Just for fun. [Archive] - Audio & Video Forums

PDA

View Full Version : Just for fun.



GMichael
09-14-2007, 11:43 AM
If I were building a pair of speakers, what do you think of these components? Oh sure, I'd still need some MDF, hardware, glue etc....


260-154 Dayton XO3W-625/5K 3-Way Crossover 625/ ..
260-304 Gold Plated Bi-Amp Terminal ..
260-330 Acousta-Stuf Polyfill 5 Lb. Bag ..
268-010 Sound Dampening Sheet 10" x 10" ..
277-020 Morel MDT-33 1-1/8" Matched Soft Dome T ..
287-020 Morel MDM-55 2-1/8" Soft Dome Midrange ..
340-255 3M Super 77 Spray Adhesive 24 fl. oz./1 ..
240-715 Dayton DSS4-BC Black Chrome Speaker Spi ..
260-525 Sonic Barrier 1" Acoustic Foam w/PSA 18 ..
300-805 Dayton SA240-B 240W Subwoofer Amplifier ..
297-022 Morel MW-144 5" Woofer ..
295-456 Dayton RSS210HF-4 8" Reference HF Subwo ..

Subtotal: $1,290.08

herm0016
09-14-2007, 09:57 PM
floor standers? i would go with a bigger woofer, i don't think you will gain that much from the 5 to the 8 with a 4 way design, and being different brands and in roughly the same freq. area may give you some weirdness. for that kind of money i think you may be better off going with a non-diy speaker, though i have never heard those particular drivers before.

basite
09-15-2007, 01:17 AM
try some ribbon tweeters and a bigger midrange speaker, maybe ribbon tweeter& mid in one, and then bigger woofers and no sub.

Feanor
09-15-2007, 04:58 AM
If I were building a pair of speakers, what do you think of these components? Oh sure, I'd still need some MDF, hardware, glue etc....

....

Fun for sure. I've whiled away many a happy hour working on speaker designs using my trusty X-over Pro and Box-pro programs.

Of all the designs I've worked up, I like this one, a two-way using Accuton drivers. On the pricey side: I'd say north of $2K with high quality crossover components ...

http://ca.geocities.com/w_d_bailey/Accuton_design.jpg

GMichael
09-15-2007, 06:25 AM
Very nice Feanor. Thanks for sharing.
Thanks to basite and herm0016 as well.

It seems like the more I learn, the more I find out that I have more to learn.

Where's Kex? This is his favorite kind of stuff.

Fred333
09-20-2007, 09:59 AM
Very nice diagrams. You must really enjoy your setup.

Feanor
09-20-2007, 03:29 PM
Very nice diagrams. You must really enjoy your setup.

Fred, I guess you're talking about my diagrams? They were created by Harris Tech's X-over Pro and Box Pro programs. They also do very nice amplitude, phase, impedence, etc. diagrams which I didn't show here.

Let me point out that I haven't actually made the speakers whose design I illustrated: just something I'd like to do some day.

kexodusc
09-20-2007, 04:28 PM
If I were building a pair of speakers, what do you think of these components? Oh sure, I'd still need some MDF, hardware, glue etc....


260-154 Dayton XO3W-625/5K 3-Way Crossover 625/ ..
260-304 Gold Plated Bi-Amp Terminal ..
260-330 Acousta-Stuf Polyfill 5 Lb. Bag ..
268-010 Sound Dampening Sheet 10" x 10" ..
277-020 Morel MDT-33 1-1/8" Matched Soft Dome T ..
287-020 Morel MDM-55 2-1/8" Soft Dome Midrange ..
340-255 3M Super 77 Spray Adhesive 24 fl. oz./1 ..
240-715 Dayton DSS4-BC Black Chrome Speaker Spi ..
260-525 Sonic Barrier 1" Acoustic Foam w/PSA 18 ..
300-805 Dayton SA240-B 240W Subwoofer Amplifier ..
297-022 Morel MW-144 5" Woofer ..
295-456 Dayton RSS210HF-4 8" Reference HF Subwo ..

Subtotal: $1,290.08

GM - That's an interesting assortment of parts you've put together there, I'm sure it would be a fun project that provides a great learning experience. With a bit of tweaking, trial and error, I have no doubt it would result in a great sounding system.

For my part, I think I would change several items on the list - especially since you're looking at that kind of money.
First, since you're going to spend $100 on a midrange driver, $100 on a woofer, you probably don't want to throw in a generic, off-the-shelf crossover. That's crazy dude. You've got some nice drivers there, but they're not going to perform as well as lesser drivers integrated with an optimized crossover. Just trust me - don't do it. Check out the driver sensitivities, there's no level matching guaranteed between the woofer, mid, or tweeter, especially with driver sensitivity coming in at 88, 90.5 and 92.5 respectively...could be a bit on the bright side. You could add your own L-pads to try and even things out, but I think you're also going to run into some impedance and phase issues in the crossover regions. Not even sure 2nd order crossover is optimal here for those units.

The Morels are nice units, but are they a good match to each other in a 3 way system?
It's been said many times, while not universally true I think it applies here - a considerable portion of the cost of a tweeter goes into creating the ability to crossover and play at lower frequencies. Most tweeters don't do so well below 2KHz. This one can probably get by at 1500 Hz, 4th order. That's pretty impressive. But you're crossing it over at 5000 Hz with that XO, which almost defeats the purpose. Or you're using a midrange driver for not many frequencies which begs the question, why use it at all? I'd downgrade the tweeter big time - just because it costs more doesn't mean it's going to sound better in that top octave or two. It probably is just more versatile than a lesser tweeter with a higher xo requirement. I bet there's some considerably cheaper tweeters that'll stack up from 2000 Hz-20,000 Hz as nicely or even better. And you really don't need to crossover low with that midrange.

I like your choice of subwoofer and amp if you're planning on making powered-towers. You definitely don't need anything bigger than a 5" woofer going that route - the subs in stereo mode will undoubtedly produce better bass anyway.

Suggestion, make them separate cabinets, but shaped so the speakers could rest on the subs and look like they're one piece - later if you decide you get better performance moving the subs around a bit, you have that option, all you need are stands.

I'm not a big fan of Morels, not because they're not nice drivers, they just seem to be overpriced compared to better offerings from Seas, Usher, and Scan-speak.

I'd probably do away with midrange unit altogether, and try two of the 5" woofers in a 2-way design, crossed over around 1800 Hz or so - multiple drivers lowers the distortion, increases the sensitivity, and lends itself to few different design possibilities that can improve sound without adding too much complexity. Doubling up woofers is sometimes a more affordable way to achieve a level of sound compared to buying a single super woofer. Maybe throw in a nice ribbon tweeter instead if you like the ribbon sound?

You've got the Acousta-stuff, sonic barrier foam, and sheet. I'd probably stick with foam, acousta-stuff, or go to pink insulation or rigid fiberglass. You don't want to use too much damping material, and you really don't need 3 types.

The cabinet is easy enough to figure out, spend most of your time on the crossover. That will yield the biggest contribution to how good the system sounds. Great drivers and a crappy crossover really doesn't sound any better than poor components.

I'm not a big fan of high priced crossover components, but I am a fan of quality components. I'd probably use higher grade caps and inductors with those drivers - 14 gauge inductors at least, and better caps - you're spending $800 on drivers, you probably want your xo components to be of a fairly tight tolerance range from spec, low DC resistance, and good power handling. I recently built a pair of small, computer desktop 2-way bookshelf types, $70 in drivers, $45 in xo parts. If I had $800 in drivers, I'd probably double or triple that budget to start.

But there's no right or wrong way, I'm sure some of the guys here could whip something up faster and better than I could do with more imagination. I like keeping things simple to limit my mistakes. I make enough as is :mad2: Whatever you do, lose the stock crossover - that's a must.

When do we start?

Luvin Da Blues
09-20-2007, 05:16 PM
Awesome post Kex!!

Feanor
09-20-2007, 06:45 PM
Super advice from Kex -- who's had some experience no doubt.

There is no need for a 2-way design, especially given the Morel drivers you've chosen. In particular the MDT-33 has a very high power handling capacity, (200 watts RMS :shocked: ) and a smooth, flat response right down to 1500Hz it is feasible to use it that low provide you use a higher order crossover.

I'd be more inclined to go with a 6" midbass despite Kex's advice, and I don't think Morel is over-priced. I'd probably choose the MW-164. It does share on one useful feature with the MW-144 which is a high Qt specification: this makes it relatively suitable for a closed box design. Closed box makes for tighter bass and is much less critical to design than a vented box.

That brings me to the cross over. For a MDT-33/MW-164 2-way, I'd suggest a 4th order cross over, (such as I illustrated in my earlier post). This makes a low tweeter crossover point, say 1500-1700Hz, quite feasible with the MDT-33, given the resonance frequency is 700; overall, it ensures that both drivers work in their most effective frequency range. Those pre-made Dayton crossovers are worthless for audiophile purposes employing, as they do, very cheap parts.

Stay tuned: in the next day or two I'll post a cross over designs for the MDT-133 and MW-164 combination.

royphil345
09-20-2007, 06:59 PM
Yeah... I agree... With a 5" woofer and a 1 1/8" tweeter... there should be no need for a midrange at all. Keeping it 2-way would simplify the crossover tuning and very likely give smoother results, especially for a beginner. Although, I think even most pros would stick with a 2-way design when using those particular drivers.

The "off the shelf" crossovers will provide a simple crossover at a given frequency. Adding resistors to equalize the output of the drivers is just about always necessary to build a "flat" sounding speaker with a good soundstage. Adding resistance also changes the crossover point of a capacitor or inductor. There are charts on the web to show what frequency different capacitors / inductors will cross over at in circuits with different resistance loads. There are also methods of altering the crossover slope, timing, etc... (like in Feanor's design) that I don't understand much at all yet. You could experiment with an L-pad... replace it with a resistor when you find out how much resistance you need to add (usually to the tweeter)... or just make the L-pad permanent... You could change your crossover capacitor to mesh perfectly with your inductor at the desired crossover point with the added resistance... or not...

Even some of the more modest store-bought speakers are using fairly complicated crossover networks these days. Most acting as more than a simple crossover and even tailoring the frequency response of the individual drivers to flatten response or widen frequency range of smaller speakers. It's hard to build something better. Although, with the very high-quality drivers you're looking at... and if you keep it simple... I'm betting the results would be very good. Sometimes simple just sounds more natural.

I know I've seen plans around for reverse-engineering some high-end speakers that use off the shelf drivers. That might be a way to go for a first project.

Feanor
09-20-2007, 07:03 PM
Super advice from Kex -- who's had some experience no doubt.

There is no need for a 2-way design, especially given the Morel drivers you've chosen. ...

Stay tuned: in the next day or two I'll post a cross over designs for the MDT-133 and MW-164 combination.

Sooner than expected :D ...

http://ca.geocities.com/w_d_bailey/MorelMTD33MW164.JPG

Feanor
09-21-2007, 03:12 AM
By the way, the above design should work for the actual crossover, however as a real-world design it lacks something, and that is "baffle step compensations".

Simply put, when the width of the cabinet becomes less than half the wave length of the sound wave being produced, amplitude of the sound produced is reduced significantly, (6dB I think). Thus this effect kicks in with a 10" baffle at about 700Hz.

This effect is compensated for using an inductor in parallel with a resistor which acts as al low-pass filter that equalizes the higher frequencies with the lower. Assuming a 10' baffle, the filter for the MW-164 would consist of a 1.6mH coil in parallel with a 6.4ohm resistor, given the specs provided. But note that if the speaker is very close to the wall or in a corner, these values would differ.

kexodusc
09-21-2007, 04:26 AM
Nice work! Where'd you find a FR/impedance plot on the Morels, or did you just have them handy? Did you subtract out the test baffle and include the response artifacts from yours? How does the bsc circuit change things?

I dont' have any data on those drivers so I can't do anything meaningful from here. :(
If you have an FRD or zma file...send'em my way...

Feanor
09-21-2007, 05:35 AM
Nice work! Where'd you find a FR/impedance plot on the Morels, or did you just have them handy? Did you subtract out the test baffle and include the response artifacts from yours? How does the bsc circuit change things?

I dont' have any data on those drivers so I can't do anything meaningful from here. :(
If you have an FRD or zma file...send'em my way...

Kex, no, I don't have those data files. My X-over Pro program does happen to have them stored internally for the MDT-33 and MW-164, but I haven't figured out how to extract them from their database. One can manually input aplitude response info into X-over Pro manually from, for example, graphs: I've done this but it is laborious of course.

In the X-over Pro output I showed, I didn't include the BSC circuit or compensate for its values. The inability to put a BSC circuit into X-over Pro is one of its unfortunate limitations. However the effects of a BSC are quite predictable on the ampitude response if not necessarily the phase and impedance behavior.

kexodusc
09-21-2007, 06:29 AM
In the X-over Pro output I showed, I didn't include the BSC circuit or compensate for its values. The inability to put a BSC circuit into X-over Pro is one of its unfortunate limitations. However the effects of a BSC are quite predictable on the ampitude response if not necessarily the phase and impedance behavior.
I've had bad luck with BSC predictions being accurate. Most drivers, even expensive ones, deviate some from spec, with these Morels I'm not too worried about that. But depending how close to the surrounding walls you place your speakers you might not need the full 6 dB. I see 3 dB as an average tossed around a lot. I usually end up somewhere in between, and it really is a an interative process.

You'll lose some sensitivity in the midrange as a result, up to around 1.3 kHz and will probably have to add some padding on the tweeter to compensate a bit. The good news for both the BSC circuit and L-pad, is resistors are cheap. Nothing a bit of trial and error can't solve.

As a suggestion, you might try a neat trick I was taught - rather than the textbook 4th order electrical filter on the woofer witha zobel and BSC circuitry, try a single, oversized coil on the woofer and maybe a shunt cap if needed with a zobel. The zobel values may change. I've used this successfully in 3 designs now. The final transfer function is the key here. Not only does the sound benefit from reduced crossover parts count, but it saves some money and cuts back on complexity as well. Off the top of my head, I'd start at 1.8 mH and go from there.. Often you can get by without the BSC circuit because of the slope the coil provides. Could save you 3-4 parts, not insubstantial. That Morel looks like it could be easy to work with.

Passive Crossover Designer is an absolutely amazing free xo tool that allows you to manipulate the element values and watch what happens to the slope.

GMichael
09-21-2007, 06:39 AM
Great information guys.
How about this? Two sealed mains using a 2-way design (6" mid-woofers), sitting on top of a pair of 10" subs (I like sealed subs but have had good luck with some ported designs as well). Then bookshelves, using the same sealed 2-way design for surrounds. How about the center? Maybe a 2-way using 2 - 5" mid-woofers and the matching tweeter?

Feanor
09-21-2007, 06:54 AM
I've had bad luck with BSC predictions being accurate. Most drivers, even expensive ones, deviate some from spec, with these Morels I'm not too worried about that. But depending how close to the surrounding walls you place your speakers you might not need the full 6 dB. I see 3 dB as an average tossed around a lot. I usually end up somewhere in between, and it really is a an interative process.

You'll lose some sensitivity in the midrange as a result, up to around 1.3 kHz and will probably have to add some padding on the tweeter to compensate a bit. The good news for both the BSC circuit and L-pad, is resistors are cheap. Nothing a bit of trial and error can't solve.

As a suggestion, you might try a neat trick I was taught - rather than the textbook 4th order electrical filter on the woofer witha zobel and BSC circuitry, try a single, oversized coil on the woofer and maybe a shunt cap if needed with a zobel. The zobel values may change. I've used this successfully in 3 designs now. The final transfer function is the key here. Not only does the sound benefit from reduced crossover parts count, but it saves some money and cuts back on complexity as well. Off the top of my head, I'd start at 1.8 mH and go from there.. Often you can get by without the BSC circuit because of the slope the coil provides. Could save you 3-4 parts, not insubstantial. That Morel looks like it could be easy to work with.

Passive Crossover Designer is an absolutely amazing free xo tool that allows you to manipulate the element values and watch what happens to the slope.

Tell me more about Passive Crossover Designer -- I don't think I've hear of it before. It would be nice to reduce the number of components. I would really like a tool whereby you could enter any topology, the speaker data, and maybe rough component values, then the program would seek the optimum solution. I had an old version of CALSOD that did that pretty much, but it was so painfully difficult to use that I gave up on it.

Of course, if you put the BSC on the woofer only, it will attenuate the mid-range but lnot the tweeter which will need more L-pad. But it is an option, if less ideal, to put the BSC ahead of the both the woofer and the tweeter.

Feanor
09-21-2007, 07:01 AM
Great information guys.
How about this? Two sealed mains using a 2-way design (6" mid-woofers), sitting on top of a pair of 10" subs (I like sealed subs but have had good luck with some ported designs as well). Then bookshelves, using the same sealed 2-way design for surrounds. How about the center? Maybe a 2-way using 2 - 5" mid-woofers and the matching tweeter?

Actually, should work really well. There is no real need to put the two midbasses in the center channel, this is more a matter of style than anything else. I built a center with just a woofer and one tweeter side by side -- check out my Gallery. It works very well.

kexodusc
09-21-2007, 07:29 AM
Tell me more about Passive Crossover Designer -- I don't think I've hear of it before. It would be nice to reduce the number of components. I would really like a tool whereby you could enter any topology, the speaker data, and maybe rough component values, then the program would seek the optimum solution. I had an old version of CALSOD that did that pretty much, but it was so painfully difficult to use that I gave up on it.

Of course, if you put the BSC on the woofer only, it will attenuate the mid-range but lnot the tweeter which will need more L-pad. But it is an option, if less ideal, to put the BSC ahead of the both the woofer and the tweeter.
PCD is a glorified Excel spreadsheet with some serious macros that a few DIY guru's have created and continued to improve over the years.
Don't be fooled. It is a very powerful program - right up there with Xover pro, speaker workshop, etc. I actually like it better than them for modelling xo's.
It's downside is the learning curve, but there's some good guides out there. You can use little sliders to tweak component values and watch a graphical display of the effect on the system as you do it.
It's limited, as well, but it beats paying $500 or whatever for Calsod for a lot of DIY-ers.
You need to either have your own measrument jigs to get frd and zma data, or you can use manufacturer data if provided. Or you can trace plots using another program, but that process is time consuming. Still, it can be effective if the plots are reliable.

Search for the frd consortium on google for the link. There's also a great compliment program that includes all sorts of xo calculators.