Hi there.
The image posted is of a Seas MCA15rcy, measured from 1 meter with no filtering applied, on axis and mounted in a 5 litre sealed enclosure. The optimum sealed volume for this driver yielding a Qtc of 0.707 is at around 1.7 litres. I have chosen an enlarged enclosure producing a Q of somewhere near 0.5 and completely full- though not packed - with dampening material.
These drivers phase track perfectly with the bass drivers from 220 to 330hz. In these enclosures however the roll off below 300 hz looks close to 12db per octave, which is very sudden. Ideally id love to cross these drivers over at 220hz however as you can see the amplitude at this point is at - 8db.
My aim is to extend the low end response of these drivers in this enclosure in order to achieve an even amplitude response of bass and mid at the crossover point.
I believe I have two choices. The first is to remove approximately 50% of dampening material. The second choice is to physically reduce the size of the enclosure.
A third more lateral approach would be to use a more gentle HP in the mid(12db) so as to combine the roll of to 24db per octave. Crossover between bass and mid is currently LR24.
Any input from you speaker egg heads would be appreciated.
The image posted is of a Seas MCA15rcy, measured from 1 meter with no filtering applied, on axis and mounted in a 5 litre sealed enclosure. The optimum sealed volume for this driver yielding a Qtc of 0.707 is at around 1.7 litres. I have chosen an enlarged enclosure producing a Q of somewhere near 0.5 and completely full- though not packed - with dampening material.
These drivers phase track perfectly with the bass drivers from 220 to 330hz. In these enclosures however the roll off below 300 hz looks close to 12db per octave, which is very sudden. Ideally id love to cross these drivers over at 220hz however as you can see the amplitude at this point is at - 8db.
My aim is to extend the low end response of these drivers in this enclosure in order to achieve an even amplitude response of bass and mid at the crossover point.
I believe I have two choices. The first is to remove approximately 50% of dampening material. The second choice is to physically reduce the size of the enclosure.
A third more lateral approach would be to use a more gentle HP in the mid(12db) so as to combine the roll of to 24db per octave. Crossover between bass and mid is currently LR24.
Any input from you speaker egg heads would be appreciated.
Attachments
Nice and flat!
The SEAS spec sheet suggests that your drivers are not yet fully broken in. It should start rolling off at 200, not 300. After break in, Fs should be 50Hz.
Too much acoustic damping (stuffing) will decrease sensitivity, compress microdynamics and lose texture. In a sealed box, you only need to add enough stuffing to reduce the resonance to the calculated system resonant freq and no more. If you are already at Q.5 then adding tons of stuffing will not help you. A 2" layer of soft FG batt on the walls is sufficient. I usually only do three non-opposing walls. Cabinet resonance is a bigger problem than acoustic resonance for a midrange cabinet, imo and the bigger the box, the worse it can get. Determine the actual Fs after break in as it will be reduced measurably as the suspension loosens up. Then determine expected box resonance, and stuff it to get down to that.
The SEAS spec sheet suggests that your drivers are not yet fully broken in. It should start rolling off at 200, not 300. After break in, Fs should be 50Hz.
Too much acoustic damping (stuffing) will decrease sensitivity, compress microdynamics and lose texture. In a sealed box, you only need to add enough stuffing to reduce the resonance to the calculated system resonant freq and no more. If you are already at Q.5 then adding tons of stuffing will not help you. A 2" layer of soft FG batt on the walls is sufficient. I usually only do three non-opposing walls. Cabinet resonance is a bigger problem than acoustic resonance for a midrange cabinet, imo and the bigger the box, the worse it can get. Determine the actual Fs after break in as it will be reduced measurably as the suspension loosens up. Then determine expected box resonance, and stuff it to get down to that.
Thanks for your input but I've already had the drivers for a good year so and I doubt if this issue is to do with run in time. As a first step I will re measure with only a single layer of dampening material on the internal walls.
Your driver has been designed as a midrange not a midwoofer. It does not have the linear travel to be crossed at 220 Hz. 400 Hz is a more reasonable crossover frequency. If you are stuck with 220 Hz because of the woofer performance then switching to a midwoofer with a substantially larger linear travel might be the way to go.
If the box loading is significantly influencing the passband of a midrange it is an indicator that things are being pushed too far. Can you tell us why you are in the position of pushing the high pass of the midrange so low.
If the box loading is significantly influencing the passband of a midrange it is an indicator that things are being pushed too far. Can you tell us why you are in the position of pushing the high pass of the midrange so low.
I would try the driver in the 1.7 liter recommended enclosure, stuffed. I don' t buy the argument that fully stuffed would lead to loss of micro details, etc. References, please. This should get you to a minus 3 dB point around 200Hz. Then add a 2nd order electrical filter, and you will wind up with acoustic 4th order. With a 4th order electric filter on the bass driver, you should be fine.
Putting this mid driver in an enclosure that is too large is plain wrong. The roll off on the low end starts too early, which is exactly what you are seeing here.
Putting this mid driver in an enclosure that is too large is plain wrong. The roll off on the low end starts too early, which is exactly what you are seeing here.
Why? A larger midrange enclosure offers a better opportunity to remove the higher frequency sound being reflecting back to the cone. It does this at the cost of volume for the woofer which may or may not be important depending on overall cabinet size.
If the box loading is influencing the motion of the midrange cone then it will be degrading the transient response in the same way as a woofer but in a frequency range that is more sensitive for the ear. For a woofer extending the frequency lower is a big win and the degraded transient response is a price worth paying. For a midrange under normal circumstances it is not but the OP has yet to tell us why he wants to do it.
Andy, I can't completely follow what you wrote, but a couple of quick points here:
- A high Q or large enclosure leads to an earlier roll off than a low Q enclosure, which has a sharper knee. The combined result is that the lower Q enclosure runs into the low end longer in a straight line, but then takes a deeper dive faster.
- In order to get the cross overs aligned in phase, you need to stay close to a Q of .7.
- Smaller enclosures have the benefit that the point where standing waves can be supported shifts upwards.
- Transient reproduction is techno babble. If you mean energy storage or decay time, small and stuffed wins.
- A high Q or large enclosure leads to an earlier roll off than a low Q enclosure, which has a sharper knee. The combined result is that the lower Q enclosure runs into the low end longer in a straight line, but then takes a deeper dive faster.
- In order to get the cross overs aligned in phase, you need to stay close to a Q of .7.
- Smaller enclosures have the benefit that the point where standing waves can be supported shifts upwards.
- Transient reproduction is techno babble. If you mean energy storage or decay time, small and stuffed wins.
If a person doesn't understand transient response how would you explain the downside of a ported speaker compared to a sealed speaker for a woofer? In our case we are looking at the difference between a small stuffed sealed enclosure and a large stuffed sealed enclosure for a midrange. Which is preferable and why in your terms of energy storage and decay time?
Phase plots of the bass and mid indicate very close alignment from about 220 to 330 hz with a small delay of 0.2ms necessary. In crossing over any where over 330 phase starts to shift quite significantly till it aligns at 900 hz. So if i could cross the mids over somewhere at around 250hz your sitting right in the middle of a long section of aligned phase. Trouble is with the early roll off of the response below 300.
Attachments
In building the three way enclosure i partitioned off a 12 centimeters section at the top of the enclosure for the mid. The five liter enclosure is over sized but offers the option of reducing the volume with timber blocks or some kind of packing. Before I go to that trouble i wish to work with what i have.
Question: Is your phase measurement with or without xo's in place?
Because as soon as you add in the xo's, they will change the drivers' phase. The art of xo implementation therefore involves not only getting the FR you want but also keeping the phase aligned. Tricks that can help you do that are asymmetric slopes, adding resistance to the parallel (ground) leg of the xo and/or moving drivers forward or backwards on the baffle (so rear mounting, adding an extra layer for 1 driver or tilting the baffle). Xo simulation software is your friend.
Because as soon as you add in the xo's, they will change the drivers' phase. The art of xo implementation therefore involves not only getting the FR you want but also keeping the phase aligned. Tricks that can help you do that are asymmetric slopes, adding resistance to the parallel (ground) leg of the xo and/or moving drivers forward or backwards on the baffle (so rear mounting, adding an extra layer for 1 driver or tilting the baffle). Xo simulation software is your friend.
I am not sure I understand your plots. How are the midrange and woofer mounted and driven? How are you determining phase? What is the woofer? Is it a passive or active crossover?
Choosing a crossover point involves a number of factors. The phase is going to change when you fiddle with the crossover components in a passive crossover. It will also change if you use a smaller midrange enclosure to use a bit of resonance to increase the low frequency SPL and degrade the transient response. So you need to get the phase to line up after you have included everything that is going to move it around. Have you done this?
To repeat what I said above, a 250 Hz crossover frequency is too low for your midrange driver because it will leave its linear range at relatively modest sound levels. 400 Hz is likely to be a more reasonable frequency unless you are trying to do something like cross to a subwoofer driver.
The woofer is a HiVi M12 Product information:M12 Bass-Midrange
Drivers are crossed over via a Behringer Ultradrive dcx. At this stage Im sticking to reliable LR4 slopes. Phase plots are measured using REW with crossovers in place at 300 hz, one driver at a time of course. The driver arrangement on the enclosure is woofer on the bottom, tweeter in the middle and mid at the top.
I feel that my best option really is to perhaps to apply an LR2 filter at the Mid high pass to give me a summed roll off of LR4 at 300 hz. This seems to make sense to me. I may the remove a quantity of stuffing to see how it effects the sound subjectively. These bass drivers are quite linear to 1000 hz however I prefer the sound when crossed over low.
Hi Geopango,
The data sheet for the Seas MCA15rcy shows a graph of the response for this driver in a 20ltr enclosure. The initial LF roll-off is significantly greater compared to the response you show in your OP.
Have you considered your slow roll-off being due to the baffle step?
I assume that your measurement was free-field, but you may be able to get a better idea of what is happening by measuring the near-field response, ie, measure the response with the microphone located within a 1/4" of the cone.
Such a measurement will not show any baffle step or cabinet edge diffraction ripple.
You may also be able to determine the response from the impedance curve if you have the facility to make such a measurement.
As for the excursion limits for this driver, you may find this link usefull.
Loudspeaker system design
Peter
The data sheet for the Seas MCA15rcy shows a graph of the response for this driver in a 20ltr enclosure. The initial LF roll-off is significantly greater compared to the response you show in your OP.
Have you considered your slow roll-off being due to the baffle step?
I assume that your measurement was free-field, but you may be able to get a better idea of what is happening by measuring the near-field response, ie, measure the response with the microphone located within a 1/4" of the cone.
Such a measurement will not show any baffle step or cabinet edge diffraction ripple.
You may also be able to determine the response from the impedance curve if you have the facility to make such a measurement.
As for the excursion limits for this driver, you may find this link usefull.
Loudspeaker system design
Peter
Ah, active. That's a different story.
If this is the unit you have, I'd think about going with steeper slopes. That's the obvious solution to a narrow bandwidth of phase alignment.
Why don't you try starting with a 48dB/octave xo on the woofer at about 275Hz. You'll be 24dB down by about 412Hz and your phase is actually pretty good right up to 400Hz. Do whatever you need to do to the HP on the mid, including eq if necessary, to get a nice smooth response. Maybe compare with the xo set just a touch lower and so forth until you get something that you can live with.
If this is the unit you have, I'd think about going with steeper slopes. That's the obvious solution to a narrow bandwidth of phase alignment.
Why don't you try starting with a 48dB/octave xo on the woofer at about 275Hz. You'll be 24dB down by about 412Hz and your phase is actually pretty good right up to 400Hz. Do whatever you need to do to the HP on the mid, including eq if necessary, to get a nice smooth response. Maybe compare with the xo set just a touch lower and so forth until you get something that you can live with.
I have posted a graph with three measurements. The green is the enclosure fully filled, the red is with virtually no dampening - just a one foot square piece- and the purple measurement is the unstuffed box with a -10db shelf applied at 400 hz as a baffle step.
As you can see the removal of dampening material has virtually no effect on the bass roll of and a large part of the response looks more ragged. I will however say that the unstuffed box is more dynamic and open. Over the coming days I'll ad dampening to arrive a a good balance with the impression now that the mids sounded slightly choked with the enclosure full of stuffing.
As you can see the removal of dampening material has virtually no effect on the bass roll of and a large part of the response looks more ragged. I will however say that the unstuffed box is more dynamic and open. Over the coming days I'll ad dampening to arrive a a good balance with the impression now that the mids sounded slightly choked with the enclosure full of stuffing.
Attachments
I prefer the sound of LR4th order . I just think it all seems to flow together and integrate better.
I didn't know there was an M12. The local distributor here supplies sizes from M4 to M8. Do you know if it has been discontinued or just not widely distributed?
Your speaker looks to be the classic 12" woofer, 5" midrange and 1" tweeter (?) monitor configuration. The midrange and woofer overlap is fine, you are using a flexible active crossover allowing different crossovers to be examined easily without cost. This all looks good.
I presume you got the 1.7 litre volume from a simulation. Are the simulated and measured transfer functions in close agreement? If not this suggests we may well be missing something significant.
If you change the delay you can align the phase around whichever frequency you want. Changing the crossover frequency is also going to change the phase. I think you may have more room for manoeuvre than you currently seem to believe.
You are in a position to experiment and learn without incurring any costs and that would be my recommendation. Try low, conventional and high crossover frequencies, 12, 24 and 48 dB/octave slopes, small and large stuffed midrange enclosures and have fun learning. Trying to work out an optimum configuration without first doing this doesn't seem right. I would also suggest getting the measured response flat and smooth for a few different configurations before letting any subjective considerations guide the process.
Thanks for that Andy. Yes i am having fun. These digital loudspeaker management systems offer so much flexibility and power you're almost spoilt for choice.
As for the M12's yes they were discontinued a couple of years ago. Ive built them into a bass reflex enclosure precisely to manufacturers spec . And the tweeter i'm using incidentally is a Peerles HDS 810921.
Your question about the modeling program chosen - Ive used a number of applets over the years but I find it always easiest going to this popular website - mh-audio.nl - Home
The sealed box calculator will is plot the low frequency curve of virtually any volume you suggest. Most other sealed calculators only calculate the optimal volume.
Cheers
- Status
- Not open for further replies.