I used to read somewhere that there might be a problem if the crossover point of the tweeter was set too high. Firstly, is it true? Secondly, what is that symptom/phenomenon called?
In addition, from my experience with experiments of the combination of 2” dome midranges and 3/4” tweeters, I preferred the upper midrange or lower treble coming from the mid rather than the tweeter. Finally, so how high is too high?
In fact, I used to have an idea to design the mid-to-high crossover for the 2” mid and 3/4” tweeter combo by using mechanical low-pass filter of the 2” mid and electrical high-pass filter at somewhere in the upper range of the tweeter, or somebody would recognize the 3/4” tweeter as a “super tweeter“. But I’m not sure about the possibility of this concept.
In addition, from my experience with experiments of the combination of 2” dome midranges and 3/4” tweeters, I preferred the upper midrange or lower treble coming from the mid rather than the tweeter. Finally, so how high is too high?
In fact, I used to have an idea to design the mid-to-high crossover for the 2” mid and 3/4” tweeter combo by using mechanical low-pass filter of the 2” mid and electrical high-pass filter at somewhere in the upper range of the tweeter, or somebody would recognize the 3/4” tweeter as a “super tweeter“. But I’m not sure about the possibility of this concept.
Last edited:
I would think that totally depends on the other drivers and their characteristics.
Maybe if the woofer doesn't like being crossed higher than 'normal' (say, if you chose 3,000 Hz for a Dayton RS180 or similar metal driver) you risk distortion, peaks, resonance and other undesirable sounds from the woofer. Cross that driver at 1,800 - 2,000 Hz, which is where many successful projects cross it, with the appropriate filter, and you don't get those issues.
It's also possible to cross a tweeter too low, but that's not your question.
Geoff
Maybe if the woofer doesn't like being crossed higher than 'normal' (say, if you chose 3,000 Hz for a Dayton RS180 or similar metal driver) you risk distortion, peaks, resonance and other undesirable sounds from the woofer. Cross that driver at 1,800 - 2,000 Hz, which is where many successful projects cross it, with the appropriate filter, and you don't get those issues.
It's also possible to cross a tweeter too low, but that's not your question.
Geoff
Hi,
the lower bandwidth transducer coverage gets narrower the higher the frequency (shorter the wavelength). This happens with all transducers also with small tweeters, just higher in frequency. This is commonly referred as beaming and is due to sound wavelength getting short relative to diameter of the transducer.
If you think woofer and tweeter at crossover frequency, the woofer having bigger diameter might already have narrow coverage, but the tweeter has wide coverage because it is smaller in size. As you move crossover frequency up, woofer coverage gets narrower and narrower while tweeter stays wide, and this results uneven horizontal coverage for the system.
Another issue with driver size is cone breakup, which usually starts right about where the coverage narrows as well, when wavelength gets smaller than the transducer size. The cone does not work as a piston anymore, but for example the surround could move in opposite phase than rest of the cone. Higher up in frequency all kinds of modes occur and parts of the cone resonate and stuff could get really chaotic. In general the response could look like a mountain view on highs, and vary to all directions quite wildly. So, in addition to coverage narrowing explained above, which is with ideal piston, the real transducers are not ideal but have the breakup in addition and the sound just gets less usable in technical sense as frequency goes up.
Tweeters, mids, woofers all have the same issues just in different frequency bandwidth, which relates to transducer size by sound wavelength. Geometry and composition affects how closely any driver behaves like ideal piston.
Similarly all transducers become small compared to wavelength when frequency gets lower and lower, so maintaining response on lows requires more and more excursion, which limits response of any transducer on the low end SPL. Thus you can basically slide crossover between any two transducers based on all these parameters, make it as high as possible for max SPL capability, but not too high where you get audible issue with system directivity or conebreakup.
the lower bandwidth transducer coverage gets narrower the higher the frequency (shorter the wavelength). This happens with all transducers also with small tweeters, just higher in frequency. This is commonly referred as beaming and is due to sound wavelength getting short relative to diameter of the transducer.
If you think woofer and tweeter at crossover frequency, the woofer having bigger diameter might already have narrow coverage, but the tweeter has wide coverage because it is smaller in size. As you move crossover frequency up, woofer coverage gets narrower and narrower while tweeter stays wide, and this results uneven horizontal coverage for the system.
Another issue with driver size is cone breakup, which usually starts right about where the coverage narrows as well, when wavelength gets smaller than the transducer size. The cone does not work as a piston anymore, but for example the surround could move in opposite phase than rest of the cone. Higher up in frequency all kinds of modes occur and parts of the cone resonate and stuff could get really chaotic. In general the response could look like a mountain view on highs, and vary to all directions quite wildly. So, in addition to coverage narrowing explained above, which is with ideal piston, the real transducers are not ideal but have the breakup in addition and the sound just gets less usable in technical sense as frequency goes up.
Tweeters, mids, woofers all have the same issues just in different frequency bandwidth, which relates to transducer size by sound wavelength. Geometry and composition affects how closely any driver behaves like ideal piston.
Similarly all transducers become small compared to wavelength when frequency gets lower and lower, so maintaining response on lows requires more and more excursion, which limits response of any transducer on the low end SPL. Thus you can basically slide crossover between any two transducers based on all these parameters, make it as high as possible for max SPL capability, but not too high where you get audible issue with system directivity or conebreakup.
Last edited:
1. the Low frequency driver will start beaming the higher you go.
2. Phase coherency in vertical axis will decrease the higher your crossover frequency.
2. Phase coherency in vertical axis will decrease the higher your crossover frequency.
Yes, excellent addition, minimum distance between the transducers relative to wavelength at crossover, where both transducers have significant output, determines interference in vertical axis. This would be audible issue especially if listening at close proximity with varying listening height. Perhaps less meaningful if listen far so that early reflections dominate. Or at fixed height at any distance, and it shoudn't matter as much.
I haven't noticed any cons. My tweeters recommended 4KHz cross overs, but I only wanted a bit of air above my full range drivers to balance my subwoofer, so I used components for a 6 KHz cross over. I'm happy with my choice.
Others have more or less covered it and I wrote about it in a blog post here.
Worth noticing, based on off-axis measurements from Stereophile that many commercial speaker makers seem to have reduced output off-axis around 2.4 kHz, a typical mid to tweeter crossover point. Focal and B&W are a couple of brands for which there are many off-axis plots that seem to demonstrate this aesthetic.
I have to wonder if this is not a deliberate choice because I once read here that reduced output ~ 2.4kHz may lead to better perceptions of imaging. I've never attempted this kind of sweetening of a design so I can't tell you how well it works or not.
Worth noticing, based on off-axis measurements from Stereophile that many commercial speaker makers seem to have reduced output off-axis around 2.4 kHz, a typical mid to tweeter crossover point. Focal and B&W are a couple of brands for which there are many off-axis plots that seem to demonstrate this aesthetic.
I have to wonder if this is not a deliberate choice because I once read here that reduced output ~ 2.4kHz may lead to better perceptions of imaging. I've never attempted this kind of sweetening of a design so I can't tell you how well it works or not.
Last edited:
Yes Erik, it does seem good intentioned. I'm inclined to suggest that it's always a good idea to keep a balance on/off. If it needs reducing, do it overall.
AFAIK, it's sometimes done in an attempt to compensate for mismatched polars through the transition band & the flare as a narrowing mid / LF unit unit shifts to a smaller dome / whatever. Not an approach that I'd normally favour myself -I'd rather get a better polar match in the first place than attempt sticking-plaster compensations, but in fairness there are / can be plenty of other factors at work, be they acoustic or commercial.
AFAIK it dates back to Harwood (often misattributed to his then-BBC employers), and was advocated by him at the time, particularly (not exclusively) as a means of compensating in particular for close-mic'd 'live' recordings of orchestras / individual instruments in an orchestra or similar, which could skew the balance unrealistically. The better solution to that last was to change the recording techniques, obviously, but that took time & wasn't / isn't always possible, so he argued some subtle shifts from a nominal flat axial balance in the reproduction chain had some merit, for that & other objects. I often fire up my Gerzon / Hafler matrixed quad rear channels* for similar reasons, even if the hi-fi police come over all faint.
* well worth anybody's time to experiment with. Even if you don't like the results, if nothing else it gives an interesting insight into stereo & how different released are recorded / mixed.
Last edited:
That's definitely possible as 2-way speakers need compromises, but still see it in multi-way speakers that would otherwise not need to do this.
Along these lines, I think he's on DIY too, Petr Koukorek did an analysis of the B&W 802 D, and the original measurements and his changes go a long way to explaining their choices. I wrote about it here but Petr's original article is here. In this case it seems a poorly performing tweeter is at least partially to blame, but given the 3-way nature of the speaker the off-axis performance is otherwise unexplained.
I recall Petr's measurements from the time; it doesn't make for particularly pretty viewing. Mind -I'm not pointing fingers as I appreciate some of the commercial realities (I've been there / sometimes am there) & I've had production speakers that don't look any prettier -which wasn't my choice, although they worked quite well within the intended context. Be that as it may -they weren't retailing for this sort of sum, and that single-cap HP is completely inadequate given the tweeter's own response -I've seen far worse HF units, but it needs hammering into some better shape than they gave the finished speaker if performance is the goal. But of course, this is [post Wilson] hi-fi, so it isn't, and hasn't been for about 40 years.
I weep for my lost innocence. 🦕
I weep for my lost innocence. 🦕
There are advantages to crossing over higher as well.
1. decreases the value of parts needed, like capacitors , inductors etc.
2. reduced parts values keeps the dynamics intact, also tends to image better... (inductors by design impede ac current and tend to round over sharp spikes in AC current ie Dynamics)
So, it all ends up being a compromise between dispersion characteristics (objective) and better sound quality (Subjective).
1. decreases the value of parts needed, like capacitors , inductors etc.
2. reduced parts values keeps the dynamics intact, also tends to image better... (inductors by design impede ac current and tend to round over sharp spikes in AC current ie Dynamics)
So, it all ends up being a compromise between dispersion characteristics (objective) and better sound quality (Subjective).
Possibly. Although depending on the specifics, you may need more components, so these things can balance out.
On that I'm dubious. Thinking off the cuff, it should be straightforward to measure though. All you'd need to do is to EQ a driver flat under a given set of conditions, and then for a given amplifier setting, measure the increase in SPL for dynamic peaks with & without a series inductor or more complex low pass in place. 2nd order Butterworth is traditionally a good standby for many comparisons & testing purposes, so should be reasonable here too. For a fair comparison, the peak / peaks measured should be at some frequency below the nominal transition band. I can't say I've really found any issues myself on this front that aren't going to be [more] apparent in other areas, but unlike you, I haven't actually performed this test, so that's only a view I'm happy to revise in light of evidence. What sort of level changes in peak values have you measured? Were they more significant at different frequencies or did the values end up roughly averaging out?
Dispersion characteristics are certainly objective, but since they are directly related to subjective impressions ('sound quality') I don't think this needs to be considered as something that's separate.
I certainly do not have equipment that is able to measure the instantaneous peaks. But I did try a lot of crossover parts and listening tests. with proper phase matched designs. Just relied on hearing and a couple of friends to avoid creator bias.
Performed tests with different gauge inductors, at different crossover points. even series and Parallel crossovers. previously mentioned pionts are the take away points of my experiments.
Performed tests with different gauge inductors, at different crossover points. even series and Parallel crossovers. previously mentioned pionts are the take away points of my experiments.
I recommend using the Vistaton boxsim software to model your speaker. You have to pick from Visaton drivers or type in the TS parameters and then enter the dome diameters to get the dispersion pattern for each driver. That will show you the trade off of dispersion and maximum SPL. As you run a tweeter to lower crossover frequencies you run into Xmax limiting the maximum SPL. As others have mentioned, as you run the midrange to higher frequencies the voice coil inductance will result in a first order roll off, the dispersion will narrow and eventually you will get to breakup. The software has excellent graphs of the radiation pattern with frequency to show what is going on. The software includes modeling of digital filters, so it is quick and easy to change crossover frequencies using those without needing to calculate part values. When the software loads it has a two way design in there. You can just move the crossover frequency of that design around to see the effect with the 5 inch woofer and 1 inch tweeter.