Those are honkin big ultrasonic tones in that study though, and produce a ~3.5kHz dissonant and unmasked IM tone only 55dB down from the others - and in an ear-sensitive hearing range. I don't see any program material exciting an ultransonic resonance anywhere near that amount. Looks to me like it only shows that people can hear IM generated tones if strong enough and in the right place!
It establishes that the phenomenon is audible and that such an extreme example is audible to 100% of a random sample of people. That's the important take away from the study and why it was presented. Science isn't all or nothing; it's small steps at a time.
The threshold of audibility is another matter, but if that random selected group becomes one that's more trained in discerning minor differences and spends hours each day listening to the speakers, say like audiophiles and DIY enthusiasts, a less extreme situation may still be audible if only in terms of increased fatigue. Even if only to 20% of the group in that case, that's still significant enough a situation to warrant consideration in transducer selection choices.
Such a high resonant peak can indeed be excited fairly easily by harmonic information and one so relatively low in frequency will be faced with more than a few potential sources of spurious excitation. Corrundum is known for it's high tendency to ring. (Weber Bar gravitational wave detectors that rely on bell like ringing of high Q materials use aluminum masses for economic reasons, but artificial sapphire was considered the ideal material in development were the funds available.) 27kHz, for instance, is only the third harmonic of the range where sibilance is most pronounced and, of course, cymbals. IMD from that could make for very spitty and hissy 's's and cymbals.
Yep, I can make a UV-curing lacquer that doesn't cure with certain wave length. That doesn't mean UV-curing lacquers are bad! Just bad engineering or comparison...
I think there is a major danger here of misinterpretation on all fronts.
To correct one misconception: I do not see anybody here expressing a 'fear' of metal dome tweeters of any kind, let alone based on expectation bias. Speaking for myself, I actually prefer hard dome tweeters on principle. However: having said that, I do not simply assume that high Q resonances close to the audio band are automatically a non-issue & that potential consequences resulting from them can be casually ignored, without regard to specific design or varying circumstances. That is not an unreasonable perspective, and one shared by a number of extremely competent designers, not least of which being Roger Russell, who is not known for holding wild subjectivist views devoid of engineering common sense. Is it likely to be a major problem in a quality tweeter? No. Is it still a factor worth considering when evaluating drive units for a new project, even if others may dominate? Yes.
To correct one misconception: I do not see anybody here expressing a 'fear' of metal dome tweeters of any kind, let alone based on expectation bias. Speaking for myself, I actually prefer hard dome tweeters on principle. However: having said that, I do not simply assume that high Q resonances close to the audio band are automatically a non-issue & that potential consequences resulting from them can be casually ignored, without regard to specific design or varying circumstances. That is not an unreasonable perspective, and one shared by a number of extremely competent designers, not least of which being Roger Russell, who is not known for holding wild subjectivist views devoid of engineering common sense. Is it likely to be a major problem in a quality tweeter? No. Is it still a factor worth considering when evaluating drive units for a new project, even if others may dominate? Yes.
Last edited:
The rolloff on this tweeter is extraordinary. At 500hz, it's only 7DB lower than at 1000hz.
This will make it *really* waveguide friendly.
With a 7" deep waveguide you could probably get it down to 500hz.
On the downside, I've always found it difficult to get metal domes to behave on waveguides. Due to their rigid diaphragm, you often get a notch around 15khz.
This will make it *really* waveguide friendly.
With a 7" deep waveguide you could probably get it down to 500hz.
On the downside, I've always found it difficult to get metal domes to behave on waveguides. Due to their rigid diaphragm, you often get a notch around 15khz.
:
Huh? How can a different material make a notch? Is it because the dome radiates differently than a soft dome? I have to check a German producer using a Scanspeak dome in a waveguide when I get back home, not sure if it's titanium. It has a totally flat response to 20kHz...
Any dome that suffers a diaphragm resonance, will at some frequency suffer breakup such that the middle of the dome is moving in the opposite direction of the outside of the dome. This will cause a dip/notch in the frequency response. The higher the Q the resonance the sharper this notch will be.
This is similar to a "surround dip" in a cone driver where the surround is flapping in the opposite direction to the cone of a driver causing a dip in the response. Some resonance modes cause peaks in the response, some can cause dips if they result in destructive interference with radiator from other parts of the cone that are not resonating.
Last edited:
Not really any more impressive frequency response than something like the SB29RDC. Ultimately what limits how low you can cross is the amount of linear excursion (xmax) not the resonant frequency or how much rolloff there is in the frequency response. A tweeter with 1mm xmax can theoretically cross at half the frequency of an equivalent driver with 0.5mm xmax. It however gets more complicated than that because there are no set standard for how xmax is measured and different drivers have different behaviours as excursion increases. Some explode into bulk distortion as xmax is exceeded while others gradually increase instead. There are already tweeters out there with 0.5mm xmax, so with 0.7mm xmax this one does not bring a heck of a lot more excursion to the table and therefore you don't expect to cross a great deal lower.
A waveguide just nets you another 6dB or so output for the same excursion, which buys you maybe a couple hundred Hz lower crossover point at the same output level.
I'm not confident that this tweeter will cross acceptably at 500Hz at a decent listening level. Already being about 6dB down at 500Hz implies a crossover which gives no attenuation above 500Hz. Pretty demanding...
In reality I think the practical crossover point for a decent output level (100dB/1metre) will be 1-1.5kHz LR4.
@peterbrorsson re #43:
no arguments on UV-curing at all, but it does mean I'll ask you about UV-blocking selectivity in various lacquers or other finishes.
there once was, apparently, an old woman who sold a UV-blocking something online,
it worked quite well at keeping Osage Orange orange long past what is normally the case (natural sunlight UV exposure=brown). Woodworker sites concluded that she must have passed, no one came forth with a name, location, or story, teeth were gnashed because the product disappeared and stayed that way. Any thoughts?
No blame whatsoever if you choose to avoid this topic, I won't be chasing you down insisting on an answer, but some might,
Quard
no arguments on UV-curing at all, but it does mean I'll ask you about UV-blocking selectivity in various lacquers or other finishes.
there once was, apparently, an old woman who sold a UV-blocking something online,
it worked quite well at keeping Osage Orange orange long past what is normally the case (natural sunlight UV exposure=brown). Woodworker sites concluded that she must have passed, no one came forth with a name, location, or story, teeth were gnashed because the product disappeared and stayed that way. Any thoughts?
No blame whatsoever if you choose to avoid this topic, I won't be chasing you down insisting on an answer, but some might,
Quard
It's actually worse than that. Excursion for constant SPL is proportional to the inverse square of frequency, not the inverse.
So to produce the same SPL at 500Hz as you were producing at 1Khz takes four times the excursion, not twice the excursion as you're implying.
So doubling your excursion capability from a tweeter that could go down to 1Khz would only allow you to drop your usable frequency by a factor of 1.4, eg, about 707Hz, not 500Hz.
As for the resonance frequency "not mattering" that is also not true. While you can EQ away the resonance in terms of frequency response, non-linear distortion is going to be highest at resonance, and that can't be disguised or removed with EQ or shaping of the crossover.
This is what causes the "buzzy" nasally distortion of many dome tweeters when used through their resonance, even if they have a resonance compensator/notch to correct the frequency response through the resonance region.
Also, distortion is always higher below the mechanical resonance of a driver than it is an equal distance above it, not just because required excursion is greater, but also because below the mechanical resonance the driver is in the "compliance controlled" region.
This means any non-linearities in the compliance of the suspension directly translates to non-linear movement of the cone and non-linear distortion in the produced sound, even if you had a perfectly linear motor that is providing a perfectly linear driving force.
However above the mechanical resonance the driver is in the "mass controlled" region. This means that it is the drivers mass and inertia that is controlling the movement of the cone in response to the force from the motor.
Once you are beyond the the main peak of the resonance on the high side of the resonance any compliance non-linearity of the suspension no longer matters and does not contribute to non-linear distortion, thus the driver distortion will drop significantly.
This is one reason I prefer to cross midrange drivers and particularly tweeters over an octave above their fundamental resonance - so that they are not producing significant SPL either at their resonance or in the compliance controlled region below it where distortion will be higher.
This is independent of and in addition to the effects of greater cone excursion being required for lower frequencies.
In the case of a midrange driver you can reduce the effect of this by choosing a driver with a very compliant suspension and then putting it in a relatively small closed box, so that the box dominates the total compliance effectively making the midrange driver an acoustic suspension design.
You can't really do that for a tweeter though, and particularly in the case of a vented pole tweeter there is no acoustic suspension effect, with the compliance all provided by the suspension of the driver, hence distortion is going to be higher at and below resonance no matter how good your motor linearity is.
Last edited:
Yes, I'm aware of this. But I've never seen a severe high q dip in the audio band of a metal dome. However that was not the issue, question was why would it be a notch when using metal domes in a wave guide...
As I understand, the notch is normally caused by bad transition at the throat and also by dome shape.
Hi Quard,
I'm afraid I can't help you. I'm working in the professional wood working paint business. We buy all of our chemicals directly from raw material suppliers. But check if you can find Tinuvin products in retail market. There are several types that are compatible with different lacquer types. Although, the binders used should also be optimised for best effect.
Wow, who uses midrange drivers in ported enclosures ?
Terrible idea!One of the disadvantages of a ported 2 way system over ported 3 way or sealed 2 way is the pipe resonance effects in the midrange caused by the ports.
It's a pretty serious problem and I actually moved the port on my 2 ways from the front to the rear because the midrange colouration from the port was so audible and easy to measure. (At the peak at 700Hz there is actually more output from the port than from the driver, despite the cabinet being well lined!)
So why would you want to do that on a 3 way's midrange enclosure where freedom from colouration in the midrange is top priority ?
Other issues would include no control of cone movement below the port tuning, so the driver would move visibly with bass unless you had a very steep electrical filter, (while a closed box acoustic suspension configuration would prevent any driver movement at bass frequencies even with a very simple filter) and you would not be able to stuff the enclosure to minimise standing waves through the midrange region, which is also very important.
My ideal midrange design, if I was using a relatively large driver say 6 or 8" would be to have a driver with very high Vas, low Fs, very low Qts, and place it in a sealed enclosure sized to bring it's mechanical resonance up to one octave below the crossover frequency, and stuff that enclosure well to absorb standing waves at higher frequencies. If the Qts is low enough you should be able to get a near flat response before mechanical roll off without a major peak.
On a driver of those sizes you'd be looking at something like 5-10 litres just for the midrange enclosure, but in a large 3 way with say 80 litre bass enclosure it would be perfectly workable.
Last edited:
The midrange cavity in the B&W Nautilus is not open or vented so is not aperiodic. The interior cavity is shaped like a teardrop with a tapering tube behind the driver that terminates some distance behind the driver, but it is most definitely sealed.
The idea behind this shape is to eliminate any direct, single reflection from a rear enclosure wall that would cause a standing wave. The back wave from the driver travels down the tapering tube reflecting at constantly changing angles, which I guess disperses them in time, hence no obvious standing wave at one frequency. I don't think there is any damping material inside.
From the point of view of the fundamental driver resonance it's just a sealed box however. The shape is all about controlling internal standing waves at higher frequencies.
Personally, while I'm sure it works I think it's a lot easier and cheaper to just use lots of stuffing! I certainly can't build something that shape in my garage.
Last edited:
Right you are, I stand corrected
While that may be theoretically true, in practice the only real significance of the Fs is that you get a 12dB/octave rolloff below it, which at some frequency begins to reduce excursion and therefore the amount of distortion. If you view the magnitude of distortion components relative to the fundamental, distortion actually peaks below Fs because this is where excursion is maximum. Depending on the Qts, this is around 1-2 octaves below Fs.
Consider this (yes, I know it's not a tweeter but it's basically an enormous tweeter and all the tweeters i've measured behave in the same way):
Fs is 250Hz or so. Distortion seems to be peaking around there right? (ignore distortion below about 80Hz, that's a windowing limitation near the beginning of the test sweep)
Now we are looking at the magnitude of distortion relative to the fundamental. Peak distortion is now at about 100Hz. If I filtered the response to flat to eliminate the low frequency rolloff, the distortion would just keep going up with decreasing frequency because the required excursion is just getting higher and higher. There will be absolutely no evidence of a 'hump' in distortion at Fs. Any 'hump' being near Fs even after filtering into a target slopes is merely a coincidence of where Fs has been engineered to be most effective. If Fs is too low, sensitivity is compromised. If Fs is too high, the frequency response is considered unusable.
To accurately compare two tweeters, you really need to filter both of them to achieve the exact same frequency response, then measure the distortion with the filter in place. That way you totally eliminate the effect Fs has on frequency response, therefore excursion and distortion. Without doing this, the best you can do is make an educated guess. If you look at Zaph's measurements of the Peerless HDS 810921 and the Dayton RS28A (link), the HDS look like it has way lower distortion than the RS. If you look at the frequency response of the HDS it's already starting to roll off at 2kHz while the RS is virtually flat to about 800Hz. That is the reason for the massive difference in measured harmonic distortion below 2kHz.
I happen to know from personal experience that these two drivers are very very similar in distortion once you filter them to around 1.8kHz LR4.
Last edited:
I wonder how the peerless compares with the new dayton aluminum dome.
Dayton Audio RST28A-4 1-1/8" Reference Series Aluminum Dome Tweeter 4 Ohm
Also the sbacoustics aluminum dome...which has unusually good dispersion
SBAcoustics-tweeters
I was thinking about cutting out my aurasound whispers and going with one of these newer domes that seem to be able to handle about a 1000 hz crossover all day long (at least with my brickwall crossovers that I swear sound better)
One of the things is that I suspect a lot of these tweeters are overdamped, which is pretty typical of tweeter design since they have less of a direct mechanical connection to the coil. Then again I haven't heard too many really nice domes so I wouldn't know.
Low volume performance is under appreciated these days.
Dayton Audio RST28A-4 1-1/8" Reference Series Aluminum Dome Tweeter 4 Ohm
Also the sbacoustics aluminum dome...which has unusually good dispersion
SBAcoustics-tweeters
I was thinking about cutting out my aurasound whispers and going with one of these newer domes that seem to be able to handle about a 1000 hz crossover all day long (at least with my brickwall crossovers that I swear sound better)
One of the things is that I suspect a lot of these tweeters are overdamped, which is pretty typical of tweeter design since they have less of a direct mechanical connection to the coil. Then again I haven't heard too many really nice domes so I wouldn't know.
Low volume performance is under appreciated these days.
Last edited:
There are distortion measurements posted of the old and new RS28 drivers and the DA25TX on the parts express forum.
Piecing together that data with Zaph audio's (assuming SB26 silk is comparable to SB26 alu), from what I can gather we are looking at a non-linear distortion ranking something like this:
DA25TX > SB29 > Old RS28 (Usher 9950) > New RS28T > SB26
What Troels data doesn't show is how the SB26 has worse high order distortion than the SB29.
Piecing together that data with Zaph audio's (assuming SB26 silk is comparable to SB26 alu), from what I can gather we are looking at a non-linear distortion ranking something like this:
DA25TX > SB29 > Old RS28 (Usher 9950) > New RS28T > SB26
What Troels data doesn't show is how the SB26 has worse high order distortion than the SB29.
None of the drivers discussed will handle a 1kHz brick wall at a high level (100dB/1m). Perhaps a 2kHz brickwall would be doable, or 1.5kHz LR4. The DA25TX or SB29 might do a brickwall at 1.5kHz or 1.2kHz LR4 acceptably.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.