A new Center Channel Design study...

[JonMarsh]

All that effort setting up this shop last year and this have really paid off...

[theSven]

Exciting times! Looking forward to see the ribs get CNC'ed and start seeing this visually come to reality.

[Evil Twin]

The proof of concept LF crossover construction has been delayed compared to the planned schedule, but is now underway.

In light of data received regarding the configuration of the initial target installation, one possibly key additional feature has been added- an adjustable LF contour to manage the interface to the system subwoofer by adjusting the corner response behavior, taking advantage of adjustments possible with modification of the impedance compensation LCR network.

It has come to light that the Customer #1 has a certain degree of attachment to elevated low frequency levels- perhaps this is conducive to his feeling the "Force" in his cinematic playback experience....

Regardless, this will be accommodated... we expect the final results to be "most impressive", but have indicated that further adjustments to the system configuration will be required to reach an optimum level of performance.


The POC junk bin parts crossover layout:

[JonMarsh]

This isn't my topic thread, but the Chancellor and Steve have brought up some thoughts and comments, and in particular believe is is both necessary and desirable to "rename" this project.

This is principally desired to bring the nomenclature in better alignment with the intent and nature of the rather radical project concept that has been developed (especially considering the technology goals and details and capabilities compared with more conventional approaches based on more modest home theater targets.

So, from this point onwards, it will be known by a simple acronym structure:


ET CCTT


Anyone having trouble figuring that out, just let me know...

I'm hearing the last POC crossover board has been assembled, and there should be test data fairly soon.

[Evil Twin]

As more information has been collected about the specific installation requirements for the Chancellor, the key client for this project, it is clear that some physical modifications will be necessary, in order to fit within the available space for this speaker system, and that more acoustical issues will need to be addressed.

A new HF candidate and crossover configuration has been selected, and some new in situ cabinet tests completed, which are leading to other modifications in the POC cabinet and test configuration.

As the moderator Jonmarsh seems overly fond of saying, "Slow work takes time".

What is left unsaid with that simple statement, is that "assiduous effort guided by the Force" defines the way forward...

[theSven]

As more information has been collected about the specific installation requirements for the Chancellor, the key client for this project, it is clear that some physical modifications will be necessary, in order to fit within the available space for this speaker system, and that more acoustical issues will need to be addressed.

A new HF candidate and crossover configuration has been selected, and some new in situ cabinet tests completed, which are leading to other modifications in the POC cabinet and test configuration.

As the moderator Jonmarsh seems overly fond of saying, "Slow work takes time".

What is left unsaid with that simple statement, is that "assiduous effort guided by the Force" defines the way forward...

Exciting to see progress for this design. Slow work is good work, like measure 2x and cut 1x type thing...

[Evil Twin]

Measuring 3 or 4 times, and buying a new ruler, is necessary at times...

[Evil Twin]

Exciting to see progress for this design. Slow work is good work, like measure 2x and cut 1x type thing...

This is a most satisfactory viewpoint... work continues today...

[theSven]

Searching the thread I realized no one has asked nor has the Empire stated the power requirements for this design. Would a 100w channel work or is more power needed? For example I have an Emotiva XPA 5, would that suffice to power this CC?

[Evil Twin]

The Emotive XPA-5 Gen 3 is rated at 250W/channel, THD <0.1%, all channels driven. This should be quite satisfactory. Note, this amplifier should be on a separate breaker feed from the rest of the system, based on the power consumption (up to 2kW) and the use of a switch mode power supply.

[theSven]

The Emotiva XPA 5 and CA Edge W are both on a separate 20 amp breaker only for this. I have another 20 amp breaker for the other gear.

[Evil Twin]

The expected system sensitivity, set by the combination of the PTT8.0X04 wired in series and the baffle step response of the cabinet is ~85 dB.

[augerpro]

Is the Faital Pro mid still on the menu for the cheaper version?

[Evil Twin]

Is the Faital Pro mid still on the menu for the cheaper version?

Due to updated information regarding space restrictions in the target environment for the Chancellor, and measurements made in the development lab under comparable acoustic conditions, as well as some basic listening evaluations, one might justifiably state that this design has been undergoing some development updates... the less diplomatic might even describe it as "design churn".

I will lay out a few highlights of this process in a following post, but at this point there are two possible configurations in consideration, a vertically oriented compact MF + HF driver set, and a coincident Satori MF+HF solution under evaluation.

The "distinct" driver MF+HF pair is "borrowed" from the Kii nostalgia project, potentially using the Peerless FSL-0512R01-08, and the Seas H1499 DXT. These were selected based on driver measurements and value proposition. Due to component size, they should minimize the vertical height and with optimized crossover achieve a reasonable height vertical lobe in the transition region from MF to HF. A Proof Of Concept crossover design has been developed based on driver measurements, and it appears that the fundamental acoustic targets are achievable, though the actual vertical lobe in the MF to HF crossover point is yet to be determined, being a Duelund design, it shows considerable promise...

This latter factor is an issue with many center channel products, and in fact is responsible for the complicated mechanical configuration and adjustability of the Wilson Audio Mezzo, which must be carefully tuned using a complex table taking into account the stand type, the distance to listener, and the listening of the listeners. One could be led to conclude they were aiming a low dispersion military laser, and not an HT loudspeaker...

A high performance coaxial design may cut the Gordian knot, by melding the midrange and high frequency wavefronts on a common axis and eliminate the narrowed vertical dispersion pattern which occurs with vertically oriented drivers. This is particularly an issue with the ever popular LR4 crossover network, which as Linkwitz himself noted, has a significant power response dip in the crossover region due to the narrow pattern on the axis between the drivers, and recommended considering a lift in the crossover region on axis to help fill this in.

While there are numerous coaxial drivers on the market place, most appear to have been the product of an aggressive marketing department with little active participation by the engineering team, OR, have rather astronomical price tags.

We have obtained samples of the SB Acoustics Satori MT19CP, a 7-1/2" diameter driver, and have reason to hope that a more useful value proposition may be achievable, via a midrange price and well above average performance.









Whether the promise of this data sheet will translate into a useful system configuration remains to be seen... I expect more than mere "satisfactory" performance.

A test cabinet is nearly finished, and there should be some hard data fairly soon...


It is expected that a larger variant of an early design proposal from Steve Manning will be pursued, taking into account developments such as overall height targets for the Chancellor's system and the driver array -



Another point which has emerged is the desirability/necessity of some tilt of the baffle on the vertical plane - in the range of 6-8 degrees.

Other details must wait pending further measurements.

[Evil Twin]

The choice of the Peerless midrange is based both on minimizing driver outer diameter and measured performance with samples on hand. The Faital driver would also be usable in this design, but may require some adjustment of vertical baffle size and crossover- however, it's flatness and extension should pose little in the way of fundamental design change.

[JonMarsh]

The task at hand is to modify a Denovo subwoofer flatpack cabinet for testing the Satori M19CP-8 coaxial driver. The one selected matches the 14" baffle width being considered for the next test build, and a new front panel was cut and epoxied to the cabinet.

First, a basic baffle edge round over was completed.




Then the driver mounting flange rebate was routed... (work being done by selected minion)





After the through hole is cut, it is most advisable to check the dimensioning, but even more advisable to check two or three times before engaging the tools...




Then, with the same router setup as the through hole cut out, a plunge hole was cut to use a follower bit for cutting out the center plug.




Then, the inner plug is cut out...





A hole diameter check followed by a test fit:





The result appears to be satisfactory...

Hopefully no unexpected issues will be discovered during testing...

[Evil Twin]

Unfortunately for the efforts of Jonmarsh, the measured results were not completely "satisfactory".




A detailed performance comparison was provided to Steve Manning, and in the end, while a more commercially oriented endeavor might have proceeded using the Satori Coax, the consistency of behavior and overall results for the midrange and the tweeter unit did not compare well with the recent value oriented project I have undertaken, internally referred to as the "Kii Kolumn".

Next to consider will be a configuration more like that of the Rockport Taurus, though with higher performance woofers and a more "reasonable" value proposition.



In consideration of that, another detailed driver + cabinet evaluation is under way drawing upon the configuration concept of the upper range of the Modula Xtreme Kii inspired.




This will serve two purposes, one for additional reasonably representative data of the midrange and tweeter for the CC development, the other mimicking the top part of the "Kii Kolumn" cabinet for the purpose of crossover development and testing.

Fabrication should begin within the hour... excuses will not be accepted!

[technodanvan]

Higher performance than what appears to be Satori TX woofers? Color me intrigued.

Edit: I'd like to see the results of the Satori coax sometime, if you don't mind.

[JonMarsh]

Higher performance than what appears to be Satori TX woofers? Color me intrigued.

Edit: I'd like to see the results of the Satori coax sometime, if you don't mind.

That can be arranged- I'll forward the email update I sent to Steve. Certainly my reasoning and conclusions are open to examination, but Steve seemed to be in agreement.

[JonMarsh]

Using LF measurements for the PTT8.0X04 woofers in the previous cabinet located at a height and distance from wall and floor representative of the specific application requirements as a starting point, a new Duelund based POC crossover has been developed with the Peerless midrange and DXT cabinet data measured in the new MT cabinet. How much baffle step will change the behavior for woofers and midrange remains to be seen, but will be evaluated soon, as dual flanking test cabinets for PTT8.0X04 are under construction.

At my insistence the LF crossover approach now incorporates the capacitive coupled LF topology used in the Modula Xtreme and now in what might be called the Kii Xtreme. This addresses inevitable interactions between the crossover filter and the impedance characteristics with a higher Fb, as well as acoustic issues due to the cabinet location. To quote the project leader designated by the Chancellor,

What is left unsaid with that simple statement, is that "assiduous effort guided by the Force" defines the way forward...

This is the modeled results, with an Fb using the older cabinet design of 50Hz, and well within a +/- 2dB window.





The impedance characteristic is fairly amplifier friendly, never dropping below 4 ohms, and being quite smooth in the critical midband, due to the LCR shunt zobel at the input.





By way of comparison, considering the normal cost pressures to reduce the component count, if the Xtreme network and the tweeter Fb and HF impedance networks are removed, and everything else remains the same, this is the expected result:





Flanking woofer test cabinets are in construction, and when completed new driver measurements will be done "in situ" for the application.

[Evil Twin]

And so, another driver investigation begins.... an overlooked coaxial which seems to have some merit based on data sheet information, including tantalizing design points, such as an underhung Neo motor for the mid woofer section, what appears to be a thoughtfully executed mid woofer surround, and an inverted dome tweeter with a better defined critical launch "environment" than is typical for coax drivers I have encountered...



And, perhaps most intriguingly, for some, a cone material derived of bamboo fibers...

• The underhung woofer design strongly suggests a very linear BL product, on a par with the efforts and results from PuriFi, though employing more of a "brute force" approach via the underhung design, which is a long gap, short coil- quite like typical dome tweeters.
• Third party measurements confirm the linearity of the TangBand motor...

Of course, the overall quality of a coaxial driver is dependent on the behavior of many component parts and technologies, but the claimed performance in the data sheet indicated that the purchase of a driver for "hand on" investigation would be justified...






It would appear that the use of a shallow cone structure and non-reflective approach to the tweeter launch may result in smoother response than has been observed in other drivers tested and found wanting, especially directly on axis. One need only examine the specification graph for the "updated" Seas C18EN002A to have an appreciation of the challenges involved...




Some preliminary impedance measurements have been conducted, to confirm some basics- due to the erratic phase plotting of the DaytonAudio DATS V3, the much slower but more precise Smith & Larsen Woofer Tester was employed.






A basic test enclosure has been prepared, and if matters proceed as expected, data should be available soon.

It is strongly desired that these results should be at a level sufficient to meet the Chancellor's expectations... this would be a preferred configuration for center channel, considering vertical polar response.

[Evil Twin]

Due to the complexity of the frame design of the W6-2313, an optimized baffle mounting for low diffraction test may take some effort.

In consideration of that, it was thought expedient to engage in some basic tests, and use preliminary data to determine if further effort was justified...

The frame design of the W6-2313 is unusual, likely driven by the mechanical configuration used in the design and an apparent need to assemble some items in a clamped configuration.

This can be inferred from a side photograph....





Unusually thick field plates sandwich the compact Neodymium magnets which provide the woofer motive force.

But the mechanical drawing makes the three piece construction more apparent.




As can be seen clearly in the mechanical drawing, the mounting flange is actually a two piece section of nearly 9mm depth.

Additionally complicating matters is the relatively narrow rear window in the upper portion- clearly the mounting technique may have an impact on the rear wave.

And for dimensioning purposes, the W6-2313 should probably have been named the W7-2313, given the true outer diameter of the mounting flange.


Regardless, the driver was mounted in a basic enclosure built for testing 7" drivers, but one not equipped with a flange rebate, so it is expected that diffraction effects may contaminate the upper range response.


As noted in the previous Impedance testing, the woofer impedance behavior is quite clean from any obvious resonant artifacts due to motor, cone, or surround.






The woofer response at various polar angles from 0 to 40 degrees tracks quite closely, and not surprisingly, the 20 degree response is nearly identical to the RMS average of the combined responses. Shown below is the quasi near field response with the HD2 and HD3 measurements.





The close tracking of HD2 and HD3 levels may be correlated with the underhung design and quite uniform BL force curve. Perhaps more notable is that even with the rising impedance curve with increasing frequency, the distortion does not exhibit the degree of rise with frequency common to drivers with that impedance characteristic. Targeting a 2500 Hz crossover would seem quite reasonable.


The tweeter curves were expected to be influenced by the step diffraction of the mounting flange, and may also have an issue which is diagnosed with the small blip in the impedance curve at about 4kHz.

The 20 degree polar curve is very close to the overall measured power response by taking an RMS Average of 0, 10, 20, 30, and 40 degrees.




HD3 is quite low, especially from 2kHz and up, also justifying a 2500Hz crossover. The inverted dome aluminum/magnesium tweeter seems exceptionally well behaved on an amplitude basis, though it may be interesting to see if a notch filter at about 35kHz might further reduce the very top end HD3 as well as HD2 in the range above 12 kHz.

Based on the preliminary measurements, it seems worthwhile to work out an optimized baffle mount and fabricate a dedicated enclosure for further tests. It can be said that the worst of the measured response "issues" are just dips, which are generally far less audible than peaks. The behavior of the two sections points to the possibility of relatively low component count filter networks.

[tktran]

It's a shame those coaxial drivers weren’t optimised for directivity control. They could be better, if they paid closer attention to geometry for directivity matching between the HF and the LF/MF elements.

Here’s one done earlier- by our own 5th element


LF element:
Normalised polar chart:




HF element normalised polar chart:




Note that as a coaxial driver the vertical directivity = horizontal directivity.

So if you crossover between the MF and HF around 2.4-2.7KHz and you get a beamwidth of 80-90 degrees for all of the treble...

Reference:
SB17Coax (5een.co.uk)

[Evil Twin]

With a step tiered seating arrangement, greater vertical beam width is actually a requirement.

With this particular driver, the expectation and desire is primarily having coincident sources for midrange to HF, and the consistent behavior through the crossover region this may afford, if properly implemented.

There are some other "interesting" considerations for this driver which extends from the three way configuration to a two way surround speaker, made interesting because of the intrinsic baffle step compensation built in, similar to the Accuton Cell aluminum drivers, and the relatively exceptional tweeter behavior compared with both the majority of standalone tweeters and coaxial ones tested to date.

[Evil Twin]

It is also worth noting that assumptions and correlated measurements of HF beam width will vary depending on the specific cone geometry- while 90 degrees included angle is common, it is not mandated OR universal... one need only compare the M19CP-8 and the W6-2313 to see large differences in design and execution.

[tktran]

To be clear, I'm talking about directivity matching when transitioning from a LF to a HF, in 3 dimensions, meaning a smooth frequency response in the reference/listening axis, AND off axis
ie. in other directions.

How far or wide the design should go, or needs to go, I agree it largely depends on the use-case, room (or lack of) and program/listening material. In a home cinema vs commercial cinema versus outdoor festival the ideals may be different.

The jury's is out to what it is preferred in all situations, if such a thing exists.

I didn't mean to imply that 80 or 90 degrees as some kind of ideal number. I meant that the directivity is controlled and has a smooth transition.

It would be nice to see off axis measurements, and normalised polar charts for drivers like the M19CP-8, or the W6-2313, to see what happens between 2KHz and 16KHz (3 treble octaves) as the HF takes over from the MF. Do you have any JonMarsh Evil Twin you can share?

[Evil Twin]

New test cabinets are in construction, for the W6-2313, and additional HF polar plots will be generated. This has already been done in preliminary evaluation, but with enclosures lacking proper diffraction control.

Issues with the M19CP-8 from initial measurements in a diffraction controlled enclosure were not satisfactory, especially in the upper octave of the tweeter, and also with regard to breakup behavior of the woofer, in consideration of the limited recommended operating range for the tweeter.

[JonMarsh]

Just the Head Minion reporting in... slow work takes time, and we don't get Thanksgiving Eve off, but the "official test cabinet" for the W6-2313 is getting pretty close...


Front baffle assembly was finished yesterday and glued on the main cabinet... today was routing the driver rebate and through hole and checking the details...


I like the dust control capability of my Porter Cable...





The inside of the pass through hole is cleaned up with my relatively ancient DeWalt, using my favorite follower bit...








At this point, we're ready for a driver check without the gasket... mind you, this has been measure three times, route once...







The W6-2313 does have an unusually thick front flange....






With the gasket, this should be sitting fairly flush...

[JonMarsh]

There is a fair amount of work in progress, and materials in the pipeline, but I think it's also worthwhile to summarize some of the challenges that arise form the basic positioning of a center channel speaker like this, and discuss in more detail the current proposed solutions.

The basic acoustic challenges come largely in part to boundary reflections we often aren't dealing with for conventional speakers. Yet, depending on the placement of a "conventional" speaker, these effects will occur regardless. However, for the CC, and this specific design serving the needs of the Chancellor, the placement is a major design consideration, and must be addressed upfront. These constraints include the distance form the rear wall and the specific necessary height limitations above the floor.

In the recording studio, when situations like this (frequently) arise, a fundamental part of the solution is usually strategically placed sound absorbers, though some may not pursue that and attempt to get by with just EQ at the listening position- but this can have a negative impact on the overall reverberant room power curve.

So, considering that, and the evolution of different form factors, one more like the Wilson Audio Mezzo (achieved by using an additional cabinet to floor baffle), and one more like the Rockport Taurus (achieved by NOT using a cabinet to floor baffle, basic evaluations were conducted for the 2x PuriFi PTT8.0 woofers and original PuriFi midrange driver.











The configuration most resembling the Rockport Taurus cabinet, but with a floor acoustic pad in front of the speaker, produces the best result, one most like an elevated half space measurement, apart from the LF bump which is also likely due to the combination of boundary loading from the floor and the rear wall.



Re the Midrange result:














An inescapable conclusion (to me) is that the use of a boundary absorption pad on the floor in front of the CC is essential to securing a usable response profile.

Two ATS acoustics panels, 2' x 4' X 4" thick have been ordered, one for the floor, and one to possibly experiment with on the wall behind the CC.

[JonMarsh]

Going from this best case response plot for the combined woofer output, the next task was improving the LF crossover functionality and reducing component count as possible. This benefited from borrowing the Modula Xtreme concept for LF tuning, as was done for the Kii inspired design. (Kii Ardent? Kii Extreme? who knows, maybe take a vote!)

😂

A reminder- the raw response of the drivers with a home brew floor pad in place, and the target response goal for the POC Duelund crossover:




With a basic 3rd order crossover for the target response:





The POC crossover with the Xtreme series cap:





And the modeled response:




This meets the target LF response goal of -3 dB @ 50Hz, and has reasonable compliance to the target curve, and good summing in the complete crossover- it's a starting point for physical construction for the crossovers, and testing and refinement. That construction is underway...

[technodanvan]

An inescapable conclusion (to me) is that the use of a boundary absorption pad on the floor in front of the CC is essential to securing a usable response profile.

Would carpet with a modest pad underlying it be an acceptable solution, considering it may already be in place? It wouldn't exactly be 2" thick though...

[theSven]

An inescapable conclusion (to me) is that the use of a boundary absorption pad on the floor in front of the CC is essential to securing a usable response profile.

Two ATS acoustics panels, 2' x 4' X 4" thick have been ordered, one for the floor, and one to possibly experiment with on the wall behind the CC.

Now I'm wondering if I need to engineer a new solution to hang the screen on the wall with an offset or whatever the word I'm looking for it to fit acoustic panels behind the screen 🤔

I need to order some of these anyhow for the side walls in my room.

[JonMarsh]

I would strongly recommend panels on the side walls relative to the Ardent D's; it's hard to say without testing whether it would be better to use the old mirror trick for locating them from the primary seating position, or simply locate them relatively close to the speakers, with the leading edge close to the front of the speaker. This would provide the most reduction of early boundary cancellations.

Were it my setup, I'd try testing with one speaker initially, and first try one panel mounted that way, and then another panel next to it further towards the listeners. Panels are 24" wide, 48" tall.

[JonMarsh]

Would carpet with a modest pad underlying it be an acceptable solution, considering it may already be in place? It wouldn't exactly be 2" thick though...

No, already had that in the setup as tested "without" any boundary absorption, as that is the default floor in the room. The "home made" pad was 4" thick.

The acoustics of CC's mounted close to the floor like this is somewhat new territory for me, I've been learning the "issues" as I go along...

[technodanvan]

Just playing Devil's advocate, but maybe a cheap wedge-like sound absorber from Amazon would be a more practical solution, as there may already be something similar in the room? The ATS isn't exactly attractive.

[JonMarsh]

To each his own, of course- I find the white linen covering on the ATS fairly pleasant, and I have similar ATS panels in storage, but they're reserved for the stereo setup on the walls. I have experience with the acoustic performance of the ATS panels. And they are far more acceptable to my wife esthetically! I ordered the beveled edge version, such as I have already.

Of course, that's almost a moot point, as it all has to go into a dedicated AV room!

[theSven]

I would strongly recommend panels on the side walls relative to the Ardent D's; it's hard to say without testing whether it would be better to use the old mirror trick for locating them from the primary seating position, or simply locate them relatively close to the speakers, with the leading edge close to the front of the speaker. This would provide the most reduction of early boundary cancellations.

Were it my setup, I'd try testing with one speaker initially, and first try one panel mounted that way, and then another panel next to it further towards the listeners. Panels are 24" wide, 48" tall.

I did the mirror trick and had someone sit in all the listening chairs and put tape on the wall for those locations. I essentially need 48" width at the front and then another panel further back for where I had planned to put the side channels for the 7.1.

The ATS are very reasonably priced option. I would go with black for the fabric color to match the walls.

[Steve Manning]

GIK Acoustics is another option. Also, if one wants to save a few bucks both places sells supplies to make your own.

[JonMarsh]

I did the mirror trick and had someone sit in all the listening chairs and put tape on the wall for those locations. I essentially need 48" width at the front and then another panel further back for where I had planned to put the side channels for the 7.1.

The ATS are very reasonably priced option. I would go with black for the fabric color to match the walls.

I think you've got it sussed!

Not surprised at the 48" width at the front, which was why I made the suggestion to consider...

[theSven]

I'll have to budget for Q1 to hopefully get some of the ATS for the room.

[JonMarsh]

Yeah, I know that feeling... and no Christmas bonuses anymore, either!

Slow work takes time- crossover division- all the on hand parts are mounted for the new POC test boards, a shipment with the remaining parts is expected on Tuesday.

LF, MF, HF.

Purely as a historical side note of sorts, I used my "vintage" PUR hot glue system on this. I say "vintage" because it's my original Steinel glue gun from 2014, and "vintage" Steinel/Titebond MP75 multi-purpose glue cartridges, cleaned out Amazon's stock. Which wasn't large.

Steve tried the newer PUR stuff, from different manufacturers, and let's just say I could hear his cursing in Maryland all the way over here in Idaho...

[JonMarsh]

I doubt that I have to tell anybody about how difficult it is this time of year to find time for things like, um, building crossovers? especially given how slow work takes time, right?

Well, I started the "Must do" holiday and home tasks at about 6AM today, and finally around 2PM had a little time for finishing the W6-2313 T POC crossover eval board...

Some build notes...

• Reminder, this board includes a general LCR impedance leveling network connected directly to the input- this could have been on a separate board or the midrange board.
• Components were mounted initially using my vintage PUR hot glue system with vintage MP75 cartridges. That worked fine, EXCEPT for the 20W Dayton resistors that have the sort of funny matte green finish. Might as well have been mineral oil, they just popped right off. Smooth and shiny parts worked just fine, though, even the Mills resistors.
• However, my other old standby, Ace Hardware high temp hot glue sticks, worked fine for attached the resistors to the original glue mound- nice and secure. Go figure.
• Soldering was done with my first serious attempt to go no-lead, with the Cardas Tri-eutectic solder. It requires a higher melting temp point, no issue for my 400W temp controlled Weller iron. Standard irons might be a problem. Note, besides setting a higher temp, take more care to heat all the items being soldered (like turret pin posts, component leads, wiring) adequately and uniformly. Mission accomplished, by magnified inspection, without any issues.

[Evil Twin]

The Chancellor contacted me today for an update on the status of this project, for which I have delegated significant portions of the work, but none of the responsibility. It is appropriate to provide a more complete overview of progress to date, with the caveat that there are many design assumptions remaining to be verified, though everything here is based on driver measurements in situ.


the working schematic plan- always subject to update, but used for the current POC crossover module builds:






Note the LCR Fb shunt zobel network, and the Xtreme configuration LF alignment.



This is for a very specific setup configuration required by the Chancellor, but which would be typical for front projection usage or any screen setup with the bottom of the screen at approximately 30" above the floor, with acoustic pad on the floor and behind the speaker.

An update on the MF crossover POC board layout:



A modular cabinet test setup should be completed before the start of the New Year...

[Evil Twin]

In the interest of accelerating the development of new test information, after a review of the past cabinet, the decision was made to create a test configuration using that cabinet with the W6-2313 coax in the place of the previous midrange driver candidates, the 6RS180 and PTT6.5M0F. While the midrange driver opening is not ideal for the new target driver, having too shallow a rebate depth, and too large, it was believed that useful information could be obtained as to a preferred configuration and optimizing the vertical listening window to the in room requirements. Furthermore, this would eliminate nearly a weeks cabinet work out of the critical path of development.

That belief proved to be realized in practice.

A detailed update has been provided to the Chancellor and to Mr. Manning- a brief overlook will be given here.


First, an assembled system was tested for the asymmetry of response above at different angles relative to the offset midrange. The expectation was that the side with the offset of the coax driver would have the most linear response offset from the cabinet centerline, and this proved to be correct.





Measuring just the coax mid and tweeter, and taking the RMS average of +20 and -10 degrees, the critical angles for the Chancellor's room setup, the coax should be located toward the top of the CC enclosure.




The full range measurements were done mimicking the in situ requirements of the Chancellor's viewing room, with height and distance from rear wall emulated with a dedicated boundary setup.

As per previous investigation, the need for an acoustical absorber on the floor in front of the CC element was confirmed, and a 4" thick by 24" wide by 48" long ATS panel was emplaced.







The first test sweep confirmed the premises of the most recent development efforts...




Lateral off axis and vertical off axis proved to be most satisfactory, satisfying the performance requirements of this specific view room layout. Some further crossover adjustment is being considered, to tighten up the conformance to target requirements, but early listening checks confirmed the usability.

[Evil Twin]

Of course, my apprentice with the overactive imagination has already suggested that clearly we have an opportunity for a "Natalie C3PO", with some adjustment to the depth of the cabinet and the crossover. For those of more impecunious circumstances, the substitution of Dayton RSS210HF would be require only a few straightforward modifications. But then, they will have to spend the savings on a more powerful amplifier.

[JonMarsh]

Fortunately ET and I don't keep the same hours (I have no idea what the time zone comparisons would be between Idaho Mountain zone and wherever he keeps his local sector headquarters and Imperial shuttle at...), so I snuck in and have done a bit of listening a couple of evenings... surprisingly interesting...

As a playback system, the dedicated test cart probably leaves a few things to be desired, but has some positive qualities- among them,

• A middling performance Freya S preamp
• A large library on the computer's SSD of lossless AIFF, FLAC, and M4A
• Fidelia playback software, which takes over playback hardware, supports high definition files, and has a straightforward library interface
• For a relatively inexpensive straight forward USB C audio interface, the Moto M4 has very high audio quality
  • ESS Sabre32 Ultra DACs
  • measured 120 dB dynamic range on outputs
  • Measured -129 dBu EIN on mic inputs (very desirable for accurate distortion measurements on high performance drivers)

On the downside,

• as configured for test, only single channel output
• No video or HT source material
• Using a "vintage" Cambridge 840W SD power amp, but then maybe that should be on the plus side?

So, I've been doing some listening tests, actually off the sweet spot axis range just to see if there are issues- using some favorite albums I'm very familiar with, things like Fourplay Remastered (a lot of well known guest artists, very good production values), Holly Cole "Temptation" album (good production values, great music for me), Emily Remmler (jazz guitar) "This is me", another well played and produced recording. And one of my favorite performers, Keb Mo, with "Slow Down". You get the idea, no heavy metal shredding, not even Steve Wilson remastered versions of Yes Prog rock (which I have, but not on this MacBook Pro). But, recordings with very well defined instrumental and vocal characteristics with good dynamic range. Should break out some Harry James and some Nickel Creek stuff next, I think.

Anyway, listening to this material on this single speaker setup was rather interesting in a different sort of way. Though measurements show that some additional fine tuning is in order, the subjecting impression is one of considerable top to bottom balance and coherency. One doesn't hear "highs" in percussion, or "clarity of midrange" on vocals, or "well balanced mid bass" so much as you just hear all the instruments and voices with substantial clarity and articulation, and free of obvious "room noise" or other artifacts interacting and degrading the clarity of the music and vocal reproduction.

I've achieved that before, but not with such a simple setup.

I'm trying to get a handle on why I enjoyed the playback so much. Certainly wasn't due to high end speaker cables, given that the crossovers are wired up to the amp and speaker cabinet with ordinary 14 gauge zip cord!

• Maybe it's the focus through the crossover region of the TangBand coax driver being used just as MF and HF- they have a full range reference design using just this driver, so I can't be working it very hard, I suppose.... and with coincident drivers, there is not a typical crossover region lobe or narrowing.
• Maybe it's the Duelund crossover, with an alignment I find that works rather well, a center frequency of 800Hz, an aleph of 2.75, which gives effective transition frequencies of 300Hz and 2kHz, and only one phase rotation top to bottom, and with all drivers being in the same relative phase (when done properly) at any working level over -15 dB in this implementation.
• Maybe it's the surprisingly clear coax tweeter, an inverted dome of an aluminum and magnesium alloy, with an ultrasonic response function more like that of a very well designed Be dome. And with a relatively flat cone waveguide, relatively low levels of interaction due to reflections from the cone.
• Maybe it's the Bamboo cone for the W6-2313 coax, which has a fairly smooth natural roll off on the top that doesn't even get a crossover per se, in the TangBand reference design, because it's also been designed to have an LF balance that provides a certain degree of baffle step compensation. The midrange part of the system crossover for this bandpass has a rather low number of parts...
• Maybe it's the "mandatory" ATS acoustic 4" acoustic panel on the floor directly in front of the system to kill the early primary reflections and their attendant comb filtering effects (peaks and dips depending on phase) otherwise resulting from the height requirement to stay below the projection screen.
• Maybe it's all of the above.

So, much head scratching and more testing will follow... If a Natalie C3PO is to be realized, and evaluated and setup to achieve similar results but with significant low frequency extension. (BTW, C3PO comes from Coax : 3 way: PuriFi : One phase rollover).

But early calculations show a possible design fork, between the PuriFi PTT8.0X04-NAB woofer, and the Dayton RSS210HF-4 woofer. If we assume doubling the LF enclosure volume, there are tradeoffs between the two drivers:

• Raw sensitivity in a 2x 4 ohm drivers in series-
  • PuriFi PTT8.0 ~ 90 dB half plane for 2.83 VRMS
  • RSS210HF-4 ~ 85.3 dB half plane for 2.83 VRMS (slightly better than Wavecor Ardent)
• LF extension sealed box
  • PuriFi PTT8.0 F3 = 60Hz by calculation without special circuitry used in CC
  • RSS210HF-4 F3 = 39 Hz
• LF extension passive radiator
  • PuriFi PTT8.0 F3 = 42 Hz with 2 RSS265PR
  • RSS210HF-4 F3 = 30 Hz, with 1 RSS265PR

It should be noted that basically these differences are due to the T/S parameters, particularly the comparison of Mms, 43g for the PuriFi, and 78g for the RSS210HF. The PuriFi achieves relatively high sensitivity, but it may be termed a monkey coffin woofer as regards required enclosure volume to reach full LF extension. The PuriFi would be better suited to a tall tower, say, 90L for the PR version, 3 RSS265PR's, F3 of 33Hz, or 70L for closed box, Qtc of about 0.6, F3 of 56 Hz.

Both have very stiff cones with extended flat response to above 3kHz before breakup modes.

[JonMarsh]

Some updates to the midrange crossover, and some more measurements.

Raw Measurements of the Coax LF without crossover in the CC cabinet in the application location confirmed a camels hump rise between 300Hz and 750Hz, believed to be due to the proximity to the floor and possibly influenced by differences between the first basic test enclosure and the CC test enclosure design. Adjustments were made to the crossover, including a series LCR bandpass control which can basically act like a tunable amplitude filter at it's center frequency by adjusting the value of the parallel resistor. With the new front panel design approved by the Chancellor, this adjustability may be useful in the final "tune up".

The primary focus for current measurements is examining the off axis behavior horizontally and vertically in consideration of the room layout, especially the location of seating elements. This dictates the need for a larger clean vertical window, to cover elevated and non elevated seating at varied distances, and lateral off axis behavior for side to side seating. Note that the accepted plan, per the Chancellor and Steve Manning, is a modest cabinet tilt of 8-10 degrees; quote off axis degrees are relative to perpendicular to the front panel. Maximum cabinet height on stand is 28". Speaker positioning is per the application requirements relative to floor and rear wall. Measurements at 1 meter.


Some examples:

On axis vertical, 10 degrees off axis:




On Axis, 10 degrees upwards vertical:




On axis lateral, 20 degrees off axis vertical (which is 30 degrees relative to the floor):




Additional adjustments may be made to the combined midrange+tweeter level, to be investigated soon.


A sample of the "production" front panel prototype needs to be fabricated along with a new test cabinet for additional evaluation.

[Bear]

It's always interesting to see the sausage being made. I learn something every time. I'll be the voice of the common man, though, and point out the obvious that the more than $1100 of price difference in the drivers buys a lot of amplifier power -- or crossover components. Would the RSS210HF really perform comparably at 4W vs the Purifi woofers being driven at 1W given the almost 6dB difference in sensitivity?

[theSven]

It's always interesting to see the sausage being made. I learn something every time. I'll be the voice of the common man, though, and point out the obvious that the more than $1100 of price difference in the drivers buys a lot of amplifier power -- or crossover components. Would the RSS210HF really perform comparably at 4W vs the Purifi woofers being driven at 1W given the almost 6dB difference in sensitivity?

Now that's a good question! Lets see what the response is from JonMarsh

[JonMarsh]

It's always interesting to see the sausage being made. I learn something every time. I'll be the voice of the common man, though, and point out the obvious that the more than $1100 of price difference in the drivers buys a lot of amplifier power -- or crossover components. Would the RSS210HF really perform comparably at 4W vs the Purifi woofers being driven at 1W given the almost 6dB difference in sensitivity?

Well, now, that's an interesting question, of course- but let me quickly iterate some key points, as the RSS210HF is being considered for the Natalie C3PO "full range"...

• The Natalie P and some of the subsequent spin off proposals like the Natalie P Ultra or Supreme are considered to be a "value proposition" family...
• The RSS210HF-4 is $129 each, when I just checked minutes ago... and I do not know of any 8" driver with a better behaved cone in anything close to that price range, per my own QNF measurements. Apart from that, it has a well designed motor by most standards (except PuriFi's, for example) with lots of copper in the gap and well controlled impedance rise. For being a "subwoofer", it out performs many drivers in the midrange.
• The PTT8.0X04-NAB is $695 each. Hmmm... that's approaching Accuton pricing, (which, BTW, has really gotten out of hand- though it makes my current stock of drivers "more valuable" in theory... they still don't perform any differently...)
• Along with that roughly 6 dB sensitivity difference, comes a huge difference in enclosure size required for a given LF extension- Hoffman's Iron Law. For a CC, we can live with just solid response to 50 or 55 Hz, and play some games with the crossover to deal with boundary issues. For a full range system with the PTT8.0, it's go big, or go home. That's why I enumerated the differences in LF extension for different enclosure configurations. The PTT8.0's are what one might call "monkey coffin" drivers in the old days, VERY different from the PTT6.5 series drivers.

Some example plots- starting with the RSS210HF QNF distortion, measured in an MDF test box:


Eat your heart out, Dayton E180HE-44.


Now, some Unibox Plots...


First, closed box, 48L, two drivers in series (8 ohm load) and 1W total power. Gotta love those easy to drive 8 ohm loads, right? Remember, classic sensitivity stuff is done with a fixed voltage drive of 2.83 VRMS, which is 1W at 8 ohms, 2W at 4, so single 4 ohm drivers look attractive, and two 8's in parallel push some SPL, but at typical real impedance of 3 ohms or less with a crossover make an amp work hard.



Same enclosure size, same two RSS210HF-4, but lets add 1 RSS265PR to the mix- not two or three, as the PTT8.0X04NAB needs for the LF tuning to work out right.




And for a little more fun, let's see what should happen if we give 50W to each driver, 100W total:





Now, I don't know about you, but I think this is a pretty entertaining value proposition for the money... given that for stereo you'd be saving $1,140, as you'd note.

Of course, the trouble is, three way speakers tend to eat up a lot of dinero for the crossover components, but this would cover a lot of that pretty well. And of course the question arises about the difference one might spend between a pretty nice 75 watt per channel amp at 8 ohms, versus, say, 200W.

You can get some low down with the PTT8.0X04, but you need 70+L sealed, and 90-120L with PR. I wonder how big a monkey would fit into that...

And then, I'd still suggest you might be ahead by going with the LF constellation of the Saint-Säens, 4 RSS210HF in series parallel, a 4 ohm load, and due to the mini array setup, 6 dB more voltage sensitivity... and still only $520 per cabinet, and with a heck of a lot of output headroom.

But, that's another project, and another story... in progress, next up after this one is laid to rest.

[Bear]

Thanks, John, as always! I'm quite convinced that the RSS210 is all that. This looks like yet another outstanding project.