Isiris Min- a definite Maybe?

[JonMarsh]

This is to kick off and share my speculation and design doodles for a scaled down Isiris- let's call it an Isiris Mini. A few folks have queried me about this possibility, so I'll lead off with what I think might be useful goals, and also an initial design pass on the basic LF configuration, which at this point is depending from the updates to the Isiris that are underway.

First off I think I have to address the idea of what and why... as well as a bit of the how.

• The original Isiris was somewhat beyond the pale for most folks, understandably, with a BOM of around $10K, and certainly some complex building requirements. Mine was build manually, Mike's part pieces were cut with CNC.
• I've had a TA6 crossover update for this in the wings since 2014, but for a variety of reasons (including the Wavecor Ardent development) no time to address it or complete. By in the wings, I mean the parts were all bought and in storage.
• Interest from another builder in Singapore has resulted in me "re-opening" the case files on this sooner than I planned... so be it. But it's still a very expensive project. Not for the faint of heart or impecunious of budget.
• There has been a LOT of progress on the performance of drivers at more modest price points- witness the Tympany DA25, for example, or the SB26CDC-C004.
• So, somewhat from just an "artistic" viewpoint, I've been pondering how much of the Isiris experience might be brought to a roughly 3/4 scale project, with perhaps a 1/3 scale budget? Not a small challenge, but possibly a worthwhile one.
• Key driver concerns are polar response, CSD response, and distortion. With a 3/4 scale effort, I expect the system sensitivity to drop somewhat, reducing the expectations for the midrange and tweeter to more manageable driver choices.
• Construction may still be a challenge, but if I can do it with a standard table saw and a good dual bevel sliding miter saw, many others probably can, too. I'm a wires and sparks guy, not an expert wood worker. I am patient and careful though.

Let's start from the fundamentals, in this case, the fundament or bass cabinet. Lest anyone not be aware, I have a bit of a soft spot for the Dayton RSSXXXHF series drivers. They rather successfully straddle the divide between woofers and subwoofers, and do so at a remarkably reasonable cost.


So first off we'll look at two possible alignments, which can be done in the same cabinet, and in fact, implemented as a "you make the call" depending on your setup and preferences.

The givens here are two RSS265HF-8 connected in parallel, which gives us a 2pi sensitivity of about 90dB for 2.83VRMS, in a 90L enclosure.




Here I'm showing the output level for 20W per driver.




There's still a fair amount of headroom for higher SPL at low frequencies, and if one uses a Cardas style placement with optimized boundary boost, the F3 at 38Hz can be substantially improved on, with basically flat response at the listening position down to 30Hz.



Excursion relative to Xmax is well controlled, an at 100 dB output you'll remain in a fairly linear part of the Xmax curve.

The predicted impedance curve looks acceptable for well designed amplifiers to cope with.





One might term this something of a purist approach...

[JonMarsh]

Boys just want to have fun

Now, how many of you remember that old Cyndi Lauper song, "Boys just want to have fun"?

I see a few hands raised around the virtual room...


So, for the Boys that want to have fun, what you do, is unplug the port built in to the bottom of the cabinet....


What does that do? (if you build it as described below...)






Selecting the right tuning frequency is critical, or it gets a bit crazy. And in this case we're turning at or just below the Fs resonance of the drivers.




Now you've got about 1-12dB of lift maxing out at 40-50Hz (rock and roll, hootchie koo!) and response down 2dB at about 26Hz. Is it party time yet?


Excursion is pretty well under control,




But you do have to keep your eye on port air speed; this is with a 4" ID port So there is some headroom, especially above 30Hz.





At the levels shown here, it should be loud, and effortless.

[Mr.Ed]

Watching this.

[Efalegalo]

Watching this.

Ditto.

[JonMarsh]

Dark horse candidate for midrange...

OK, we've already mentioned the SB15CAC as a possible midrange candidate... more testing will be done, specifically for the 4 ohm version soon.

But... there is a "dark horse" candidate. And I don't mean the SB17 that was gracefully posted... I'm reaching back to the past, to a classic Pro Audio builder, and a driver I've been sitting on for a while, waiting for what might be a good opportunity.


Eighteen Sound is a secret love of mine... I have quite a few of their drivers that have been acquired in the last few years, passed initial tests, and been stashed away for further development. One a 6-1/2 driver described as a bass midrange driver, or mid woofer, but which I see as with potential for a larger high performance midrange. That's the 6ND430- A Neo motor part that has been out for some time and is still readily available.








It has a very smooth, though sloping upwards response curve, and good off axis behavior and low distortion. And it's not hideously expensive, either. Extra work was done in the development to control resonant radiation of the cone surround, so it has an unusual looking surround shape, especially for when it was designed.

This is a 1 meter capture, mounted in a small ported enclosure.




Here is a polar series, on 15 degree increments, in the same enclosure.




For me, this indicates very solid behavior for a crossover in the 2-2.5kHz range, if 3rd order or higher.

[JonMarsh]

One more thing...

I'm going to leave you with one more plot, to tease your interest- this is a semi near field response plot (microphone at about 10") so as to see more of the driver intrinsic quality without room reflections, and with the 2nd and 3rd order harmonic distortion plotted, too.




As you can see, 2nd harmonic is down about -48 dB, and 3rd harmonic more than 10 dB lower. Between -60 and -70dB. Not shabby.

Definitely worth giving this one a closer look soon.

Very solid potential in the desired range from about 250-300Hz and up.

[Efalegalo]

Jon,

Just an FYI - I just tried to PM you, but your inbox is all full and cannot accept messages.

Thanks,
D

[Efalegalo]

Zaph Review

That's the 6ND430-

It appears that Zaph tested the 16ohm version in 2008 and had this to say:
"Value: Excellent performing 16 ohm pro midwoofer. Smooth response and an extremely clean midrange. Very usable T/S parameters and excursion though somewhat limited bass extension for what is meant for woofer service. Very sensitive. It would make a fine midrange also. Lightweight neo magnet assembly with cooling fins. A couple of these paired up for an MTM and used with a subwoofer would make a reference level system. High value for this level of performance."

[technodanvan]

Promising beginnings...

[JonMarsh]

I’m on relocation work travel the for three days, starting today. I’ll fix the PM in box soon, probably this evening when in CA.

[Bear]

I'm going to leave you with one more plot, to tease your interest- this is a semi near field response plot (microphone at about 10") so as to see more of the driver intrinsic quality without room reflections, and with the 2nd and 3rd order harmonic distortion plotted, too.


[ATTACH=CONFIG]31240[/ATTACH]

As you can see, 2nd harmonic is down about -48 dB, and 3rd harmonic more than 10 dB lower. Between -60 and -70dB. Not shabby.

Definitely worth giving this one a closer look soon.

Very solid potential in the desired range from about 250-300Hz and up.

I've played with this one in models, going back to some posts from John K regarding Duelund-style crossovers. This was one of the midranges that looked like it could actually hit a first order acoustic HP (Qtc = 0.5) with a simple sealed box. That's obviously "bench racing", but it did look like an interesting, if not especially attractive, driver for use as a mid. It also has no shortage of sensitivity for use as a hifi mid. The truncated frame is a touch of a PITA, but it does help with the CTC spacing -- though Kimmo's posts on that subject have challenged what I thought that were some relatively clear design guidance around that.

[JonMarsh]

I think it’s going to be interesting taking an even close look - for now, I could even build just a new top module and pair it with one of the existing bass modules for evaluation.

And the mounting will be just as much of a PITA as the C79, if not more!

And the PM in box has been opened up!

[Efalegalo]

If considering a more efficient mid-range driver (e.g. 6ND430), would the RSS315FHA-8 be a more suitable bass-duty driver?

In regards to the RSS315HFA-8 vs RSS265HF-8, solely based on a quick review of spec sheets:
- The 315 has less "humped" bottom end than the 265, but maybe that favors the 265 if less BSC is needed
- The 315 has a smoother FR between 100hz - 1200hz, a less benign breakup (but still severe nonetheless)
- The 315 has an ever so slightly flatter impedance curve than the 265. Not sure if this is relevant, but I thought I read somewhere that a flatter impedance curve is an indication of a good motor design.
- The 315 SPL rating (85.7@2.83V/1M) is more indicative of the FR-SPL at the target x-over frequency (250Hz-300Hz) than the 265, on which the spec sheet states 84.2@2.83V/1M, but the FR indicates 82db at the target x-over frequency.

The 315 is only marginally more expensive than the 265 ($200 vs $170), but there may be a bigger delta once you factor in additional cost from a larger cabinet.

[dwk]

If considering a more efficient mid-range driver (e.g. 6ND430), would the RSS315FHA-8 be a more suitable bass-duty driver?

.

Not to speak for Jon, but if you check the 'other' Isiris thread, you'll see that the RSS315HF-A is being used/evaluated as an alternate bass driver for the *full size* Isiris. I think what Jon is going for in this 'mini' version is not so much cost saving on the bottom end but instead a smaller cabinet and slimmer footprint - the full sized Isiris is a pretty large beast. Granted, he seems to be starting off looking at 90l vs 115l which isn't overly significant, but I suspect the RSS265 version can be squeezed a bit.

The cost saving aspect seems to be in play with the M/T components though. Of course, around here many a 'budget' initiative gets corrupted on the way to realization, so we'll have to see how this one plays out

[JonMarsh]

What he said- quite observant. The RSS315HFA-8 is for the updated original Isiris- already have 4 here, cabinet is being updated a bit, and CAD drawings to match- have another member in Singapore very interested in this, and he has the JDT1024 tweeters already.

Yes, this is intended at this point to be a slightly scaled down physical version, and a substantially scaled down cost version. Hopefully without giving up hardly anything. Since I”lol have both on hand, I’ll be able to make measurement comparisons and listening comparisons.

[JonMarsh]

Also, dropping back to the RSS265 helps with the sensitivity matching, and it also gives two interesting response profiles to choose from in moderate size enclosure. The RSS315HFA-8 is a bit too fat ported.

[Efalegalo]

Not to speak for Jon, but if you check the 'other' Isiris thread, you'll see that the RSS315HF-A is being used/evaluated as an alternate bass driver for the *full size* Isiris. I think what Jon is going for in this 'mini' version is not so much cost saving on the bottom end but instead a smaller cabinet and slimmer footprint - the full sized Isiris is a pretty large beast.

Duly noted:-) though I'm may be one of the select few that wouldn't mind owning a speaker the size of the full Isiris. I owned a pair of Klipschorns as well as a pair of JBL 4345 Studio Monitors at one point. While these speakers are not as tall as the Isiris, they do still take up quite a bit of floor space.

.... Of course, around here many a 'budget' initiative gets corrupted on the way to realization, so we'll have to see how this one plays out

Hahaha - So true, dwk! So true!

[JonMarsh]

Well, maybe I should hint that the “full” Isiris is going to be investigated for further updates, though it might require building a new top module. I do have a pair of zC168-890’s on hand- but whether they will be a good replacement for the C173-6-090 is an open questions. Accuton wants to stop making the 090, it seems, but it’s one of the best drivers they’ve ever done. Sill waiting for published data from Accton on the -890; at this rate I may have my own before they get around to it, just like some SB parts.

I wrote too quickly, without checking again. Data is up, and it is not quite what I was looking for.. flatter than the -990, but only about 85dB sensitivity. Definitely not a C173-6-090 replacement. Going to have to measure the C168-9-990 I have carefully…

[JonMarsh]

The "mini" is mainly intended to be much more "mini" in price... options remain-

OK, let me toss out a few more comments or thoughts, in consideration of PM's I've gotten-

• The biggest focus on the "Mini" is making the Mini more price conscious, and attempting to get as close to the same performance as possible, but at half or less the BOM cost.
• Even with the RSS265HF, the mini would not be very "mini"- same cross section is planned, but it would be shorter, reducing the enclosure volume.
• If one wants to do a reflex, the RSS265HF-8's in 90L are flatter than you can get with the RSS315HFA-8 based on Unibox modeling- and a big problem for the RSS315HFA-8 if you wanted to go ported is the diameter of the port needed to keep port speed under control at the higher exucsion levels. In fact, that would result in a primary port resonance below 200Hz, so sealed is the way to go for that driver.
• For a variety of reasons, the top module design will not be changed between the "maxi" and the mini, as regards the cabinet dimensions.
• The low frequency drivers are crossed over to the midrange at about 350-400Hz, and both the RSS265HF-8 and the RSS315HFA-8 are very smooth out to 500Hz. Both require ellptic filters to supress the cone breakup and resonance amplification of distortion product. The higher sensitivity of the RSS315 would likely make it a challenge to match up with almost any midrange 6" other than the C173-6-090, but the proof is in the pudding and actual measurements. This will be easy to do, once an LF cabinet is modified, and I have another top module built for the final driver configuration.
• For the midrange, my target is a flat distortion profile down to 300Hz, and wide polar response up to 2kHz or slightly beyond.
• For the tweeter, my desired distortion profile is flat down to 2kHz, and no inductivity rise above 10kHz.
• There are less tweeters to chose from for this project than one might realize, as I intend to stay with a classic 104mm frame diameter (this rules out the DA25, unfortunately, which is much larger at 120mm.
• Another sourcing problem for the tweeter is that many potential candidates are out of stock in North America, from well known high volume suppliers like ScanSpeak- by out of stock, I mean not available at Solen or Madisound, and listed as out of stock, and a call says they don't know when they can get some.
• I have a candidate tweeter driver on hand, a pair, which I need to test; other tests on it show the desired distortion profile may be possible with using an elliptic breakup filter, and has very low 3rd harmonic at 95dB output (available from Solen, in stock). Price is higher than I would like, but only 1/4 the Jantzen diamond tweeters.
• I will be retesting the 6ND430's more extensively soon, partly for verifying sensitivity characteristics in comparison with "known standards". The polar data I already have shows it exceeding the target, with promising behavior past 2kHz.
• I have NO reason to discourage anyone from doing a full size Isiris Mini; let's call that an Isiris MkII for now, given that the C173-6-090 is difficult to procure, though currently in stock at Madisound, as of April 2021.
• It's looking like the driver cost BOM will be in a similar range to the Wavecor Ardent, for the "Maxi Mini", around $1K per cabinet. That's a lot different than the $3500 for the original. And it's coming in on the target I set for myself.

6ND430 Polar response 0 - 50 degrees, 10 degree increments. This may change with a different cabinet profile.

[technodanvan]

What is the rationale for sticking with a 104mm OD tweeter? Aesthetics and fitment in the top cabinet?

I'm onboard with this one, as soon as hardware choices are firmed up I'll start submitting the project piecemeal to my budget manager. May choose cheaper design materials (primarily MDF) to go along with fiberglass for this.

[Efalegalo]

[*]There are less tweeters to chose from for this project than one might realize, as I intend to stay with a classic 104mm frame diameter (this rules out the DA25, unfortunately, which is much larger at 120mm.

Jon,

Would the Satori TW29R work? It has a frame of 103.8mm, and has developed a reputation for an excellent sounding tweeter. I might have a pair here at the house.

HiFiCompass testing:

https://hificompass.com/en/speakers/measurements/satori/satori-tw29r

I also have a pair of CSS LD25X (frame = 74mm). These were the earlier version, but they have been reintroduced. Apparently, the new one has identical performance to the prior version, with maybe just a slightly higher extension.

The late Mr. Bagby did perform some testing on them when they were first released:

https://www.box.com/s/sooaxdil8sjjvcl1d61p

Current production specsheet: https://www.parts-express.com/pedocs...spec-sheet.pdf

For mid-range consideration, would the Satori MW-16P-8 work for the Iris-Mini (or 4ohm version for full Isiris MKII)? It's about 2db less efficient than the Accuton at the target x-over range, but the parallel RSS265 is 2db less efficient than the RSS315 at the target x-over range (as indicated in your PM).

HiFiCompass testing:

https://hificompass.com/en/speakers/measurements/satori/satori-mw16p-8

If any of the aforementioned drivers are of interest, let me know and I'll be glad to send them to you for your testing.

[JonMarsh]

What is the rationale for sticking with a 104mm OD tweeter? Aesthetics and fitment in the top cabinet?

I'm onboard with this one, as soon as hardware choices are firmed up I'll start submitting the project piecemeal to my budget manager. May choose cheaper design materials (primarily MDF) to go along with fiberglass for this.

Hey Danny,

My rationale? Steve would say I'm just a stubborn cuss; if I wanted to make a joke, I'd say I'm too lazy to do a complete new 3D plan for this, but the justifications I'll stick with for now are...

• If the JDT1024's go south or I need to replace them with an "upgrade", I'd like to have some possiblities waiting in the wings... and a proven filter
• I don't like the effect that a wider top has on the ethetics or the baffle loading behavior; get out your copy of "Edge" and simulate the tweeter response position with different widths and shapes, if you're curious.

But maybe, in the end, Steve is right!

:T

(What I didn't mention was that I'd be happy to use a "suitable" smaller form factor tweeter if one was available, but that is proving to not be the case either for performance reasons, or the supply just drying up due to what I suspect is pandemic supply chain issues.)

[JonMarsh]

Jon,

Would the Satori TW29R work? It has a frame of 103.8mm, and has developed a reputation for an excellent sounding tweeter. I might have a pair here at the house.

HiFiCompass testing:

https://hificompass.com/en/speakers/measurements/satori/satori-tw29r

I also have a pair of CSS LD25X (frame = 74mm). These were the earlier version, but they have been reintroduced. Apparently, the new one has identical performance to the prior version, with maybe just a slightly higher extension.

The late Mr. Bagby did perform some testing on them when they were first released:

https://www.box.com/s/sooaxdil8sjjvcl1d61p

Current production specsheet: https://www.parts-express.com/pedocs...spec-sheet.pdf

For mid-range consideration, would the Satori MW-16P-8 work for the Iris-Mini (or 4ohm version for full Isiris MKII)? It's about 2db less efficient than the Accuton at the target x-over range, but the parallel RSS265 is 2db less efficient than the RSS315 at the target x-over range (as indicated in your PM).

HiFiCompass testing:

https://hificompass.com/en/speakers/measurements/satori/satori-mw16p-8

If any of the aforementioned drivers are of interest, let me know and I'll be glad to send them to you for your testing.

Thanks for taking the time to make these suggestions and kind offers... I should give you some more background info.

First, I tested a boat load of SB Acoutics part back in 2019, including the MW16P-8, the MW16PNW-8, the TW29BN, the TW29DN, the SB29BAC, and the SB26CDC-C004. I also tested the LDX25 at that time, and the SDX10. The SDX10 is a keeper, the LDX25 didn't bring joy for me.

Since the 70's and working with a few different soft dome tweeters and hard dome tweeters, I've developed a bit of a bias against soft dome- many of them, well, for me, perhaps most of them, just sound a bit fizzy, even when measured response seems to look OK. This was something I used to discuss with Charlie Hansen when he worked for me as a tech, while going to CU in the late 70's, and we revisited that topic after he founded Avalon Acoustics and we were both still in agreement. The TW29DN didn't change my mind, and I find tweeters that use a center pole, or pin down, and the whole tweeter is basically acting like a surround, which, to make a slight joke, is for me, a romper room no-no.

This is not to say that people can't make a fun speaker with a soft dome and enjoy it and what it does with the music they like- that's even why I joke about making a ported version of the RSS265HF-8 Isiris Mini, because some might find that coloration to the response profile a lot of fun with the music they enjoy. And in the end, this should be about fun, and (pardon the meme) sparking joy.

But I've got my habits sacculated over 50 years of doing this, and at this point I tend to stick with them, with occasional side investigations and reality checks.

One thing I have to acknowledge is that I rarely do infinite baffle tests for comparing drivers- because unless you do a built in, the actual driver loading on the baffle will look quite different. For tweeters I usually use a very simple setup, with a 1ft square box with baffle with full round-over edges to reduce overt diffraction. So I'll share some measurements, but don't compare them directly with measurements done differently. What I do is a lot like what Seas does, which is to provide measurements on what amounts to a typical application baffle size. That may not be your preference, or others, but it's worked pretty well for me over the years.

So I'm going to share some of those results next. One thing I have to frequently remind people about driver manufacturers, is that there is not necessarily a correlation between technical excellence of the performance of the driver, and commercial success of the company. Some interesting parts come and go, but without the right promotion and support, go nowhere commercially, and get discontinued. Having a good business model is just as difficult as having a good product design.

[JonMarsh]

For those not familiar with my test setup,

• I have a couple of ACO Pacific PS9200 Kits and preamps, which are good out to about 50kHz, and have very low noise and wide dynamic range. This includes the PS9200 power supply and mic preamp buffer
• The current audio interface is a Motu M2, but I also have a Focusrite Scarlet two channel as backup, since my TC Konnekt died on my this year.
• The controller computer is a 2014 MacBook Pro. If it ain't broke, and runs like new, why replace it? SSD of course. Never had a laptop with an SSD at the multi-billion dollar company I used to work for- too expensive for their taste.
• Power amplifier is a Cambridge 840E
• Voltage levels driven to the DUT is closely monitored by a Teledyne T3DSO1204 digital storage scope.
• Test Software is Fuzzmeasure, now sold and updated by RØDETest. I have been using it since the earliest version by the original developer. It has three different operating modes, including a focus on speaker measurements, and a complete set of architectural acoustic tools. Only available for Mac.
• I have a CLIO Pocket version 2, which runs with Mac or PC and is completely self contained, does acoustics and impedance measurements and just interfaces with USB to the computer. But it is slow and futzy for doing some things that are very easy with Fuzzmeasure. So it's more of just a backup or a very compact portable tool for on location basics. I used to use the main CLIO product years ago.
• I used to use Liberty Instruments Praxis years ago and have a computer with that installed somewhere in storage; that was the tool used for the original Isiris development, especially distortion measurements. It was last updated in 2011.

Let's look first at some of the shiny bits from SB Acoustics...




we'll start off with the TW29DN...





Here's the basic SPL plot as captured by Fuzzmeasure; it's normalize to the electrical system level, not to a calibration level. I do that separately when I want using an ACO Pacific level calibration generator.




Here is the captured polar response, at 10 degree intervals out to 50 degrees, using a turntable to rotate the test cabinet on a stand, and the front baffle aligned with the mechanical center of the turntable.




Here is the measured HD2 and HD3. KISS.




The TW series tweeters have a pretty high grade Scanspeak like motor; which means quite low HD3 and much higher HD2. What's interesting to do with drivers that have highish HD2 is to do some dual tone spectral analysis with the acoustics model for my Audio Precision APx555. There's a lot of low level spikes around the primary tone, that low HD2 motors don't have to the same degree. Is that "fuzz"? I don't know. I know it's less objectionable than high levels of HD3.

There's some pretty good performance here for a $155 dollar tweeter, but maybe we should expect more in some regards? I'm not liking the flatness of the SPL response, or the polar behavior all the much, but it is just a $155 dollar part...

[JonMarsh]

Let's move down market to $50...

Let's look at a part I've used in several recent designs. Same test setup, same test cabinet design.






Basic SPL plot on axis:






Now, the polar plot in 10 degree increments out to 50 degrees:




My best guess as to why the overall response smoothness and the polar behavior are so nice is a combination of the profile of the dome, a very low profile surround, and a nearly absolutely flat front plate that doesn't create diffraction problems.


And rounding things out, a distortion plot- note that with a series elliptic filter for the breakup resonance, the HD2 climb from distortion product resonance amplification can be eliminated.





Now, some of the dip at around 1800Hz and bump up around 1kHz is due to the baffle step characteristic of this test box. But you can see above 3 kHz the DA25 is nicely behaved. in both amplitude and distortion characteristics. The resonant breakup is easy to deal with, and what you have left looks very nice, as seen in the measurements of the Calliope. Anyone that knows a better way to spend $50 on a tweeter, please let me know.

[JonMarsh]

Tw29bn-b-8

Let's post one more this evening, the TW29BN-B-8. This one is closing in on $375, same turf as the 6640 used in the Wavecor Ardent.






First, the basic SPL response:




And the polar response out to 50 degrees:





Theres's more separation of the response and change in the profile with angle than I would like to see in a tweeter this expensive. Contrast this with the behavior of the DA25.


And last, the distortion plot for HD2 and HD3





The motor is very much typical Satori, which is mostly a good thing.

The tweeter has a "minimally invasive grille" and no phase shield in an attempt to improve off axis response. This is another wide surround design, with a very shallow cone effect in the faceplate.

Sometimes I wonder if this design doesn't create problems with diffraction and the response behavior as seen in the 0- degree on axis behavior, and to some degrees in off axis. I'm only speculating.

[JonMarsh]

Css ld25x

Let's take a look at the CSS tweeter...

I was interested in this for a center channel application






I don't know how much the basic response of the tweeter is modified by the short waveguide baseplate. Certainly the top octave may be. The factory spec sheet shows a top octave 5dB bump in the response centered at 15-16kHz.

The factory curves show a slight broad dip centered at roughly 2kHz, and this is seen to a greater degree in the sample measured.

The LD25X is the only tweeter I know of claiming an XBL2 motor, but they don't give any specs for Xmax or how an XBL2 motor will improve performance compared with the underhung motors typically used for tweeters, which already have pretty flat BL curves in the normal operating range.




Here, the polar response in 10 degree increments is presented from 0-60 degrees.





Like the factory curve, the 30 degree response rolls off somewhat sharply above 15-16kHz.


Distortion is shown here:





A low spread between HD2 and HD3, which I like. Steadily declining HD3 indicates distortion at 2.83VRMS drive is directly related to excursion, so I'm questioning if their choice of an XBL2 motor is working out as they hoped. At this drive level one often sees a flat HD3 curve from 2 or 3kHz up.

[JonMarsh]

SB26CDC-C000-4 is a tweeter I was planing to use in a new project scheduled for later this year for Steve... unfortunately recent checks seem to show the it has gone the way of all things of the flesh.

The product page is removed from Madisound Speakers, though there are traces of the part lingering if you know where to look...

For the record, I'll review the characteristics, then move on to another unobtanium part.





As you can see from the picture, this is a tweeter using a phase shield; I modified one part and tested in both configurations, but I'm only presenting the "stock" configuration here.

The basic on axis SPL plot:





The polar response, in a different format: 15 degree increments, to 45 degrees.





The distortion curves at 2.83VRMS.





For a hard dome driver, the ceramic/aluminum combination works pretty well; there is a break up peak which the phase shield hides from the on axis reposes, but you can see that the distortion pilots look pretty nice, and no resonance amplification of distortion products is evident.

Note that this is NOT in their series that use a wide surround to "increase output" - is there a chance this correlates with the overall lower HD2 to HD3 ratio?

Interestingly, with the phase shield removed, there is a 10 dB peaking in HD2 at about 13kHz, but it's relatively mild compared with what you see in many hard dome drivers.

[Efalegalo]

Very interesting! Thanks for posting the measurements above, Jon. I guess I never really paid any attention to the dispersion profiles of tweeters. I would certainly think about the polar response between the tweeter and mid, for directivity matching at x-over frequency, but never really focused on dispersion above (say) 8Khz.

[JonMarsh]

Sb29bac-c000-4

The next part is not the other unobtainium part I mentioned, but another "midrange" SB Acoustics tweeter, the SB29BAC-C000-4. This is also a Beryllium dome tweeter, and with a wide surround design, but not technically in the Satori series.







On axis Basic SPL plot:





The polar set here is a bit more limited, on 15 degree increments to just 30 degrees. I didn't have the heart to go further...






Here is the HD2 and HD3 distortion sweep.






HD3 is quite low, as I'd expect from a typical SB Acoustics tweeter motor, but the HD2 is a bit "quirky". Large surround issues? Hard to say. Maybe a resonance in the motor system that contributes to the bump at 4500Hz. The actual dome construction seems good, but does it have a well damped resonance just at 27-28 kHz? Seems low for a Be tweeter. The problem with budget Be tweeters is that due to the cost of the dome, this tweeter still sells for over $300.

[JonMarsh]

Jantzen JDT-1024

And now, we delve into the history books a bit more... the Jantzen JDT-1024.

Part of the attraction for a diamond tweeter is moving the break up resonance up so high that it has no relevance to the audio band, and no impact on factors such as resonance amplification of distortion products. Still, you have to have a good motor and overall design. The Jantzen was one of the least expensive diamond tweeters offered, and was used by some OEMs.





These measurements were done on the Isiris top module, and so should not be directly compared with others I've shown, at least not in the lower frequency ranges, as the baffle step characteristics are quite different.





Note, this tweeter also uses a phase shield, and this probably has something to do with the less than stellar smoothness of the on axis plot.

Now, think about this- for the overall power response, the 15 degree or 30 degree characteristic covers a much larger part of the forward hemispherical power response than directly on axis.


Here's a plot of the 30 degree off axis response:




Compare this to the 30 degree curve for the SB29BAC above.


And here is the combined 15 degree plus 30 degree plots:

[JonMarsh]

And here we have the distortion profile, presented for 30 degrees:




HD2 is pretty dang low... HD3 isn't shabby either.

Anyway, this should give some insight into the performance characteristics that I'd like the next tweeter choice to exceed, not just match. But I could settle for matching, I suppose...

It's not as easy as you might think.

[JonMarsh]

Today I'm picking up supplies and cabinets from my storage, and will begin setting up test cabinets for the 6ND430 and for a fairly well known newish tweeter, which I know is well stocked in Canada.











This will get a new custom rebated test enclosure, as is usually the case for each tweeter tested unless it matches a previous part dimensionally.

I'm scrounging stores in the area to see if I can come up with LBL for a new upper module build soon... I don't want to disassemble the existing set. If the measurements on my "standard" test platform look good, then I'll hopefully proceed with a cabinet build soon.

I have some of the T25 variants which I've tested, but tests so far indicate that a 2kHz crossover is not really a good choice for them, based on HD measurements and SPL measurements.

Shortly I'll run down some data on midrange drivers, past and present candidates. Just yesterday I found out one very nice one, in consideration for another project, has apparently been discontinued, and is closed out at PE, the TangBand FW135H. I'll show my data for that part, too. It's a shame, and evidence that it takes more than a good product to be successful.

[xandresen]

SB26CDC-C000-4 is a tweeter I was planing to use in a new project scheduled for later this year for Steve... unfortunately recent checks seem to show the it has gone the way of all things of the flesh.

The product page is removed from Madisound Speakers, though there are traces of the part lingering if you know where to look...

SB still has the SB26CDC product page up on their website, linked through the menu: Drivers/Tweeters/HARD
Meniscus and Solen are accepting back orders.

I use it in my current design, hope we don't have to say RIP so soon...

[JonMarsh]

SB still has the SB26CDC product page up on their website, linked through the menu: Drivers/Tweeters/HARD
Meniscus and Solen are accepting back orders.

I use it in my current design, hope we don't have to say RIP so soon...

I'm with you on that, I'm going to check some other vendors next.

.......

Solen has 4 in stock, with backorders allowed. Curious... maybe Madisound is just cutting back on their offerings, or perhaps they're redoing some pages. The picture link was available at Madisound, and typing it it into the search field seemed to generate a hit, but resulted in a "page not found" error.

[JonMarsh]

OK, some discussion about midrange drivers-

For an Isiris MkII, and assuming that Accuton finally does close the tap on C173-6-090 (like they have for the C79) what options are left? Viable options, that is...


The logical one, and the one used in some form or another (sometimes with diamond diaphragms!) is the C168-6-990. Now, mind you, I like the C173-6-890, and have a pair on hand, but they're intended as a bass-midrange, and have about 6 dB lower sensitivity. Not in the right ballpark.




This is a fairly early release part right after the introduction. The current factory photos look more "shiny".

There is a fair amount of controversy about this driver and it's suspension technique, and whether that allows for rocking couples and resonances affecting distortion. I certainly don't have the tools to do a mechanical simulation, so all I can do is measure the results. Please note my usual disclaimers at the bottom of this and every post...





These measurements were done in a PE 28L cabinet, with a custom fabricated mounting plate. The construction of this driver is such that (in my opinion) it must be mounted on a thin strong plate so that the back wave from the cone can exit the holes through the voice coil former and into the enclosure. I question that just having the spider area for back wave is sufficient. So, I expect for everyone except Steve Manning this will raise the Olympic degree of difficulty should you chose to use this driver in a build.





Here is a free air impedance plot:




You can see that for the intended operating range (for my purposes) between 350 and 2200 Hz the impedance curve is pretty flat- makes life a bit simpler.

[JonMarsh]

C168-6-990 measurements

OK, a few measurements, done in the 28L PE cabinet with special mounting plate.


Nominal on axis SPL:






Polar response, at 10 degree increments:




Possibly the shallower cone is responsible for improved high frequency dispersion compared to other similar sized drivers from Accuton.


Distortion at 2.83VRMS, 92 dB max output:





The obvious comment of note is the resonance amplification of 2nd order distortion products from the 7250 Hz breakup. It remains to be seen if an elliptic network will suppress that, but past experience with other hard cone mid woofers indicates it most likely will. This is not stellar performance, but we've all seen much worse.

This driver is being used in the Osiris build, and I'll be testing it further in a week or two.

[Efalegalo]

I use it in my current design, hope we don't have to say RIP so soon...

From the various measurements Jon posted, it may be amongst the highest value tweeter. The response 15-30 degrees off-axis is surprisingly flat. Also equally impressed with the Peerless.

[JonMarsh]

For the various measurements Jon posted, it may be amongst the highest value tweeter. The response 15-30 degrees off-axis is surprisingly flat. Also equally impressed with the Peerless.

Yeah, I did a lot of tweeter testing in 2019 and that was the conclusion I drew then- Steve and I have a concept for a center channel speaker using the SB26CDC-C000-4, and I even have a test box fabricated for it, based on Anarchy mid woofers for the bottom end.

As to the Peerless, this was the first system built with it, the two way using the MW19P.




The bump in HD2 at 1300Hz or so was due to the woofer, not the tweeter.

[Efalegalo]

Ss18w/4434g00?

Certainly not in the league of the Accuton's, but since budget drivers are being considered, how about the SS18W/4434G00 - 91.5db/2.83v.

I have no personal experience with this driver, but I've noticed that Troels keeps coming back to this driver (the 8ohm version, that is).

His comments on the 8ohm version:
"I want to start out with the most important driver here, the 18W/8434G00, on the given baffle.
Now, not since the Ellipticor 18WE have I seen anything like it. Smooth like a baby's bottom.
Not the slightest hint of any disturbance around 800-1200 Hz, rubber surround resonances, which is all too common.
Extended range up to 7 kHz! And a smooth roll-off."

Measurements on an actual baffle (sorry - no polars), per Troels, attached.

[Bear]

Today I'm picking up supplies and cabinets from my storage, and will begin setting up test cabinets for the 6ND430 and for a fairly well known newish tweeter, which I know is well stocked in Canada.

If this is the target droid, I'd be happy to offer up my pair of T34A's for testing for a slightly lower price point option. But let's be realistic, you probably have eight of them in stock that you threw in to fill up the rest of the container in your last order from "up north".

I'd be curious to hear more about the T25. I have a singleton in stock that I'm planning on using for a more classic center channel.


I've got a pair of a pair of SW223BD03s that are also in inventory waiting on a project. I'm still moderately fixated on more of a SF Strad idea, though.

[JonMarsh]

Bear, I'll see if I can't dig up that info about the T25. I acquired some of them pretty much for that purpose, in the manner you speculate on, for a CC, but for doing a CC in the Duelund style, they didn't quite have the coverage needed. For a more conventional crossover approach, they'd probably be fine... But clearly, the one to have would be the Diamond, though the Be one isn't bad either- some significant differences in linearity of frequency response that leave me scratching my head a bit...

[JonMarsh]

Certainly not in the league of the Accuton's, but since budget drivers are being considered, how about the SS18W/4434G00 - 91.5db/2.83v.

I have no personal experience with this driver, but I've noticed that Troels keeps coming back to this driver (the 8ohm version, that is).

His comments on the 8ohm version:
"I want to start out with the most important driver here, the 18W/8434G00, on the given baffle.
Now, not since the Ellipticor 18WE have I seen anything like it. Smooth like a baby's bottom.
Not the slightest hint of any disturbance around 800-1200 Hz, rubber surround resonances, which is all too common.
Extended range up to 7 kHz! And a smooth roll-off."

Measurements on an actual baffle (sorry - no polars), per Troels, attached.

Now, I never would have looked at this part, because of SS's published data. We ARE talking about the Scanspeak Discovery 18W/4434G-00, right? A relatively budget 7 woofer that goes for $71 at Madisound?




It has the classic impedance bump at 1kHz, and the discontinuity in response with a big on axis bump up starting around 1100 Hz, with a classic Scanspeak/SB Acoustics plateau peak starting at about 3kHz.

Frankly, this "classic" behavior for paper cone drivers is one of the reasons Evil Twin did the NatalieP speaker with the RS180's way, way back... there is none of the impedance discontinuity from mechanical resonance of the cone edge and surround, and the response is well controlled and tracks pretty well off axis to 45 degrees out to about 1800Hz. And sensitivity is about 89.2 dB. So, for a budget implementation, based on factory data, I'd be inclined to trust the RS180-4 a bit more- and I know well how to deal with the breakup at 7 kHz.

Something always to keep in mind with any manufacturer data- or other DIY site data- is a favorite saying a former colleague and I used all the time... "In God We Trust, all others we verify..."

Just recall the recent fandango with the Dayton RS52...

I'm not a well known reviewer like Troels or HiFi Compass- I get my parts not from the distributor or manufacture, but off the retail shelf, just like you guys.

So, if you are inclined to follow up on this anyway, let's discuss off line. I even have a "spare" CLIO Pocket setup I could send you, which only requires a USB interface on your computer and the supplied software- running down and duplicating both companies measurements would be an interesting experience. If you want to proceed, I'll get an RS180-4 and send it to you.

[tktran]

Hi Jon,

Pleases me to read you're back active on the forum after some quiet years. Hope you're in good health.

From what I can gather 18W/8434G00 is a good but basic woofer; and measured by Hobby Hifi in 2012:

http://website.audiocomponents.nl/download/scan-speak/18W-8434G00%20-%20Hobby%20HiFi%202012-2.pdf

There's definitely a dip there around 1Khz, and corresponding blip in the free air impedance trace.

On a typical 8-9" baffle, baffle step may cause peaking around 1Khz and thus the dip may flatten out. Whether it's perfectly flat depends on the resolution of one's measurements. With the usual far field/windowed measurement of even 10ms, that's only a resolution of 100Hz. So between 800-1600Hz, one may not see the dip.

For instance, to get resolution of 33Hz, doesn't one need a time window of 30ms (1/f)? If my math is correct, that's a mighty high distance for the speaker and mic to be above the ground!
One really needs an anechoic chamber, or a Klippel Transfer Function Measurement module with Klippel’s In-Situ Compensation module to generate anechoic-quality measurements in the far-field with a resolution small enough to detect high Q dips and peaks.

In summary "smooth as a baby's bottom" may be due to the gated far-field measurements and 1/12th octave smoothing employed.

[JonMarsh]

Exactly!!!

I hadn't had my coffee yet and wasn't prepared to go into so much detail- but for this sort of application I'm usually using a 50msec window with Half Hamming filtering so I can't avoid some room bounces, but with care (and a high ceiling and large room) the effects can be minimized on a 1 meter and 1/2 meter measurements. But this is also why you see me normally using minimum filtering, 1/48th octave.

I don't post all the tests I do, but this involves near field and far field and long basic gates that are then tuned to look at the range I want to examine and plot. I don't mean to pick on Troels or anyone else, but when I see baby smooth plots like that, I have a tendency to discount their usefulness to me. Hence, the "In God We trust, all others we verify!" quip. Seriously, that was the outlook Anders and I had.

I'm retired (finally - (I'm a late learner/bloomer) and though there are many practical and logistical issues to deal with about relocating to a new state, getting a home construction project underway, cleaning out the last of my storage units in CA (UHaul is a disaster in CA- had to rent trailers up here) I'm slowing freeing up time for speakers and audio.

But as my former colleagues at work note (and I was lending a helping hand last week in preparation for the virtual APEC conference this week) I have a lot of deferred gratification on tap...