Large vs. Small midrange

[boris81]

I would like to have a discussing on why 8" drivers are often preferred for midrange duties over 4" ones.

The superior performance of large area drivers in terms of sensitivity, distortion and power handling is obvious so it's no use to prove that point here.

I'm mostly interested in the issues of driver lobing and maintaining good frequency response off-axis. Using a smaller midrange is the best way to achieve uniform dispersion through the crossover region, yet it's hard to ignore a number of successful designs that use large midranges.

Let's try to identify how design priorities should be laid out in a midrange-to-tweeter crossover.

 

[454Casull]

2-3" dome mids seem to be the sweet spot - many of them have fantastic distortion performance and power handling, with flanges that are not especially large, allowing for tight C-C. Their limited excursion can be made a non-issue by crossing to a midbass, rather than a bass unit.

 

[jerome69]

Very interesting debate !

There are subjective things in this discussion. The basic comparison is the size of the instrument and how big sounds the driver.
My taste is 220cm² minimum for a speaker, below it misses matter to my ears. And i play loud

Objectively small drivers make higher distortion than the bigger drivers, moving more air. I think it 's a beginning of explanation

We can define priorities ? here we talk about multiway

- High sensitivity
- Perfect uniform response
- Small size
- Low distortion
- Low bass response
- High power handling
- Type of baffle
- Type of crossover passive/active
- Marriage of the drivers
- Price or range of the drivers

Other suggestions are welcome

I think it's very important to see/hear when you change a parameter, how the sound changes.
The priority of a parameter gives very different results.

 

[DougL]

If you are crossing to a Horn Tweeter, you can match the horn coverage and the driver coverage at the crossover to minimize power response issues.

In general, 3" or 4" mids makes sense for dome tweeters, 6" to 8" mids for Compression drivers.

Its all about compromise and design tradeoffs.

 

[Bob Richards]

I like the idea of a vertical line array of 3 inch drivers, maybe 3 ft. high, maybe open baffle, and a 180 or 360 degree horizontal array (a ring so to speak) of 3/4 inch dome tweeters a the top, with energy below 100-150 Hz being handled by separate closed box woofers with EQ to be flat down to 20HZ. I like the more horizontally narrow drivers because they work better at delivering accurate timing cues in the lower midrange, where that's how we determine image location.

Plus, the more physically displaced drivers you have putting out the same thing at the same time, the more the cancellations from room acoustics get filled in by each others varying cancellation frequencies. Plus, a vertical array will be very directional at the higher frequencies of its range on the vertical axis, so considerably less damage from floor or ceiling bounce (which is often a real big deal). I'd cross the 3 inch drivers over to the ring of tweeters at 6-7kHZ; just above the frequencies where we are hyper sensitive to amplitude comparisons for image location, and in the range of frequencies where images get real diffuse sounding anyway. Omni treble above 7kHZ sounds significantly more real to me. If the speakers can't be out from the front wall by at least 3 ft., then I'd do a 170 degree half ring for the tweeters.

If you go with a small number of large diameter midrange drivers, you'll have bigger lobing problems and foggy imaging in the lower midrange, if that matters to you. It could be argued that if you don't have an inter-aural cancellation thing going on (electronic or acoustic) (which I do), then imaging timing cues in the lower midrange will be relatively un-usable anyway. Power handling would be the weak link with the 3 inch drivers. You'd have to jam in as many as would fit in that 3 foot high tower. Linkwitz knows how to do the math on that. It's on his website. It would be real nice to take the 3 inch drivers all the way down to 100HZ, but it's a tradeoff with power handling.

 

[CharlieLaub]

An 8" driver is about the largest that you can use and still cross over to a tweeter around 1.5k-2k Hz. Above that frequency, the 8" driver starts to beam. Larger (e.g. 10" and up) drivers will beam at even lower frequencies, making a crossover point to the tweeter impractically low.

For the same SPL output, the larger the diaphragm the smaller the excursion that is required. As excursion increases, properties like suspension compliance and motor strength start to vary with position enough to cause a rise in distortion over "small signal" conditions. Thus, the smaller excursions required of the larger (e.g. 8") driver should give rise to better performance, all other things being equal.

Smaller diameter drivers tend to have lower sensitivity and lower power handling compared to larger drivers with a larger diameter voice coil. So often you need to deliver more power to the smaller driver for the same SPL, while at the same time voice coil heating will be greater, which can lead to power compression after the system is running for awhile. If the power compression is not happening equally for all the drivers in the system, the spectral balance will shift compared to when the system is "cold".

These are some of the motivations for using an 8" midrange that I can think of...

-Charlie

 

[jerome69]

These are parameters we must deal with : size, bass extension, efficiency



Line array is an interesting subject, i don't know well. The concept is interesting but you run into high cost and a lot of wood working.
The area of a 3" driver is near 25cm², to have a good soundstage, we can use 8 per speaker ?
I have seen an article on the linkwitz's website : http://www.xlrtechs.com/dbkeele.com/PDF/Keele (2005-10 AES Preprint) - CBT Paper 5.pdf
The presentation is very cool
http://www.linkwitzlab.com/Keele - ...(AES-ASA-BAS presentation, Jan. 14, 2010).pdf

 

[Havoc08]

Interesting thread Boris. I'll follow it with great interest, but do not have the knowledge to contribute much.
I know that breakup is usually an issue with 8+ drivers making it hard to cross to high.

One thing I would like to read more about is lobing which I presume relates to beaming, but other than that, I have only seen simulations of dispersion patterns and have not idea what influences them other than they are frequency dependent (someone had a link once of a simulator that visualized the dispersion pattern).
Is it because a too large driver will act as a horn at higher frequencies or what happens? (links or input appreciated).

 

[Havoc08]

An 8" driver is about the largest that you can use and still cross over to a tweeter around 1.5k-2k Hz. Above that frequency, the 8" driver starts to beam. Larger (e.g. 10" and up) drivers will beam at even lower frequencies, making a crossover point to the tweeter impractically low.

For the same SPL output, the larger the diaphragm the smaller the excursion that is required. As excursion increases, properties like suspension compliance and motor strength start to vary with position enough to cause a rise in distortion over "small signal" conditions. Thus, the smaller excursions required of the larger (e.g. 8") driver should give rise to better performance, all other things being equal.

Smaller diameter drivers tend to have lower sensitivity and lower power handling compared to larger drivers with a larger diameter voice coil. So often you need to deliver more power to the smaller driver for the same SPL, while at the same time voice coil heating will be greater, which can lead to power compression after the system is running for awhile. If the power compression is not happening equally for all the drivers in the system, the spectral balance will shift compared to when the system is "cold".

These are some of the motivations for using an 8" midrange that I can think of...

-Charlie

How about loading the room? A 8" midrange will of course move more air than a 3" if both are fed a X Hz signal, which I presume makes them midrange more "palpable" and the sound more homogeneous in the whole room instead of just within 3-5 feet of the speaker?


Another question for the discussion is:
Wouldn't a larger diaphragm make it less likely that the driver behaves as an ideal piston compared to a 3" driver?

Does the gained SD outweigh the increased moving mass (thinking of transients etc. here)?

 

[CharlieLaub]

Interesting thread Boris. I'll follow it with great interest, but do not have the knowledge to contribute much.
I know that breakup is usually an issue with 8+ drivers making it hard to cross to high.

One thing I would like to read more about is lobing which I presume relates to beaming, but other than that, I have only seen simulations of dispersion patterns and have not idea what influences them other than they are frequency dependent (someone had a link once of a simulator that visualized the dispersion pattern).
Is it because a too large driver will act as a horn at higher frequencies or what happens? (links or input appreciated).

Beaming/Lobing is the same thing. It occurs off axis because the difference in path lengths from the various points on the surface of the radiator (the cone) start to become comparable to the wavelength of sound at higher frequencies and the resulting difference in phase (the relative phase) at the listening point causes interference (causing increasing levels of cancellation) as you move off axis. Every non-point-source radiator suffers from this effect.

-Charlie

 

[boris81]

These are parameters we must deal with : size, bass extension, efficiency

I'd like to keep bass extension out of the scope of this discussion. Let's assume that there is a dedicated subwoofer handling the low end. This theoretical midrange should be able to handle the low hundreds Hz but nothing crazy.

I don't want to put a limit on the size. Some Econowave designs use a 12" as a midrange and that's fine. If bigger speakers are shown to sound much better, then so be it.

Let's focus at what happens at the crossover mid-tweeter crossover.

 

[CharlieLaub]

How about loading the room? A 8" midrange will of course move more air than a 3" if both are fed a X Hz signal, which I presume makes them midrange more "palpable" and the sound more homogeneous in the whole room instead of just within 3-5 feet of the speaker?

Palpable? What is that? I don't think your statement about moving more air is correct, in general. If two drivers have the same sensitivity and you feed them the same signal then the SPL output is the same. Room effects are a completely different topic than what has been discussed so far.

Another question for the discussion is:
Wouldn't a larger diaphragm make it less likely that the driver behaves as an ideal piston compared to a 3" driver?

Does the gained SD outweigh the increased moving mass (thinking of transients etc. here)?

What is "SD" that you refer to?

You are correct with your piston observation. But keep in mind that moving mass alone is not determining the transient response. It's a combination of mass and motor strength.

-Charlie

 

[CharlieLaub]

I'd like to keep bass extension out of the scope of this discussion. Let's assume that there is a dedicated subwoofer handling the low end. This theoretical midrange should be able to handle the low hundreds Hz but nothing crazy.

I don't want to put a limit on the size. Some Econowave designs use a 12" as a midrange and that's fine. If bigger speakers are shown to sound much better, then so be it.

Let's focus at what happens at the crossover mid-tweeter crossover.

Even though the topic is about a midrange, that pie chart still applies. It's known as "Hoffman's Iron Law". Instead of "bass extension" you can think of it as "low frequency cut off". Even a midrange driver is still essentially an acoustic band-pass filter, so there is a lower limit to it's operating range. For some dedicated midrange drivers, getting it to work down to 200Hz is a real stretch. Some midrange domes can's do below 500Hz for instance.

Also "size" should really be "enclosure volume". The term "size" could be interpreted as many things.

-Charlie

 

[jerome69]

Ok Boris we will not talk about bass but low end extension
You don't have the same low end with 5" or a 8"
Very important in the mid ! Try a 5" and a 8" mid down to 150Hz ... no doubt which one is more at ease.
True midrange have high efficiency but low xmax : compare a MCA15 and a CA15 from SEAS. Less mass, less inductance for the mid.

 

[Loren42]

It really depends on what other drivers you are using in the system, where the intended crossover points are, and what is the directivity of all the drivers in the whole system.

I would like to have a discussing on why 8" drivers are often preferred for midrange duties over 4" ones.

The superior performance of large area drivers in terms of sensitivity, distortion and power handling is obvious so it's no use to prove that point here.

I'm mostly interested in the issues of driver lobing and maintaining good frequency response off-axis. Using a smaller midrange is the best way to achieve uniform dispersion through the crossover region, yet it's hard to ignore a number of successful designs that use large midranges.

Let's try to identify how design priorities should be laid out in a midrange-to-tweeter crossover.

 

[boris81]

Obviously larger drivers do better on the low end.

Why is low end extension more important than good directivity in the midrange?

 

[Zuhl]

The size of unit you use depends on your design needs. For me, I've never been happy with crossovers in the 1.5-3khz region for a 8" plus 1", I always feel a large driver to small driver is a mismatch as you are stretching both to their limits. My room has a severe dip at 60hz - a sub sounds 'odd' stretching across that and is best kept below 60hz. My main speakers would then have to reach down to at least 60hz. I can't get away with a sub + a 4" and 1" two way.

So with that in mind I have a three way + sub. 12" sub, 8" bass, 2.1" dome mid and 3/4" dome tweeter. Crossover points are 60hz, 850hz and 5khz. Meets all my objectives, design wise and, the most important thing, listening to music.

 

[jerome69]

Obviously larger driver do better on the low end.

Why is low end extension more important than good directivity in the midrange?

Good question
I think it is a density of informations/energy, there are a lot of energy in the range 100-300Hz. 100-300Hz is the territory of large midrange. A small driver can't cope this amount of energy even if the directivity is better, they yell.
Midrange is the range 160-1280Hz or 2560Hz ? A 8" can handle 160-1280Hz easily. And fundamental frequencies in voices are in the range 60-1800Hz...

What i notice about directivity, it could be not annoying because the tweeter compensates a little the mid and the beaming of the mid could be ear friendly. We listen on axis ?
A good response off axis gives a more clean/natural sound.
It is difficult to make good directivity, high Sd if you don't multiply the drivers (4 ways ? or CBT line array ). The problem is to find a good compromise.
Note we can't talk about directivity without talking about crossover (order and type).

 

[Moondog55]

When I sit in the sweet spot I cannot tell the difference between an 8inch and a 5 inch midrange.
If the change in diameter is so important then why do some of the early Japanese 2-ways that use an 8inch mid-bass plus a cheap cone tweeter crossed at 5k sound so good??

 

[boris81]

Off-axis dispersion is important, especially in smaller rooms. Our perception is primarily influenced by the direct sound but reflections play a role too. A serious disturbance in the off-axis frequency response will be audible at the sweet spot even if the direct sound is perfect.

If you are crossing to a Horn Tweeter, you can match the horn coverage and the driver coverage at the crossover to minimize power response issues.

In general, 3" or 4" mids makes sense for dome tweeters, 6" to 8" mids for Compression drivers.

Its all about compromise and design tradeoffs.

Studies on the topic have discovered that humans are much more sensitive to directivity on the horizontal axis than the vertical. As far as I understand it's important for loudspeaker designs observe the horizontal dispersion but some compromise can be had on the vertical axis.

3" or 4" can mate with a dome tweeter for a wide horizontal dispersion and 6" to 8" woofers can match a corresponding size waveguide for a narrow directivity design. In this case the issue becomes a tradeoff between increased sensitivity and low end extension versus small listening angle on the vertical axis.