HiFi build using RAAL, Scanspeak, & Morel drivers!

If you purchase all the parts from Madisound, they can design a crossover for you. It won't be 100% as they don't have your cabinets in hand, but it will be a step in the right direction. From there you can pick up an Omnimic and DATS, and learn how to fine tune the design.

As for your expectation of the small RAAL, I wouldn't cross it anywhere below 3K. 4-5K may be even better, which would necessitate a smaller mid such as the SS 10F or any 12cm model. As for that woofer, I'm not familiar with it, or any other Morel driver. What are your size or budget restraints?

Face2 is right on both points.

The RAAL spec sheet suggests a 4th order crossover no lower than 2800 Hz. Higher would be better.

Before you can design crossovers, you need to have a cabinet. Mounting a driver on a front baffle of less than infinite width will significantly change its response compared to the manufacturers spec sheet. Compare the spec sheet of an Aurum Cantus G2 ribbon tweeter to the first figure below. Diffraction interference from the cabinet edges creates that prominent dip. You have to be able to see that before you can make an appropriate crossover filter.





Similar things happen for the midrange.

Why do you want two midrange drivers? At 4 ohms, using two will present problems. Also, that driver has a very uneven frequency response above 1 kHz. There will be problems getting that to go high enough, and smooth enough, to cross that ribbon tweet. Consider a different midrange.

That woofer, if its similar to other Morel woofers, probably works best in a sealed cabinet. It will probably not be nearly as sensitive as the other drivers, and you may need two woofers. Again, consider other 8" woofers, such as Seas, Peerless, or SB, all in a similar price range.

Crossing a 8" woofer to a mid at 200 Hz seems unnecessarily low. With a passive crossover, 200 Hz is unthinkably low. Even with an active system, why that low?

It seems like you've never built or designed a speaker before. That's OK, everyone has to begin somewhere.

Don't throw money away on expensive drivers if you are just starting.

Stick to 2-ways, they're much simpler.
Edited by R Swerdlow - 2/20/13 at 4:28pm

Marty-

I'm sure you have seen my posts on the active DSP thread. When I talk about people trying to use active DSP as a shortcut to designing a speaker I am referring to situations like this.

Picking expensive drivers off of a website and pushing some buttons on a DSP will not make for a high quality speaker. In fact, as nice as these individual drivers are, they would sound terrible together.

Designing speakers is a skill you have to develop. Using active DSP does eliminate the need to understand electrical circuits but that is probably the most rudimentary skill in speaker design.

Using active DSP is kind of like using a graphing calculator vs solving on paper. Either way you still need to understand the principles even if the fancy tool is doing some of the dirty work. A graphing calculator doesn't eliminate the need to understand math and an active DSP doesn't eliminate the need to understand speaker design.

If you want to learn how to design a speaker start small. Stick to a dirt cheap 2-way. The high end drivers you chose are notoriously difficult to work with.

In fact, the best way to learn is to build a 2-way passive kit. You can take your own measurements and design your own active DSP to replace the passive network. This gives you a benchmark to listen against. Once you get a handle on that you can graduate to more complicated designs.

You can alsonread sites like Zaph Audio and HTGuide's DIY section. They focus on passive.designs but the principles are all the same. Just click on one of Zach's builds and it will walk you through his whole process. I would also suggest googling "Loudspeaker Design Cookbook". It is a good starting guide. You might.find a free download but I wouldn't know anything about that.

I triple agree with the guys here. I don't even like to use my 140-15 much lower than 3KHz, so the 70-10 ought to be kept above that point and ideally, over 4KHz for it to "play nice". It's a tough tweeter to use right because of that limitation. The OEM fairs much better at 3KHz, but unless you get lucky, they are not for sale to DIY.

If you still have your heart set on the 70-10, maybe an MTM arrangement is not best? Rules are meant to be broken, but.....those Morel are not that great up high, so now you're trying to blend two drivers that are right at absolute edges of their working ranges. That's a tough one to get right. I'd say choice smaller mids or step up to the 140-15 Raal and use one mid.

And remember, just picking some great drivers is not a guaranteed recipe for success. The Raal and Scan Rev15W are still some of the best out there, but you gotta use them right. Do you have some measurement gear to use? If not, I suggest getting some and playing around with it for a while and reading up more. Based on your questions, it seems as if you may be on the beginning stages of a learning curve. These drivers are quite a financial commitment, so I'd hate to see you be disappointed with the results. And as you pointed out, most of the guys on here are a little more into the HT side of things, so you might want to spend some time at the HTGuide Mission Possible forum and over at DIYaudio both of which has oodles of resources in old threads for hi-fi DIY systems.

Greg

No problem Marty. You can do this (as can anybody IMO) so long as you don't take shortcuts. The Loudspeaker Cookbook will help you with all of your questions. It doesn't contain every possible answer but it will get you pointed in the right direction on crossover design, box design and measuring.

As far as choosing drivers, there are a number of factors. For starters, things are easier with direct radiators as opposed to horns. You have to find drivers that can comfortably play a certain range while maintaining certain directivity criteria. How you judge "comfortably play" is debatable. Directivity is primarily a function of driver size with direct radiators. For instance, if you had a 10" woofer, it will collapse from omnidirectional to a 90deg pattern from about 1khz to 1.8khz. Above that it will start to "beam" which means the directivity gets very narrow. A 1" tweeter is fairly omnidirectional up to about 12khz IIRC. If the tweeter can only play down to 3khz comfortably that means the 10" woofer will need to play up that high. The problem is that the woofer will be beaming severely at this point.

Like the guys stated above, sometimes it is as simple as two drivers that can't comfortably play with enough overlap to cross them. I've not use the RAAL but if it only plays down to 3khz and the mid you want to use only plays up to 2khz it won't work even if directivity isn't an issue.

Then there is the issue of factory measurements. These only tell so much and their reliability varies by manufacturer. There is no substitute for your own or trusted 3rd party measurements. A site like Zaph Audio can be helpful since he has a large database of drivers.

This is just the tip of the iceberg. Speaker design can become extremely complex as you get into more advanced speaker types and attempting to reach greater levels of refinement. I would suggest sticking to a simple cheap 2-way dome with 5-6" mid to start. SEOS designs are also not too tough with an active DSP. Like I suggested earlier I would buy an existing passive kit and then try your own tweaking with the DSP.

As an alternative to the Loudspeaker Design Cookbook, I like Speaker Building 201 by Ray Alden. It's written more clearly and has a different organization that I preferred. The first 3 or 4 chapters of SB 201 provide (in my opinion) a clear explanation of how to go about designing a proper enclosure based on a woofer's Thiele/Small parameters. It requires a familiarity with high school algebra/trig. Yes, there are web sites and spreadsheets that do all the math for you, but to avoid making mistakes, you must also understand the process. You mentioned that you know how to do this for subwoofers. The process is essentially the same for any woofer.

Both books provide good information and overlap somewhat, but the LDC provides more info about making speaker measurements.

Coctostan already spoke about this when he described a speaker's transition from radiating sound omnidirectionally to beaming sound in a narrow pattern. Imagine a 2-way speaker, and start with the woofer. Beaming varies with the woofer's cone diameter and frequency. If you look at the manufacturer's measurements (I linked the chart for a Seas ER18RNX 6½" woofer from Madisound), you see three measured traces.



The dark black trace is measured with the microphone on axis in front of the woofer (0° off axis). The other two are measured 30° and 45° off axis. See how the woofer's off axis response gradually drops as the frequency increases? Above roughly 2500 Hz this speaker has lost enough off axis response to beam. So, that's a good frequency to crossover to a tweeter (in a 2-way design). I wouldn't go any higher, but there might be benefit to going as low as 2000 Hz.

It's important to design a crossover that suppresses any sound from this driver at frequencies beginning below 4000 Hz. That is probably what's known as break up, but a frequency response curve won't tell you whether those peaks are a louder sound or if its undesirable noise. Just as a first guess, I'd use a 4th order crossover to avoid that break up noise. It's possible that additional filtering with an LCR filter might also be needed.

With that info, you can then look for a tweeter capable of performing well, without significant distortion, as low as roughly 2500 Hz. A general rule of thumb is to look for a tweeter with a resonance frequency (Fs) three times lower than the intended crossover point.

For a brief but interesting explanation of some of the thought processes involved in designing a crossover, read this link, Crossovers 101 by Dennis Murphy
Edited by R Swerdlow - 2/21/13 at 11:40am

Good post Mr. Swerdlow. That is another good book. These books are just a way to get started. Neither is comprehensive. Reading design discussions and white papers written by designers on the web is how I learned...along with doing it.

Speaker design is not exceptionally difficult, but it does require a certain level of knowledge and experience. I don't know of any way around that.

Thanks. Call me Richard . I rarely post here on AVS, and usually hang out at Audioholics.

I probably learned more from talking to people at DIY speaker builders' meetings that occasionally happened in the DC/Maryland/Virginia area where I live. I met a few very experienced and friendly designers who were glad to show and explain things to new guys.

When I read Marty's post, I recognized enough of myself to want to respond.

If you want something that really sounds good and HiFi, you should pick a peer-reviewed and well respected design.

I've heard really good things about this kit:
http://meniscusaudio.com/triton-pair-p-1346.html

Uses really good parts and I like Jeff Bagby's stuff. It probably would benefit from a subwoofer.

Mini Statements:
http://meniscusaudio.com/mini-statements-pair-p-1342.html

Another highly regarded design, uses a ribbon like you were looking at. No subwoofer needed with 2x 7" dayton reference woofers per tower.

And lastly:
The finalists. A really exceptional midrange driver is used in these, with a unique application quasi-open baffle mids.
http://meniscusaudio.com/finalist-pair-p-1364.html


I do like Meniscus, I bought the Mandolin kit from them and they are really great sounding speakers. They generally solder the crossovers for you, but you'd have to find someone to do the woodworking for you.

The Fs of 3x the the crossover point is a general rule, but not absolutely required. It has been years since I have played with a dome tweeter so I'm not sure what the current crop of drivers require. Lots of old rules of thumb are due to the days before modern measurement and modeling software. There is no replacement for measurements.

Generally speaking, yes and this applies to a 2-way or a 3-way.

1. You can estimate where a woofer starts to beam by its actual cone diameter. Basically measure from surround edge to surround edge. Where that is equal to one wavelength you will typically be in the range of a 90deg dispersion. It is best to measure using a turntable.

2. You can see breakup in the measured acoustic response. You can also usually see it in distortion plots. You will notice a suddenly ragged response.

3. You have to measure to know a drivers off-axis response. The baffle will also impact the off-axis response on tweeters significantly.

4. Yes, the larger the larger, the lower in frequency its directivity will begin to narrow. This applies to all drivers. Rectangular drivers work this way too. A taller than wide driver like the bigger Raal will beam vertically well before it beams horizontally...but like with anything there are caveats. They use foam to delay and attenuate the upper and lower portions and help widen the vertical directivity. Rules of thumb are only generalizations.

You can generally use the manufacturer or 3rd party measurements online to get an idea about a driver's capabilities and caveats before buying. Most drivers have already been used and finding someone's build and explanation of their process can help too. Also, don't trust a spec sheet that says a driver can be used over some frequency range. That is just written by some marketing guy.

Ultimately, you should get into speaker design because you want to get into speaker design not because you want to build a single set of speakers with drivers that don't have a kit. You can always find some design out there that is close enough and the designer has already done all of the work for you.

Where are you located Marty? You have an invite any time you are in Central Florida. Check over atHTGuide's Mission Possible forum. Browse through the completed projects sections for tons of info and some really high-end designs

Also, Troels has a large website with many many projects and descriptions of his designs:

http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm.