Need a little design help..

[Blazin]

Hey all, newb on this forum, but I know you guys may have some answers for me.

I am starting to look at ideas for a Senior EE project for this coming year, and I have really always been interested in speaker design. I'm pretty positive the whole project would be enough work to constitute a project; however, through all of my reading I'm still lost a tad due to cost limitations.

Project Plans:

Design:
- 3 or 4 way speaker system
- Powered with 110W RMS x 2 @ 8 Ohms
- Design/Build Custom Passive Crossover utilizing (most likely) 2nd-order Butterworth filters for a 12db/octave slope
- Doesn't need to look pretty, and I'm not a carpenter so It won't - haha
- I want it to not sound like crap - duh

Cost:
- I'm at a crossroads, if I take the school's money I'm limited to a budget of $200 I believe (barely enough to complete the project) and they keep my project at the end of the year. If I spend my own money, it's out of pocket but I get to keep a (hopefully) great sound set of floor speakers for my HT setup.
- Out of Pocket budget: ~$400


Questions/Suggestions:

- 3 or 4 way setup, I'll have 3 months basically to design the crossovers, speaker cabinets, and pick drivers.
- At one point in time, I had actually thought of creating some basic 2-way speakers with a high SPL: > 90db/w/m and then creating a basic 15w Tube Amplifier, but I'm not sure how much work would go into both of these projects.


One last tidbit: I have 1 semester to design, block diagram, plan, and basically order all of my parts, then I have my 2nd semester to build, troubleshoot, and get my project working. Each semester is approximately 15 weeks long, or a little over 3 months.

Coolness factor:

Tube amp powering custom built 2-way speakers 10/10



SOOO WHAT do you think?

/Ramble

[TacoD]

Budgetwise a 3-way is more realistic. Also a 3-way will give you enough troubles . Look at Dayton drivers and you will stay in budget and can get great results.

[Blazin]

I was looking at getting the following speakers, based on budget, power, and spl ratings.. but maybe I'm incorrect..

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=270-035&scqty=2

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=280-020&scqty=2

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=290-055&scqty=2

High efficieny drivers = less power required?

Realistically I'm never going to be powering these things with more than 110W RMS.. so I want whatever I build to sound great with that supply. By Great I mean crystal at all volume levels including loud as all hell

[Dennis H]

Does the school have a calibrated and/or pretty flat microphone you can borrow for measurements? That will be necessary for the best results and will also help you document what you've done for your report. Speaker Workshop is a free download and will produce some nice graphs showing how you measured the drivers, optimized the crossover for the target acoustical (drivers + filters) response, and measured the final result.

I suggest a Linkwitz-Riley acoustical target rather than Butterworth as the LR will sum flat. You may find it easier to go for LR4 rather than LR2 depending on the drivers' natural response curves -- sometimes a 2nd order electrical filter will give you a 4th order acoustical response. Also, if you want to go with tubes, you need to consider the amp's high output impedance. You'll need some Zobels to flatten the speaker's impedance curve so the sound won't change too much when you switch between tube and solid-state amps.

[joecarrow]

You're not on a bad track with those drivers based on your budget and goal, but you should keep in mind that your budget is quite small for a pair of 3-way speakers. The Xmax (a critical parameter for most bass speakers) for that woofer is only 2.9 mm. It won't stay crystal clear at high output- but then, neither will the midrange. The tweeter stands a fighting chance of hanging together at high volume, but it also won't start out as nice as the tweeters a lot of us use.

What amp are you using to power them? Can you quantify how "loud" you like it? I've used an SPL meter, and found that I often listen to music that averages a bit below 70 db, with peaks as high as 90 when I'm in a rock 'n roll mood. Overall, this really isn't that loud. To me, "Loud as Hell" is when it starts getting over 110 db- easily loud enough to cause hearing damage, and loud enough for everyone to know you're having a party. My present speakers (Modula MT with 100ish watts amplification) probably wouldn't make it past 105 db, really not much past 100 db before distortion starts to become an issue. To put it in perspective, I think that I'd have to spend at least $600 on commercial speakers to beat them.

Also, how much do you care about bass?

[augerpro]

I'll second what everyone else has suggested. You'll need a measurement setup, I'm sure you already have a stereo receiver you can use, a computer, suitable soundcard (M Audio Transit is a cheap usb option), but you'll need to get a mic and mic preamp, and download Speaker Workshop. Behringer ECM8000 and Xenyx 802 are popular but that is about $100 out of your budget. But a good investment I think.

My advice would be to build a pair of high quality 2 way bookshelf speakers. If you need bass you would be better off adding a subwoofer than trying to build 3 way your first time out. You have alot work ahead of you and believe me you will need those 2 semesters to finish this.

Having said that I would use the remaining $300 on some partsexpress small enclosures, like these: http://www.partsexpress.com/pe/showd...number=302-700

The crossover will probably be about $75. With the remaining money I'd consider these drivers: http://www.partsexpress.com/pe/pshow...number=264-850
for a woofer.

For the tweeter either of these: http://www.madisound.com/catalog/pro...oducts_id=1767
or

http://www.partsexpress.com/pe/pshowdetl.cfm?&Partnumber=275-075

or even the pressfit Dayton ND20.

Add a cheap powered sub if you have money left. Or not, either way I think you could have a very sweet little system (much better sound quality than the other drivers you listed are capable of) that should impress the classmates and professor, looks great, and you can show them the design process using the software. I'm thinking A+

[Blazin]

Also, how much do you care about bass?

Well theoretically upon completion I'll be replacing my current (HTIB) front pair of 2 way speakers.

I have a powered subwoofer already, so I will not be in need of a seperate subwoofer at all.

I've seen a lot of suggestions about speaker design software.. in my mind I'll be using some industry standard programs to design my crossover filters.. Matlab, Pspice/Electronics Workbench.. basically, don't worry about my filters those will be a lot of work; however, I have the proper tools to ensure the filtering is done properly.

Speaker enclosure build will require more investigation, maybe those programs supply enclosure dimensions based on specifications?


I really really really want a crystal clear set of front for my setup.

I have an Onkyo Reciever 110W RMS x 7 @ 8 ohms.. I have a powered sub @ 230W.. So yes, bass not important for these, just clarity.

I'll be building everything from scratch.. so Budget = cost of parts.

I don't understand the microphone testing, can I just use a Oscilliscope hooked to the terminals coming off the Xover?

I could do a 2-way speaker, but this project is supposed to be really in depth, and this will not be my first filter design project.

[Blazin]

Oh one last thing, the design time behind the tube amplifier and the related cost.. have made me decide against it, cool, but not exactly cheap.

[Blazin]

Ok looked at some of the Dayton Drivers:

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=275-075&scqty=2

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=295-308&scqty=2

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=295-485&scqty=2

Puts me to about $120 for drivers. I had previously looked at that tweeter before it was suggested, I like the sound of soft dome rather than the titanium tweeters.

Any suggestions to those would be welcomed.

[JonMarsh]

I've seen a lot of suggestions about speaker design software.. in my mind I'll be using some industry standard programs to design my crossover filters.. Matlab, Pspice/Electronics Workbench.. basically, don't worry about my filters those will be a lot of work; however, I have the proper tools to ensure the filtering is done properly.


I'll be building everything from scratch.. so Budget = cost of parts.

I don't understand the microphone testing, can I just use a Oscilliscope hooked to the terminals coming off the Xover?

I could do a 2-way speaker, but this project is supposed to be really in depth, and this will not be my first filter design project.

Since you're an EE student, we can recommend some deeper resources that will hopefully kick start you into getting this project off the ground and understanding the typical design/work flow that is used. I think that while you have some solid training background in analog electronics, from things you say you probably don't have a lot of past experience in speaker design/development, which is a very specialized type of analog design. Half of developing a filter for a crossover is knowing what target transfer function to develope, which means understanding the tradeoffs in different acoustical filter functions as well as the efforts needed to realize them, considering the combination of the driver response (which is not flat, not in a typical enclosure/baffle, even if the driver exhibits very ideal nearfield behavior). Are you familiar with concepts like baffle step response, directivity vs frequency for radiating diameters, etc?


One suggestion would be to go back to the AES anthologies on Speaker Design, but a quicker and more practical approach might be to get a few of the more useful books on this topic, and go through them in some detail.

Suggested reading:

"Testing Loudspeakers" by Joseph D'Appolito (available at Madisound) (covers a lot more than just testing)
"Loudspeaker Design CookBook" 7th Edition, by Vance Dickason, Madisound or PE

For free, Jay Butterman's manual for Speaker Workshop (a freeware program for speaker measurements and design) will give you a good overview of the whole process, including embedded software utilities to help with various aspects of speaker design. It can be downloaded here.

Speaker Workshop is available here.

I use LspCAD for crossover filter and system modeling development; if you can develop good electrical models of the drivers and their measured response and impedance, then doing this in MatLab, MathCAD, or a SPICE like simulator is feasible, but probably more work. I use a Praxis system for measurements at home, and a Fuzzmeasure setup on my Mac with a Firewire connected preamp for field work, both with an ACO Pacific mic; much lower cost measurement solutions are available, using electret low cost capsules, or low cost microphones like the Behringer ECM8000.

You won't be able to do a good job of designing and measuring loudspeakers just by using measurements with an oscilloscope at the driver terminals- these references will explain the process, including the importance of driver measurements in your actual cabinet design, developing a target transfer function for the filter including the effects of the driver response mounted in the enclosure, and of course, the pro's and con's of different acoustic target functions.

Of course, you could read some of the design threads here and get some overview too, but it's too large a topic to do a "tutorial" in one post.

Conventional filter design software and simulation programs like SPICE are better than no tools, but are not optimized at all for the typical speaker design flow. Given your budget limitations, downloading Speaker Workshop would be a good idea, as regards making measurements and developing your network design based on that. Separately, of course, you may want to pursue a more detailed mathematical approach using MathCAD or similar tools, but as a practicing engineer of 35+ years, only occasionally for specialized tasks do I use tools like MathCAD- read the "Three Way Design Study" thread for an example.

If you're set on doing a three way design, review that thread and consider Steen Duelunds approach to developing as three way acoustical transfer function. There's some good math behind it, and perhaps you'll find it interesting.

There are many details to developing a good speaker design, and figuring out how to do a good two way is generally a sufficient AND NECESSARY prerequisite to doing a good three way. Note, I said "good". Many of my two way's demolish commercial three way systems within their power handling limits- the M8ta is a good example.

[Blazin]

Well with all that said, in terms that I actually understand (wow).. I will pick up those books, I will propose the idea when it comes around to design a 2-way speaker, MTM possibly.

I was thinking about using Speaker workshop with a homegrown microphone, using some small mic i can get off Digikey and then pre-amping it with a low power Op-Amp for said purpose. $10 speaker testing is better than nothign


I'll take a look at some of the 2 ways on this forum and I will most likely end up designing a 2 way system, as I have found that there are very few DIY 3-ways found online for example, in my mind because they are seemingly an added step of confusion that is unnecessary when proper methods are use in design a good 2-way.

To answer your previous question.. I have no experience in speaker design at all, I have done analog and digital circuitry; however, baffle step response, directivity vs frequency etc are not in my toolbox yet.

Thank you for a more precise answer.



In terms of budget, any suggestions on percentages I should look to spend on the areas?

IE:

70% - drivers
20% - enclosure
10% - crossover components

Thanks!

[cjd]

drivers/crossover average 50/50 split for me. Depends on the specific price of all components, so it's really tough to give you a fair estimate. But I can say 70/10 is almost surely not going to be accurate. Well, unless you're in the camp that likes their distortion and selects drivers based on how pleasing the distortion is.

Box - no way to estimate. I mean, you could use $10 in drivers and crossover components but choose to veneer in burl...

C

[Blazin]

Oh yea distortion - vomit.


I guess when it gets closer to project time I may have a better understanding for my needs, but for the time being.. any suggestions on drivers that will be within my budget?



Dayton sure, but which - hah.

4 - mids
2 - tweets

$150-200

[cjd]

Well, just as an example the MTM's I've done based on the Dayton RS150S-8 runs ~$300 for everything (including crossover) - though the exact price depends on what tweeter and what crossover you pick. But the RS drivers want at least a 4th order slope IMHO.

If you want a softer cone, Silver Flute makes some pretty darned nice sounding budget drivers. I think they're only slightly more forgiving than the RS drivers, but a little can go a long way.

Um. Lots of drivers to choose from out there, but...

C

[Blazin]

Well, just as an example the MTM's I've done based on the Dayton RS150S-8 runs ~$300 for everything (including crossover) - though the exact price depends on what tweeter and what crossover you pick. But the RS drivers want at least a 4th order slope IMHO.

If you want a softer cone, Silver Flute makes some pretty darned nice sounding budget drivers. I think they're only slightly more forgiving than the RS drivers, but a little can go a long way.

Um. Lots of drivers to choose from out there, but...

C

I'm assuming you mean crossover components..

[cjd]

Hmmm? No, I mean crossover... I have about, oh, dunno. eight or ten variations.

C

[Blazin]

Oh, ok.. I'm building my crossover from scratch.. aka designing and such.. for the project

[Dennis H]

I just want to reinforce what Jon (he's a professional EE, BTW) said that using Spice or something like that to design the crossover without considering the acoustical response is a recipe for failure. The simple reason is that drivers don't have flat frequency response and they don't have a flat impedance curve. An "8 ohm" driver may vary between 5 and 30 ohms as the frequency changes. An otherwise flat driver won't be flat anymore when you put it in a box due to baffle effects -- low frequencies radiate in all directions but high frequencies radiate mostly forward.

When you set a target, like 4th order Linkwitz-Riley or 2nd order Butterworth, that's an acoustical target, not an electrical one. The electrical transfer function is whatever it takes to get the desired acoustical response.

Take a look at some the threads in the Mission Accomplished section or follow the 'Speaker Camp' thread in this section to get an idea of what's involved. Here's Jon's idea of a 'simple' 3-way. When you look at some of those big component values, it's easy to see how the crossover can cost more than the drivers.

[Dennis H]

It's easy to get bogged down in the minutia of the design process but I think the overview from the SW manual is a pretty good big picture.

STEPS TO DESIGN AND BUILD A SPEAKER OVERVIEW
(DIRECTLY FROM THE SPEAKER WORKSHOP SITE)

Designing a speaker can be broken down into a few steps.

Select Speaker Type
The most common types of speakers are 2-way (woofer and tweeter), 3-way (woofer, midrange and tweeter) and subwoofer (just a big woofer in a box).

Select Drivers
Based on the type of speaker, pick drivers to attain a flat amplitude vs. frequency response as well as a good impulse response. You can prototype low frequency response (woofer and midrange) using the Enclosure commands before buying these drivers.

Buy the Drivers
Purchase the drivers and wait for them to arrive. Create Driver resources to hold the measurement data.

Measure the Drivers
For the woofer and midrange measure impedance in free air and in a prototype box. Calculate T/S parameters for the woofer and midrange. Match drivers as closely as possible for stereo pairs to improve imaging.

Design the enclosure
Use the T/S parameters to design a final enclosure for the drivers. Build one.

Measure the Drivers
Mount all of the drivers in the final enclosure and measure their impedance and frequency response.

Design the Crossover
Using the measurements from the real enclosure use the Network commands to prototype a crossover.

Test the Loudspeaker
Use the Signal commands to verify operation of the crossover with a load resistor, then attach the crossover to the drivers and test the frequency response of the system, splicing near field and far field measurements.

Test the System
There are two things you can do here: either evaluate in room measurements from a listening position (custom for the room) or evaluate distortion and time and phase domain measurements in a pseudo anechoic environment to assess for general accuracy of the speaker.

[joecarrow]

Definitely don't get bogged down by thinking that your first set of speakers are going to blow you away. This is a school project, and if you're careful you'll probably make something that's quite easy to listen to, and better than the truly cheap stuff- but you're not going to compete with the top shelf stuff on a tight budget and your first try.

Don't get paralyzed by analysis, don't get your hopes up *too* high, and keep in mind that an electrical engineer makes enough to build some really fun speakers after graduation. I know I spent about a year fretting over driver choices and design ideas when I was working at my first startup and couldn't really afford to build anything. A couple of raises later, I built Jon's Modula MT. Now that I'm finally working at market rate (mechanical engineer) I'm getting ready to design a set of speakers to really compete.