listen to other peoples work in a critical environment where we know the strengths and weaknesses of the tools that we use, the critical link in the signal chain, the loudspeaker. In my design there were many different decisions that had to be made based on size, budget, SPL capabilities as well as low frequency extension. In my design I attained the best possible results by setting forth guide lines to aid in enclosure design, driver selection, amplification source, and crossover design in order to limit and push myself to make the best design possible.

Goals.

Often at work I am surrounded by loud, in your face, systems that care more about SPL than how they blend to create an acoustic stage. My design focused on creating a pair of speakers to be used in a home theatre, easy listening setting. The speaker will be used in a medium sized living room, where you are seated approximately 8-14 feet away from the source and will be used to play a wide verity of music. Since people will be seated around the room these speakers will need to have good off axis response, sounding good in multiple locations across a wide angle.

Box Design.

For my design I will be building a bass reflex enclosure. I decided to go with this design because for a given river you can get more extended bass response. Also for a particular f3 and enclosure size you can get greater efficiency.[2] When designing and building a bass reflex there are three main variables, cabinet volume, length of the port, and area of the port. Below the tuning frequency of the port air moves in and out of the port under the influence of the driver, since there is no air loading to hold the speaker, it becomes easy to over excurse drivers.[3] For my design my internal cabinet volume is approximately 50 liters or 1.7 cubic feet. I will have 1 port that will have an id of 2 in a surface area of 3.147 in2 and a length of 4.212 in. My port will be flared and placed on the back of the cabinet to avoid hearing any port noise.

Size.

Since they will be sitting next to the tv I don’t want them to be too obtrusive, but quality over takes size. Meaning that if to get more low end I need to make my box bigger I am willing to make that trade off. I feel that tower speakers would be difficult to pack and move, so I am building speakers equivalent to the size of large portable studio monitors. My current dimensions from my drafting size my speaker as being 20” wide (at the widest point), 17” deep, and 30” tall.

Materials.

My cabinet design employs a 4 part, constrained layer design. The outer most layer will be ¾ of an inch of 13 ply Baltic birch plywood, inside of that will be ½ of an inch of MDF, inside of that will be 1/8 of an inch of mass loaded vinyl, and the inner most layer will be 1/8 of an inch of masonite. “The MDF and plywood panels are rigidly coupled, and the resulting panel is substantially stiffened. The Baltic Birch plywood also offers additional stiffness because of its grain and highly layered construction. Stiffness is proportional to the square of panel thickness, therefore the composite panel is at least four times as stiff as a single MDF layer.”[4]

In order to add rigidity to the cabinet and increase the panels fundamental resonance frequency my cabinet will be braced. “The correct material to use for internal bracing is multidirectional plywood. Plywood is very strong in any direction under tension, because wood does not stretch along the grain. Although not strong under compression, it is still superior to soft wood and far superior to composite board.”[5] My braces will be a solid shape lining the inside of the cabinet and screwing into the inner layer of my design. In addition parts of the cabinet will be lined with acoustic foam to dampen the cabinet, a acoustic foam wedge will also be put onto the back wall in order to further break up waves inside the box.

Shape.

My design will utilize a six sided triangular shape where the points of the triangle would be are cut off at 45 degree angles. By using this design I hope to avoid many problems that more traditional rectangular boxes encounter. These speakers will use the construction materials and angles in such a way in order to minimize panel vibrations, internal cabinet resonance and standing waves as well as edge diffraction.

Diffraction.

Diffraction occurs in the low frequency range of loudspeakers in enclosures that

are located in the open, away from walls or other surfaces. The essence of it is this: At high frequencies the speaker is radiating into "half space" i.e. it is only radiating into the forward hemisphere.[6] There are three things that I will be doing to combat diffraction. First, my drivers will be mounted at different distances from the edges of the baffle. Second, I will be flush mounting the drivers, which will create a solid plane on the front baffle. Lastly I will be having my front baffle attach to my sides at 45 degree angles which allows the waves to roll over the corner and there not be a sudden shift in direction.

As seen above, the sphere has a very smooth diffraction curve, as waves are unobstructed as they roll off of the center point and encounter no hard edges to bounce off or change the waves impedance. More angular enclosures like the square and rectangle suffer from a 6dB step along with a series of response ripples that are depended on the placement of the driver with respect to the baffle edges.[7] In my design the front baffle connects to the front side panels at 45 degree angle, and the rear sides connects to the front sides at 70 degree angle. My design addressing diffraction problems early on in the design phase so they will no be an issue when testing and tuning.

As mentioned earlier driver placement is a key variable in minimizing diffraction. My woofer will be put on my front baffle 4 ½ inches off of the bottom of the baffle centered left and right. This leaves approximately 1 ½ inches on either side of the driver before the edge of the baffle. My tweeter is placed in the upper right hand corner of the baffle. It is 2 ¾ inches below the top and 1 inch from the right side of the baffle. I feel that locating the drivers in this way as well as my box shape and design will minimize diffraction effectively and efficiently.

Finishing.

For finishing I am planning on using a stain with a gloss clear coat. This means that I will have to make sure that my speakers are free of dust and debris when beginning the finishing process. My design currently calls for no end grain to be showing except for the top, bottom and where the back panel attaches. I do not want to see end grain, so what I am planning on doing is making this an accent point, much like old cars had chrome accents, this will be the place on my speaker where I can give it some color and flair. I am planning on using a metal of some type and wrapping it around the top and bottom so it looks deliberate instead of end grain left unfinished.

SPL.

I want these speakers to be able to act as a stand-alone set, covering the whole frequency spectrum. My speakers will have clear defined highs and solid low end that can act without a subwoofer, but the addition of one would compliment them nicely. After doing some testing, having my bass go down to 45Hz and lower would be amazing, between 50-62Hz is okay, a range that I could deal with and like, 60-80 Hz is fair but I think if that is what I am left with I will add a sub sooner rather than later, and above 80Hz simply will not do. I would really like my design to have an f3 of around 40Hz.

Since I will be using these for movie playback and general music listening I would like them to comply with THX reference levels. THX reference level is that, you are able to reproduce 85dB SPL with 20 dB of headroom, and have minimal distortion.[8] This fits much nicer with the levels I was referring to in my last paper of being able to meter at K – 14, where 83 dBC SPL is zero, which would allow me to have +14 dB of headroom which would be 97 dB. Since at my listening position I want it to be 80 dB which is 4 meters away, making my 1 meter source level at 92 db, and I want to be at K – 14 standards, my speakers will need to be able to produce levels at 106 dB. This means at a distance of 4 meters I would like do have a SPL of 80-85 dB.

| | | |Desired Distance = 4 M |

|Sensitivity (1W/1M) |Distance (M) | |SPL |Wattage |

|88 |1 | |76 |1 |

|82 |2 | |79 |2 |

|76 |4 | |82 |4 |

|70 |8 | |85 |8 |

|64 |16 | |88 |16 |

| | | |91 |32 |

| | | |94 |64 |

| | | |97 |128 |

| | | |100 |256 |

| | | |103 |512 |

As you can see, I modeled my tables on a woofer with a 88 dB sensitivity, at a distance of 4m. This means at 1m I will have 92db, plus and additional 14 db of headroom, bringing my SPL at 1m = 106 dB. To attain this SPL I would need to get an amp with wattage of greater than 64 per channel.

[pic]

As you can see from this dB vs. W graph, it is clear to see that you have a minimal gain of SPL above 128W. This is the point where the wattage needed to drive your speakers an extra 3 db starts to get much larger and much more expensive. For my design I am using a Onkyo TX-NR708 Receiver, this receiver boasts 110 watts/channel, THX Select 2 Plus certification, Dolby certification, as well as on board Audyssey EQ. At a distance of 4 meters it will allow me to attain between 94 and 97 dB.

Low Frequency Extension.

“A little bit more bass extension costs you a lot of efficiency or box size... you must pay mother nature for the extra octave of bass, there can be no cheating.”[9] One of the biggest things you will hear people talking about is bass, more specifically how much and what type. For my design the two most important things for me are bass response and SPL. Since currently there is no plan on adding a subwoofer in the near future I would like these speakers to be able to handle the low frequencies and reproduce them in a clean and efficient manner. I am designing around having an f3 of 40 Hz, which would leave my bass response in a range that I would really enjoy.

Drivers.

When I first started looking at drivers I spent hours meticulously going through frequency plots, looking at manufacturers specifications, calculating f3’s and box sizes. I was on the quest to find the best drivers to make the best speakers possible. What I found after all that time and effort was that there was no perfect driver, no combination that would be magical and make my life easy and be able to couple my desired spl and low frequency extension. That’s when the compromise really started to happen. This compromise has lead me down some interesting paths looking at drivers that I never expected to use but ultimately were the best choice for my project..

Tweeters.

After much time reading about different tweeters the first main decision came down to hard dome or soft dome. Many people on forums were voicing their opinions about it and what it summed to was that it is less of the material it is made of and more of how it is implemented into your system. The sound of your tweeter is more dependent on your crossover and not the material it is made of. Now that is not to say you can not hear the difference between the two, what I am saying is that material is not the only deciding factor. One post on home theater forum struck a chord with me, they wrote “I prefer a high quality well utilized metal dome tweeter. Compare the best metal to the best soft and I'll go for the metal every time. Now since that's hard to achieve most of the time I'll take a soft dome over a mediocre metal dome setup.”[10] Based on my research and discussion with some recreational speaker builders at 8th day, I decided to go with a soft dome.

I wanted a tweeter that was accurate, and had the ability to have a low crossover point. In the 1k range, this would allow me to preserve the vocal integrity and not put my crossover in the middle of my vocals. One of the other distinguishing factors about my tweeter was that I wanted a great off axis response. Since the purpose of these speakers is home theatre use, I wanted every seat to be the “good” seat. Also I wanted my drivers to have a visual interest to them. These factors became my primary decision criteria.

The tweeter that met my design goals the best was the ScanSpeak Illuminator D3004. Highlights from madisound are listed below. [11]

1" soft dome,

AirCirc Magnet System,

Rubber painted aluminum die casted faceplate

The frequency response curve of this woofer taken from manufactures spec sheet shown below.

[pic]

One thing that this ScanSpeak had going for it that allowed it to beat out the competition was the off axis response. The green line is 30 degrees off center and the red is 60 degrees off center. As you can see going off axis it is able to preserve the majority of the intelligibility region. This tweeter also performs at low frequencies staying flat until it slopes off at 800. This gives me plenty of room to do a second order crossover at 1.5kHz with no problem. This driver fit exactly what I wanted to do, and has a look to it that is appealing.

Woofers.

Once I had found a tweeter I shifted my focus to woofers. I looked at woofers that had a diameter in the 6-8 inch range. I eliminated drivers that were not recommended for a vented box system. I then looked at low-end frequency response. This was one of the bigger deciding factors in which driver to go with. Since I wanted an f3 of 40 Hz or lower I eliminated drivers that either could not go that low, or the required box to get that low of a frequency was too large.

After this cut was made I was left with 3 drivers remaining that were still in the running. I modeled them all using WinSpeakerz so I could see how they would perform in different box types, and the results were convincing. On average in a sealed box of equivalent size I was getting an f3 that was 20 Hz higher than with a vented enclosure. This extra low end was what convinced me to build a 4th order vented box. A few of my woofers really stuck out head and shoulders above the rest. These three were then carefully considered in 4th order vented boxes with both SBB4 and QB3 alignments. These three were then modeled using the QB3.

My first driver considered was the ScanSpeak 21W/8555-0, details from Madisound website.[12]

Hard paper cone,

low loss rubber surround,

SD-1 motor

|Driver Specs | | |SBB4 | | |QB3 | |

|Vas = |136 | |VB = |57.23 | |VB = |52.42 |

|Fs = |20 | |fb = |20 | |fb = |25.496 |

|Qts = |0.31 | |f3 = |33.844 | |f3 = |30.14 |

[pic]

[pic]

After looking at how this driver responds in a QB3 alignment I really liked the performance. But the further I looked in to it, the more I didn’t like the jagged response between 20 Hz and 100 Hz, and with an f3 I knew I would have to contend with that as well at the impedance peak 30 Hz and the breakup mode at 4 – 5 kHz. I was also very discouraged to see that that it started to roll off at 200 hz.

The second woofer that I seriously considered was the Seas Excell W22EX, details from Madisound website.[13]

Precision cast and surface treated magnesium cone

Heavy copper rings mounted above and below the T-shaped pole piece

Extremely stiff and stable injection molded metal basket

A solid copper phase plug enhances the performance of the copper rings

|Driver Specs | | |SBB4 | | |QB3 | |

|Vas = |98 | |VB = |50.06 | |VB = |49.75 |

|Fs = |25 | |fb = |25 | |fb = |29.25 |

|Qts = |0.34 | |f3 = |36.015 | |f3 = |33.15 |

[pic]

[pic]

I really liked how this driver preformed. Looking at the frequency response, I am going to get a calculated f3 of approximately 33 Hz in a box that is about 50 liters. I like how in the frequency response plot it is a smooth low end that could be extended very nicely with a port, shown in the WinSpeakerz plot. One thing that worries me about this driver is the large breakup mode that occurs at 4.5k but I feel with the correct crossover design at 1k this breakup will not be an issue.

The last driver that I seriously considered was the ScanSpeak 21W/8554, details from Madisound website.[14]

Kevlar cone

foam surround

SD motor

|Driver Specs | | |SBB4 | | |QB3 | |

|Vas = |160 | |VB = |33 | |VB = |25.65 |

|Fs = |23 | |fb = |23 | |fb = |40.66 |

|Qts = |0.22 | |f3 = |67.9 | |f3 = |52.44 |

[pic]

[pic]

With this driver I did not like how in the low frequencies it is very jagged from 20-100 Hz. I like how with this driver I would be able to get an f3 of 52 Hz with a box size that is about half of what I am working with now, 25 L. I also like how the breakup mode is smaller but at a lower frequency, which would force me to move my crossover, lower which would put more stress on my tweeter than I want to. I also really didn’t like that this driver is yellow. It’s pretty ugly actually.

After looking through all of the drivers I feel that the 8” Seas Excell will be the best match for my design and what I am shooting to attain through it. Based on the WynSpeakerz analysis I will be able to get 104 dB, with 40 watts of input power.

[pic]

Crossover.

When you think of crossovers and system DSP there is of course a big debate that occurs weather to use passive or active crossovers. Each type has its pro’s and con’s. In my design I gave very little weight to looking at active crossovers because of my application in home theatre. It would be possible to use a receiver with an external crossover by using the pre amp outputs, but due to cost constraints I will be building a passive crossover.

There are two main problems that plague designers when they start to look at crossovers. First, in the range where the frequency responses of the two drivers overlap there would be too much output from the system. Second, drivers will usually have different sensitivities, meaning one driver will play louder than the other.[15] Typical crossovers have slopes of 6, 12, 18, or 24 dB per octave. There are many different types of crossover types, but with any passive crossover you run into problems because your load is reactive instead of a simple resistor. For best results in your crossover, your drivers should be selected so that their frequencies combine to produce a flat frequency response, and that there is a good deal of overlap so you can have room to put your cross over in and still attain a flat frequency response. Below is a graph of my two drivers superimposed together to show the overlap region.

[pic]

Since I am using an 8” woofer, my conservative upper cross over limit is 1.23k and the highest recommended frequency is 2.46k. It is my design goals to be able to cross over my system fairly low. In order to get more power to the woofer I would like to cross it over low somewhere in the 800-1k range, this means that my tweeter crossover will be around 1k – 1.5k this keeps it out of the middle of vocal intelligibility and will result in the tweeter having the power and freedom it needs. Murphy has some guidelines that are useful to keep in mind when looking at component selection. Capacitors: Plastic film type, tolerance +/- 10% rated for 100 v or higher. Inductors: Air core type, +/- 5% tolerance, DC resistance less than .25 ohms. Resistors: Wire wound ceramic type, +/- 5% with power ratings from 5 to 60 watts.[16]

To get my crossover point placed in an optimum location first I looked at the tweeter. It has a natural roll off that begins at 700 Hz. Trying to crossover my tweeter this low would lead to problems and blown tweeters, which contrary to popular belief do not sound so good. I decided to start my crossover by putting a second order Bessel filter 12 dB / octave filter on my tweeter starting at 1.5k.

Components needed for tweeter crossover are taken from formulas below[17]

C1= .0912 / Rh f L1=.2756Rh / f

C1= 1.52E-5 L1=7.349E-4

To get a 24 dB roll off for the woofer, I chose a Linkwitz-Riley 4th order filter.[18]

L3 = 0.3000RL/f L4 = 0.1500RL/f

L3= 2.4E-3 L4=1.2E-3

C3 = 0.2533 /(RLf) C4 = 0.0563 /(RLf)

C3= 3.166E-5 C4=7.037E-6

Due to the differences in sensitivities of the drivers a 4dB pad was designed for the tweeter using the cart in Eargle’s Louspeaker Handbook[19]

R1= 1.5 R2= 6.8

Finally, a baffle step filter will be prepared to go at the input of the circuit. The essence of [Diffraction Loss] is this: at high frequencies the speaker is radiating into “half space” i.e. it is only radiating into the forward hemisphere. No significant energy is radiated to the rear of the speaker. At low frequencies the speaker is radiating into both the forward hemisphere and the rear hemisphere. That is, at low frequencies the speaker radiates into “full space” or “free space”. Because the “energy density” at low frequencies is reduced there is a loss of bass.”[20] This circuit as well as a zobel network will be added as needed in the testing and tuning phase. In the end this is what I calculate my crossover point to look like. Where blue is the woofer actual, red is the tweeter actual, yellow is the woofer crossover, green is the tweeter crossover, and purple is the summation of the two crossed over signals.

[pic]

Conclusion.

Good preparation and understanding will culminate with a result that you can be proud of. I feel that I have put together a solid design that will surpass expectations and the time I have put into this project in planning will shine through as we move into construction, testing, and tuning. I feel that my design satisfies my goal of creating the building block of a home theatre system. From my calculations I will have an f3 of 33 Hz and have a relatively flat frequency response up beyond 20 kHz on axis. I will be able to exceed my desired SPL capabilities with the 110-Watt receiver. Over all I feel that I have been making good design choices which will lead to a good final product that I can be proud of.

Speaker Recap:

SPL : meter at K – 14, where 83 dBC SPL is zero, 1 meter source level at 92 db, and I want to be at K – 14 standards, my speakers will need to be able to produce levels at 106 dB

Frequency Response: Flat, f3 33 hz

Use: Home theatre

Power Source: Onkyo TX-NR708

Size: 11” wide, 14”deep, 22” high

6 sided

Finished wood, 13 Baltic birch

Design: Vented box

Diffraction: Rounded edges, 2” 45 degree angled pieces off of front baffle

Size, SPL, Bass: My speakers are going to a hybrid between Hi-Fi and a 3-way system and be designed primarily for bass and SPL. As long as I can keep my box size fairly small I will be happy.

X-Over: lows crossed at 1k

tweeter crossover at 1.5k

Bibliography

DeBoer, Clint . " THX Certified Home Theater Program. 18 Jan. 2011. .

Dickason, Vance. The loudspeaker design cookbook . 7th ed. Peterborough, N.H.: Audio Amateur Press, 2006. Print.

Eargle, John. Loudspeaker Handbook . New York : Kluwer Academic, 2010. Print.

Katz, Bob . "Level Practices (Part 2) (Includes the K-System)." Digital Domain. N.p., 1 Sept. 2000. Web. 19 Jan. 2011. .

McCarthy, Bob. Sound systems design and optimization : modern techniques and tools for sound system design and alignment. 2nd ed. Amsterdam: Focal Press/Elsevier, 2010. Print.

Moulton, David. Total recording . Sherman Oaks, Calif.: KIQ Productions, 2000. Print.

Murphy, John L.. Introduction to loudspeaker design . Escondido: True Audio, 1998. Print.

Newell, Philip Richard, and K. R. Holland. Loudspeakers for music recording and reproduction. Amsterdam: Elsevier/Focal Press, 2007. Print.

Short, George . Cabinet Handbook. Old Forge: North Creek Music System, 1992. Print.

"THX Reference Level « ." . N.p., n.d. Web. 26 Jan. 2011. .

-----------------------

[1] Murphy, John L. Introduction to Loudspeaker Design. (2 ed. Andersonville, TN: True Audio, 1998.) p.11, 7

[2] Murphy, Pg 27

[3] Newell, Philip, and Keith Holland. Loudspeakers for Music Recording and Reproduction. (1 ed. Burlington, MA: Focal Press, 2007.) p.73

[4] North Creek Music Systems, The North Creek Music Systems Cabinet Handbook. 2 ed. Old Forge, NY: North Creek Music Systems, 1992. Pg.8

[5] North Creek, Pg 10

[6] Murphy, Pg 68

[7] Murphy, Pg 70

[8] THX Levels, DeBoer, Clint . " THX Certified Home Theater Program. 18 Jan. 2011.

[9] Murphy, Pg 57, 58

[10]

[11]

[12]

[13]

[14]

[15] Murphy, 97

[16] Murphy, 113

[17] Dickason, Vance. The Loudspeaker Design Cookbook. 7 ed. Peterborough, NH: Audio Amateur Press, 2006. p166

[18] Cookbook 166

[19] Eargle,John. Loudspeaker Handbook. 2 ed. Norwell, MA: Kluwer Academic Publishers, 2003. p. 112

[20] Murphy 68

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download