DIY Mystery Speakers

[reganzi]

I got the midrange and subwoofer ("Woofers") drivers from an Altec Lansing ACS300.1 computer speakers system that broke. Both are 4ohm, 4" and I think the drivers from the subwoofer are long-excursion (or so I was told, I'm not sure if such a thing exists.)

Currently I have them wired directly to my amp without a cabinet or crossover and they sound better than pretty much anything else in my house. I bought some 3/4" MDF to make a bookshelf-speaker sized cabinet. I thought that as long as I am building it I should add a tweeter. (The set used to have tweeters, they blew. The set was taken apart and eventually I ended up with the parts.)

Since the speakers already sound pretty fantastic and I don't have any specifications, I figured I'd leave them alone as much as possible by using a high crossover freq. So, like 1st order Butterworth at 6400 Hz give or take.

My ideas on the config are:
1) Find a 4ohm tweeter. Not as common as 6/8ohm so it's a little tougher. Can go with 2-way or 3-way crossover setup.

2) Maybe, a 2-way system like Mid+"Woofer"=8ohm series + 8ohm Tweeter in parallel.

3) If I'm using such a high crossover freq anyway most tweeters are way above nominal impedance so maybe I can get away with a 6/8ohm tweeter without a big SPL difference?

I'm sure you can tell, I've done a good amount of reading but haven't got much experience with audio so any suggestions or help is appreciated.

[chris661]

Ok - I'll go in the same order as you to avoid confusion.

"High Excursion" does exist, but it's generally used to simply show off and has very little value. Example: I took apart a 2.1 system, with a sub that I would call long throw (it was 4" with +/- 4mm Xmax). But there's bigger speakers out there with a lot more excursion that I wouldn't class as long-throw...

The impedance of the tweeter won't matter, as long as you account for the decibels lost/gained by different impedances.

You may need a woofer tester for this project, as it may be difficult to come across the TS parameters for these drivers.

How about a 2.5 way system? Tweeter + 2 woofers, one only playing bass, the other playing midrange aswell?

Please could you specify exactly how many of each driver you have - it is difficult to tell from your post.

Thanks
Chris

[reganzi]

I have two of each type of speaker, so enough for a stereo set.

http://reganzi.com/other/speakers01.jpg
http://reganzi.com/other/speakers02.jpg

The speakers are stamped with the following: A-1481 on the plastic one and A-1498 C19506 on the other. I've done extensive Googling and there is no information to be had.

Since I haven't got any tools, what if I get a tweeter with a sensitivity of about 91db (assuming the ones I have are around 89-90db) and adjust the difference using a variable L-Pad?

[PeteMcK]

download ARTA and measure them, then you won't be guessing:
ARTA http://www.fesb.hr/~mateljan/arta/download.htm
ARTA Jig - http://zobsky.blogspot.com/2008/01/s...t-jig-for.html

[chris661]

It would be better if you fed all the speakers with 1W, see how many decibels come out, and work from there. If not, trail and error using cheap tweeters and hearing would work, in the end.

Looking at the drivers and your requirements, a 2.5 way system would probably do best. Using just the midrange or just the sub would lead to either: little bass, or little in the midrange. Using 2.5 way would mean the drivers the manufacturer intended to be used together, would be used together.

Chris

[reganzi]

Originally posted by chris661
How about a 2.5 way system? Tweeter + 2 woofers, one only playing bass, the other playing midrange aswell?

I went with your 2.5 way system idea, so thanks. I used a series crossover for the mid/tweeter pair and a parallel crossover for the woofer. 6400Hz and 800Hz, respectively. I got an MDT-29 for the tweeter.

I haven't built the cabinet yet so it's just connected for testing but it sounds fantastic. I probably could have gone a bit lower than 6400Hz but it's my first project so meh.

Keep in mind I've had a $60 5.1 system from Logitech (X-540) for the last like 6 years so pretty much anything is a massive step up in quality.

Also, I tried the Arta Limp software but it didn't work. It calibrated fine but I couldn't get a proper curve. It always came out with totally unreasonable values on each computer i tried. Maybe my jig was no good. I tried a few different ways though. I wish it worked because I have some inductors of unknown value I harvested from some old equipment. Maybe when I'm not busy I'll try again.