Help crazybastard build speakers

**cjd** · 29 October 2007, 09:15 Monday

I'm curious. You're working with drivers YOU don't even have all the specs for. You're working without much by way of measurements. You're looking for "a large, good sounding, efficient full range open baffle" - large is easy to hit, you're there I think. Efficient you're possibly going to hit by using large and efficient drivers. Good sounding? It starts with driver selection and ends with crossover implementation.

Best of luck, and have fun with it. But I do wonder, what kind of help can we lend? Our ears aren't there, and you're starting off with a lack of data so we can't really even help there (unless someone happens to find it).

**joecarrow** · 29 October 2007, 10:26 Monday

I think the only hope of something like this working out well without impedance measurements and baffle diffraction simulations would be the use of a DCX2496 or similar as crossover. In this case, you're free to do a LOT of trial and error without building a lot of things. Since they're pro drivers and very efficient, you can get away with really low power (maybe T-amps, or similar).

**crazybastard** · 29 October 2007, 10:27 Monday

Thanks CJD. The Radians were given to me used. I looked at the crossover that came with it. Seems like a 2nd order at ~1500, and some attenuation of the HF driver. Do you think the 2 18" will work OK? I'll proably use a plate amp and boost the low end a bit.

**crazybastard** · 29 October 2007, 11:02 Monday

Thanks Joe. I'll probably pick up the dcx2496 at some point. Is there a way to simply simulate baffle diffraction? I figure it's pretty hard w/o t/s for the radians.

btw, I have a stupid question. I was at the RMAF and noticed that most speakers sound remarkably similar. If they all do 20-20k +/-3, they all sound the same right? I mean it doesn't matter if it's metal or ceramic driver since a 1K signal is the same not matter which driver is making the sound?

**ThomasW** · 29 October 2007, 11:10 Monday

Originally posted by crazybastard

btw, I have a stupid question. I was at the RMAF and noticed that most speakers sound remarkably similar. If they all do 20-20k +/-3, they all sound the same right?

They all tended sounded the same there (mostly bad) due to the terrible room acoustics

I mean it doesn't matter if it's metal or ceramic driver since a 1K signal is the same not matter which driver is making the sound?

This is analogous to saying all V-8 engines perform the same, nope...

Different materials make for different sounding drivers.

**cjd** · 29 October 2007, 11:35 Monday

Distortion in the driver(s) (or elsewhere in the playback chain - see tube amps) can dramatically change how 'flat" sounds. Distortion comes in many flavors, and really it includes room interaction.

**crazybastard** · 29 October 2007, 12:07 Monday

Originally posted by ThomasW

They all tended sounded the same there (mostly bad) due to the terrible room acoustics
This is analogous to saying all V-8 engines perform the same, nope...

Different materials make for different sounding drivers.

I'm not sure how a 1k tone will sound different. Discounting distortion, noise, etc. Looking a pure physics, why would this sound different on different driver? If you put a mic in front of a driver, and it measures 1k, then this signal should be the same not matter if it comes from an 18" woofer or a 1" tweeter? 60 miles an hours is the same in a Honda or a BMW right?

**cjd** · 29 October 2007, 12:14 Monday

Originally posted by crazybastard

Discounting distortion, noise, etc. Looking a pure physics, why would this sound different on different driver?

Well, when you specifically eliminate the reasons it would sound different...

Though dispersion will be different for a 1k tone from a 1" tweeter and an 18" driver.

**crazybastard** · 29 October 2007, 12:25 Monday

Originally posted by cjd

Well, when you specifically eliminate the reasons it would sound different...

Though dispersion will be different for a 1k tone from a 1" tweeter and an 18" driver.

Am I correct in saying that if a perfect speaker does 20-20 flat, then that's the holy grail of speaker design? All good speakers should then sound the same if you sit right in front of them? The fact that there are differences in sounds refect the effects of distortion, differences in amplitude of the driver at different volumes? The most 'different' sounding speaker that I heard were the MBLs. Maybe I'm used to hearing vocals coming out of a singer's mouth directed towards me. I though I didn't like the sound of the MBL because it's freaky because of the 360 degree dispersion - makes it seem like the singer's voice came out the back of her head!? But in reality just different room interaction? Or perhaps there were lots of distortion? Does that mean a good sounding speaker have harmonic rather than disruptive distortion?

**ThomasW** · 29 October 2007, 12:31 Monday

Originally posted by crazybastard

I'm not sure how a 1k tone will sound different. Discounting distortion, noise, etc. Looking a pure physics,

We're talking pure physics here (but not operating in a vacuum....

) and that's why different materials sound different.

The behavior of the materials reproducing the sound 'colors' the sound.

A 1kHz tone from a planar doesn't 'sound' like a 2" metal dome midrange

Am I correct in saying that if a perfect speaker does 20-20 flat, then that's the holy grail of speaker design?

No that's but one aspect of loudspeaker design.

**crazybastard** · 29 October 2007, 12:47 Monday

Originally posted by ThomasW

We're talking pure physics here (but not operating in a vacuum....

) and that's why different materials sound different.

The behavior of the materials reproducing the sound 'colors' the sound.

A 1kHz tone from a planar doesn't 'sound' like a 2" metal dome midrange

No that's but one aspect of loudspeaker design.

I think that John at zaph audio discribed the 'timber' in ribbon and planar as distortion. I kinda like this sound, makes music seem 'brighter'. Maybe this is 'harmonic' distortion, rather than the disruptive kind.

It's interesting how some manufacturer go about designing speakers. I talked with Tyler of Tyler Acoustic a couple of years ago. I liked the way his speakers sound in general. I recalled him saying that he basically have an idea of a cabinet/driver arrangement, then adjusts his crossover to compensate for what ever is severely wrong.

I might take this approach. Basically build the baffle, then play a pink tone and compensate for any sever peaks/dips?

**crazybastard** · 29 October 2007, 13:18 Monday

Found some info on the Radians.

link

**crazybastard** · 29 October 2007, 13:37 Monday

any thoughts about using this woofer as a 'fullrange' up to around 2.5k?

selenium woofer

**cjd** · 29 October 2007, 14:00 Monday

Why?

**Deward Hastings** · 29 October 2007, 14:59 Monday

That’s an interesting choice of drivers, enough out of the ordinary that you’re going to be pretty much on your own designing for them. The Selenium drivers, based on specs anyway, should do well up to 200 Hz, maybe a bit higher. If you mount them one facing forward and one to the rear the 2nd harmonic distortion will be considerably reduced and, again based on the specs, you’ll get plenty of low distortion output on the low end. The baffle is going to have to be rather large, though, and you’re going to have to come up with some creative way to keep the MT driver from being too high for normal seated listening (since it kinda hasta be above the two 18 inch Seleniums, which puts its center 4 feet high if they are stacked vertically). These will be large speakers . . .

The Radian driver raises other issues. One of the design goals/advantages of dipole mounting is relatively uniform polar pattern/power response. The Radian will control its own directivity from the point that the 12 inch driver starts beaming (well below 1000 Hz.). In general, in fact, it is the Radian that will control the “sound” of this proposed design . . . if you don’t like the way it sounds the whole design will fail, if you do like it then you will be able to make a complete speaker that you like. I’d listen to them for a while, either in boxes or on flat baffles, before going too much further . . .

Working out a crossover to handle dipole correction and driver integration is going to be your biggest challenge. I’ll second the vote for using a DCX2496 . . . there’s just not enough data to do an analytical design except in the broadest sense, and in any case you'll want to tune these for your specific listening room. You’ll have to do it by cut-and-try engineering. So you’ll need an easily programmable digital crossover, and a good set of measuring tools (mic, pre-amp, PC and measurement software).

There’s a lot of on-line literature available about dipole design . . . here in the various dipole threads, and especially at the linkwitzlab site, but your proposal is different enough that while the general guidance will apply the specific examples given will be rather different (which will challenge you to more fully understand what is being discussed). As you determine more specific design goals (and narrow the design to match) it will be easier to solicit advice . . . expect many of the replies to be in the form of “did you consider this?” rather than firm answers . . .

Good luck. If you follow through on this the next year or two will be quite the “learning experience” . . .

**crazybastard** · 29 October 2007, 16:35 Monday

Originally posted by cjd

Why?

increased efficiency of 3db over the radian woofers. Not very expensive either, if it is indeed an upgrade over the radians.

Thanks for the input Deward. I do like the radians other than the lack of bass, so adding the woofer should do some good. I'm curious to see how this will turn out. I'm sure much eq'ing will be needed. Ultimately though, I'd like to use passive crossovers for the mid/high, and use a plate amp for the woofers. It's going to be a large pair of speakers, that's for sure. I'll put the radians on a hinged baffle so it can tilt down a bit if necessary.

Hey Thomas. Is there a diy club in CO? I should could use some inputs locally.

**Dennis H** · 29 October 2007, 17:00 Monday

You're going to lose a lot of sensitivity in the dipole woofers after you build a filter to flatten the 6dB/octave dipole rolloff as well as the drivers' natural LF rolloff. Depending on how low you want them to play, you may lose something like 12-18dB from the woofers. So there's no need to look for more sensitive mid drivers -- you'll probably end up padding them down anyway to match the woofers.

**crazybastard** · 29 October 2007, 18:10 Monday

Hi Dennis. Thanks for your input. I'll stick with the radians then. I'm hoping to get the woofers to play down to around 40hz. My subs can pick up from there. My baffle is going to be about 20-22" wide, at what freq. will they start to roll off? I'm going to have to do more reading, but the regular job and kids + lack of brain power makes it difficult. The Partexpress amp have a parametric eq, which should come in handy.

**ThomasW** · 29 October 2007, 18:56 Monday

Originally posted by crazybastard

Hey Thomas. Is there a diy club in CO? I should could use some inputs locally.

There's the Colo Audio Society... but it's not all that DIY oriented.

Things like the Iowa meeting have just been organized by regular guys interested in getting together. It might be a good idea to do this in Denver.

I used to have get-togethers at my house since it has two separate listening rooms. Unfortunately neither is large enough to accommodate more than a handful of people.

If someone has a church or similar organization affiliation (I don't) those are the best places to get a free hall/room.

Edit to add....

If you're interested in hearing the Emerald Physics CS2 loudspeakers, Clayton Shaw will be demoing them 11/17 at the Tweakgeek showroom in Arvada. Contact Tweakgeek.com for info.

**crazybastard** · 29 October 2007, 20:04 Monday

I wish I can hear the CS2 again. Got to work that weekend. I'm sure I can find a place for diy here in Colorado Springs if there's enough interest. I don't know if anyone want to go all the way down here though.

**crazybastard** · 03 November 2007, 20:38 Saturday

Can some one advise how high up can I use this driver?

link

**Deward Hastings** · 03 November 2007, 21:46 Saturday

Still talking dipole, for music reproduction? I’d say 200 Hz. absolute max using an LR4 crossover, and that’s really pushing it. By 300 Hz. you’re seeing major cone breakup, so you want that well down the curve. While the distortion looks quite acceptable, even good, for below 200 Hz. above that it rises rather than falls, in a region where even moderately good drivers do far better.

The crossover point, wherever below 200 Hz. you put it, should be determined by the low frequency performance of the mid driver. Keep the signal on it as long as it is performing better than the Selenium. I’d target a crossover at maybe 120-150 Hz. if using those as the dipole woofer.

**crazybastard** · 03 November 2007, 22:44 Saturday

How can you tell that distortion goes up beyond 200? I'm looking at the manufacturer pdf file. seems to do well up to 2K. Is there something I'm overlooking?

**ThomasW** · 03 November 2007, 23:21 Saturday

Both the impedance curve and the frequency response start going to pot from 200Hz up.

**Deward Hastings** · 03 November 2007, 23:29 Saturday

Look at the 3rd harmonic curve, and how it jumps when the cone starts breaking up. And look at the scale. More than 40dB down is not bad distortion for a speaker at 100 Hz. Less than 40dB down is *terrible* at 1000. And the wild peaks and dips of the distortion curves suggest that it's not motor distortion being displayed, but cone breakup.

The concentric rings embossed in the cone are there to progressively decouple the outer portions of the cone as frequency goes up. Without them the cone surface would look like a wave tank with a family of frogs playing hop scotch in it. But they introduce their own problems as soon as the cone stops moving as a unit, and all the curves show where that starts happening.

**crazybastard** · 04 November 2007, 00:07 Sunday

Thanks guys. Didn't realize that was a 100db scale! I'll get my glasses checked.
I'm looking at the ficaraudio.com q18 as well. Last time someone commented that due to high Le, they are not useful beyond 80hz. I don't see the Le on the manufacturers site. Is there a way to derive the Le? Will it be awful if I use it up to 200hz?

Since I'm going in circles with my woofer choice, would someone help me decide what to use with my radians? I wan to use 2 woofers per side
1. 2x 15" subs from Kevin
2. 2x 18" selenium sw1000 (these seem to have less distortion than the sw600)
3. 2x 18 ficaraudio q18
or any other woofers....I'm open to suggestion.

Thanks Thomas again for the info on the cs2 speakers. I heard them at the RMAF, and having read the specs on the components, I'm real inspired to do my own design. Seems like they use so-so components and eq the heck out of them....to very good effect.

Help crazybastard build speakers

Help crazybastard build speakers

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment