Home-Made IEMs - Page 7

diagram
http://www.ultimateears.it/pics_prod...expl_large.jpg

real pictures - 7th, 8th & 9th pictures down
http://images.google.com/imgres?imgu...a%3DN%26um%3D1

I just got an e-mail from Tom Hazlett over at KA. I'd asked him about what drivers KA considered its top contenders as either woofers or tweeters. While suggesting that it's a "very subjective question," he said he's seen the TEC, CI and FK models used as woofers, the ED and FK models as tweeters, and combinations of TED + ED, CI + ED, CI + TWFK (triple) and TEC + TWFK (triple).

A couple of observations are in order. First, I think we've all seen the CI used as a woofer. That seems to be a no-brainer. Lots of pictures of customs, when blown up, show the CI on the outside, with its model number fairly visible. Second, it's interesting that the FK would be used as a woofer, in some designs, and as a tweeter in others. This is suggestive of the shapeshifting nature of BA's. There really aren't any obvious woofers or tweeters. It's what you do with these wide-range drivers that makes all the difference. Third, in dual-driver configurations, the ED may be a popular second driver, probably as a tweeter. Fourth, you'll notice that the TWFK is popular in triple-driver configurations. That's likely because the TWFK is a dual driver, the world's smallest.

When I piggy-backed my TWFK on my CI-22955, it was small enough to put the entire apparatus into a de-cored foamy. I was going to do this with the other driver when I had my soldering disaster. One thing I noticed was that there was ample room to piggy-back a second TWFK, to create a five-driver setup - and small enough to fit into (not onto) an eartip. With a 2.55mm minijack (let alone a two-prong wire connector), this would make for an amazing set of tiny shell-less earphones. Even without an "integrated passive crossover" or any filters to tweak the sound, I got amazing sound from my CI-22955 + TWFK combo. I found I had to position them right, to get a good seal, but once I did, it was an impressive combo.

Point of correction. I was able to wire up the CI-22955 for both ears (which worked wonderfully). I was not able to get the combo in the second ear because of the soldering disaster, forcing me to buy more drivers. What I did notice in the one ear (with the CI/TWFK combo) was a very cool, dynamic, sound.

This could be the beginning of a beautiful friendship. I am not a moldsmith. If you can mold shells, I can wire them up, or at least design the configuration.

By the way, the sound tubes, which I got from Microsomic, are there to compensate for recessed drivers. There is truth to their use as a frequency tuner - longer and narrower lowers the high frequency while shorter and wider opens the high end -but they are not a necessity. The challenge with dynamic drivers is size.

The tweeter was ED-29689, which sells for $24.49 through Mouser. That's a good price because Mouser will sell in single units. The price of the CI-22955 (through Mouser) is $26.19. So, not counting shipping, the price of the drivers for a pair of Sm2's is $101.36. The amount of wire used is nominal. The amount solder used is nominal. The cost for an SMD capacitor, especially the single-driver "crossovers" typically tied to a single tweeter, is also nominal (less than $5). The tubes, available from Microsonics, are a few bucks at the most. Once we figure out what filters sound best with these drivers, the cost of two pairs of acoustic dampers is approximately $6 (not counting shipping).

So think about this. The drivers on your SM2, which sells for 280 Euros (approximately $393) cost $101.36 - that is, if they were purchased retail. That's about a quarter of the price of an Sm2. Even when you include the other components - the little bit of wire, the few drops of solder, the few millimeters of sound tubing, the six dollars' worth of acoustic dampers, and the cabling - there's a lot of potential for building an Sm2 equivalent for a song and a dance. Right now, my weak link is the acrylic casting. I don't think injection molding is really needed for something whose per-unit value makes it unnecessary for an automated process capable of producing a huge batch of plastic housings. This can be done cheaper and with less efficiency - and still remain worth doing. But that's the process I'm trying to get a handle on now.

Scroll down past the pretty girl (easier said than done):

Google Image Result for http://lib.store.yahoo.net/lib/yhst-94024159501758/1se530blow.jpg