Active 2-way design with cheap drivers... Madness?

[Riku-78]

Disclaimer: This is my first post here, so if I say something impropriate or violate some forum rule X, just try to understand... My English can also be really bad time to time, sorry about that.

The other day I was browsing different speaker and electronic parts dealers' websites, and got this weird idea to build a pair of active two way monitors (bi-amped with some readily available GainClone based designs). Now, before I start ordering parts and do something REALLY stupid, I would like to ask other DIYers what they think.

First of all, I found this 7" driver:
http://www.partsexpress.com/pe/showd...number=290-308
There are both 8 and 4 ohm versions, but the link is to the 4 ohm version, because there were no response graphs on the 8 ohm version product page. If I'm actually going to use this driver, it will be the 8 ohm version.

Now, I know Goldwood isn't exactly Scan-Speak, but what I found amazing is that this driver has 82dB sensitivity at 20Hz!!! Using an active filter section it should be possible to get a relatively flat frequency curve from 20Hz to the point where the tweeter takes over. With its 80W RMS power handling, we get about 101dB continuous SPL (80W = ~19dB, 82dB + 19dB = 101dB), right?

Now this sounds just too good to be true using a 7" woofer. What did I miss?

[wigginjs]

One of the things that graph doesn't tell you in anything about that drivers linear and non-linear distortion. When you send that driver 20Hz I'm almost positive there will be some very severe reproductions of second and third harmonics. This isn't to say the driver is a bad choice, but I wouldn't ever think that using a 7" driver to try to accurately reproduce 20Hz is a good idea. Perhaps a 2-way speaker with a subwoofer, or a traditional 3-way speaker with a large woofer would be a better choice to truly full-range reproduction. Just some ideas...

Also, which tweeter were you planning to use and where were you planning to cross it? I took a look at the response graphs for that woofer and it seems to start breaking up around 1KHz, this is REALLY low.

BTW, your english is really quite excellent, perhaps better than mine and i'm a native speaker =).

[tiroth]

Don't forget xmax. At 2.5mm and 7", you are not getting 20Hz.

In addition, I'm not sure the PE graphs can be trusted below 150Hz or so.

[Riku-78]

Thanks for the reply...

I'm planning on building a subwoofer to support the *lowest* bass. However, I'd like to have the sub crossover as low as possible to have a decent stereo image.

I haven't quite decided the tweeter yet. There are some very good tweeters out there, but with this woofer the tweeter should get pretty low, like you said. I like the sound of a tweeter playing the upper midrange, so the low crossover point is not really a problem for me.

I would like to use a LM3886 chip to drive the tweeter while using LM4780 with both channels wired in parallel to drive the woofer. Using LM3886 makes it necessary to have a relatively sensitive tweeter.

Vifa DX25TG-05-04 looks like a great option. It has rather flat frequency response down to 900Hz and -3dB point somewhere around 700Hz. Of course, it costs more than three times the price of the Goldwood woofer, but it will play about half of the audible frequency range (~ 4.5 octaves). Its frequency range also extends well over 20kHz.

It just feels stupid to couple a 35 € (~ $42 US) tweeter with a $12 woofer...

[Riku-78]

Don't forget xmax. At 2.5mm and 7", you are not getting 20Hz.

I think that might be the answer to my question: "What did I miss?"

Ok... Well, if I'm going to use a Danish high-end tweeter anyway, maybe I should look at Danish high-end woofers, too. The overall cost of a project like this is so large pile of cash, anyway that buying decent woofers doesn't really make a big difference.

[Riku-78]

Peerless CSX 176 H looks like it might work pretty well in my design, just have to make the box a little larger than I thought. And it doesn't cost a fortune.

The cabinets I'm going to build are actually so experimetal that they will probably not work at all, so I guess I'll need to build a few pairs before I'm satisfied.

Hopefully not too many.

I hate woodwork.

[Mark25]

if you haven't already seen it, the ESP site offers some good info for active beginners. One thing not to miss, would be a capacitor to protect your actively driven tweeter from thumps, bumps and DC your power amp may decide to throw at it.

Good luck.
Mark

[audio-kraut]

Hi, re tweeters: I am using - and have blown - quite a fewD27TG 05. Silkdomes, and very smooth - no tweeter hash at all. I would also consider looking at vifa tweeters.
I run two 12"vifa - number escapes me - in compound loaded config, and they go to 20Hz @ - 6dB. Raised it a bit with a behringer deq (app. +2 db) and get excellent clean powerfull bass with great attack, sustain and fast decay - no loitering there.

[Riku-78]

Thanks for the info.

I have seen the ESP site, though. I'm actually probably going to use the crossover circuit presented there, and possibly the power supply circuit (modified to satisfy the needs of the 120W + 50W amps), too. Have found other useful designs on the web, too, but I think the ESP circuits are the most reliable, been built and used by so many DIYers around the world.

[Riku-78]

D27TG 05 06 appears to be very similar to DX25TG 05 04 at least on paper. DX25TG has lower impedance (nominal 4 ohm) and rolls off a little lower. It also has -3dB point at almost 40k...

Am I going to fry these things if I use an LM3886 based amp? Gain Clones should have some protection stuff, but I think it's to protect the chip, not the speaker, right?