Measured the new Tang Band ceramic tweeter

[Mazeroth]

I visited Paul K. recently to listen to his fantastic speakers and while I was there picked up a pair of the new Tang Band 25-1744S 25cm ceramic dome tweeters to measure. I figured this would be a great chance to take some measurements with my new ECM8000 that I had calibrated by Kim Girardin. The tweeters were flush mounted on an 8" wide baffle that had a 1/2" roundover, something close to how they would be used in the real world. Four 22 uF capacitors (88 uf) were wired in series to protect the tweeters. The 1 meter frequency measurement were taken around 80 dB to protect the tweeters as I wanted to get an accurate low end. Distortion measurements still used the 88 uF of capacitors but started at 500 hz. for protection since they were receiving much more power.

https://www.parts-express.com/Tang-Band-25-1744S-1-Ceramic-Dome-Tweeter-264-866?quantity=1

The first measurements I took were nearfield frequency. After taking measurements for both tweeters and doing an overlay I was shocked. The second tweeter looks to be 3-4 dB less sensitive than the other. Not letting a single measurement be the proof I remeasured them both again and got the same result. Not good:



Up next is a 1 meter measurement at around 80 dB:



Here is the distortion measurement from 1 meter at around 90dB for the more sensitive tweeter. The F3 is higher down low, which could be due to the higher output at the same drive level:



And now the less sensitive tweeter's distortion measurement:



You can see how things get a little funky around 3 khz. with this tweeter. I remeasured both tweeters just to be certain the first measurements were accurate and both were very close. It looks for certain that something's not right with the second tweeter, which is a shame as I was hoping these would be a great performing, small flange tweeter.

Sorry, Paul. Probably not the news you were looking for but at least we know something needs fixed before you tried using these.

[Rick Craig]

Sorry to see that. I recently used the same tweeter but with the phase plug.

[Mazeroth]

As I was putting these back in their boxes I noticed something. The tweeter that's less sensitive seems to have the hexagrid cover much closer to the ceramic dome, by about 1 mm. I checked both tweeters' hexagrid cover's convexity and they're identical. It looks like the less sensitive tweeter is almost touching the cover with a little room, where the other has about another 1 mm or so gap. I'm not sure if this could affect anything but I figure I might as well post it

[Jed]

Thanks for posting this, Christian. I was interested in those tweeters... WAS-interested until now.

[fjhuerta]

I was going to order these, and the new TB Berillium tweeter on Monday.

Guess I'll buy a pair of Vifa DQ25's instead.

[jkrutke]

As I was putting these back in their boxes I noticed something. The tweeter that's less sensitive seems to have the hexagrid cover much closer to the ceramic dome, by about 1 mm. I checked both tweeters' hexagrid cover's convexity and they're identical. It looks like the less sensitive tweeter is almost touching the cover with a little room, where the other has about another 1 mm or so gap. I'm not sure if this could affect anything but I figure I might as well post it

That could actually mean that one of the voice coils was partially out of the gap, which would explain low sensitivity, high Qts and higher harmonic distortion in one.

[Dennis H]

Except the 3rd harmonic is worse on the one with higher sensitivity. Very strange.

[eyekode]

Hrm... so we all know TB's graphs are optimistic. But they show the response rising at 20k. Your graphs are nose diving at 10k. I would not expect that from a ceramic dome...

[Mazeroth]

Hrm... so we all know TB's graphs are optimistic. But they show the response rising at 20k. Your graphs are nose diving at 10k. I would not expect that from a ceramic dome...

Not sure if you were saying my setup isn't working correctly or not but I did just measure three Peerless 810921 tweeters a few minutes ago (doh, left 1/3 octave smoothing on!) and they look almost identical to Zaph's:





Looks like I is good! :T

Note: They weren't optimally measured as I used a baffle from an RS28 and the Peerless flange was sticking out quite a bit from the baffle.

[eyekode]

Mazeroth, thanks for doing the test. I definitely appreciate it. In fact I think you just saved me ~70$ . I was just trying to correlate your findings with the spec sheet.

I know the details are never the same, but don't the overall shape/trend of the graphs usually match the documentation?

You know this tweeter originally did not have the hexagrid cover... I wonder if they did more significant redesign but failed to update the spec sheet? Still... don't they use ceramic's because they are light/stiff and don't start to give out until > 20k?

[Sakura]

What about the other Tang Band Ceramic tweeter?

This one: http://www.partsexpress.com/pe/showd...number=264-865

Anybody tested that yet?

[Rick Craig]

What about the other Tang Band Ceramic tweeter?

This one: http://www.partsexpress.com/pe/showd...number=264-865

Anybody tested that yet?

Yes - I just finished a center channel with one. Not too bad for the $.

[Jed]

Yes - I just finished a center channel with one. Not too bad for the $.

What model and could you summarize the performance? THD, FR, or a graph?

[owdi]

Four 22 uF capacitors (88 uf) were wired in series to protect the tweeters.

Four 22 uf caps wired in series is 5.5 uf, not 88. That's why you measured a bump in response at 700hz, then a dip. If you want the caps to sum, you have to wire them in parallel. Your measurements are invalid.


Dan

[Jed]

Four 22 uf caps wired in series is 5.5 uf, not 88. That's why you measured a bump in response at 700hz, then a dip. If you want the caps to sum, you have to wire them in parallel. Your measurements are invalid.


Dan

He wired 88 uf in series in the circuit. The 22uf caps were wired in parallel to eachother to make that value. His measurements are valid.

[Mazeroth]

Four 22 uf caps wired in series is 5.5 uf, not 88. That's why you measured a bump in response at 700hz, then a dip. If you want the caps to sum, you have to wire them in parallel. Your measurements are invalid.

Dan

If I had wired 5.5 uF in series you would see a drop off in response starting around 5 khz. if my memory serves me.

:T

EDIT: Yep, I was correct. Pulled a 6 ohm tweeter in PCD real quick and added a 5.5 uF cap in series. Sure enough, around 5 khz:



With the 88 uF it looks like this:

[owdi]

Four 22 uf caps wired in series is 5.5 uf, not 88. That's why you measured a bump in response at 700hz, then a dip. If you want the caps to sum, you have to wire them in parallel. Your measurements are invalid.


Dan

I guess I was wrong. I made that mistake once with a tweeter and my results looked similar. Oh well.

Mazeroth - do you have the impedance of the tweeter loaded in PCD? The rolloff from the series 5.5uF cap looks way too smooth.

Dan

[Mazeroth]

Sure did. I loaded the Seas T25-001 that comes bundled with PCD, since it and the TB are both ~6 ohms. I didn't measure the impedance of the tweeter as Paul was only interested in FR and distortion.

I guess I was wrong. I made that mistake once with a tweeter and my results looked similar. Oh well.

Mazeroth - do you have the impedance of the tweeter loaded in PCD? The rolloff from the series 5.5uF cap looks way too smooth.

Dan