So, I’ve gone on a bit of a housing journey the last two years. The place we were renting was taken for eminent domain reasons. Not for space bypasses sadly, but for a roundabout. We downsized as much as we could and everything else went in storage while we rented something small and threw ourselves into the housing market. We’re moved into our new house, and there’s LOTS to do. In a good way, with millwork and cabinetry.
But now it’s time for tunes, and so out come my wife’s trusty JBL L96. They may fit the aesthetic of that room and are period for the rest of my gear. So I’m giving them an overhaul. Bypassing the L-Pads at the very least, and quite possibly whole new crossovers. Generous cab buffing. And fixing the tweeters.
Something didn’t sound right with one of them. I had always blamed the L-Pads being dumb, but I came to learn that the little foam pad that sits behind the dome is urethane, and we know how that gets with time. It sloughs off little bits of itself that get into the VC gap. Maybe this is it.
So I purchased an aftermarket kit that replaces the foam with felt.
Trying to separate the dome’s thin bakelite board from the magnet was harrowing. I had no clue how much force was OK and how much was not OK. But I managed not to screw things up. Cleaned some oxide whiskers out of one of the tweeters and shot them both with compressed air. Put on the felt pads and started buttoning things up.
There’s a lot of slop to the mounting holes in this assembly. But the DATS turned out to be super useful in getting the voice coil aligned. I just adjusted the placement until I got the best possible curve. Here’s the overlay of the two tweeters. There’s still something “not great” about one of them. We’ll see what the other measurements look like.
I’ve taken apart many tweeters, and the method I’ve been using to get the voice coil centred for tweeters without alignment pins is to use a multi-tone signal from REW’s generator. The multi-tone makes it really obvious if there is any rub/buzz occurring from the voice coil touching the motor. It’s much less obvious with a single tone.
Yep I’ve done it with ARTA but probably same thing. (With LIMP box)
It uses white noise type tone instead of woops and you can set the test cycles up high enough that you can play around with the diaphragm in real time for a couple minutes and watch the results as you go.
Sat down and took some measurements. Tweeter and mid out to 45, though that was more from curiosity. Also took measurements of the speaker with the stock crossover, and modeled the same in Vituix:
So this is the crossover modeled from JBLs posted schematic. I figured out how VCAD library blocks work, and simulated the variable L-Pads. “Twiddling the knob”, this was the smoothest curve I could make. The light grey dashed line is the measurement of the stock speaker with a sketchy L-Pad.
Trying to decide whether I do much more with the mid/tweeter. I’ve got way better phase response now, and the crossover point is about the same, but the crossover point is still a little high to make the most of that phase. I’ve got another solution that pushes them closer to 2.5khz, but there’s no ferrofluid in there, and I really don’t want to smoke them. LOL
