12-inch drivers for sub

chaudio wrote: Problem I find with bandpass cabs as one box bass solutions is that they never seem to integrate well with the mid-tops. You can never get away from the fact that you're listening to the bass and the mid-tops separately.

I would only use bandpass cabs for real sub. And given for a BP6, the furthest you really want the front and rear tunings apart is probably an octave and a half. In simulations I've always found around an octave spacing seems to give the best results.

My experience differs. To get a bandpass to integrate cleanly as a single box solution you:

- Set a wide frequency range, say 40-150 (which is nearly 2 octaves). You want to overlap with the tops to integrate properly, so tops should get to say 120Hz.

- Get the group delay to be roughly constant in the upper half of that range (eg 80-150). A group delay peak at the top end destroys the temporal coherance of drum hits, giving the classic mushy bandpass sound.

- Obviously, time align using a controller, just like you would with a horn bin.

These may slightly compromise efficiency (per cab volume and lower cutoff) but the subjective increase in output more than compensates - instead of just "rumbling" it actually seems to "play" the bass, and you get much more physicality from the drum hits.

In order to get all this in simulation, you have to be patient and experiment. 8th order gives you more flexibility to meet all the requirements, but since there are more variables it takes more experimentation. I don't think anyone would build the 20 plus prototypes that would be required to find a good 8th order bandpass design by hand, which is probably why bandpasses are not popular among those who use this approach.
Edited by: jsg

I never made a true 6th order BP but designs thatwere similair in effect/sound, which indeed benefit from overlap. Usually in the order of tops high passed at 60 Hz, subs low passed at 100 -110 Hz (smaller parties). People usually do not seem to notice thereactually is a sub in the first place. In one such case with the port pointing upwards people thought for real that it was a trash can

Regards

I've used bandpass a lot in other situations (BGM, low level, home, etc) and agree completely with overlap.

The thing that irks most with BP is group delay, and transient response (such a horrible word - I define transient response as being how well a speaker can reproduce a signal's attack and decay times so we can agree on things).

However, it can be ironed out in many ways. Still not my favorite enclosure type.

I've used bandpass designs up to 200Hz. The success of this depends on the rest of the system. I think one of the downfalls of home built systems is that people tend to look at each passband in isolation. Cabinet X with drive unit B is best for sub. If you don't use this comp' driver horn combination the highs will sound crap, etc; but then don't think about how it all integrates together
Obviously the success of any design is ultimately subjective, which is why I asked why bandpass and what the expectations of the cabinet were. The typical forum members here and at the original speaker plans seems to consider a compact sub anything too small to live in and the 36 litre cabinet design I came up with would be something to usefully supplement the bass from their mp3 player.

The above is not a criticism of any particular design, and I've had a number of discussions whether the ultimate system should be sonically neutral or designed to complement the genre of music it is amplifying. If you amplify mainly electronic instrumental music that consist mainly of sub 150Hz THUNK, THUNK and ultra 6KHz TCH TCH. Do you need to worry about reproducing the subtle nuances of vocals or violin? Or should every system design strive to be absolutely linear?

Meandering off topic even more, this has lead me to ponder occasionally whether it is possible to allocate a rating for a bass cabinet based of size, maximum spl and low frequency extension.

Back to the BMS 12” the following graph shows the maximum output from a 40ltr bandpass (reducing the volume slightly has only a minor effect on the output) and a 44ltr bass reflex. It is flat down to about 35Hz with a maximum of about 114dB. This is a nice match for some 8” BMS coax drivers I have ordered from Tony.



The other graph shows the response of a tapped horn.






To give some comparison of outputs I had a look at the measurement results of the Tuba 24 on the Audio Round Table site. At frequencies below 80 Hz the maximum output is about 118dB with a response down to 40Hz. This is in a cabinet 216 litres. The Tuba is a lot more efficient above 80Hz where the maximum output goes up to about 128dB. This used the Eminence HL10 which is probably about the same price as the BMS.

This comes back to the topic of rating cabinets. Is it worth building a cabinet over 5 times as big to get an extra 3-4dB output at 40Hz

www.audioroundtable.com/ProSpeakers/messages/250.html


Link to Audio Round Table

Deadbeat wrote: The thing that irks most with BP is group delay, and transient response

If you can get the group delay to be constant, then the behaviour is no different to a horn path delay, or simply standing further from the speaker.

So the problem is *varying* group delay.

I believe that pretty much all speakers have a group delay hump near the bottom of their range. What makes the *average* bandpass different is they also have a hump near the top end. I think the hump is much more disturbing for the listener in the 50-150Hz range than in the sub 50Hz range.

But you *can* design bandpass cabs to iron out the top-end hump. You're basically going for a Bessel-like rather than Butterworth-like upper knee shape. I've done this, built the cabinet, got the T-shirt.
Edited by: jsg

Deadbeat wrote: However, it can be ironed out in many ways.

Exactly [img]smileys/smiley32.gif[/img]

Deadbeat wrote:

Deadbeat wrote: However, it can be ironed out in many ways.

Exactly [img]smileys/smiley32.gif[/img]

Yep, make the cabs from the finest hand selected Russian Birch ply making sure all the joints are fully rebated etc. Finish off with 6 carefully applied coats of aldcrofts speaker paint and then wait till Nov 5th and burn the sh*tty things [img]smileys/smiley2.gif[/img]

You may not be able to polish a turd but you can roll it in glitter but at heart its still a turd.

I have no problems at all with BP6 on Sub but why anyone would want to run lower mids through them especially as there are much simpler ways of doing it is beyond me. Just because it can be done is no excuse for actually doing it.

Steve looking at your graphs will you be looking to cross at about 100hz?

Tony
Edited by: AJW

[img]smileys/smiley32.gif[/img] = Well done John - you might have noticed the rest of my post. Sorry, too lazy to post a full response.

Not exactly rabid promotion of the concept, but hey - it did make sense.

Every designer makes their choices....blah blah blah.

Edited by: Deadbeat

Tony,
Yes 100Hz will be a good starting point. The exact configuration will depend on the measurements.

One of the disadvantages of using linux on my house computer is having to copy all my winisd, hornresp and measurement screen captures across from my laptop. Which is a long winded way of saying why I haven't uploaded a response plot that shows how the 8” coax drivers match well with the 12” sub. Either bandpass or reflex.

............................

The subject of group delay is complex. Group delay is defined as the negative of the rate of change of phase with frequency. The quantity has the dimension of time, but the question is, what time does it represent? If the phase slope is positive this would mean a negative group delay yet that would imply the signal comes out of the system before going in.

Group delays are associated with any form of filter including loudspeakers. The bandpass like its electronic namesake has the characteristics of a high-pass filter and a low-pass filter hence the two peaks in group delay. Unfortunately there is no getting away from it.

To understand why it occurs consider yourself to be a low-pass filter with the requirement to pass signals below 1000Hz and stop those above 1200Hz. Now consider that you are presented with a complex signal, but you are smart enough to separate it into individual sine waves “a”, “b” and “c”.

Signal ‘a’ is 500Hz, signal ‘b’ 900Hz and signal ‘c’ 1300Hz.

First we look at signal ‘a’. We are only interested in seeing if signals are below 1kHz to pass, so by 1ms we can be sure that this is a low-frequency signal and pass it. Signal ‘b’ is takes longer, by 1ms we have almost seen a cycle but we are not 100% sure that it is below 1kHz so we wait a bit longer to check. By 2ms we are confident that it is below 1kHz and so let it pass. It is now delayed by 2ms. Signal ‘c’ is similar to ‘b’, except that after we keep it for around 2ms we decide to stop it passing through.

Real filters behave in a similar way, albeit with a more complex mathematical description, and the sharper the roll off of the filter the larger the group delay. With loudspeakers you have to consider the natural roll off of the loudspeaker plus the roll off of the cross-over filter. With a bandpass loudspeaker there is already a sharp roll off at the upper end of the frequency range which is compounded by the use of another electronic filter. The combined group delay could easily be audible.

With a minimum phase system there is a relationship between amplitude and phase which allows a complete determination of phase from amplitude. If a loudspeaker is a minimum phase system phase response, and hence group delay, can be calculated from frequency response. Dick Heyser (he developed time delay spectrometry) expressed the opinion that loudspeaker drivers are 'largely' minimum phase systems.

If there is no getting away from group delay the question is at what point does it become audible? One commonly referenced reports on the audibility of group delay is: 

Blauert, J. and Laws, P "Group Delay Distortions in Electroacoustical Systems"

The bad news is that the figures I have seen only go down to 500Hz. The good news is that others have extrapolated downwards and converted the results from a time delay into a cycles of delay by dividing the group delay by frequency. For example 1000Hz goes through 1 cycle per millisecond. A 2msec delay corresponds to 2 cycles.

Obviously listening conditions will influence audibility of any artefacts, but one banded about figure for the upper limit of audibility of group delay is 20msec. This corresponds to 1 cycle at 50 Hz. Blauert and Law's figures suggest 3.2msec or 1.6 cycles at 500Hz, 1msec or 2 cycles at 2KHz and 1.5msec or 6 cycles at 4KHz. If researchers were trying to hear the effects of group delay were they listening under ideal conditions and how would this compare with a typical gig?

The other problem with both bass reflex and bandpass enclosures is that they rely on resonance to work. Resonance is also associated with ringing. The amount of ringing depends on the amount of damping in the system. Then there is the room. If your system is in a space with a 5 sec reverb time, will you hear a few milliseconds of overhang from the loudspeaker?

Steve,
I hope you and jsg realise that I post a lot of my messages very much tongue in cheek.

I also hope that jsg goes on to build a stunning box that performs way beyond his wildest hopes and he can raise a single digit and proudly acclaim SEE you "Black Country Twat".

Thank you for the very comprehensive reply regarding the problems associated with bandpass boxes however I have for a number of years come to the conclusion that the simplest way of getting lower mids to sound natural is a direct radiator performing above the band that is influenced by any reflex enhancement. I have yet to be convinced otherwise.

As there are a number of transducers available that will do this perfectly well I do not see the need to add a compromise element into the equation.

As you have hit upon in an earlier post the final application has to be taken into account i.e. if it is to be used for mainly electronic music maybe it is not much of a compromise but orchestral with a bass singer then perhaps not.

Regarding the last paragraph I have been trying to find a link to the phenomenon known as "The Hayes * affect" which explains why we are much more conscious of the shorter direct signal than a later farther away version. This probably explains why a lot of the affects which are difficult to measure may have some profound auditory results.

Its a funny thing the brain when listening to music!!!

Tony

* not sure if this is the correct spelling




Edited by: AJW