SLA build

[jimbo7]

Finally bit the bullet and ordered the parts build a pair of SLA's per plans (6 mid x 9 tweet). Already there's a problem. One of the tweet's diaphram's seperated from the magnet during shipping.


Can I put a few drops of epoxy on it?

The tweets are Goldwood GT-302 and mids are: http://www.parts-express.com/5-paper-co ... m--299-006

Stay tuned for updates

[Grant Bunter]

I'd send that pic to PE.

While you could epoxy it back in place, you would need to (carefully) cut out the dustcap in order to shim the voice coil and magnet so they are in the right position as the epoxy sets.
No way you can do that otherwise.

See "warranty" on this page:
http://www.parts-express.com/why-Pe-No-Hassle

If it fell apart during shipping, IMO it's defective...

[jimbo7]

I sent them an email. I'll need to make another order soon, so maybe I'll get a credit. Should be able to continue on.

[67baja]

I ordered 50 of those tweeters a couple months ago and 23 of them had the magnets separated in the box. I phoned PE and explained the situation to them, and asked if other people had problems with these tweeters recently. They said no, but credited me for the bad tweets. Several others of the 50 I ordered had the magnets fall off in the cabinets. Simple fix is to hot glue them back on. Sounds like they work fine now.

[Hackomatic]

Since there appears to be some indexing nibs for alignment, shouldn't be too difficult to glue it back together. If it was me, I'd scrape off the old "non-adhesive" first. Epoxy for sure. Hot melt glue might hold 'em for a while, but it's pretty brittle when cured, and if you're planning to take these on the road, don't know how long they's hold up to all the bouncy bouncy . . IMHO.

[Grant Bunter]

Given a greater than 50% failure rate with 67baja in terms of magnet separation, I'd question that choice of tweeter.

I guess you could epoxy them, without taking of the dust cap, and testing for VC rub as you went, but you wouldn't want a rapid setting epoxy so that you could move them a little if rub was evident. The tolerances would be tiny though. And you'd have to work out how to hold them in position as the epoxy set.
Let us know how you go...

[Hackomatic]

Given a greater than 50% failure rate with 67baja in terms of magnet separation, I'd question that choice of tweeter.

Agreed!

[Bill Fitzmaurice]

I ordered 50 of those tweeters a couple months ago and 23 of them had the magnets separated in the box.

That failure rate tells me that the box had been subjected to high heat, which weakened the epoxy. You'd be very surprised to see how hot it can get in a vehicle left out in the summer sun.

[Hackomatic]

You'd be very surprised to see how hot it can get in a vehicle left out in the summer sun.

. . or a SeaLand container languishing on the ocean.

[jimbo7]

PE shipped out a new one today. They were very helpful and hassle-free.

There are little tab thingys on them, so I'll try epoxy when I get the new one. The SLA's will be in my basement most of their life, so heat wont be an issue. The glue does look thin and weak on these tweets. My main issue was how they packaged them. They layed about 5 or so in a line and rolled them up then layed some more and so on. It was the fact the ends were not taped so it was easy for them to slide out the ends. Like a burrito.

Tonight I'll run a freq sweep to test them.

[jimbo7]

So I tested all 12 mids and 18 tweets and only 1 tweet was bad. Got lucky I suppose. The tweet that separated in shipping works perfectly after carefully epoxying back together. Got lucky again and came out even. This weekend I'll cut out all the pieces and hopefully get them glued up. Not looking forward to wiring all this, since soldering all the tabs alone took forever.

Here's the players getting acquainted:

[Hackomatic]

Glad to see I'm not the only one using the kitchen counters as essential speaker assembly areas . . .

[jimbo7]

Glad to see I'm not the only one using the kitchen counters as essential speaker assembly areas . . .

There's an outlet at the bottom of the pic (end of island) where I usually setup my soldering station. But only when nobody is home

[jimbo7]

So I finally have the boxes built. It was such a pain to cut out all 30 holes with a jigsaw. I need a router and Jasper jig bad. Now I'm wiring them up. I knew it was gonna be a big undertaking and it's kicking my butt.

Since I'm not using the recommended mid's https://www.parts-express.com/5-paper-c ... m--299-006, I had to design a custom crossover for them. Here's the schematic and other scribblings:



Here's my problem: I've wired them per plans, but it sounds "tinny" with no bass at all. Even when I turn up the mixer, it's still not loud at all (when it should be deafening). Did a freq sweep in Audacity and the crossover seems to work fine and it is louder. Could it be out of phase? My wiring is a rat's nest, so I'm sure I messed up somewhere. I also noticed the a tweet "cracking" when wiring them together. What could that be?

[CoronaOperator]

I see a flaw in your crossover calculations. You have 1.5 x 8 ohms = 12 ohms, yet when you use the Zobel network (impedance equalization circuit) you no longer need to do that. Also, once all drivers are wired series/parallel as per the plans (6 midbasses, 8 ohms each) your midbass load drops to 5.3 ohms. That is the number you should have used to calculate your low pass for the woofer instead of the 12 ohms that you used. Measured impedance may even be a bit less than 5.3 ohms with the Zobel network in place. For now, removing the Zobel network will help a bit until you get the correct value components.

Crossover woes aside, that still doesn't explain the lack of volume which sounds like a polarity/phase issue to me. Take a 1.5 volt battery and hook it up to the speaker terminals. All the woofers should move the same direction, if not then correct the wiring on the offending woofers.

Edit: I also noticed that you calculated the impedance equalization circuit based on one driver (Re = 6.45 ohms, Le 0.72 mH) and not the whole system. That is fine if you use a zobel network on each midbass driver (the parts are cheap so why not) however if you are just using one zobel filter then it should be calculated for the whole system. Re(system) = 4.3 ohms, Le(system) = 0.48 mH therefore Cz=16uf and Rz=5.3 ohms. Its probably not too critical but if you have to rework the crossover then you may as well do it.