Eminence 3015 mid-loaded cab

[5StringPocket]

Originally Posted by kringle77

I personally would go with a pair of eminence 3015 drivers and a faital pro 5" mid. Cross over to the mid at around 1.5khz and you will a clear but punchy cab. There are technical reasons not to but, I'd also have a switch available to run the 15's fullrange as well since you would get more of a regular bass cab sound with better dispersion from the 5". If I ever have a cab built with 15's again, this is exactly what I would do. Those 3015 (non lf) are just amazing drivers.

I snagged this from the Lowdown Sound Club thread but thought it deserved some further discussion. Based on my own experience with mid-loaded cabs and several builds with Eminence Kappalites, I think this idea would work nicely. Here is a sketch of the cab I'd like to try, based on the Eminence website medium vented design. The single tri-port has a bit more cross section area than the 25.12 sq-in of two 4" circular ports and would work better with the driver layout shown.
http://www.eminence.com/pdf/Kappalite_3015_cab.pdf
The Kappalite 3015 with this loading should really punch with lows that are solid and clear but not too deep. I like this voicing for my 5-string basses. I also like the idea of a switch to run the 3015 woofer either full range for a darker vintage voicing or mid-loaded for a more modern extended-range voice. It's like having two different cabs in one. Two of these would make a monster modular stack.


From its response curve, I think this Faital 5" driver would be a great match with that woofer.
http://www.faitalpro.com/products/LF...p?id=101010100
It's pretty flat between 1-2khz with a dip between 2-3khz and a rising response from there, peaking just shy of 6khz. Nice.

A mid driver 2-3dB less sensitive than the woofer pairs well as seen with the 93.6dB Alpha-6A/95.5dB 3012LF (Fearful 12/6) and 95dB Faital W6N8-120/97dB Custom Eminence 12 (Baer ML112). The 8-ohm equivalent sensitivity for the Faital M5N12-80 (99dB at 12 ohms) is about 97.5dB. This is a little over 3dB down from the 3015 (100.8dB at 8 ohms) which ought to match well with good presence without being too much.

One of the key components is the crossover and I'd like to open it up a bit here. I'd like to see a low/mid crossover point high enough to let some of that grit starting at 1.1khz through but low enough to limit the rise of that 1.1-2.3khz swell to about 2dB so it's slightly smoother but with that extended mid/high response. Would a 2nd order crossover at 1.5khz provide this or should it be a bit lower around 1.1 or 1.3 khz? If someone would be willing to have a look at this and post some crossover response curves that would be great.

Also, when choosing components for a crossover do you use the Re given with the TS parameters rather than nominal impedance, and is a Zoebel network to keep woofer impedance flat around the crossover point a worthwhile improvement to add?

[morebass!]

subscribed

[fdeck]

Here's my thought about the crossover. I'd use the actual impedances of the woofer and midrange at the crossover frequency, taken from the impedance graphs on the datasheets.

[rpsands]

In this box alignment I think you'd get better results from the Faital 15Pr400 (it'll probably be marginally easier to cross over too; less peaky in the crossover region).

* regarding Xover, it's worth your while to learn to use the passive crossover designer worksheet. It will turn out some pretty spectacular results with some investment of time.

[R Baer]

I would not use a single port of this type with the Kappalites. We had some issues with the early Kappalites making a strange clicking sound that took forever to track down. I went so far as to sending some cabs to Eminence for testing. Their impression was that the air moving through through just one port located off to the side, was causing "cone rocking". Basically, as the cone would move back, the air rushing out of the cab in one direction only was pushing things slightly to one side and causing the voice coil to rub. If you remember, our very early proto cabs had one single corner port. We redesigned the cabs with two ports and the problem went away. I believe Duke redesigned the TC cabs as well to use more than one port for the same reason, if I'm not mistaken.

[DukeLeJeune]

Originally Posted by R Baer

I would not use a single port of this type with the Kappalites. We had some issues with the early Kappalites making a strange clicking sound that took forever to track down. I went so far as to sending some cabs to Eminence for testing. Their impression was that the air moving through through just one port located off to the side, was causing "cone rocking". Basically, as the cone would move back, the air rushing out of the cab in one direction only was pushing things slightly to one side and causing the voice coil to rub. If you remember, our very early proto cabs had one single corner port. We redesigned the cabs with two ports and the problem went away. I believe Duke redesigned the TC cabs as well to use more than one port for the same reason, if I'm not mistaken.

Yup, I try to get approximate symmetry of ports with long-excursion woofers, especially when shoe-horning them into a small cabinet. Credit to engineer Jerry McNutt of Eminence for the symmetrical port suggestion. He explained to me that woofers with unusually long voice coils can have that clicking problem from cone rock, whereas a shorter voice coil wouldn't end up far enough over to tap against the side of the magnetic gap. All else being equal, larger-diameter cones are more susceptible to cone rock than are smaller-diameter cones.

Armadillos?!

As for the crossover design, I've almost never done a crossover design that ended up being what I expected it to be. Rpsands' suggestion is an excellent one. Don't be afraid to experiment, as sometimes a higher or lower crossover point just sounds better, or something non-textbook like asymmetrical filters or either "too much" or "too little" overlap. Also try the midrange driver with both polarities - there's a roughly 50% chance that the optimum polarity is not what you'd expect because there are phase rotations going on that are very difficult to visualize. Finally, if possible keep an eye on the impedance curve, because you don't want to end up with a nasty impedance dip if you plan to later add a second identical cab and run them in parallel.

[chienmort]

So two or more ports are needed but how do you calculate the triangular port tuning dimensions and how well do edge/corner ports work? I ask this because so many sources say keep port ends away from cabinet walls.

Getting back on track, the design looks really interesting.

[Arjank]

Originally Posted by DukeLeJeune

Finally, if possible keep an eye on the impedance curve, because you don't want to end up with a nasty impedance dip if you plan to later add a second identical cab and run them in parallel.

+1
Good point

[Arjank]

That cone rocking can be a real problem.
It also is more of a problem with longexcursion woofers and large diameter cones that have a not so stiff suspension and single spider.
Some (expensive) woofers have multiple spiders to prevent this problem (to some extend), I don't know if the 3015LF has such a construction.
This one e.g. has a double spider for enhanced stability
http://oberton.com/index.php?option=...m-loudspeakers

[farboozle]

Oh man, I have been thinking about a cab like this for months. Pair either the 3012HO or the 3015 with the faital mid. I have a 3015lf/6 cab, so I know I don't need the lows that the lf drivers produce.

I struggle in comparison between almost equivalent designs using the 3015 and 3015lf with the mid. With the mid, the existence of upper frequencies from the woofer are moot. Even if I don't think I need the lows, why not use the lf driver? It slightly more lows even in a non-optimized box.

regarding Xover, it's worth your while to learn to use the passive crossover designer worksheet. It will turn out some pretty spectacular results with some investment of time

Also try the midrange driver with both polarities - there's a roughly 50% chance that the optimum polarity is not what you'd expect because there are phase rotations going on that are very difficult to visualize.

I've been playing around with PCD. I can get .zma and .frd files from graph digitizers, but those files have 0s for phase. Is that ok? Isn't phase important for crossover design? I understand fundamentals of signal phase, phase shift, etc, but I don't understand if a driver in free air has a phase plot. Does the signal only get phase shifted when the crossover components are involved?

[will33]

Nice idea. I had also looked at that 5" a while back to run with my 12's...also about 100db worth of woofers. Didn't follow through with it but one thing to consider....the empedance and response curves are going to change some from the datasheet (measured in wall baffle or free air) once you stuff it in a small chamber.

If possible, it'd be best to measure the impedance at your chosen crossover frequency while it's in the box....I think you can do that by passing a sine wave and taking a measurement at a resistor placed across the terminals or something, I'd have to look it up. Otherwise some programs can model it pretty well and others can take an actual measured impedance sweep of the whole system.

Barring that, I would use the impedance reading for the woofer that is in the eminence cab model pdf rather than the driver datasheet, that should be accurate. For the mid, use the datasheet impedance reading at that frequency and tweak by ear from there.

Cool design, man.

[will33]

Also +1 to the porting. I'd port in the top left in addition to what's shown. Might mean shoving the woofer into the corner, but would be better overall under high power.

[astack]

Love it. I was just flipping through this forum thinking there's been a drought of cab design threads like this. Thanks for throwing this out there.

Probably be another very great cab. I'm really interested in checking out the 3015 and might next time I see them on sale. I might also lean slightly toward Will's idea with a pair of Deltalites with a mid over the single 15. There's magic in a pair of twelves, I tell ya.

Re: the clicking and asymmetric loading, I think I may have been hearing that with the 15PR400 in my box. The cab is pretty similar to the Bassic layout, though, and people don't seem to have a problem there, so maybe not. I used 3" ports instead of 4" ports (shortened as required for fb ~ 40Hz), so maybe that could add some problems? Hopefully it's not something fatal (zing) to the driver. /OT

[5StringPocket]

Very constructive feedback. Thanks so much!

Cone rocking from assymetrical port loading makes sense so I incorporated two diametrically opposed ports. I think this should provide more equal backpressure loading through the centerline and eliminate the chance of cone rocking. I also shifted the bracing to allow for an inset handle.
18mm birch ply for the baffle and 12mm Okoume for the rest of the box should allow the weight to come in around 37 lbs with a perforated steel grill.


Triangular port tuning matches a similar sized round port according to what I've read and Eminence Designer lets you toggle back and forth between the two. Diameter vs sq-in toggles but port length stays the same.

Using the actual impedances of the woofer and midrange at the crossover frequency, taken from the impedance graphs on the datasheets also looks like good advice. The impedance graphs actually show pretty close to 10 ohm impedance (as opposed to Re of 7.2 and nominal of 8 ohms) for both woofer and mid around 1100 hz but woofer impedance climbs with frequency. This would change the LC component values and suggests a Zoebel network to maintain woofer impedance at 10 ohms in the crossover region would improve the low/mid transition performance.

If the 3015 crossover is a problem the Faital 15PR400 could definitely be used but I like the higher power handling and touch of grit the 3015 provides. Hopefully if some combined response curves surface that call can be made. I know it would be worth my while to learn to use the passive crossover designer worksheet but I've got too much on my plate right now so I'm hoping to lean on the expertise of others to nail this.

I'm thinking impedance dips and whether or not to flip the mid driver polarity won't be sorted until the response and phase curves become available. Thanks for the heads up on that - should be interesting.

[makohund]

Originally Posted by 5StringPocket

I also shifted the bracing to allow for an inset handle.

Oh, good. I was gonna mention the handle type in the first drawing and the brace just inside it might not agree with each other.

Could maybe use an active xover or amp w/dsp to dial it in close, while planning/designing the passive xover.

[Chef]

great concept; great thread.

[rpsands]

Just use a 12mm baffle dude. It's not necessary to go thicker. Use birch instead of Okoume if you want to have a few more plies for screw holding or something.

[5StringPocket]

Agreed, 12mm birch would probably be fine. That marine grade Okoume is 9-ply so no difference in the number of plies but the birch is more dense and stiffer - definitely the best choice for a baffle.

Using an active xover or amp w/dsp to dial it in close, while planning/designing the passive xover could be quite useful in listening for the best sounding crossover point. You could fiddle with it without having to build anythin, then just build what sounds best. I have a Crown XTi 4000 with built in DSP so it wouldn't be hard to set up a 2nd order Butterworth crossover. I wonder if True RTA would be useful for comparing frequency response?

[will33]

TrueRTA would be of help. You'd be able to take soloed measurements of each side, low and high to see the effect of the crossoverw rolloff. Things like rises or valleys in the impedance or peaks/dips in response will affect the passive filter, sometimes givng you something other than a smooth rolloff that's right on 12db/oct. or move the corner frequency up or down. Then of course you can measure the whole system, looking for good summing of the 2 halves at the crossover point and avoiding any big peaks or dips in impedance there.

Agreed in doing it actively to get it zeroed in, then working up a passive to mimic that.

[Passinwind]

Originally Posted by 5StringPocket

I wonder if True RTA would be useful for comparing frequency response?

Sure, and IME it's also useful for looking at distortion components. Problem with that is, once you port to a passive crossover the distortion components tend to move somewhere else.

My crossover building workflow:

1) Testing with active filtering, including EQ and delay.
2) Measurement of individual driver impedances and FRCs in actual cabinet. "Measuring" FRC involves my ears as much or more than using an RTA. IMO it is useful to look at distortion components with your RTA, listen carefully to them, and also to think about which ones you like and want to retain/enhance, and which ones you want to avoid if possible.
3) Repeat step 1.
4) Design, build, and test three passive crossover candidates.
5) Discard two and refine third one until patience and/or money are exhausted.
6) Repeat step 1.
7) Profit???

Looks like a cool project, best of luck.