BSM 4592

[mark1101]

23 hours ago, ClaudeJ1 said:

Looks like you'll be upgrading to the Celesion Axi's soon..............now that they are available in the USA, which they were not about a year ago.

 

Those look to be about $1k apiece.  I only paid $600 each for my four TAD drivers.

[ClaudeJ1]

1 hour ago, mark1101 said:

 

Those look to be about $1k apiece.  I only paid $600 each for my four TAD drivers.

You got a steal deal on those TADs. I guess if you want 300 Hz perfomance you have to go bigger than TAD. Ad former TAD owner, I can say you don't need to upgrade anything.

[OO1]

On 1/14/2021 at 4:33 PM, ClaudeJ1 said:

Looks like you'll be upgrading to the Celesion Axi's soon....

this driver seems to be very hard to find ------no inventory anywhere

[Chris A]

Just a few relevant comments on the Celestion Axi2050 performance vs. the TAD TD-4002 and BMS 4592ND 2" compression drivers: 

 

I haven't gotten one of these new Axi2050 drivers yet to test, and I haven't yet dialed one in.  But I see that the only real advantage of this driver is the 320-410 Hz performance over, say, a TAD-TD-4002.  The cost of the Celestion Axi2050 is $1K (USD)--this is about the same cost as the Radian 950BePB beryllium dome 2" driver.  The TADs went for $2000-$2500 (USD, street price) new when they were still being manufactured.  That's the advantage, and if you focus only on the performance, having an extra 100 Hz of pass band isn't a lot.  I can tell you that the TADs will have better HF output and overall smoothness of SPL on the top end.  If you already own TAD TD-4002s, I really don't recommend selling them because their sound quality (IMHO) can't really be topped.

 

The only advantage of the Axi2050 over the BMS 4592ND (bi-amped) that I can see is just that, the need to bi-amp--which is merely an added complexity that you get with having a three-way loudspeaker instead of a two-way, but without the polar lobing and need for a separate driver and a separate horn aperture. 

 

And this ability to bi-amp using the BMS dual-diaphragm drivers is not really a disadvantage, because they have much lower FM distortion vis-à-vis a "full range" 2" driver.  In other words, when playing recordings having a lot of midrange and higher frequencies playing simultaneously, the dual-diaphragm 2" drivers (properly dialed in) at higher SPL will have more clarity from the lack of FM sidebands at higher frequencies.  These FM sidebands make the output sound more "opaque", and are created by the driver producing more than one frequency at a time, i.e., the reciprocating motion sound of Doppler distortion. 

 

Both drivers require a full-range horn to realize this wide SPL response band--the only one of which I'd really consider is the K-402.  The pictures and videos of Celestion using large diffraction slot horns (the red ones) aren't really my cup of tea, since these will tend to promote the "frying bacon" sound of higher order modes above 4 kHz--the cut-on frequency of 2" width diffraction slots.  This frying bacon sound shows up at higher SPL--probably around 90-95 dB at 1m. 

_____________________________________________________________

 

Finally, one should be clear about why the Celestion Axi2050 is useful in a Jubilee application, using the K-402 on top of a bifurcated (two mouth) "W" section bass bin.  At higher frequencies, the Jubilee bass bin begins to beam, particularly in the horizontal direction, and above 210 Hz, this behavior becomes more and more apparent, with only the output from the crossing K-402 with 2" compression driver to widen the horizontal and vertical polars between 210 to about 500 Hz.  This isn't a problem if the bass bin has only one mouth instead of two.  The BMS 4592 can also be crossed at a lower frequency than the TAD TD-4002, and taking it down to 300 Hz isn't a problem using DSP crossovers.  If this is your concern, on an affordability and lower modulation distortion basis, I'd actually recommend the bi-amped dual-diaphragm BMS 4592 than using a single-diaphragm 2" compression driver.

 

In the K-402-MEH, there is but one (large) horn mouth, so the need to cross over below 400 Hz isn't really necessary.  These 2" compression drivers really like to be crossed above 400 Hz. The 5" diameter voice coil of the Axi2050 helps down low, but the physics of a 2" throat compression driver still has an overwhelming influence over the low frequency cutoff point. 

 

Below is the Axi2050 on-axis SPL and electrical impedance response on an (unspecified) 90x40 degree horn as tested by Celestion itself in one of its earlier brochures:

 

 

Chris

[Chris A]

A picture of the diaphragm of an Axi2050.  It doesn't look inexpensive to me.  The voice coil is perpendicular to the diaphragm at mid-point radially.  Technically, it is a ring radiator diaphragm, just like the two BMS4592ND diaphragms, except this is titanium, and the BMS diaphragms are polyester:

 

[mikebse2a3]

2 part Videos talking about the technology and challenges in Wideband Compression Driver Design - Dr Jack Oclee-Brown

 

Part 1 This is just a 2 minute introduction for the meeting.

Part 2 Shows research (special attention at ~13 minutes into Part 2 Video) demonstrating the issues dual diaphragm compression drivers like the BMS 4592 have and why Celestion went with the Axi2050 design.

 

This is very impressive research and development IMHO..!  

 

miketn

 

 

 

 

[ClaudeJ1]

1 hour ago, mikebse2a3 said:

2 part Videos talking about the technology and challenges in Wideband Compression Driver Design - Dr Jack Oclee-Brown

 

Part 1 This is just a 2 minute introduction for the meeting.

Part 2 Shows research demonstrating the issues dual diaphragm compression drivers like the BMS 4592 have and why Celestion went with the Axi2050 design.

 

This is very impressive research and development IMHO..!  

 

miketn

 

 

 

 

I agree. Great watch and justification for taking a different path, which Roy agrees with.

[Chris A]

So modal decoupling...and the realization that thick/stiff diaphragms aren't the solution--rather "axi-periodic" repeating patterns of changing shapes are the solution to keep the mass effects at bay--are the techniques used.  (I would have thought these techniques used in the aerodynamic fields for perhaps 50 years or more would have long ago penetrated into acoustic driver design--but apparently not.)  Good on Celestion to raise the bar, so to speak, and force the other compression driver manufacturers to do a better engineering job.  However, Celestion needs to patent this (and it's already been over a year, so the patent application had to already have been submitted a couple years ago or more now). 

 

Some observations on the YouTube videos (above):

 

1) In the end, he's showing lower second harmonic distortion at the lowest frequencies relative to dual diaphragm drivers (and I'm assuming he's talking about the Axi2050 vs. the BMS 4592ND).  That's interesting, but there are three points to take notice of:

 

     a) second harmonic distortion, while not desirable, is also, in itself, not very audible.  PWK said this many times.  The fact that second harmonic distortion eventually turns into modulation distortion, albeit, not usually at a rate that is audible in amplifiers, and there are a lot of people that (fallaciously) believe that second harmonic distortion somehow "sounds better" (it doesn't, however). 

 

     b) the lower second harmonic distortion shows up at "cinema" levels.  This level of SPL operation is way, way higher than home hi-fi, even when someone is creating 105-110 dB in a small listening room.  In a commercial cinema-sized room, the SPL could easily be 20 dB lower at the bulk of the audience listening positions, and the listeners are hearing that distortion virtually all of the time.  That's a problem for commercial cinema, not as much for home hi-fi, IMHO.

 

    c) I couldn't see at what frequency this breakpoint in 2nd harmonic levels begins, and at what SPL.  If this is a function of low frequencies below, say, 700-1000 Hz, perhaps this is something that would influence you to raise the lowest frequency cutoff a bit in order to let the lower-frequency driver (a woofer) take that load.  Usually, I see the harmonic distortion levels for woofers in that range are not very bad.  This is something to think about.

 

2) He never talks about the effects of FM distortion build-up of a single diaphragm driver over a dual-diaphragm driver.  This is the dominant form of distortion that is recorded at higher frequencies (via Klippel data), and is the most objectionable form of distortion for the human ear to digest (without compliant).  In other words, the "dear doctor" just ignored the most dominant source of bottom line: how it sounds, especially for home hi-fi.

 

3) He never talked about the shape of the mating phase plug surface and whether or not the phase fronts of the combined wave fronts at the throat are severely distorted at the highest frequencies (10-20 kHz).  You won't see that using harmonic distortion plots, but you'll hear it in terms of phase distortions as you go up in frequency.  If you can do your FIR filtering to flatten the overall phase response, you can compensate somewhat, but the question remains on how much the human hearing system can detect those phase disturbances, and whether or not higher order wavefront disturbances are propagating down the throat and into the horn, and whether they exist the horn mouth.  Just like diffraction slots, these disturbances cause the generation of higher order horn modes.

_____________________________________________________

 

I guess the bottom line is like "the proof in the pudding"--the tasting (listening to them).  I hope that some form of level A-B blind tests can be rigged up and staffed.  Most important is using essentially flat on-axis SPL response (taken at 1m in room) all the way up to 20 kHz, and controlling for early reflections in the listening room (of which I see few listening rooms being treated for early reflections).

 

I'll stop there for the moment. 

 

Chris

[ClaudeJ1]

35 minutes ago, Chris A said:

2) He never talks about the effects of FM distortion build-up of a single diaphragm driver over a dual-diaphragm driver.  This is the dominant form of distortion that is recorded at higher frequencies (via Klippel data), and is the most objectionable form of distortion for the human ear to digest (without compliant).  In other words, the "dear doctor" just ignored the most dominant source of bottom line: how it sounds, especially for home hi-fi.

I'm pretty sure PWK once wrote about his analysis of Co-Axial drivers using, what I assume, is the Altec-604 or one of the derivatives. He used a spectrum analyzer to measure Intermodulation Distortion, technical papers from which he won an AES Silver Medal of Honor back in 1978. As I recall, in the specific case of testing a 2-way co-axial woofer-tweeter combination, concluded that it had greater IM distortion than separate drivers, like in all Klipsch speakers. I presume this was a result of the treble being acoustically modulated by the woofer as it exited the tweeter horn as opposed to mechanical modulation when coming from a single diaphragm.

 

In my teens, I heard a stereo pair of Altec 604's in Bass Reflex cabinets with McIntosh Tube Amplification and they sounded quite good. So was PWK right or not?

 

Also, keep in mind that 99.9% of Celestion customers will be using them for Theater or PA applliations and NOT the home market, which is reserved for DIY horn freaks like us. This represents a very small group of people that are probably less than a few dozen across the USA and/or Europe, aye?

 

So your point about home use would not impress their sales/marketing departments whose efforts are dollar based Return on Investment.

 

 

 

[ClaudeJ1]

39 minutes ago, Chris A said:

guess the bottom line is like "the proof in the pudding"--the tasting (listening to them).  I hope that some form of level A-B blind tests can be rigged up and staffed.  Most important is using essentially flat on-axis SPL response (taken at 1m in room) all the way up to 20 kHz, and controlling for early reflections in the listening room (of which I see few listening rooms being treated for early reflections).

 

I'll stop there for the moment. 

I'm pretty sure there's a Klipshead that bought the first pair to hit the USA from Parts Express last year and put them on his K-402s with a positive feedback post on this forum.

[Chris A]

Compared to the tools we have today, PWK labored with stone axes and flints to make fires by comparison.  He did almost all of his JAES work before 1970.  In 1978, I was a newly graduating ME with specialty in finite element analysis.  I know what was generally available then.  I don't believe that anyone other than the huge corporations and federal government could afford them.  Even in the geophysical exploration field--which was floating on money--the resources required for these tools was extremely limited at that time--big computers costing many hundreds of (1978) dollars per CP hour. 

 

I think PWK did a really great job, but some things had to wait until both the analytical tools (simulations, etc.) and really good acoustic and accelerometer measurements became much more ubiquitous and affordable.

 

As far as my views today are concerned, if you've got the measurement gear and the analytics to model what's happening (guided by the tests), then that should help you form opinions.  If you're far away from the tools available today, you tend to form opinions that may not be "separable" from other issues. 

 

30 minutes ago, ClaudeJ1 said:

As I recall, in the specific case of testing a 2-way co-axial woofer-tweeter combination, concluded that it had greater IM distortion than separate drivers. I presume this was a result of the treble being modulated by the woofer as it exited the tweeter horn.

I think PWK was referring to a device whereby the motion of the woofer (~5x greater than a horn-loaded woofer of the same size for the same SPL) would tend to create AM and FM distortion issues that would be large by today's standards, some of which were separable into separate distortion sources in his day, some of which weren't.  He apparently wasn't able to separate AM and FM distortion sources, and also the effects of other driver nonlinearities.  If he had access to a Klippel near-field system (NFS), I think he would have advanced the state of the art even farther. 

 

In any case, do you think two horns and two drivers are better than one?  Even PWK could quickly answer that question.  He apparently always wanted to migrate back to two-way Klipschorns, but the technology of the day wouldn't let him do it--until right at the end of his life when the effects of age were beginning to take their toll on energy levels and cognition.

 

PWK was a giant in the audio reproduction field.  We're all here because of him.  But the measurement and analysis capabilities keep getting better--and so do the technologies available in terms of drivers, horns, and crossovers/correction networks.  I don't believe it's fair to ask the type of question implied.  Look at the measurements today, and then listen for the differences, and look at where those issues came from.  Then you can make sweeping generalizations--if you still feel up to making them.

 

Chris

[ClaudeJ1]

3 minutes ago, Chris A said:

I don't believe it's fair to ask the type of question implied.  Look at the measurements today, and then listen for the differences, and look at where those issues came from.  Then you can make sweeping generalizations--if you still feel up to making them.

I don't know about "sweeping generalizations" in talking about audio, since so many people make them. I was just trying to communicate by using a form of CONTRAST, which is how we all communicate. Contrasting between 1978 and 2018 (which is about the time Celestion got on board with this new product development), is not a bad thing to do.

 

If a college kid send a letter home to ask for more money from his parents and he writes it on a white piece of paper with white ink, he won't get the check. Lack of contrast.

 

I agree with you that the better modern tools for measurement, math, and simulations allow for better/further developments. After all, isn't that what progress is all about?

 

In either case, using a a TAD driver, a BMS, or an Axi2050 on a K-402 still represents the pinnacle of performance above 300 Hz, where room modes do NOT dominate.

 

It's not an argument but a comparison, which still a generic form of contrast. Your reply reminds me of 3 rich car guys arguing over Maserati, Lamborghini, and Ferrari being the "best car," which to a driver simply means they can ALL do over 200 MPH, while leaving the technical discussions to the guys that have to do their tune ups or engineer them.

 

[Chris A]

Pardon me.  The form of your question was questionable to me in the way it was posed.  I don't trust those type of questions.  They don't advance the state of the art, I find, but they do inflame emotion--needlessly.  And they only look backward, not forward.

 

Please--let's not do this. I'm asking--openly.

 

Chris

[ClaudeJ1]

10 minutes ago, Chris A said:

Pardon me.  The form of your question was questionable to me in the way it was posed.  I don't trust those type of questions.  They don't advance the state of the art, I find, but they do inflame emotion--needlessly.  And they only look backward, not forward.

 

Please--let's not do this. I'm asking--openly.

 

Chris

How else was I supposed to learn your thoughts on the matter without asking a question? Since principles of audio have been laid down over almost 100 years, acknowledging the evolution by comparing the historical moments is a perfectly valid way to feed further discussion. 

 

Aside from that, while I have yet to DIY my own, I have actually OWNED 23 and still own 7 factory built MEH horns, which represent the best of all worlds IMHO.

 

So my commitment goes way beyond simple discussions here and has involved more money than time.

 

So I don't understand why you feel you have to react so defensively and view my questions so negatively when they should only serve to continue the learning process for all who spend time here and read your posts.

 

In case something else is not obvious to you, I thinks you are one of the sharpest people here, and I have worked with the top brains in Silicon Valley for the last 40 years, so I speak with other disciplines besides audio, but it's a big world out there, let's not forget.

[ClaudeJ1]

55 minutes ago, Chris A said:

In any case, do you think two horns and two drivers are better than one? 

I did and do. Based on the video I just watched and having had a curiosity about this new single driver solution by Celestion, I think it deserves a listen and could be a game changer.

 

Turns out that Ricci, of Data-Bass fame and excellent DIY subwoofer designer, got the first pair I know of about a year ago to replace the K1133's he sold to me. He had to find a dealer in England to ship them here because they were not yet available in the USA.

 

After 3 months he still had not included them in his system and was still using "elephant cheek" JBL horns for a top end. I will further enquire about his listening impressions.

 

I agree with you that while theory and measurements are fine, the proof in the pudding is how do they sound. If they were not at least close enough to the BMS offering, then why would "Roy Delgado select them for the upcoming flagship Jubilee II that will be released this summer? Price and management influence are always a factor in engineering, but we all know that much.

[billybob]

Seems that what was then and now shows the expertize and fortitude to get the Klipsch speaker before release as best perfection in sound as was practical for the public at large.

 

 Think that shows, as emulating the Klipsch success at the time, and now.

Engineers it seems regardless of tools then and now still on the hunt and their curiosity and energy still abound to now and the future.

Cool cause, some like both your findings, and the sharing of same.

 

[OO1]

5 hours ago, Chris A said:

A picture of the diaphragm of an Axi2050.  It doesn't look inexpensive to me.  The voice coil is perpendicular to the diaphragm at mid-point radially.  Technically, it is a ring radiator diaphragm, just like the two BMS4592ND diaphragms, except this is titanium,

 

 The Titanium  composition is one key element that will make the Axi2050 perform better under high stress versus the BMS 4592  , and as Durability increases 2x vs Polyester , so will there be less failures overall - 

[Chief bonehead]

Probably need to move to separate thread but I will say a couple of things. In later years, which is when I came along not really nothing anything about what Mr K said or wrote, Mr K did consider that even order harmonic distortion were the ones that were the most annoying. These tended to be related to mostly physical irregularities while odd order were the ones that most people considered “pleasing”. Mark Dodd is quite a good resource and I always look out for his contributions to the acoustic art. 

[Chris A]

18 minutes ago, RandyH000 said:

The Titanium  composition is one key element that will make the Axi2050 perform better under high stress versus the BMS 4592  , and as Durability increases 2x vs Polyester , so will there be less failures overall - 

If you watch the video, the apparent goal of the AXi2050 diaphragm design was to isolate the "mechanical modes" (flexible motion of the diaphragm) from the acoustic modes (what happens in the air cavities inside the chambers, phase plug channels, and throat).  The specific stiffness of the material is critical to that approach--especially in moving a large diaphragm design like that one they settled on.  Using titanium was one way to get there.  Using beryllium probably would have made it a bit easier (it has about 7x the stiffness/weight ratio of titanium--by various combined measures--pushing the issues up at least one octave above 2" titanium diaphragm issues).  Titanium is, however, the economic alternative.  I understand that, and applaud them for achieving that performance.  The price is a little high right now, relative to the BMS driver--for the same performance. (Engineers usually are concerned about performance and price at the same time.) 

 

Polyester diaphragms are using a completely different approach--one that isn't nearly as dependent on decoupling resonant modes as the Axi2050 design.  Will they "wear out"?  I don't know--I haven't worried about that.  When my BMS 4592ND diaphragms wear out in home hi-fi duty, I'll let you know. 

 

Don't hold your breath, though.  I believe that will take a very long time.

 

Chris

[Chief bonehead]

30 minutes ago, ClaudeJ1 said:

"Roy Delgado select them for the upcoming flagship Jubilee II that will be released this summer? Price and management influence are always a factor in engineering, but we all know that much.

I did have quite the task of comparing most of the state of the art compression drivers. As the gentleman in the video eluded to, you have to know what you want. I didn’t choose the axi2050 blindly. All the measurements that I did led me to one thing......it measures very well. Now how does it sound......