Tweeter power handling ?

[5th element]

Originally posted by Dave Bullet


I was talkng about the older 812687 HDS tweeter. My statements are purely subjective with other tweeters I have, and also Zaph on his site mentioned "finally Peerless bring out a world class tweeter" (or words to that effect) - essentially meaning previous Peerless tweeter attempts have been sub-par.

PS: I think "burning in tweeters" at frequencies below their usable range is pointless. Isn't that like using 20KHz+ tones to burn in a woofer? (I mean - what is the point in burning in a driver outside the range in which it is used anyway?)

Cheers,
David.

Peerless have always made decent tweeters afaik, its just they have all been 'cheap' affairs. Certain models have been very well received, but nothing has ever really stuck out when the SEAS 27TFFC has always been so highly recommended - People have generally gone for that then trying something else. The HDS is at a much higher price point then the tweeters peerless has released before, and its stats do justify 'world class' certainly in comparison to what they have brought out before, and in comparison with much of the highly regarded benchmarks such as the SSD2905/95. World class, comparing with the best there is, probably not, but then you get what you pay for, even if it is that last 15% of performance.

RE burning in - I wouldn't try something that could be potentially destructive to a device. People often 'burn in' woofers with frequencies well below ones they would be able to reproduce, 20hz tones on 6.5" mid/bass drivers. This will serve to soften the suspension, but you can easily guage how much power you can throw at one as the excursion is easily visible, so destroying something is unlikely. I wouldn't mind betting however, that surround compliance is also somewhat frequency related, so a burn in at 300hz on a tweeter might get you 50% of the way there, but it would require being played with music to finish the job.

[youyoung21147]

Software signal generators can, when turned on and off, produce small turn on and turn off thumps. These, if applied directly to the tweeter, without any protection from DC, could EASILY exceed the given xmax or x'damage' parameter and thus cause the tinsel leads to break.

I've used no protection for DC at all, because I get no troublesome offset (~19mV) with my current setup. It is true what you say about thumps ; however they are quite well controlled on my generator and the tweeters didn't blow at this moment, but short after the signal starts to play ( something like one second)

You have calculated max rated SPL for a given xmax figure at any given frequency, however this is constant level

Yes, it is the max. level you can reach, and is not time-dependant. Transcients cannot exceed the max level your soundcard can put out. Let's say the Xmax/SPL numbers quoted are the absolute maximum you can achieve, and that even a transcient cannot exceed.

PS: I think "burning in tweeters" at frequencies below their usable range is pointless.

This will serve to soften the suspension, but you can easily guage how much power you can throw at one as the excursion is easily visible, so destroying something is unlikely. I wouldn't mind betting however, that surround compliance is also somewhat frequency related, so a burn in at 300hz on a tweeter might get you 50% of the way there, but it would require being played with music to finish the job.

Yes, it was to soften suspension. Burn-in was finished with normal music played on the tweeters. It is this way I have blown the Seas tweeter : I increased the volume higher than the Xmax-limited level, and thus it blew on a transcient.

The Seas tweeter, though it has a lower Xmax, is better designed because the leads are allowed to move with less stress and are attached to a foam piece rather than a rigid plastic piece.

[ucla88]

I think that you need to rethink your philosophy of testing.

running full range signals or a gated (0.5s) 300 Hz transient is really not at all the right way to do it.

Sorry

Why is a 0.5s 300 hz tone a bad idea? Well, because gating the tone introduces a wide spectrum of frequencies. If you base your xmax on a 300 hz tone, you are incorrectly simplifying the analysis.

see my- http://www.markk.claub.net/linear_distortion.htm

If the gating is gentle, it might not be so bad. A sharply gated signal would introduce multiple transients at a very wide range of frequencies. Bad idea. Hard to draw any conclusions about power.

Full range music is even worse. You really don't know the distribution of frequency energy and so don't really know exactly how far past xmax you're going.

Lastly, it's just not a real world signal. Extrapolating from your results and drawing conclusions about how much power you can get out of these at 1.5k is not appropriate.

Bottom line is that you need to make at the very least a simple 2nd order filter or higher, at 1.5k and drive the tweeter to realistic levels to get an idea of how the tweeter will perform.

Based on my tests, the Seas can be crossed at 1.5k. I don't think you can acheive 110dB transients, but 100-105dB at 1m single filtered unit is probably acheivable. I don't have quite enough info to say whether the peerless can do this...

[Rounder]

I have been watching this thread closely as i have plans to build a peerless two way with the HDS tweeter and the 6.5 inch hds exclusive woofer. My goal is to cross the tweeter at 2khz to avoid allowing the woofer to start beaming higher up in frequency. For the tweeter at two khz i was hoping for a first order electrical filter. Of course there would be a resistor for spl matching to the woofer and perhaps a notch to fix the impedance peak at fs. Based on everyone's experience is it safe to assume that the tweeter (HDS) can handle this low and shallow of a crossover? Oh and the first order filter would be based on the solen split not the butterworth. The distortion graphs say yes, but how about the excursion and thermal limits. If not then at what spl or wattage would i expect to harm the tweeter?


Thanks in advance.

[DSP_Geek]

Originally posted by Rounder
I have been watching this thread closely as i have plans to build a peerless two way with the HDS tweeter and the 6.5 inch hds exclusive woofer. My goal is to cross the tweeter at 2khz to avoid allowing the woofer to start beaming higher up in frequency. For the tweeter at two khz i was hoping for a first order electrical filter. [....] If not then at what spl or wattage would i expect to harm the tweeter?

First order filters eat tweeters for a couple of reasons:

1) Excursion goes up as the inverse square of frequency. For example, excursion at 1 kHz is four times that at 2 kHz for the same output. Your first order filter will be -3 dB at the crossover point, -7 dB one octave down, so combining the two factors above means the tweeter excursion at 1 kHz for your filter will be about three times that for 2 kHz, where typical tweeter power ratings are specified.

2) Spectral power distribution typically peaks around 120 to 250 Hz, rolling off at -3 dB / octave above that region. As a result a first order filter isn't rolling off power at -6 dB / octave, but with typical music the rolloff is closer to -3 dB / octave. Combine that with the excursion discussion above, and the tweeter really gets pounded.

You also need the tweeter and woofer to be quite flat at least two octaves, and preferably three, beyond the crossover point, otherwise the summation will not be as flat as you'd hope.

Most vendors specify their tweeters with second order filters, which have a fighting chance at controlling excursion, but a first order filter means your tweeter will have roughly triple the excursion for every octave down from the crossover frequency. It'll roll off before hitting the 250 Hz energy peak, however, and cause a dip in the mids when it does. I'd also expect the midrange energy to produce all sorts of amusing intermodulation products in the passband, too.

Long story short: I wouldn't do it.


Cheers,
Francois.

[sreten]

Hi,

note the excursion described above depends on the acoustic roll-off.

A ferrofluid tweeter with an Fs of 3K, Q ~ 0.7 (very common) and a
1st order filter (very common) works because the acoustic response
falls at 18dB/octave below 3khz, this applies to lots of cheap speakers.

Low tweeter Fs and low order c/o with low c/o point do not mix well.

/sreten.

[youyoung21147]

ucla88 : you're probably right, I've been asking too much to those tweeters.

The 300Hz tone was calculated to reach Xmax at 300Hz, that's true. Max music power was calculated to reach Xmax at 10Hz, and no tweeter blew with the music being played at the appropriate level. (but this level is so low : 4.2V RMS, roughly 2 or 3 watts)

When I have the opportunity again, I will test these units with useful frequency ranges (i.e 1500Hz and up) to see what they are worth in a real speaker.

Rounder and DSP_geek : The speaker in which I use these tweeters is exactly what you plan to build ! Peerless 830883 and tweeter. I have to choose which one fits the application the best.

My choice was a 24dB/oct 1880Hz active crossover. Why 24dB ? Because with 12dB, the highest excursion point is situated below the crossover point, thus leading to unuseful excursion. As I want the speakers to put out high SPLs, I couldn't choose 12dB/oct 1500Hz for reliability reasons.

I still have to figure if "domestic" tweeters can put out 110dB though... That would involve several dozen "real" watts input !

[AndrewT]

Originally posted by youyoung21147
...............I still have to figure if "domestic" tweeters can put out 110dB though... That would involve several dozen "real" watts input !

Hi Youyoung,
if the tweeter has a sensitivity of 102db/W then 6W gives about 110db SPL.

Many domestic tweeters are available in the sensitivity range 100 to 105db/W. They get you into the ballpark.

If you need high SPL then go horn loaded.
An example is the pepper pot horn fitted to the Tannoy tweeter. It works down to 1kHz and has a sensitivity reputed to be about 125db/W, but that is at the most sensitive frequency due to a very non flat response. A usable sensitivity after removing the peak is about 110db/W. Many other horn tweeters can match or even exceed this example.

Then you can go commercial!

[youyoung21147]

Many domestic tweeters are available in the sensitivity range 100 to 105db/W. They get you into the ballpark.

Which models are you thinking about ?? I'm really curious because most high efficiecy tweeters I know have high distorsion and a bad CSD.

I have bought the two dome tweeters because they are the lowest distorsion and cleanest CSD affordable tweeters I have found, despite their average sensitivity.

[DSP_Geek]

Originally posted by sreten
Hi,

note the excursion described above depends on the acoustic roll-off.

A ferrofluid tweeter with an Fs of 3K, Q ~ 0.7 (very common) and a
1st order filter (very common) works because the acoustic response
falls at 18dB/octave below 3khz, this applies to lots of cheap speakers.

Low tweeter Fs and low order c/o with low c/o point do not mix well.

/sreten.

You're right, as usual. I had the big Morels and Scan-speaks in mind when I wrote that - those have Fs well below 1000 Hz, so a first order xover at 2 kHz would beat them up pretty good.

As for the 3 kHz tweeter you mentioned, it makes sense to have the Q above 0.7 so the crossover could be a true 3rd order Butterworth with one pole contributed by the 1st order passive. Hmmm. That's a cool way to save a buck and still have it sound decent. Your manufacturing has to be quite consistent, however.


Francois.