Three Way Evil Design Study

[Jed]

I know the distortion measurements from KT weren't that great on the Ti100's but the SQ was awesome.
Jim

Jim- I'd have to disagree with you on the performance of Ti100. The distortion looks quite good to me if using it as a mid from 300-2.5K. Also, all the TB specs I've seen have been quite bad in contrast. Though I haven't seen specs in Klang and Ton on the TB titanium driver yet.

Image not available

[JonMarsh]

Thanks for the explanation regarding Darren's quick measurements, Jim. One thing possibly appealing about the TB is that all holy heck doesn't break loose until somewhat higher than the Visaton TI 100.

Well, however it turns out, it's not huge money, and it should be interesting to see how it does- my interest is basically 400 Hz and up to 2-2.5 kHz. 3 mm might be enough... (cross fingers) depending on the linearity of the motor system.

[LoudandClear]

Here's a couple crossovers with the TI100

Image not available

Image not available

[Jim Holtz]

Thanks for the explanation regarding Darren's quick measurements, Jim. One thing possibly appealing about the TB is that all holy heck doesn't break loose until somewhat higher than the Visaton TI 100.

Well, however it turns out, it's not huge money, and it should be interesting to see how it does- my interest is basically 400 Hz and up to 2-2.5 kHz. 3 mm might be enough... (cross fingers) depending on the linearity of the motor system.

Hi Jed,

I wasn't implying that the Ti100 was bad, just that it might not be at the same level as the Accutons Jon is testing. Actually, I found them to sound superb as I indicated in my post.

Jim

[capslock]

The W12CY001 is not without it's problems... including low overall sensitivity, frequency response which begins drooping above 2 kHz (due to cone flex, I believe, as the W22 does above 1 kHz), and a wonderfully wide band cone break up mode commencing at 5 kHz with significant output to above 15 kHz. It is most curious how Seas doesn't provide an impedance curve for this component that extends above 5 kHz- that's undoubtedly where the issues begin to arise.

Reading this made me realize what you might be pointing at. The W22 has a hump beginning just above 1 kHz and then droops above say 1.5 kHz. The W12 exhibits the same phenomenon, just a little higher in frequency.

The theory I heard about this from an experienced driver designer years ago was that this is not cone flex but rather a resonance off the air trapped in the cavity formed by cone and surround. I don't know if it is true, but I looked at a bunch of HobbyHifi driver reviews at the time, and the trend was clear: the hump was more evident the harder and deeper the cone, and its position roughly correlated with the diameter of the cone.

[JonMarsh]

Reading this made me realize what you might be pointing at. The W22 has a hump beginning just above 1 kHz and then droops above say 1.5 kHz. The W12 exhibits the same phenomenon, just a little higher in frequency.

The theory I heard about this from an experienced driver designer years ago was that this is not cone flex but rather a resonance off the air trapped in the cavity formed by cone and surround. I don't know if it is true, but I looked at a bunch of HobbyHifi driver reviews at the time, and the trend was clear: the hump was more evident the harder and deeper the cone, and its position roughly correlated with the diameter of the cone.

Well, I'm sure not an experienced driver designer, and I have no where near the knowledge of driver design which you do, Eric.

Maybe the hump in the response results as you say- I also think rear wave reflections can play a role, such as in the C79. But the overall issues in many Seas Excel and Dayton RS drivers result (IMO) because of the cone profile and the resulting behavior as you push higher and higher, which is quite similar for these cones. Let me illustrate some of the measurements which have led me to conclude this (erroneaously or not).

Exhibit A, 1 meter response of W22.


Any experienced DIY guy would look at this and recognize the W22 response profile.

Next up is a close nearfiedl measurement in the center.


The notch/dip is (IMO) a symptom of a bending mode which is initially resulting in a dip in the repsonse; as one goes higher in frequency, you hit the all heck breaks loose mode, which is masked somewhat in this measurement because of the change in wavelength and drop off in response above 2 kHz due to arrival time phase differences.

One easily can question whether that supposition/theory is correct-

The 8" Peerless 830884 behaves simiarly, but it's dip is at a more typical frequency for the first decouping/flex for an 8" woofer, 900 Hz.

Exibit B: HiVi M8a 8" midwoofer at 1 meter.


It's not as high tech a cone composition as the Excel; more mass, and mostly aluminum, with it's breakup peak at ~2.1 kHz. BTW, having measured many of these, they're at least quite consistent for the cone behavior, unlike, say, Accuton's. Better CPK on the process.

Here's the HiVi M8a nearfield response in center- at the edges it looks very little different, except for roll off of the 2 kHz and above area due to path length.



This is why I like concave spherical diaphragms, probably why Accuton and Revel and many others do, too.

Wish I could buy a driver with an improved version (metallurgy) of this cone, and an W22 or RS225 motor. Of course, a C220-T6-220 is pretty close, or more than pretty close, but the price is rather dear. Still may try a set in my M8ta some day. But the motor is not as symmetric as a W22 or RS225- higher 2nd order distortion.

[jkrutke]

Tang Band is a company that could be pretty scary if they really got their act together. They make a lot of stuff for other folks (can you spell Adire, Ascendant?), I'm also curious how the 25-1414SC tweeter plays out, for the small form factor applications.

I've got my eye on the 25-1414 TB also. Like you, I need a small format tweeter for a couple apps. I think I'll pick up a pair. So far however, assuming the D26NC55 will run out, I'm liking the Seas 27TAFNCD. Particularly the mounting method. A little 1/8" surface mount flange can easily overlap a mid driver flange and get super close without tedious odd shaped countersinking or flange modifying.

If you grabbed the W4-1337S Titanium 4", I'll anxiously await your comments on it. I just can't bring myself to blow $50 on a 4" Tangband. Of course, I blew $80 on a 10" TB a couple weeks ago... not the best $80 I ever spent. 8O

On the subject of Tangbands, the W5-704 isn't too bad. Great response curve and average non-linear distortion performance. I could probably forgive the non-linear performance, since it's a smooth extended driver that can cross over relatively high to a cheap tweeter. (assuming one is willing to blow off all that stuff about operating in the pistonic range.) :B

Edit: if the 25-1414 uses the same motor as the 25-1166, that's probably not a good thing. I guess I'll find out.

[cjd]

So, why not more Aura... NS4-255-8D1 - distortion too high? Sensitivity too low?

C

[JonMarsh]

I ordered the W4-1337S yesterday, should be in by the middle to end of the week. I get you about blowing $50 on a 4" Tangband, but who knows...

That little Seas doesn't look too bad, especially if you do keep it above 3 kHz or so. Something to consider...

I wish the D26NC55 wasn't going away, I hope they actually DO have a hundred or so at Madisound. It's a bargain for three way system performance. Thinking about a low buck three way using the D26 and this Tangband if it pans out.

Unfortunately, it looks like the same basket on the 25-1414 as the 1166.

[JonMarsh]

So, why not more Aura... NS4-255-8D1 - distortion too high? Sensitivity too low?

C

No, it's pretty good above 200, and yes, at 4 ohms, not 8 ohms. stamped frame, not cast. Top end gets ragged above 5 kHz. Considered and discarded it. Not that it couldn't be a pretty good cone mid in a smaller three way, say, 400 Hz crossover; anything with dual midwoofers would be too sensitive for it, though. Say, with a single RS225, and something like a D26 brought in above 3 kHz. But you have a three way crossover; I'd rather spend the money on making a good two way, with a better tweeter. Three ways with 4 ohm midranges usually result in hard to drive speakers.

[Brian Bunge]

That little Seas doesn't look too bad, especially if you do keep it above 3 kHz or so. Something to consider...

Sorry Jon, but which Seas is that?

[WillyD]

I believe Jon was referring to the 27TAFNCD that John Krutke mentioned.

http://www.seas.no/Prestige%20Diskanter%20PDF/27TAFNCH1397.pdf

[cjd]

No, it's pretty good above 200, and yes, at 4 ohms, not 8 ohms. stamped frame, not cast.

Yeah, I see stamped frame but the 8D1 is an 8ohm version... or are you saying you would only use the 4ohm (4D1)?

I, of course, have my eye on it for use in a HT setup, MTM with a larger driver covering below up front, and probably just straight MTM for surround duty with the NS4. Haven't found much that offers the clean response and the travel to actually do well enough stand-alone as surround. Still looking.

C

[JonMarsh]

Hey, I don't mean to pick on the little guy. I think they're recycling the FR graphs, and I really doubt that the 4 ohm version and 8 ohm version have identical sensitivity- if they do, someone is doing something wrong...

A baby MTM with the 8 ohm version, crossed to something like the D26CN55 could be real cute- could tolerate 3 kHz or even a little higher. One thing that's tough with basket designs like this (there was a 5-1/4" Seas PP cone driver back in the 80's) is rear wave masking when you try to mount them in a 3/4" baffle. I got around that using machined aluminum plates to mount all the drivers, mounting that on the baffle, but that's a little beyond the ken for most people, and raises the ante a bit.

What I don't recall seeing on the Aurasound specs was the Xmax- do you know what it is? I'd hope at least 3 mm.

Even if I use the Tang band W4-1337 which seems to have a better basket design, I'm already considering aluminum panels for the same reason. Fortunately they're not too pricey from McMaster-Carr.

Now if only I could get someone to support me, so I could BUILD every idea that comes into my head.

[augerpro]

I wonder if any of these would be a good replacement for the D26NC

http://tymphany.com/datasheet/printview.php?id=364

http://tymphany.com/datasheet/printview.php?id=341

Also I the Morel MDT22 seems decent as a small tweet. I just bought one, I'll post measurements hopefully this week.

[dawaro]

Even if I use the Tang band W4-1337 which seems to have a better basket design, I'm already considering aluminum panels for the same reason. Fortunately they're not too pricey from McMaster-Carr.

If this is the road you go let me know and I will donate them. I have access to 1/4" aluminum plate at work and a buddy with a water jet. We can call it payment for all the advice and the Nat P request I had.

[Dennis H]

Looks like plenty of Xmax on the 4" Auras. Gap = 16.7, coil = 5.8.

I still think a line of the 3" or 4" Auras would give the big BG planars a run for the money as a line-source midrange. I can't see anything the BGs do better -- performance, cost, width to get them close to the tweeter line, whatever.

[JonMarsh]

Now, why didn't I see the gap spec? Must have lost my glasses today, or forgot to put them on... 5 mm sounds pretty reasonable.

I'd say the NS4 covers a bit different range than the big BG planars- the RD's are optimum in the 500-5000 Hz range (too high for the NS4, IMO) the NS4 looks pretty killer up to maybe 2.5 kHz with a steepish crossover. OK, here's your wacky line array concept, NS4 in a string column up to 1 kHz, the Tang Band up to 6 kHz, and Fountek CD2 or my old original's from 5-6 kHz up. Still not cheap- let's see, for 60" line we'd need 18 NS4, that's $450 a side, a similar number of TangBands (~$900 a side) and the ribbons...

There's gotta be a better way to do that...

[Dennis H]

the NS4 looks pretty killer up to maybe 2.5 kHz with a steepish crossover.

Right and they could probably reach down to 250-300 or so. A line of JPs should be able to handle above 2.5K with no problem and a line of 10-12" should be able to handle up to 250-300 so you've got a real 20-20K high-output 3-way.

Things that make you go hmmmm.....

[cjd]

Hmmm, aluminum eh? Now there's an idea!

XMax specs I've seen suggest 5.25mm on the NS4.

Also, looking at the raw T/S data Madisound provides (not in the PDF) they do indeed seem to be about the same sensitivity for the 4ohm and 8ohm versions.

C

[TacoD]

What about the new peerless full range? It featueres a neodymium motor an aluminum dome and copper in the motor. One problem, however, it is an OEM driver.

[JonMarsh]

Pretty remarkable range, but I get nervous when they don't publsih an Xmax, power ratings, and the sensitivity number is that low. Still, I'm sure they're not building it unless they have a good market for it- Upgrades for Bose HT Satelites, anyone?

[capslock]

12 cm² cone area, that's hardly more than a tweeter.

[TacoD]

Also a 3" is coming out in february, yes the product is aimed to accompany their LAT offerings. I just wanted to inform you on the newest offerings of Tymphany.

[Hank]

Looks like Smith and Larson are doing some interesting things... just what I don't need, more things to spend money on!

Yes, that Speaker Tester (ST) is mighty tempting: http://www.woofertester.com

[Evil Twin]

Aluminum plate at reasonable prices- delivered promptly - and I did not need to resort to rebel smugglers.

Unless McMaster-Carr has alliances I am not aware of...

[cjd]

http://www.onlinemetals.com/

C

[Evil Twin]

http://www.onlinemetals.com/

C

Interesting... most interesting.

[Paul W]

www.metalexpress.net is also a great source. ME is actually a network of suppliers around the country so nothing is ever "backordered". I've ordered custom cuts 15-20 times over the past couple of years with delivery as quick as 3 days order-to-doorstep.

[cjd]

What I like about onlinemetals is not only the variety of options and their cost structure, but the fact that you can get custom cuts (within reason).

Their plastics selection is also very interesting. Very very interesting.

I can see the Emperor demanding 3" Titanium for the baffle now!

C

[Hank]

following an ad in CES Daily, I discovered this little interesting driver:

http://wave9systems.com/hi-x.pdf

[JonMarsh]

73 dB /1watt sensitivity!?!?!?!?!

SHhhheeeeeeshhhh! That makes the Arvo Part sensitivity look like a K Horn in comparison!

[dlneubec]

What I like about onlinemetals is not only the variety of options and their cost structure, but the fact that you can get custom cuts (within reason).

Their plastics selection is also very interesting. Very very interesting.

I can see the Emperor demanding 3" Titanium for the baffle now!

C

Has anyone ever tried something like UHMW polyethylene or just HDPE for a baffle? It might be an interesting material to try. I've purchased some industrial HDPE from these guys for pretty cheap for another use in the past and that stuff is heavy, dense, easy to cut, drill, machine, no sawdust problems,etc. I don't know if you can glue it at all, so it might only work for bolt-on type baffle. These guys use it for cutting boards and even have it in some colors.
Cutting Board Factory

[Hank]

I just opened an old link to see what it was, and found this:

http://www.ascendantaudio.com/poly_6.5_page.html#Technology

Have I missed a post on this? The fr graph looks good. Prodigious Xmax. Of course it's not a metal cone... :roll:

[jkrueger]

Under Budget Maybe?

There have been a lot of drivers suggested for the mid of this cool project. Every single one of them has been substantially less expensive than either accuton mid suggested at the beginning....I'm wondering if this project might come in under budget???


:fart:
Do you smell something?


Jon

[cjd]

I just opened an old link to see what it was, and found this:

http://www.ascendantaudio.com/poly_6.5_page.html#Technology

Have I missed a post on this? The fr graph looks good. Prodigious Xmax. Of course it's not a metal cone... :roll:

Yes, poly sounds muffled...

[JonMarsh]

There are really two projects here, it seems, the M10ta, which is more "budget" focused but with similar cabinet design (if you so choose), and Evil Twin's more "no holds barred" M12ta, with one of the Accuton mids, C30N tweeter, and the NS12 Aurasound woofer.

Now, if there's a surpise upset in driver performance, who knows? Maybe the M10ta should be specifically pitched as "how close can we get for a whole lot less money?"

It might come in much more like the M8ta with Rs28a as regards component cost, but the cabinets still won't be a walk in the park. But, imagine an RS270 up to 400 Hz or so, possibly the Tangband or RS125 up to 3 kHz, and any number of reasonably low distortion tweeters above that- H1189, D26CN55, your choice... just needs to be fairly clean at up to 104-106 dB in the required range.

The RS270 would require a fairly large cabinet, so same basic 75 liter design as the M12ta; - 6dB anechoic with Fb tuning of 26 Hz with 3" ID flared port is about 30 Hz, with significant output (up to 100 dB anechoic per cabinet) down to ~24 Hz. Not as good as the M12ta projects, but... a bunch less bucks and more sensitive. With 5 dB of BSC, overall sensitivity should be ~85 dB.

[jkrueger]

Don't bother

Thanks for the quick respons John

This feels like a stupid question....but I will ask it anyway.....Is the Mt10a going to be a 3-way? I only ask because I have looked at the FR chart for the Aurasound 10". I would imagine that you could probably create a 2-way out of it.


Jon

Edit* Too late and a few drinks=stupid ops:

[JonMarsh]

The M10ta will be three way; the cone control of the Aurasound appears remarkable as does their NS12, but there's also beaming with rising frequency relative to the larger cone to consider. For the approximately 6" diameter cone in an 8" woofer, 1000-1200 Hz is OK; for the 7-1/2 to 8" diameter cone in a 10, an upper limit of maybe 900 Hz is appropriate, which to keep some good transistion band behavior, means maybe 600 - 700 Hz max in practice?

The Dayton RS270 has a reasonable looking on axis FR up to 1 kHz, but energy storage behavior indicates one might prefer to restrict it's range to 600 Hz.

I think you could do a two way possibly with the Aurasound, but it would require something like a BMS 4540 driver and a DDS waveguide working down to 1 kHz, which is a bit of a stretch even for those pieces.

To get a nice room energy power response, the mid range and even the tweeter will be operating in close to 4 pi space at the lower end of their range, to improve 60 degree off axis capabilities. This is a function of driver selection, and cabinet and crossover design.

[jkrueger]

8O Hello? Anybody home? 8O :tv:

[crackyflipside]

What is the guesstimated budget for this new 3-way?

$500 / pair?

$750?

$1000?

[Evil Twin]

The parameters of the design establish the basic budget because of the limited cost/performance ratios of available drivers.

In this case, the expected driver budget per pair is about $1500-$1600.

While some designs have been known to have crossover costs well exceeding the cost of the drivers, that is not expected in this case, though it is a three way, and wil require solid impedance compensation networks as well. An approximate estimate would be $500, perhaps $750 with some upgrades or unforseen issues.

Then, there is the cabinet- BB ply and HDF primarily; choice of veneer will establish the upper cost limits. A reasonble total estimate is under $3K.

That is a fairly high cost for a DIY system, but understand the target is systems in the $20-$30K range like the Avalon Eidelon.

This is for the M12ta. An M10ta with RS270 and alternate midrange and tweeter might be half that cost.

Moving the Imperial labs this week has limited available time for testing, but that may be remedied this coming week.

[crackyflipside]

That proposed budget even looks reasonable for a piece of Bantha pudu like me.

I have great confidence in the Dark Lord of the Sith.

[Evil Twin]

Regarding the Tangband Midrange

The Emperor is not pleased with the delays, but accepts that the relocation of the Imperial laboratory has some unavoidable consequences.


First, measured SPL and phase on a 2' X 2' baffle with short window (1.6 msec) to avoid room and baffle effects.


Clearly a potential candidate for an extended top range mid driver, rivaling dome midrange units.



Next, HD2 measured at 1 Watt with an 8 msec window, so that lower frequencies can be seen (using Praxis Flex window mode).


Not as low as might be liked, but probably usable.


Last, HD3 level, which gives more clues about how hard it's working at 87 dB output...


Not at the levels which a Peerless 830883 achieves over a narrow frequency range, but nonetheless impressive... most impressive. Probably very usable wtih 3-5 kHz crossover with LR4 or better. Certainly an extended range for a cone device in it's price range.

[Evil Twin]

Accuton C79 basic measurements

SPL measurements made with short window on 2' X 2' baffle to reduce room effects, with flex windowing on.



As is the case with most published C79 measurements, there is something of a lift in the 1-2 kHz area, possibly because of the limited rear window.


HD2 with 8 msec window at 1 watt drive


On this alone I would suggest a uesful range from ~500 Hz to no more than 4 kHz.


HD3



HD4


Altogether, possibly a useful operating range from 350 Hz to no more than 3 kHz.