Weird Impedence Graph-Titanium Tweeter

[rogoll]

Can someone tell me what's going on in the area of the graph enclosed with the red line? This is the WT3 graph of TW-3060 Titanium Tweeter. I ordered 5 of these buyout speakers from the previous flyer. Out of the 5, I have two that arrived with the plastic grill intact. When I contacted PE they were already out of stock, so they just refunded the cost of all 5. I'd like to use the two good ones if possible, but maybe I just need to toss them and learn a lesson from the experience.

Thanks in advance for your suggestions an input.
Leonard

[Jeff B.]

Can someone tell me what's going on in the area of the graph enclosed with the red line? This is the WT3 graph of TW-3060 Titanium Tweeter. I ordered 5 of these buyout speakers from the previous flyer. Out of the 5, I have two that arrived with the plastic grill intact. When I contacted PE they were already out of stock, so they just refunded the cost of all 5. I'd like to use the two good ones if possible, but maybe I just need to toss them and learn a lesson from the experience.

Thanks in advance for your suggestions an input.
Leonard

It is probably a reflection from the internal cavity back to the dome that just happens to correspond with the tweeter's resonance. If crossed above 3kHz they may perfrom fine.

[rogoll]

Have you seen dual resonance peaks before, like this?

[dubbreak]

It is probably a reflection from the internal cavity back to the dome that just happens to correspond with the tweeter's resonance. If crossed above 3kHz they may perfrom fine.

So could he stuff them?

No harm in trying to mod them as they are now effectively free.

[charlielaub]

I have seen this a few times with tweeters. In fact, annoyingly so, because my woofer tester could not figure out how to assign Fs with such a messed up impedance plot. This is not so bad, any you can just guesstimate Fs based on the general shape of the impedance and the zero crossing of the phase.

I would avoid using the tweeter in this area, but since it is at 1k Hz I doubt that you had planned this to be in the passband!

-Charlie

[daver]

Are the drivers being measured secured in place?

I saw something similar, it happened because the driver was not firmly supported (e.g. clamped in place). The whole tweeter was probably vibrating, although I don't have high speed camera or accelerometer to capture that.

To remedy I made a fixture that (1) clamps the driver in place, and (2) keeps the driver up off a table top at least a foot to alleviate boundary effects from the back of the cone. I use it for tweeters, mids, and woofers.

I recently measured a 4x10 driver, and really had a lot of problems with the impedance peak. Ended up using bar clamps to stiffen the fixture in certain places, and added some weight to the table top to dampen vibrations in the table legs that the fixture was setting on.

Don't need a fancy fixture. For a tweeter, I'd think you could easily cut a hole and mount it on a small baffle -- then clamp the baffle tightly somehow.

As someone else noted, it's not that critical to know Q for a tweeter. All you really need is a reasonably close value for its Fs, its DCR, and its impedance an octave or so above Fs on up to 20 kHz. You'll generally crossover higher than that.

Hope this helps.

[mattk]

Another possibility, however remote. I notice that the impedance analyzer is set to a range of 20 ohms, and that the two peaks are at 18, with the trough between them suspiciously looking like it would, if inverted, go up to around 24 ohms, which is above the 20 ohm max setting. I just wonder if the setting were changed to 30 ohms or higher if you wouldn't see a normal impedance graph with a peak around 24 ohms.

[badman]

It's a cavity resonance. Open 'em up and play with damping.

[marlboro]

Use an electronic crossover. Your amp driving it directly will never notice it.

[rogoll]

Are the drivers being measured secured in place?

I saw something similar, it happened because the driver was not firmly supported (e.g. clamped in place). The whole tweeter was probably vibrating, although I don't have high speed camera or accelerometer to capture that.

To remedy I made a fixture that (1) clamps the driver in place, and (2) keeps the driver up off a table top at least a foot to alleviate boundary effects from the back of the cone. I use it for tweeters, mids, and woofers.

I recently measured a 4x10 driver, and really had a lot of problems with the impedance peak. Ended up using bar clamps to stiffen the fixture in certain places, and added some weight to the table top to dampen vibrations in the table legs that the fixture was setting on.

Don't need a fancy fixture. For a tweeter, I'd think you could easily cut a hole and mount it on a small baffle -- then clamp the baffle tightly somehow.

As someone else noted, it's not that critical to know Q for a tweeter. All you really need is a reasonably close value for its Fs, its DCR, and its impedance an octave or so above Fs on up to 20 kHz. You'll generally crossover higher than that.

Hope this helps.

I had it clamped in a vice in the vertical position (as if mounted on baffle) and the vice secured to a large table.

I would set the impedence range higher, but then WT3 would automatically reset it back to 20 Ohms.

[rone]

That is some impressively complex behavior for a tweeter. You might want to take a careful frequency response measurement on that. A cavity/pipe resonance at ~1kHz would need to be on the order of 6" or so.

[rogoll]

That is some impressively complex behavior for a tweeter. You might want to take a careful frequency response measurement on that. A cavity/pipe resonance at ~1kHz would need to be on the order of 6" or so.

Well, I guess that shoots that theory down then, because this is a typically small tweeter.

[rone]

I wonder if you might have a leak in construction somewhere? There also seems to be some odd resonant behavior there aside from the dual peak. What does it sound like? Taking it apart and damping the pole piece (a piece of felt if bare/solid and felt/stuffing if hollow) may improve things, or you might just break it in disassembly.