Adire Extremis MT

[Davey]

Here's a photo of the Extremis sitting next to a RS180:

http://home.comcast.net/~dreite/Temp/IMG_0542.jpg

The difference in size and weight is dramatic.

Davey.

[TacoD]

Here's a photo of the Extremis sitting next to a RS180:

http://home.comcast.net/~dreite/Temp/IMG_0542.jpg

The difference in size and weight is dramatic.

Davey.

Wow, those neodym magnets are really small :E . Jon, I am looking forward to your modification of the 26T tweeters.

[GrahamT]

What are you planning on doing Taco?

[TacoD]

What are you planning on doing Taco?

A friend of mine has ordered a pair of lpg tweeters. He wants to compare those tweets with his MBQuart tweeters. Those MBQuarts sound nice, but I am a softdome kind of person. I really like the Morel Supreme 110 and Scanspeak 7000. Here in Europe those Supremes cost a fortune, at Madisound they only cost $250 a pair (compared to $500 here in the Netherlands). I would consider those as an alternative for the Seas millenium.

In the near future I will try those Dayton 7" woofers in a 2.5 way in combination with a Vifa XT25. This will be my 3th project with a Cauer filter a la Jon . I will also try this fitler in a 2.5 way Seas Excel W22 with Eton ER4 (a really nice tweeter and the cabinets are almost finished).

[JonMarsh]

If money is an issue, consider the 27TDFC as an alternative to the Millenium Excel.

It has a comparable resonant frequency, only slightly higher LF 2nd and 3rd harmonic, and lower higher harmonic distortion -




This is a comparison of nonlinear distortion using three tones clustered around the nominal test frequency (1 kHz); the red is the performance of the Millenium.

Objectively, the 27TDC and 27TDFC have very, very good measured linear and nonlinear distortion, ESPECIALLY for their price class.



I've got both (that's where the measurements came from), and will be evaluating both. The Milleniums shure are purty, though...

[TacoD]

I agree, those graphs look nice, but am I bit sceptic if those "cheaper" tweeters are really as good soundwise. Some tweeters somehow are more natural as others and measure almost the same.

[JonMarsh]

Measure similarly on amplitude, yes. Distortion, that's another matter- which is why I posted this.

A big factor in my opinion, besides the motor design (which doesn't look to be all that special on the 27TDFC, but still works, is the diaphragm. To say that there's more than a little similarity between the two is something of an understatement- same Sonotex, same surround.

But I can also understand pride of ownership, and again, the Millenium is very nice looking and well made with the die cast front panel. But I always self enclose the tweeter, so there's never any backpressure from the midwoofer- so front panel construction (which is pretty stiff even on the 27TDFC) isn't an issue.

~Jon

[Davey]

Jon,

I'm not positive the diaphragm's of those two drivers are as similar as you might think. Maybe there was a design change somewhere along the way to the Millenium because the pair I have (which are a few years old) don't have the same brownish color of those in your photo and the 27TDFC. Also, the surrounds look different and the dome "feels" slightly different when I touch it. Hmmm.

I'm not sure how hard Mark K. (I see 107db at 7khz, but is that a calibrated measurement?) was stressing these drivers when testing, but when I used both tweeters with a 1500Hz fourth-order electrical (active) crossover and cranked up the volume quite high the difference is certainly audible. I think this is obviously the result of the higher Xmax capability the Millenium tweeter has.

Cheers,

Davey.

[JonMarsh]

That's good feedback, Davey. Your comments jibe somewhat with Mark's results, because he noted lower 2nd and 3rd harmonic distortion when testing at low frequencies- in the 1000-1500 Hz range.

These samples were just bought from Madisound- it's curious that you note a difference in the older Millenium diaphragm from these. I was suprised to see that they were so similar; I even looked at them under a low power microscope, and the weave, material, etc, seems remarkably similar- I would expect some difference possibly in doping and treatment, but perhaps also they came on what they consider a "universal" approach which they make in bulk. Just speculation...

I don't want to tear them down and investigate internally yet, though that would be the next logical thing to do. The older standard Excels have a different color dome, too. Is the Millenium supposed to have a Faraday ring? I don't see anything in the specs about that. Many of the better tweeters with low distortion in the upper end do- because of the voice coil inductance reduction. This is true of my Hales, for example- a much flatter impedance curve than the Millenium's. That may be the only really critical feature missing from the Millenium's. Even without that, they're a great sounding tweeter. I think most of the folks that go gah-gah over the concept of a ribbon tweeter perhaps have never heard the Millenium Excel.

~Jon

[GrahamT]

Maybe there was a design change somewhere along the way to the Millenium because the pair I have (which are a few years old) don't have the same brownish color of those in your photo and the 27TDFC.

Here is a pic of a 27TDFC that has a different coloured diaphragm. Maybe it's just the camera though, I don't know.

[cjd]

Unless the camera is *really* poor quality, that 27TDFC does look to have a different dome material than what I have (and Jon has, since his look like mine). There is a distinct pattern to the fabric that I do not see, even if the color is off. The shape looks the same. Not sure.

C

[Mark K]

A couple of points.

Unless otherwise noted, all distortion measurements are at 106dB at .25m, or, roughly, 94dB@1m. This is calibrated. Twice. The mic is a mitey mic with an original cal file from Joe D'Appolito and the mic/preamp combo was recheck by KimG.

Second. The domes of the 27TDC/TDFC that I have and the millenium look the same. This doesn't mean they are the same. They just look the same.

Third, I don't doubt the Millenium and the 27 sound a bit different at higher levels. I keep meaning to post this page on my main site

What the heck are these nonlinear distortion graphs anyway?

Image not available

The 27TDC and millenium have different distortion patterns. I suspect the millenium sounds a bit more detailed, and does have very good/slightly better 2nd and 3rd products.

Whether the extra low level higher order distortion is a good thing, even though it can sound good, is a separate question.

One of the things I've noted is that almost universally, tweeters regarded as "detailed" have more higher order distortion. It's probably the auditory analog of the distortion that occurs when you "sharpen" an image in photoshop. It looks better from a distance, but if you zoom up, you can see the distortion along the high contrast borders. This could be the equivalent of distortion "brightening up" the transients which are 15-20dB higher and thus, have more nonlinear distortion.

It may sound better under the right conditions. Whether or not you want your loudspeakers to "sharpen" everything is debatable.

Is the perceived detail related to moving mass, transient response, etc. Well, maybe, but probably not. Transient response is a function of frequency response linearity. If a driver has good linearity it will have good transient response, regardless. This is Fourier's doing. John K and Feyz have converted me on this. If people perceive "detail" on a driver that has fair to poor FR linearity, then I'd be willing to wager money that if I test the driver, it will have a fair bit of higher order distortion.

Of course, take all of this for what it's worth. :

[Paul H]

Thanks Mark, very interesting.

Having run these sort of tests and likely having done a/b listening comparisons, do you have a "favourite tweeter"?

Paul

[JonMarsh]

I'm curious to hear the answer to that question, too, if there is one! I have suspicions that Mark has measured a lot more drivers than he's listened to - he seems to have turned into the Grand Moff of speaker testing for the DIY community!

[Mark K]

I'm curious to hear the answer to that question, too, if there is one! I have suspicions that Mark has measured a lot more drivers than he's listened to - he seems to have turned into the Grand Moff of speaker testing for the DIY community!

Absolutely. I can't possibly listen to all these drivers! And I don't think there is a "best driver."

The test difference between many of these drivers are small.

I used to think the OW's and ribbons were the cat's meow. I do think they sound "more detailed" than the high quality 1" domes. But lately I've gravitated towards the well built 1" domes. They sound more "natural." I can't help but wonder if the detail of the smaller domes and ribbons is truely more detail, or just some increased apparent detail.

But, it's hard for me to distinguish between say a 9500, and an Usher 9950. I'm never sure it's the implimentation vs a tweeter difference that is responsible for something I hear, or think I hear.

I think the xt25 is a great driver when used right. The only system I've heard the millenium in, the Orion, sounds fantastic. No tweeter problems there.

My goals for testing aren't really to pick a "best" tweeter. More to highlight certain aspects to allow one to optimally use a driver and to understand, in some small part, why a driver might sound a certain way.

I actually haven't done any extensive listening to the 27 series. I'm working on it, but you know, why build a two way with an RS180 and 27TDC when I know Jon will get to it and he'll do a better job? My only contribution might be a unit that sounds 95% as good as Jon's and has a cheaper xover cost.

Also, part of my rationale for doing these tests is to decide which drivers to work with. I can't possibly build systems with all these drivers. It's much easier to test them, then start designing around the ones that look the most promising.

I've avoided any flat out name calling. While most of the drivers I've tested are at least "good," there are a couple of real dogs buried in here, in my opinion. But I'll avoid explicitly naming them.

[TacoD]

I'll try those Seas 27TDC tweeters, but before I jump to conclusions, I have to listen first. Thank you for all your measurements Mark, it really helps to relate measurements to what your actually hear.

Jon, sorry if this is asked before, but how much distortion a x-over filter adds to a speaker? I assume the moving parts are the main source for distortion.

[JonMarsh]

A well designed filter with low DCR air core coils and film caps should add no non-linear distortion and negligible linear distortion (energy storage).

The main sources of non-linearity in drivers are the motor itself (magnetic system and voice coil), the driver suspension, and non-pistonic behavior of the diaphragm.

I myself have been correctly labelled as being quite willing to spend rather large sums on tweeters, so as one of my New Years resolutions I'm trying to investigate and understand the capabilities of the best performing inexpensive tweeters... Mark K has been a big help in that regard, steering me in some very good directions.

Fortunately that's not as much a stretch, as, say, Michael Shumacher trying to figure out what the nicest handling intermediate class Ford rental car is...

[cjd]

Had you not picked Ford, it would have been a fairly simple task. THere are some *exceptionaly good* vehicles that are ultra cheap, IF you can handle being seen in them. I'm aware of at least one $20k American compact car that went out and beat a whole pile of BMW M3's in a stock class - meaning, sticky tires, otherwise like it came from the factory - and this was a Divisional, so it was unquestionably some of the best drivers in this country in said cars. But it's not a Ford.

And, building stuff that folks without large lumps of change to throw at a speaker that sounds great is always good. That last few % of performance is something that a lot of people won't hear - the significant improvement that is available by carefully picking drivers in lower price categories, when compared to what a lot of people currently use, can be huge. Trust me here - I'm in that group. I'm designing on the shallower side of the price/performance curve where each dollar spent makes a bigger difference. But I've moved from a point where $50 in components (drivers, XO) was about right (for a 2-way) to a point where I don't think twice about dropping three to four times that isn't an issue. More, I'm not sure I would hear. But I'd be curious.

It's not all about the elite golden-eared folks. You get a few of them appreciating your hard work. Or droves of folks sporting fine silver ears in the late evening sun where they look vaguely golden.

C

[JonMarsh]

Well said, Chris.

Well, I mentioned Fords because I wanted to make it a little harder...


What you're getting at is at some of the core to what we've been working on since about 2000- it's why I've spent as much time messing around with 8" two ways as I have, because of the bang for the buck issues regarding overall balance and performance, and the ability to scale the design over a wide range of cost.

One of the guys I've been in contact with through the forum is doing M8ta's also; his cabinets are pretty far along;




for now, we think we can probably get away with the 27TDFC, but the upgrade to Millenium Excel is possible without a significant crososver change. It's kind of like the original M8 "bookshelf" with the Vifa XT25; you can slide in an SS9800 for nothing more than the price of the tweeters; another $80 per side gets you a worthwhile film and foil cap upgrade. ThomasW didn't think it would be worthwhile, until he heard it.

But you can get most of the basics at the starting point... its fortunate how some of these things have worked out.

~Jon

[markrpaulson]

Thanks for posting the photo, Jon-

I was initially considering the Excel W22 and the SS9800, which would have run ~$800. Fortunately, I came across this forum, and underwent this project when, serendipidously, the Dayton RS225s and the 27TDFCs were being tested--and getting great results. Now we're going to see how much performance Jon can squeeze out of them via crossover design. It somehow seems more exciting to me to go with the underdog drivers. I can always swap out Milleniums down the road if it'll help me sleep at night--until then, considering the total driver cost was <$150, now I can consider some serious crossover components.

[TacoD]

Nice cabinets Mark, how long did it take you to build these?

Jon, I always use air-core inducters and MKP caps. But if the value of the inductor is rather large I use a transformer core inductor (don't know if that's the right name for it ...).

[GrahamT]

Mark those cabinets look great. You might even get them finished before Jon does. :T :lol:

[JonMarsh]

Yes, Mark has done a fine job- especially considering he got them to this point in since starting in earnest around the beginning of November, if I remember correctly!

I've spent a lot of time prepping my new old stock rosewood veneer since the beginning of Christmas- softening and flattening it, then drying it out; I haven't finished all of it, but enough that I think it will work and I'm going to start veneering one cabinet soon. Actually, I originally bought Walnut paper backed and solid veneer, and started veneering the sides with that; I finished both cabinets just for consistency. If I was just doing the Walnut, (for which the raw veneer is nice and flat), I'd be all done by now.

But I'm hoping that the rosewood will look nicer when I'm done- (this is something of a gamble for me, as I've worked with Walnut and Oak before, but never rosewood- I did a lot of research to get ready for this.)

Here's a closeup shot of the trimmed side:




Today is "Mondo Measurement Day"; that's going to include new M8ta measurements, with 27TDFC, probably Millennium Excels, too, to finish design checks. I already have a "provisional" design 27TDFC using scaled impedance data from the Hales measured on the cabinet. Now I just need final measurement data - the idea being a massive data collection effort the next two days, then "digest" it all while at ThomasW's.

[JonMarsh]

Nice cabinets Mark, how long did it take you to build these?

Jon, I always use air-core inducters and MKP caps. But if the value of the inductor is rather large I use a transformer core inductor (don't know if that's the right name for it ...).

Yes, we've used laminated steel core inductors in some projects- such as ThomasW's Avo's. On mine, though, being the luddite I am, I used AWG 12 air core inductors. No hysterisis distortion there! Getting an iron core inductor that will really handle 400 watts peak without any distortion is difficult. ThomasW's is partly active crossed over, which makes things a skosh easier!

~Jon

[JonMarsh]

Jon,

I'm not positive the diaphragm's of those two drivers are as similar as you might think. Maybe there was a design change somewhere along the way to the Millenium because the pair I have (which are a few years old) don't have the same brownish color of those in your photo and the 27TDFC. Also, the surrounds look different and the dome "feels" slightly different when I touch it. Hmmm.

I'm not sure how hard Mark K. (I see 107db at 7khz, but is that a calibrated measurement?) was stressing these drivers when testing, but when I used both tweeters with a 1500Hz fourth-order electrical (active) crossover and cranked up the volume quite high the difference is certainly audible. I think this is obviously the result of the higher Xmax capability the Millenium tweeter has.

Cheers,

Davey.

Interesting thing I noticed today, Davey- doing tests in another room, which has "natural light" flourescents, the diaphragms look more grayish, not brown. The other room where the pictures were taken had halogen lighting- that must be making a difference in the blue spectrum, and how the camera sees them as well as how they look to the naked eye. In the room where I'm testing, they both look gray- and still quite similar.

~Jon

[Davey]

Jon,

Yes, I've seen a newer set of Millenium tweeters and the color of those domes look identical to the 27TDFC. I didn't touch 'em though...they weren't mine.

Anyway, back to the topic at hand....I'm trying to do my best Mark K. imitation by looking at the non-linear distortion of the Extremis driver relative to a few others I have. I don't have Praxis, but I'm using SpectraPlus and some shaped bursts I built with Goldwave. (Kind of a poor-man's Praxis setup.) The distortion levels of the Extremis look superior to two others I've checked, but I'm not finished yet. I'll post some further information when I get a chance.

Cheers,

Davey.

[GrahamT]

Thanks Davey. These drivers had so much hype, they better have low distortion. And I hope that the upper end ( at least until 2k) response looks better on future tests.

[JonMarsh]

I'm sure there's a lot of people in the DIY community that will be happy to see your data, Davey, including me. I did some playing with them last night; though the surround doesn't look so special they will truck on the Xmax! I'm off this week, and testing continues, though not as fast as I'd hoped last Saturday.

I have the LPG26T rebuild done on one tweeter, with some intersting results- I'll post shortly this week, probably Wednesday (tomorrow is travel day to Colorado). I don't think I'd use these tweeters without doing this... folks'll understand why.

~Jon

[GrahamT]

I look forward to seeing the results of the tests and pics of the mods. I'm glad that the mod made an improvement.

[JonMarsh]

Just a few data points

OK, guys, I'm not going to pour more gasoline on the flames over at Madisound. I don't see any need to get in an arguement with folks about what it all means.

Just for those interested, here are some measurements I took Monday, 1 meter as I would use for design, "nearfield" like Davey's, and an impedance curve.


First, the conventional measurement.






This measurement was made in the same small sealed Woodstyle cabinet I also made RS180 measurements and Peerless 850439; around 18 liters if I remember correctly. The point of the measurements wasn't LF behavior, but overall response from 100 Hz up, for designing the crossover and baffle step (same baffle width as target cabinet). Because of the stand height and gating, there's normall a small dip at 200 Hz on systems measured in this room. Due to the long gating window for LF resonse, there is roughness in the curve due to room reflections. Which you'll have when you listen in a real room. Still, it's a pretty decent response profile on axis at 1 meter.

The roll off due to baffle step seemed a bit less than for most midwoofers.





This measurement was also made in the same enclosure, but at a mic distance similar to Davey's, roughly planar to the surround. Curiously, the driver response seems to slope down about 3-4 dB between 100 Hz and 1 kHz. I do not think this is a measurement error; I think it's how the driver raw resonse works- perhaps a function of it's XBL2 technology and long Xmax? I'm only speculating. Most other midwoofers I measure near field this way are quite flat up to the first cone mode. Note the relatively large nearfield peak at 1.5 kHz. I don't think this is a measurement artefact, because some other drivers, like the M8a, show no sign of this behavior, and have near field and farfield measurements which agree fairly closely. I wonder if there's something else at work, like cone profile?

Last is impedance curve:




This agrees pretty closely with Adire's curves. The curious thing about this curve, compared with say, SS SD1 motor drivers like the 7" midwoofers, is how the impedance is has a slow rising slope until about 7-8 kHz, then takes off with a much faster rate of rise. The SS drivers have a near straight line impedance rise. Not that this should make any difference in the range we'll use this driver.

[Davey]

Jon,

What's that little double-hump on the resonant peak in the impedance trace?

Davey.

[GrahamT]

Thanks Jon. The far field and impedance graphs do somewhat resemble the Adire ones. I dont see anything too scary there, am I right?

[JonMarsh]

Jon,

What's that little double-hump on the resonant peak in the impedance trace?

Davey.

That's what I wondered, too, Davey, but I was in something of a "production mode", and after I noticed it figured I would go back and investigate later. Later my daughter broke her toe, and also informed me she had an interview she'd been waiting for six months on Thursday, so I had to reschedule her deaparture flight for Denver to Friday from Tuesday (yesterday). So, things didn't quite go as expected on Monday!

[JonMarsh]

Thanks Jon. The far field and impedance graphs do somewhat resemble the Adire ones. I dont see anything too scary there, am I right?

I've seen many other drivers with nearfiled behavior like this- including the RS180. OTOH, I've got some that don't exhibit this behavior. HiVi M8a, M6a, for example, track very well on farfield tests and nearfield. These are usually the driver's I've stuck with in hopes of having true pistonic operation in the midrange up to the crossover point.

I think we'll have a clearer picture of the driver behavior after I get one sent to Mark K for ETC tests.

For my main reason buying these, these look very interesting- they are capable of some pretty long throws! I wish the basket rear opening took more advantage of the small magnet- it's not really as open as the first photo's I saw led me to believe. But perhaps that's necessary to have an adequately large spider.


I'll have a nominal crossover design done within a couple of days, I expect- the only thing slowing me down is having several things like this happening in parallel (the two MTM designs, three if you count the high efficiency one, which is a lower priority), and final tweaks on the M8ta, plus posting the LPG26T tweeter rebuild/alignment (coming soon to an HT Guide thread near you!).

~Jon

[GrahamT]

I'll have a nominal crossover design done within a couple of days, I expect- the only thing slowing me down is having several things like this happening in parallel (the two MTM designs, three if you count the high efficiency one, which is a lower priority), and final tweaks on the M8ta, plus posting the LPG26T tweeter rebuild/alignment (coming soon to an HT Guide thread near you!).

~Jon

Thanks Jon, I appreciate it as do the lurkers here, judging by the views.

That is a pretty full plate you have. I have been reading the Madisound threads and I am looking forward to seeing Marks testing results. It's fun to watch the process as we learn more about these drivers.

[TacoD]

I am looking forward to your designs!

[sbolin]

Thanks for the info, Jon, look forward to the designs.

[GrahamT]

Welcome sbolin,

Are you doing one of Jon's Extremis designs or one of your own?

[sbolin]

Welcome sbolin,

Are you doing one of Jon's Extremis designs or one of your own?

I have my own designs, but want to see what Jon comes up with. I need help with the x-over design, so like to see what others come up with. I am planning on doing a TL design, both MTM and MT using the Extremis. Frankly, the FR plots are scaring me, though. I planned on crossing over around 2.8kHz, but the response is so ragged I am not sure

[GrahamT]

I know what you mean. Jon's crossover will be as low as possible, while keeping tweeter excursion and distortion as low as possible. I dont know of anyone else who can get such steep attenuation and this driver appears to need it. I'm confident that he can make these things sound sweet. Frankly, even though the FR is ragged, that is a price I am willing to pay for a 7" stand mount two way that will troll down in the 20s. It will be fun to see what people come up with for this driver.

[Mark K]

I'm looking forward to testing this driver. I'm not so worried about the FR irregularities high, as a low xover makes sense for this driver. It would be a very good candidate for a TL, or ML TL. In Martin's mathcad worksheets, usually a cross sectional area of 2.5-3.5 x Sd works well. It makes for a very large cabinet, say for an 8" driver. So an extremis based 2 way could have a very small footprint and still pump out bass.

It had better have good distortion numbers, since you'll be sacrificing some sensitivity, and the driver is a bit expensive, compared to the RS180. (Two RS180's in jon's mtm are still cheaper than one extremis, so you get better sensitivity at a smaller $$. The extremis only works if it can do as well as two RS180's or better with respect to nonlinear distortion.

[GrahamT]

I'm looking forward to testing this driver.

I'm looking forward to the results. You do some good work my friend. :T

Dan has posted the FR of his design. The Xover is around 3 kHz I believe and it's 1/3 octave smoothed.


Image not available


It is encouraging that he has achieved such a flat response.

[JonMarsh]

Well, as you know, my measurements aren't that smoothed, so they don't look as "pretty", being 1/12 octave resolution, not 1/3.

Anyway, working on that in the test cabinet, it's coming together I think.




It's a little hard to compare directly. This plot shows down to over -50 dB from nominal. The HF breakup is pretty well suppressed.

This is NOT the ported optimal cabinet, just my small Woodstyle test box, which was

A) Available and ready to load with the right size hole

B) Not ported, and pretty small - but close to the right size for getting the midrange crossover (nominal 1400) OK.

I'll post more details tomorrow- too tired for now!


~Jon

[GrahamT]

Looks really good Jon! These are completely different designs I agree. I dont think there is any comparison.
I'm going to start to mod the tweeters today. I'll let you know how it turns out.

[JonMarsh]

Yeah, 3 kHz is a little high, IMO. The wavefength issues with driver spacing and the lobe in the crossover become kind of problematic for getting good power response- plus, I haven't seen a 7" driver that has good off axis at 3 kHz. It would be interesting to compare the off axis horizontal and vertical plots for the two designs- I think that has a large bearing on the sonic character and transparency in room. But I expect my design will have the disadvantage of more expensive crossover parts.

Note that the tweeter here is down 25 dB at 1 kHz in our design, almost as much as the for the 3 kHz crossover - by 800 Hz or so, the attenuation will be higher than for the 3kHz crossover. :B

I'll have a start an official "Design/construction" thread by tomorrow in the same format as for the MTM; I'll have to fill in the details on the first page as I go along.

Here's the Crossover schematic.





Regarding off axis and power response, this is LspCAD 6 predictions based on driver diameter, spacing and crossover frequency at 15 degree, 30 degree, and 45 degree off axis, plus net power response.



I've had very good agreement between LspCAD and measurements on these parameters in the past. I think you can see how even a 1400 Hz crossover isn't completely optimum- 3kHz would be much worse.

~Jon