Dipole midrange driver recommendations

[Saurav]

I'm wondering about different midrange drivers I could try in my speakers. Rough requirements:

• Woofer is an AE OB12, usable up to about 400Hz on an OB, but ideally I'd like to cross around 250Hz or below (then I could try an H-frame and get more bass out of it)
• Tweeter is a Neo3, cross at 2kHz or higher
• I plan to continue my current narrow / no baffle approach, so I won't get help from the baffle when it comes to excursion. Not sure how well MTM will work with a Neo3 between the midrange drivers. TMM might work, if I can keep the whole thing a 3-way
• I'll use the DCX, so efficiency matching isn't a concern
• Something reasonably priced would be nice

The current driver is an Audax PR170M0 - a hi-efficiency (mostly) pro-audio paper midrange that I've had for many years. While I like it's sound in general and its 'punchiness' in particular, I'm pretty sure I can do better when it comes to resolution and refinement.

[Space]

I'll recommend the 18Sound 6ND430 for this range. It will extend easily to 250 Hz but not really below 200 Hz in an open baffle. It begins to sound "papery" if crossed too high, but 2.3 kHz or so is doable. This driver should have a similar level of sensitivity and snap as your Audax but adds more usable suspension travel. And the thermal limits are probably untouchable for home use. Reverse mounting (as in these pictures) is a good idea to reduce masking. That is, assuming you have a baffle...

If you're not in a hurry to get these I could part with my test pair, assuming my next big idea works out as planned (should know that within a few weeks).

A runner up at a lower price would be the SB Acoustics 6".

[gbegland]

John @ AE has a new TD6 almost ready to go. There is a dipole version as well as a higher sensitivity "M" version. Check his forum.

Greg - www.phatplanetstudios.com

[cjd]

I've been really curious about some of the B&C mids - both 6.5" and 8" - you may even be able to use the 8PE21 crossing to the Neo3.

[ttan98]

The a number of high efficiency mids you can look at are,

1. B&C 6" driver similar to what Dan was using
2. PHL 1120 6.5" mid range or if you prefer PHL 2460 a mid woofer both at 96dB at mid range.


With your neo I suggest B&c or PHL 1120. You can get PHL from E-speakers
not cheap at $175 each. The latter has very good review and very dynamic. I have the 2460 version which I am very happy with it.

[Saurav]

Thanks, those are all good suggestions. I should mention that I'm not locked in to a high-efficiency mid, just because my current mid happens to be one. My personal biases make me believe that I'll prefer the sound of a higher efficiency driver (all else being equal, which of course means very little when comparing drivers), but I'm not dead set on that.

[Saurav]

Did some reading.

AE TD6 - there's a thread on the PE board that shows not-too-great measurements on an OB. My first thought was, I wonder if the driver was front / flush mounted, and how it'll look if mounted on the back. The dip around 1.5kHz reminded me of my current driver, and how the response changed when I mounted them to the back of a (much smaller) baffle.

B&C 6MD38 and 18Sound 6ND430 - these seem pretty close. The comments on Zaph Audio seem positive for both. Dan used the B&C in his BaSSLines on a pretty narrow baffle, and that's got excellent reviews. However, his XO point is around 450Hz, i.e. slightly higher than my current XO point, and I'd like to push it almost an octave lower. The 18Sound seems to have more output around 200Hz, and it has 5mm Xmax vs. 2mm for the B&C.

PHL - The 1120 would seem to be the 'spiritual successor' to my Audax, I've read many comments about how they sound pretty similar, with the PHL being more refined. Someone else suggested the 2460 as more of a mid-woofer to get more bass. Haven't done too much research on these. They're more than 2x the cost of the B&C though, with the 18Sound in the middle (unless Space gives me a good deal on his pair ).

Haven't done any reading on the SB Acoustics driver, that's next. So far I'm sort-of leaning towards the 18Sound being the one to try, unless there's a big hole in my logic or something I'm not thinking about.

[HareBrained]

The SB15-4 is on sale at Madisound for $42. Fairly efficient if your equipment can handle the lower impedance. Of course, a pair would be a very good option.

[Saurav]

What are the implications of having 2 midrange drivers? I'd put them in series if they were 4 ohm drivers, I'm pretty sure I can't handle 2 ohm loads. Assuming I have the same crossover on both midrange drivers, what are the pros and cons of MTM vs. TMM? Better vertical lobe control with the MTM. CTC distance issues - in one case the M-M distance will be high, in the other case it'll be the T-to-lower-M distance. Which is worse? I don't have enough channels on the amps or the DCX to go 3.5/4-way, or I could try a 2.5-ish thing on the 2 M's.

[CraigJ]

Saurav,

If I were to build another dipole speaker today, I'd take Jon's advice and use the same driver pair as used in the Modula MT MkII. The SSD2608/9130 with waveguide and Seas ER18RNX on a narrow baffle, combined with your AE OB12 IMHO, would be a very nice system.

CraigJ

p.s. you could even mount your Neo3 facing the back wall.

[Saurav]

Yeah... I tried the dipole + W/G tweeter thing, and didn't really like it. Everything measured fine on-axis, but the sound was always a little disconnected. What I haven't tested is how much that would be improved by just sending some treble energy to the back, vs. the so-called 'true dipole' response that I've attempted with my current design.

I'll do some reading on the ER18RNX.

[CraigJ]

Everything measured fine on-axis, but the sound was always a little disconnected. What I haven't tested is how much that would be improved by just sending some treble energy to the back, vs. the so-called 'true dipole' response that I've attempted with my current design.

PaulW mentioned he had issues when using waveguides and dipole mids. At the time I didn't know what he was referring to until I put my Isiris Jr. on a PVC pole. This system also measured well, but I think the large waveguide with the mid in free air just didn't sound right. Perhaps without the baffle, the waveguide becomes too directional for the other drivers?

My Phoenix clone with dual SS8545s+SSD2608/9130 with waveguide on a narrow baffle has excellent in room power response and sounds very good and is easy to measure flat. Also, if I remember your systems correctly, I don't think your horn was similar to the shallow MCM waveguide. Just food for thought.

Craig

[Saurav]

Those are all good points. Food for thought indeed.

[pixelpusher]

Hi,
I've been using a PHL 1120 with the Neo3pdr. While I am happy with the sound, it is not an ideal match. The PHL 1120 must be padded down quite a bit to match the sensitivity of the Neo3. I plan to experiment with using two Neo3s on each side.

[Saurav]

I have the same problem with my current speakers, but I use an active crossover (and the amp has gain controls on individual channels, though that may change soon since I'm looking for a new amp), so that should not be an issue. Once I settle on a set of drivers I can add passive inline attenuators as needed to balance out the signal levels, and avoid large amounts of digital gain / attenuation.

[cjd]

Well... You're treading around the waters I'm playing in, vaguely, so I may as well through some more stuff out there to make this even harder.

I'm currently (intermittently - haven't touched it in a couple months!) playing with custom dipole waveguides for a Neo3+Neo8 pairing, with the goal of running that straight over to a 12" or 15" lower midrange, which will cross into some woofage, and finally some sub-woofer (which will be monopole I am sure). I assume at this point that it may be WtmMW or so.

While you may not wish to go this route specifically, it may allow you to go mtMW or tmMW with a Neo8+Neo3 and a larger (10"?) lower midrange driver - more driver overlap on frequency range which may or may not be of use, but... IMHO it would be for dipole. As frequency goes lower there's something about having surface area behind the sound vs excursion (I say this particularly keeping in mind my recent experiences with the Anarchy...)...

C

[Saurav]

I'm all for surface area vs. excursion. There are many others who feel you really need a 4-way to do a full dipole. A ~ 6" mid is probably right on the edge of what I could stretch between a 12" woofer and a tweeter. If I went 4-way, then I'd probably be better off with a 15" woofer to push the XO to the sub lower, and then a couple of mid/mid-woofer drivers, probably 10" and something smaller, like you said.

But that needs even more amp channels, and another DCX. Unless I go partially passive... which I will suck at. I can't see any of the drivers not needing a lot of work to handle the dipole effects.

[Dennis H]

If I were starting from scratch doing the narrow/no baffle thing, I'd start with a Neo3 and go with a 4-5" mid, maybe one of the new budget Scans, to get the MT XO high enough while preserving directivity. You could cross that to something like a TD15M with no problem, 2 octaves below the MT XO. Cross all that to a sub at 80 or so and you've got a full-range 4-way.

(Easy for me to say, I'm not the one spending the money. )

[Saurav]

I could kinda do that with the OB12, except for a deep notch in its response around 500Hz. My current 350Hz XO bypasses that somewhat... and it's difficult for me to trust my measurements down in that frequency range anyway. That's partly why I'd like to move the M-W XO down to the 250Hz region. But that's asking a lot from the midrange driver, 250-2k on no baffle.

[Dennis H]

Yeah, the TD15M is clean up to 2K so all you'd need to worry about is directivity, from both cone size and baffle width. I don't know if the 4" Scans (featured in JonMarsh's speaker porn thread) would play low enough to work with the OB12's notch.

http://www.scan-speak.dk/datasheet/pdf/10f-4424g00.pdf

http://www.scan-speak.dk/datasheet/pdf/10f-8424g00.pdf

Edit: try a close-mic'd measurement to see if the OB12 notch is in the driver or if it's floor bounce.

[Saurav]

It's the driver, I took nearfield measurements back when I first got it. The dip is in John's measurements too.

http://www.htguide.com/forum/showpost.php4?p=408663&postcount=122

http://www.htguide.com/forum/showpost.php4?p=410223&postcount=125

[twest820]

Yeah, six or maybe six and a half inches seems the preferable woofer size for a three way dipole. Smaller and they get pretty closed up for the back wave. Bigger and you'll end up running them above the dipole peak. My experience is vented pole pieces help a lot with the back wave, but I don't know of any six inchers with vented poles. I'd avoid Seas as their drivers' designed in directivity narrowing from 500Hz up is generally undesirable in nude or minimally baffled dipoles. The 6ND430 does this to some extent as well. The Neo8 doesn't have the necessary excursion---though the Neo8S might be workable in an MTM if you can actually get any from BG---and the Neo10 is awkward to cross to a Neo3.

My suggestion in this space would be the SB17NRXC35 for its low Rms and good excursion; efficiency's not bad either and it has SB's typical low distortion. Since EQ's no big deal another low Rms woofer to consider is Selenium 6W4P, though the SB17s are cheap enough I don't see any reason to prefer it. Both require a fairly steep cross at 2kHz to avoid cone breakup (LR6 would be ideal, but I believe the DCX doesn't support it).

Don't forget the complexities of phase management in a four way. If you go that route I'd recommend a linear phase PC crossover over another DCX. More precise sound and no hassles with phases in the different passbands interfering with each other.

[Space]

... vented pole pieces help a lot with the back wave, but I don't know of any six inchers with vented poles.

See post #2, the top image. The SB17 also has a large pole vent. None the less, there is measurable masking beginning in the 2 kHz region from the back, for both of these drivers.

Another note about nude drivers and 18Sound: the 6ND410 has a slightly different basket design without the cooling fins on the back, and there are three tapped screw holes. These are intended for adding a heat sink but could be useful for a magnet mount design. However this model would not be a good choice to reach as low as 250 Hz.

[Saurav]

I decided to run some numbers through Linkwitz's spl_max1 spreadsheet. It takes Sd and Xmax, and I used a path difference of 1/2 the frame diameter + 25mm (loosely, a baffle that's 1" bigger than the driver on each side). The SPL numbers are at 250Hz. Numbers are Sd, Xmax, D, SPL

Audax PR170M0: 139cm2, 0.5mm, 120mm, 89dB
18Sound 6ND430: 133cm2, 5mm, 106mm, 108dB
B&C 6MD38: 132cm2, 2mm, 110mm, 100dB
SB15NRXC30-4: 82cm2, 5mm, 100mm, 103dB
SB17NRXC35-8: 118cm2, 5.5mm, 110mm, 108dB

No surprises on the Audax, I know it can't do 250Hz on OB. The others were a little surprising, mostly on how close they were. I thought the SB15 would end up lower than it did. Still, as Chris said, that's getting the output by excursion, not surface area.

So it looks like 6ND430 and SB17NRXC35 would be the ones to research further.

[twest820]

D is the effective path length separation? If so, the numbers look low---don't have my data handy as I'm on the road but it's usually comparable to the driver diameter. The method StigErik and I hashed out over on DIY was to simulate the driver in Edge and plug its dipole peak into SL's spreadsheets. I've only used it with a couple drivers so far, but Edge has predicted the dipole peak within the limits of measurement accuracy.

I've remarked on this over at DIY, but rated xmax is 10% distortion. If you're interested in clean SPL the usable xmax is lower.

The SB17 also has a large pole vent.

All the better. :T

[Saurav]

Yeah, I could use EDGE. I thought D was 1/2 the 'baffle width'. The text next to the cell says 'effective path difference'. In that case D would double for all these measurements, which would push all the SPL numbers up by 6dB.

If you're interested in clean SPL the usable xmax is lower.

Agreed. So maybe they all need to be 6dB higher. Then add 6dB from 2 channels. Then these levels are 3dB down if the crossover is at 250Hz. I think peak SPL will be enough for my 12x15 room

[Saurav]

The EDGE simulations for the SB17 and the 6ND430 are very close. Going from no baffle to a small square baffle drops the peak from 1550Hz to around 1300Hz.

[Paul Ebert]

Well... You're treading around the waters I'm playing in, vaguely, so I may as well through some more stuff out there to make this even harder.

I'm currently (intermittently - haven't touched it in a couple months!) playing with custom dipole waveguides for a Neo3+Neo8 pairing, with the goal of running that straight over to a 12" or 15" lower midrange, which will cross into some woofage, and finally some sub-woofer (which will be monopole I am sure). I assume at this point that it may be WtmMW or so.

While you may not wish to go this route specifically, it may allow you to go mtMW or tmMW with a Neo8+Neo3 and a larger (10"?) lower midrange driver - more driver overlap on frequency range which may or may not be of use, but... IMHO it would be for dipole. As frequency goes lower there's something about having surface area behind the sound vs excursion (I say this particularly keeping in mind my recent experiences with the Anarchy...)...

C

Two things. First, could you explain more about your experiences with the Anarchy? Second, I'm curious about the custom waveguide for the Neo3+Neo8. Could you share a picture of it?

[Paul Ebert]

Perhaps the Vifa NE180W should be considered. Pricey, but obstruction of the back wave should not be an issue.

[Space]

Here's a post from John Janowitz, just today on diyAudio. He promises the 6" is available. Those AE speakers are very nice, so maybe this would be a winner. In a more perfect world, they would hire someone to keep their website up to date.

[Saurav]

I saw that. This is very interesting:

My goal was for this driver to be able to cover the range of 300hz to about 3Khz where they can then be crossed over to any number of dome tweeters, smaller ribbons, etc.

I'll keep an eye on that one for sure. Hopefully more measurements will start to show up.

[JonMarsh]

Something to always keep in mind that I think frequently gets overlooked is the need for the driver to be usable and controlled to at least an octave beyond the crossover frequency unless you're using steep cut off digital filters. (that is a tic for another long discussion, I suppose)

So for a working range of 300 - 2500 Hz, I look for clean behavior in the 150 Hz to 5 kHz region. Not easy to fulfill.

[Saurav]

That's a good point. I'm currently using the midrange between 350Hz and 2kHz, and I'm mostly looking for something that can go lower. I don't think I can move the M-T XO point any lower - others have used 1800Hz LR4 XOs with the Neo3, but to me it always sounds better when crossed at 2kHz or higher. I'll probably retry that when I get a new mid.

[twest820]

Slope matters, though. My experience is the one octave rule is aggressive at LR4 and slightly conservative at LR6. So the Neo3 can be OK lower if crossed steeper. For me the non-PDR is marginal at best with 2kHz LR4 but 1.8kHz LR6 is manageable, though 2kHz would be better. As I've commented over at DIY I would expect 1.7 or 1.6kHz LR8 to be OK. Wouldn't expect PDRness to matter much on the low end; if anything Zaph's PDR distortion measurements are better on the low end than what I got on the non-PDR.

Typical three way dipole compromises; only thing you stand to lose if throwing the DCX at it doesn't work out is some time.

[Saurav]

Can any generalizations be made about the consequences of higher order crossovers? I believe no one really goes above LR4 with passive XOs because of complexity and parts count, although I believe Jon's CE filters follow slopes close to LR8 for an octave or more. I've browsed through John K.'s article on crossover slope, and AFAIR, the power response null gets sharper (and deeper?) as you go to higher orders. I don't remember if he said anything about whether that sounds better or worse. I remember some article where someone tested XOs on identical drivers and I think they decided B3 sounded best to them, but I don't know how far that test can be generalized.

LR8 will probably be easier to implement with a DCX than LR6, since the latter will require LR4 + notch filters. AFAIK there's no built-in LR6 filter block.

[Dennis H]

Someone, (maybe dlr or Feyz?), plotted tweeter excursion vs. F, and LR4 actually stresses the tweeter less than LR6 or LR8 with the same Fc. It's because the LR4 has a more gradual roll-off in the shoulder region while the steeper ones stay flat lower and then roll off more quickly. Max excursion happens in the shoulder a bit above Fc.

[twest820]

LR4 actually stresses the tweeter less than LR6 or LR8 with the same Fc

Do you happen to recall if that result assumed a uniform power spectral density, a 1/f density, or something else? I'm guessing it's probably based on a uniform PSD, whereas 1/f is usually more representative. I've not worked with LR8 as Allocator (unfortunately) doesn't support it, but my experience with LR4 and LR6 is LR6 sounds better and allows slightly lower crosses. Though, as that's a subjective result, it's hard to disambiguate between tweeter excursion and all the other things which also change when crossover slopes and center frequencies are manipulated.

Can any generalizations be made about the consequences of higher order crossovers?

At least for me, subjective impressions of quality are dominated by how well the crossover avoids driver limitations. The steeper the slopes, the easier that gets. In particular I find it's not until four way there's enough latitude in choosing the midrange that it starts to make sense to cross the tweeter someplace other than as low as possible. However, I've not completed a four way yet and the one I'm working on is a Neo3/Neo8 where's still advantageous to cross the Neo3 as low as possible to get better vertical polars from the Neo8.

AFAIK there's no built-in LR6 filter block.

Likewise. Disclaimer: I don't own a DCX, so I have to read the manual. :B