Browsing for a tweeter for a new build. Notices many spec sheets stop at 20K, so their breakup behaviors are not shown. To me, this is almost a scam. A lie by not telling the full truth.
We can mitigate it some by use of a Zobel or more aggressive LP filter, but I still want to know what I might have to deal with.
Why does a 22K breakup matter? Harmonics, IM and resulting audible HD.
We can mitigate it some by use of a Zobel or more aggressive LP filter, but I still want to know what I might have to deal with.
Why does a 22K breakup matter? Harmonics, IM and resulting audible HD.
Last edited by a moderator:
Another point of view is that only children can hear to 20 kHz. Most adult's hearing stops at 16 kHz. Men over 40 are doing good if they can still hear above 12 kHz.
All HF drivers have breakup modes, some are above 20k, most are below 20k. If a buyer is not aware of this fact, whose fault is that? not the manufacturer's.
A manufacturer has a duty to the customer to be factually correct in their specs and published literature. They have a duty to their employees and shareholders to present the facts in the best possible light. That is not lying, that is selling. They do not have no obligation to point out how their product is inferior to the competition, or to talk a potential buyer out of buying...
All HF drivers have breakup modes, some are above 20k, most are below 20k. If a buyer is not aware of this fact, whose fault is that? not the manufacturer's.
A manufacturer has a duty to the customer to be factually correct in their specs and published literature. They have a duty to their employees and shareholders to present the facts in the best possible light. That is not lying, that is selling. They do not have no obligation to point out how their product is inferior to the competition, or to talk a potential buyer out of buying...
Well, I am male in my 60's and can still hear well into 16 to 17K, and can to 18 but that is way down there. I protected my hearing. Maybe someday they will have a pill that thins the inner ear fluid back down restoring our HF hearing.
Of course all drivers have breakup modes. My point is some manufactures are not disclosing them. Some do.
For instance, I have never liked Morel tweeters, Not sure why, but they are one who stops their spec @ 20K. You can look at SS, Seas, SB, Peerless and even Dayton and they tell you what you need to deal with.
I think a major point is missed here. It is not hearing a harmonic in the 20's, We cant. ( I could when I was 20) It is the IM resulting in HD down where we can hear. Kind of the opposite of the doppler distortion issue with woofers. This is why passive crossover LP filters work.
Of course all drivers have breakup modes. My point is some manufactures are not disclosing them. Some do.
For instance, I have never liked Morel tweeters, Not sure why, but they are one who stops their spec @ 20K. You can look at SS, Seas, SB, Peerless and even Dayton and they tell you what you need to deal with.
I think a major point is missed here. It is not hearing a harmonic in the 20's, We cant. ( I could when I was 20) It is the IM resulting in HD down where we can hear. Kind of the opposite of the doppler distortion issue with woofers. This is why passive crossover LP filters work.
Breakup modes can cause sub-harmonics to appear, which are more audible.
However, music rarely has any content at 20 kHz to excite this in the first pace.
However, music rarely has any content at 20 kHz to excite this in the first pace.
Said the bat which hears up to 200 kHz 😉
For us Humans? ... not that high.
20kHz covers any Hi Fi interested Human.
I prefer to find independent tests of drivers. HiFiCompass (LINK) and Timothy Feleppa (LINK) are two popular sites that do extensive driver tests. A couple other sites are zaphaudio.com and audioexcite.com. The only caveat is that you can't really compare between sites since each of them have their own test methodologies. But much, much preferred over the manufacturer data.
NOT music content. Harmonics. HARMONICS as I said in post 1. I discovered this issue when understanding the differences in miller compensation vs dominant pole compensation in power amp design and how the interaction with the speaker can influence the choice in design.
I use specs as a purchase decision. I then do my own measurements. I can't afford to buy a dozen drivers that look good @ 20 to see what I am dealing with above. I want a clue. John measured a lot of tweeters, but not all I am interested in.
I use specs as a purchase decision. I then do my own measurements. I can't afford to buy a dozen drivers that look good @ 20 to see what I am dealing with above. I want a clue. John measured a lot of tweeters, but not all I am interested in.
That is an old dispute, between metal domes and soft domes. Soft domes never have that breakup, or very negligible. Some metal domes has it in tens of decibel. How audible is it? Not sure, but I can not stand metal dome. All metal dome speakers I ever built are gone. Gone. Most soft dome speakers I built I still have.
Ah, no. That is folklore. Some soft domes have horrible breakup. Likewise, some metal are bad, some much better. In my current living room, I replaced the Seas soft dome with the Al/Mg version of the same motor and it is much smoother. My next build, I am eyeing both the CSS soft and the Peerless ceramic. All drivers have challenges. It is how you use them. Have you heard the same metal dome with and without a LP filter? You may change your mind.
Well, that is why there are more than two tweeters on the market. We each can have our pick. Others will never use a fabric dome because of "soft dome sizzle" Also, a feature of individual drivers not of the material. In the out-of-my-even-looking-at-for-fun range are Be domes. Beryllium is a very "dead" metal. Material is just one choice and not an absolute end factor. I suggest taking each driver on it's own merits. Much easier if we have a plot to at least 30K and not masked by misleading cutoffs. That is my point.
I understand that subharmonic and IM distortion is possible in theory when a diaphragm is in resonance, but do we have any measurements of real drivers to confirm this?
Now hi-fi news is doing that in speaker reviews. Scary, huh? Would you really want a tweeter that does that?
https://www.hifinews.com/content/monitor-audio-silver-500-7g-loudspeaker-lab-report
Well there is Dual Ring Radiators that are Flat to 40K and the datasheet shows response to 40K
And there is numerous ribbon tweeters with response to 40k and datasheet also shows 40k
actually some ribbons have datasheet with response to 50k
and then you will see that they go to 40k and response drops off very quickly after that.
Same with all the usual " 20k " tweeters response drops off very quickly around 18 to 22k
so pointless to show on datasheet.
Far as harmonic distortion, some datasheets have it, but usually at 1 watt power.
And there is numerous ribbon tweeters with response to 40k and datasheet also shows 40k
actually some ribbons have datasheet with response to 50k
and then you will see that they go to 40k and response drops off very quickly after that.
Same with all the usual " 20k " tweeters response drops off very quickly around 18 to 22k
so pointless to show on datasheet.
Far as harmonic distortion, some datasheets have it, but usually at 1 watt power.
With multi-tone testing. One needs to contrive a representative signal. You could quantify it, or as I suggest, block it. The difference you hear is with high level instruments, like a trumpet, full string section, or really hitting the cymbals having more "glare" or "harshness" in the 2 to 4K region. Subjective imprecise words but that is what we hear. I know if I block it, it sounds smoother. It is only a little annoying to me, but my wife is super sensitive to it. Unfortunately, that is only a half fix as the tweeter itself produces the harmonics too and "self excites" So, we can filter harmonics from source, DAC, amp, but we need lower distortion tweeters to start with. The trend is lighter and more rigid diaphragms so as to push the breakup higher and higher. Note the Monitor is in the 30K region where most traditional down in the 22K or so. he higher the breakup frequency, the lower the harmonics that might excite it and just a guess, the IM being up out of our most sensitive range. Big guess.
The CLS plot is informative if you know enough about drivers. A bit misleading, or terrifying, if you do not.
With all respect to Mr. Pass, yes the first Watt is important, but 1W distortion numbers being good are meaningless. Bad are, well bad. 5 or 10W are far more informative for a tweeter, if it can handle it. I measure my systems at the level that should be 1/2 X-Max for woofers and midranges as I know by experience, that is about where the whole curve starts to climb. If you reach Fs it will have skyrocketed. In a living room, 1/2 X-Max is really loud and above that would be transients I am not sure we could hear if distorted. My unscientific design rule of thumb is 1/2 @ 96dB, 1M. That will usually give you the low frequency use target.
RE: some do.
Yes SOME provide the data. I said that and provided examples. Some do not and that is what I consider deceptive by being incomplete.
I really don't care what $800 ribbons do or don't do. Their distortion is too high for us in the first place.
Curious, which ring claims to go to 40K? I have used the XT25 and the SB.
RE the Monitor tested.
No, I would not want a speaker that does that If it is a complete speaker system, I can hear it. ( unless you are brave enough to buy a speaker without hearing it. . If DIY you can't hear it until you build it into a system. I want to know what the driver does that so I can choose to design around it, or choose a different driver.