2021-2023 Isiris Update

[JonMarsh]

It wasn't me! Honest! I didn't do it!!

But I did buy the next to last adjustable height Husky work table in all of the Treasure Valley Home Depots this morning... it's a winner for what it cost in build quality and functionality!

[Juhazi]

Excuse me, but I've been wondering what step response of Isiris and Update version is? Typically we need setback of tweeter to get good phase match at xo to midrange...

[JonMarsh]

Excuse me, but I've been wondering what step response of Isiris and Update version is? Typically we need setback of tweeter to get good phase match at xo to midrange...

[ATTACH=CONFIG]31577[/ATTACH]

MEB prepared a nice drawing back in 2012, explaining this as part of the original Evil Three Way Design Study, which became the Isiris project- he's also done it for the Osiris.


Due to how tall this system is, it's desirable to literally create a focused array "aimed" at the seated listening position. The measurements done on this page reflect that concept, as regards microphone location, about 38-40" above ground (well, a six foot tall person in a not too short chair...)






So, no, these are not designed to be listened to standing up at some distance- BUT, due to the driver positioning, and the crossover alignment currently used, standing works surprisingly well.* (3rd order -6dB at crossover point- think of it as being like a 3rd order Linkwitz-Riley, though no such thing exists; however, some other folks are finally recognizing the utility of this concept which I've employed since the original Module MT and NatalieP, and you may see it in a few design tools).

And, though some earlier design proposals for the Isiris have used the Duelund concept, this current build is based on the 3rd order -6dB, as is the Wavecor Ardent.

Why this alignment?

• This deveopment build of the Isiris is focused on improving the "bang for the buck" in every aspect possible, which means hopefully improving performance while reducing cost
• This 3rd order quasi LR crossover winds up having pretty reasonable cost performance tradeoffs, as regards the component values and number of components.
• Compared with a traditional LR4, it has a wider performance window in the crossover overlap region than the LR4- in fact, the LR4 has such a narrow window that it's often recommended to peak the on axis response somewhat to compensate for the power response dip in the crossover region.
• It's my personal opinion and experience that this 3rd order configuration is an optimal set of "compromises" compared with the LR2 (wide overlap and slow roll off rates) and the LR4.
• YMMV, of course- explore and investigate. I've done a lot of both. There's nothing that beats first hand experience and measurements and listening.

I have some asymmetrical crossover concepts for minimum phase behavior I'd like to try in a future project, but it's not amenable to this enclosure configuration. But for a future design targeting the Dutch & Dutch 8C, it may work out well. But that's another story, and another day.

[tktran]

Hi Jon,

Your in-room response looks really nice, it seems about +/-2dB between 300Hz and 10Khz

It's great to read along this build/update and learn about your processes along the way. To hear and see how designs decisions are made around a specific objective eg. indoor space, eg. 2.5m listening distance etc.
Too often we see 1m measurements, or, the latest spec the spinorama/CTA 2034A, which everyone over at ASR seems to think that that's all they need to know, and thus guide them towards a purchase decision.
But a 1m anechoic measurement doesn't take into account the finer details IMHO, like how a speaker actually measures in-room, or how the drivers' amplitude and phase responses overlap at the crossover points at a typical listening position like 2.5m.

When I've taken an in-room measurement, ungated and smoothed, I get a healthy dose of reality, but I do find it can challenge me.
Is this what I'm hea

[JonMarsh]

well, there are more updates coming, parts coming in tomorrow and Thursday! I'm expecting to be able to tighten things up a bit further, perhaps adjust the voicing a little more.

We'll see.

I hope that explaining what I do and why as I go along is helpful for others, even if they're doing a rather different kind of system, but say, another three way. Sort of like having a checklist, I'm working down mine for this one!

[Finleyville]

I hope that explaining what I do and why as I go along is helpful for others, even if they're doing a rather different kind of system, but say, another three way. Sort of like having a checklist, I'm working down mine for this one!

Your annotated thought processes have been very helpful with this build so far. Please keep them up if you wish.

What are your thoughts on the only "blip" in the response around 150Hz? What could be causing this? Does this anomaly even need to be addressed?

I have spent many hours listening to another one of your 3rd order crossover designs, NatalieP's, in WKHanna's home. How would you describe this design's sound signature compared to that one?

With how tall this design will be, will I have any great audio response problems placing these in my basement with only a 7' 8" ceiling?


Your efforts to increase the performance while reducing the overall build cost from the original design are to be applauded. However, I have only 1 more speaker build left in me. So I will not mind building an ultimate version of this design if you think of any "questionable" upgrades to your final published design. In other words, I am looking for my forever speaker pair without having to shell out a literal small fortune for my bucket list ones.


Oh, and those adjustable work tables are nice! I almost bought one myself, then remembered I had no room for it! lol

[Efalegalo]

What are your thoughts on the only "blip" in the response around 150Hz? What could be causing this? Does this anomaly even need to be addressed?

I wonder if this is just floor bounce cancellation?

[JonMarsh]

Your annotated thought processes have been very helpful with this build so far. Please keep them up if you wish.

What are your thoughts on the only "blip" in the response around 150Hz? What could be causing this? Does this anomaly even need to be addressed?

I have spent many hours listening to another one of your 3rd order crossover designs, NatalieP's, in WKHanna's home. How would you describe this design's sound signature compared to that one?

With how tall this design will be, will I have any great audio response problems placing these in my basement with only a 7' 8" ceiling?


Your efforts to increase the performance while reducing the overall build cost from the original design are to be applauded. However, I have only 1 more speaker build left in me. So I will not mind building an ultimate version of this design if you think of any "questionable" upgrades to your final published design. In other words, I am looking for my forever speaker pair without having to shell out a literal small fortune for my bucket list ones.


Oh, and those adjustable work tables are nice! I almost bought one myself, then remembered I had no room for it! lol

Yeah, I hear you about the space thing, but I have a secondary work room with no tables, and I moved the Costco portable folding table from my main workroom to the secondary room- voila! space available!

• Regarding the dip at 150Hz, this is a function of the speaker location for testing and room acoustic- these aren't setup in my normal manner (Cardas room setup) for these tests, as I was not worried about room modes.
• Below is the response previously measured for the woofer section alone, no crossover, relatively near field.
• I am a bit rabid (by HTG standards) about room setup and acoustic control- have been since the mid to late 70's. No, not 1870's! Just seems that way...
• If you know of a 12" woofer that will do 86-87 dB and a pair measure like this in 110 liters sealed, please, PLEASE! let me know what it is.

• I think the reception to the NatalieP over the years kind of speaks for itself about the succsss of the design decisions; the "special" 3rd order crossover is an integral part of that, and works very well in an MTM configuraiton. Of course, if you do your homework about conventional MTM design at that time and earlier, even the 3rd order Butterworth was recommended by many "authorities" as the best choice for an MTM. My main gripe with the Butterworth class of filters, is that though they summ electrically just fine, they actually operate in what is called phase quadrature at the crossover region, and so produce lobes above or below the crossover point- this is true for all Butterworth crossovers, and is occasionally discused in the literature, but I discovered it empircally in the '70s- yes, 1970s' not 1870's!
• I don't have measurements of the NatalieP with the tools I've been using since about 2014 (Fuzzmeasure, but same ACO Pacific mics I've been using much longer), but what I do have shows the same kind of flat on axis repsonse and very smooth and consistent off axis response, due to both the RS180 characteristic and the relatively low crossover point (1800Hz) which the RS28a tweeter (a product made by Usher) could readily support. The newer Dayton products just aren't in the same performance ballpark, sadly.
• I have some ideas for a NatalieP Ultra, and drivers and baffle simulations and what not, but haven't had time yet to get to hardware- that will wait until moving into the new house under contract.
• For the NatalieP Ultra, the 6mm Xmax PuriFi 6.5" woofer is the leading contender, for all the obvious reasons.
• The 6.5MF PuriFi is similar, but further optimized in many ways for the midrange task- and in fact, there is infomraiton I have about it I'm not allowed to share yet. But that's another story.

• I appreciate your inputs about costs and wanting to do a sort of "final solution"
• I looked at a LOT of drivers before picking these, and while cost did have some bearing, I was still looking closely at performance. if there were two drivers with nearly identical performance, but one had a clear cost advantage, that was the choice. In the case of the midrange, there was no other contender. I really couldn't find a good contender for the woofers also- my choices for three ways has usually been toward parts that straddle the design divide between conventional subwoofers and woofers- the Aurasound I original used did this; the woofers for the Wavecor Ardent did this, and these RSS315HFA do this rather well, too. If there was a XBL2 12" woofer with smooth flat reposes to beyond 1kHz and monotonically decreasing distortion to 1kHz, reference sensitivity of 86 dB (low frequency) I'd likely jump on that regardless of the price, but so far I haven't found one- the closest would be the CSS SDX12, which falls a bit short in several areas. And it's massive and pretty expensive, too, almost double the RSS315HFA-8.
• At this point, I belive the PuriFi is the best midrange solution I could find meeting my performance expectations, and measurements to date, plus my initial quick listening check, have only confirmed that belief. Sorry Accuton. The newer cell drivers are interesting, as are the C168 variants, but frankly, they don't measure as well.
• Now, on the tweeter side, I also have a pair of BlieSMa Beryillium dome tweeters, (in a currently inaccessible location in storage without a huge effort) the T34B, and they do cost more, and have a breakup mode at a higher frequency. But so far, with several recent experiments, including the Tymphany DA25 and now the T34A, it seems a properly designed breakup filter can nearly completely suppress any resonance amplificaiton of distortion products lower down, which is how we actually "hear" tweeter breakup.
• This is borne out in the initial distortion checks I've done for the tweeter- I'm posting a cropped view of that; note I'm replacing inductors in the midrange crossover and woofer crossovers which appear to be contributing to what I call a distortion "noise floor" at about -50 dB with equal HD2 and HD3.
• My point being, if the final measured response (SPL and distortion) are identical with the crossover, is their any point in buying the Be version? Ask Sven about his experiences listening to the speakers with the DA25 and similar crossover approach, versus some other well regarded tweeters.

• With your ceiling configutation, I would STRONGLY recommend that you consider installing acoustic absorbtion panels in the ceiling to reduce strong early reflections that will cause comb filter effects at the listening position. Where? Get a friend to hold a mirror up to the ceiling, and while you're seated in your normal listehning position, have him move the mirror (that is flat, flush with and parallel to the ceiling) until you can see the midrange and tweeter drivers in the mirror.
• This brings up another point that is rarely discussed on HTG- it's probabably going to get you better results to have a $2K speaker system and $500 of room treatements, than to have a $3k speaker system setup with less care.
• And while I'm mentioning some of these peripheral topics, don't forget the impact associated gear in your system has- everyone that has built Wavecor Ardents has commented on how "responsive/revealing" they are of other system updates- source components and amplification. This is something I even hear freuqently from NatlieP owners.
• I could almost go so far to say, if you always listen at relatively modest SPL levels, a set of well built NatalieP's with good quality crossover components (especially caps) might benefit from a much larger investment in electronics than you might expect.
• (and yeah, I've been pondering the idea of a quick NatalieP-A update with switching to the DA25 tweeters but with RS180s and new network- could be a real low budget killer)
• Lets remember my expectations tend toward the outsized at times- I'm an ex professional/performing musician (worked my way through school that way- first paying gig when I was 14)(I've been recreating my old setup with some upgrades, which means I am back to having a full Marshall stack and some smaller setups but with several amplification chains I can use with it, and Vintage 30's in the bottom cabinet, and Greenbacks in the top cabinet), I have designed some small recording studio setups, including the monitors and patch bays, and I had a spell in the late 70's running live concert sound for a local Boulder company.
• And I'm a classically trained musician, and have done live classical recording, too, so my expectations are a little more extreme than the average bear...
• my idea of a successful design is like what happened with the X1 SLAMM Klones- you could play laid back music like Michal Hedges at louder and louder levels, but it didn't "seem louder", just more of it there, until you got out the SPL meter and realized we were well over 100dB- for just one test cabinet. And then you noticed a teensy bit of ringing in the ears... that is taking it too far, but that's what a successful design allows you to do. And for that level of SPL and clarity, speakers like the Ardents and up, size and complexity wise, are needed.
• So for me, a solid "budget" system for the NatP's or the Wavecor Ardent or the Isiris is a Denafrips Terminator DAC, a Benchmark LA4 line preamp (and this was JA's testing summary at Stereophile: Benchmark's LA4 is the widest-bandwidth, widest-dynamic-range, lowest-noise, lowest-distortion preamplifier I have encountered.—John Atkinson )(I use one in my acoustic test system), and a Cambridge Audio Edge W. (the Cambridge is designed as a very high current amplifier, using their Class SD, (analog) and with just one custom cable can be converted into a high performance mono block- you just need a reversing phase patch through cable- tested it myself)- OK, so you might substitute a Hypex or PuriFi Class D at lower cost, and lower power. I have an Audio Precision APx555, and have tested all these items; some are posted on our Bench forum section.
• And yeah, I do have stuff beyond this "budget configuration" in storage- Halcro preamp and a pair of Halcro mono blocks, a TotalDAC D2 balanced (completely electrically balanced), Brainstorm DCD-8 re-clocker, rubidium reference oscillator, Sonore Signature Rendu, etc. But the budget setup can come pretty close, unless you're looking for the last word in imaging and definition- but if you haven't done your acoustics homework, you're better off investing on that side first.

[dar47]

I concur these builds benefit from a good front end. I recently lost a power supply in my dual mono Hypex NC500 build and thought I would try a new amp. Put in the NAD C298 with the new Purify module in and found it to be a vary nice replacement for the Hypex build at a reasonable price.

If I am not reading into to much are you considering a Purify Nat P design after the budget build?

[JonMarsh]

I concur these builds benefit from a good front end. I recently lost a power supply in my dual mono Hypex NC500 build and thought I would try a new amp. Put in the NAD C298 with the new Purify module in and found it to be a vary nice replacement for the Hypex build at a reasonable price.

If I am not reading into to much are you considering a Purify Nat P design after the budget build?

Yeah, I have PuriFi drivers, and I've tested a setup with Visaton waveguide and T34A tweeters, and done a baffle analysis using edge; I would likely switch to the 6mm Xmax parts from the long throw ones. To really optimize the wavefront launch it wound up having a bit different baffle size and shape than you might expect...


This is what came out to be the most optimum configuration using Edge.

[tktran]

If you know of a 12" woofer that will do 86-87 dB and a pair measure like this in 110 liters sealed, please, PLEASE! let me know what it is

I wonder whether this was a rhetorical question, but I did simulate the RSS265HF-8 in the same sized cab, and it's about 1.5dB less sensitive, which is probably a deal breaker for some. 86-87dB isn't all that sensitive.

What other woofers did you consider?

[Bear]

I wonder whether this was a rhetorical question, but I did simulate the RSS265HF-8 in the same sized cab, and it's about 1.5dB less sensitive, which is probably a deal breaker for some. 86-87dB isn't all that sensitive.

What other woofers did you consider?

I'd expect that the main interest for the RSS265 is for cabinet size reduction. The per-unit pricing differences are not quite a rounding error at this level of price and performance.

[Scottg]

So far, things seem to be on track for bringing this in at about the same component BOM cost as the Wavecor Ardent, about $4k. That's a huge reduction form the $10K of the original version, and I think this one will have some performance advantages.

Cross fingers...

You might consider a 20%/80% value with *Auricap XO/Clarity Cap CSA for "everything" (in signal) but the tweeter. ..and for the tweeter signal cap: MiFlex KPCU-03.

If you've got a tweeter signal "padding" resistor I'd strongly suggest Dueland resistors (..I think it's the most critical element in a passive crossover, even more than the lead-in tweeter cap).

*Auricap XO comes up on sale as well - increasing the value proposition, the Dueland resistors are also often on sale.

[JonMarsh]

You might consider a 20%/80% value with *Auricap XO/Clarity Cap CSA for "everything" (in signal) but the tweeter. ..and for the tweeter signal cap: MiFlex KPCU-03.

If you've got a tweeter signal "padding" resistor I'd strongly suggest Dueland resistors (..I think it's the most critical element in a passive crossover, even more than the lead-in tweeter cap).

*Auricap XO comes up on sale as well - increasing the value proposition, the Dueland resistors are also often on sale.

Thanks for the suggestions, Scott! Balancing out the cap choices is one of the most critical things- especially since the Clarity MR caps are no longer available.

That's an interesting note on your part regarding the Duelund resistors- I've usually been using the top of the line Mundorf, with what seems to be satisfactory results- but am experimenting with some other things. Also looking closely at inductor construction and performance again- I'm on my 3rd rebuild for the midrange crossover at this point. And there are some "special" parts on the way for the woofer crossovers from Parts Connexion.

At this point the tweeter crossover is all Jantzen Alumen film and foil- the dissipation factor is extremely low (0.0001) and internal resonance seems non-existent. My current favorite for tweeters, and for notch filters, since you can tune the Q more accurately with this low DF cap.

I'll look into tghe MiFlex KPCU-03, though.

You know what they say about crossover development- it's not over until it's over. Or is that the baseball season? I get them mixed up sometimes...

[JonMarsh]

OK, checked on the MiFlex, and they are in stock at both Parts Connexion and HiFi Collective (in the UK- use them a lot for some things).

And of course, the price is approaching the level of, "If you have to ask, you can't afford them!" - but only approaching- not like Duelund caps, for example.

well, this gives me an "upgrade" path to consider next year, which is when I'll have all the fancy-schmancy equipment setup again (networked music sever on isolated network, and rubidium clocked digital signal path).





This diagram does not include the Brainstorm DCD8 clock system or rubidium oscillator after the Alpha USB, or the following DAC, or the Sonore Signature Rendu I now have to replace the micro Rendu.











With this stuff up and running, and other retirement topics settled, taking a 2nd look at the tweeter crossover, and also building one for the T34B's I have tucked away in storage might be just the ticket!

[Scottg]

The Alumen is the cap that Tony Gee recommends for a 20% value fill-in with the Clarity CSA (80%) for the net cap value. Still pricey though for only a moderate increase in subjective performance.

(partial) Pricing example:

Ex. from post #7 (..though I realize the design is still evolving)

Series/Signal components (targeting the pricey components): using current partsconnexion sale prices..

Auricap XO's are of the 200Vdc variety unless otherwise noted.

Tweeter: 6.2 uf = .1uf, 2.2uf, 3.9uf all Miflex KPCU-03 ($203)
3 ohm R10 = Duelund Resistor 3R Ohm 10W Graphite-CAST Series ($33)
5.1 uf x 2 = .22 uf Miflex KPCU-03 & Auricap XO Capacitor 10uF 400Vdc XO Series ($49)
$285 subtotal

Less expensive Alternative: change 6.2uf = 2.2 Miflex KPCU-03 4uf Auricap XO ($106)
$188 alternative subtotal

Though not in series, maybe spend a bit more for a Jantzen waxed multi-strand inductor for that 1st parallel inductor..


Midrange: 33uf + 15uf + 15uf = 20 uf Auricap XO + 33uf Clarity CSA ($89)
Mundorf Resistors are fine
Copper Foil Inductor from *Erse/Goertz/Jantzen/Mundorf
56uf + 15uf + 1.2uf = .22 uf Auricap XO 400Vdc + 20uf Auricap XO + 22uf Clarity Cap CSA + 15 Clarity Cap CSA + 15 Clarity Cap CSA ($123)
Copper Foil Inductor from *Erse/Goertz/Jantzen/Mundorf
$212 subtotal


Woofer: Copper Foil Inductor from *Erse/Goertz/Jantzen/Mundorf
Copper Foil Inductor from *Erse/Goertz/Jantzen/Mundorf
10uf (in parallel with inductor) = Auricap XO 10uf ($40)
$40 subtotal


..and then going back in and pricing-out those series inductors.

Everything else = cheap! (..or at least as inexpensive as possible for good quality.)


*and Erse Foil inductors are usually less expensive than the others. If I were to spend more, it would only be for the Jantzen waxed foil inductors.

[JonMarsh]

This is a good example of how to approach this and balance cost/performance considerations.

And yeah, it’s been evolving a fair amount, as I get back into the swing of things- Steve has been wondering why I’ve been doing 3 parts orders a week! Did another one today. But that’s how this phase always goes- measure, ponder, think, recalculate and simulate in VituixCAD, update a trial or build a new board completely- I’m just modifying the midrange board again (ripped about half the parts off) but I’ve ordered parts for my 3rd woofer board build.

Still on the first tweeter board build, it’s all Alumen, but I’m putting a post it note up on the calendar for next April or May to look into a new build with KPUCU-03. I’ll probably go whole hog, not doing fractional groupings. Might as well see what it has to offer, right? And then the next logical thing would be to do a ratio’d build and compare. We’ll just have to see if that comes to pass!

One problem I’ve noticed with some of these parts (particularly the more exotic inductors) is supply chain issues…

[Scottg]

..Steve has been wondering why I’ve been doing 3 parts orders a week!

8O That's a LOT!

-sadly, supply chain issues are a problem over a lot of industries. The "..vid" certainly screwed-up the economy, and our gov't has substantially exacerbated it by basically (and continuing to) paying people to stay home = Inflation (ex. my typical lunch's eating-out have gone up by about 20% since 2020 because they can't get enough staff and if they do the turn-over is really high, and other sectors are just taking advantage of it and raising prices like the cost of Beef or Timber - from collusion). My guess is that more problems will occur globally with China's recent policy's (maybe not as much with respect to their direct actions, but rather with respect to people/business/money pulling-out of China and moving where it's less oppressive).

Makes me think of the derived Chinese proverb/curse: "may you live in interesting times". I think it's going to get more "interesting" over the next decade or so (..and given bitcoin's absurdity, a fair number of others think so as well).


..anyway,

If you proceed with the full-signal Miflex copper cap's, you might as well go all the way "quasi-Dueland" and spring for the wax Jantzen Inductors along with of course the Cast Dueland resistors to "complete the package". (..maybe something like Panasonic EZEP for non-critical cap's in whatever Vdc.)

[JonMarsh]

I COMPLETELY agree about the "interesting times" quip, and have used that frequently to complain now and then about being my "expected retirement experience" being stolen.... it's changed cost structures, access to experiences, and all kinds of stuff in ways I never would have guessed could happen 3 years ago!

A few of my former colleagues and I read WAY too much news to be good for out mental health and discuss it a fair amount via messaging; already got my early start today before 6 AM with my still employed colleague from Boston; we usually sign off around 9 PM mountain time from our intermittent bursts of interchanges. These days I foresee many issues likely to result in "interesting times" based on developing trends. I still think I made a reasonable relocation choice though, considering the near term future.

Yeah, a build like you describe would be fun, and if housing hadn't jumped so much in the last two years I'd probably not bat an eyelash about doing it- most of available on hand resources have been set aside for the new house construction that should be ready to occupy in March. And that was started in June, so it gives you an idea of lead-times. Meanwhile, I'm living in our second rental here that was bought as an investment property by a CA couple last August (had been under contract since early this year) (who plan to move into it in a few years when they retire, according to the landscaping guy)(we moved here in September, after the owners of the place we were living in decided they wanted to sell it) and it's we're in the same basic floor plan as the place we're buying, minus some options we went for. But over the last two years housing has gone up about 30% out here in Idaho... No one expects the Spanish Inquisition or price increases like that (I'd been watching this market about 10 years).

Ah well... health is good, especially for my age (I've outlived my parents and all members of FOO) (I do a lot of hiking and walking, and have rehabbed well from my two hip surgeries- even my doctors have commented on that). And I've done all my moving stuff, including heavy appliances, DIY- good exercise. Of course, I believe in working smart, not just hard, and have a nice collection of appliance and moving dollies.

Have a plan, and work the plan.

[JonMarsh]

I wonder whether this was a rhetorical question, but I did simulate the RSS265HF-8 in the same sized cab, and it's about 1.5dB less sensitive, which is probably a deal breaker for some. 86-87dB isn't all that sensitive.

What other woofers did you consider?

Well, the appropriate question would be, what other 12" woofers did I not consider?




You see, based on the original LF system concept (derived from the Etons used in the Avalon Isis) and the previously used woofers/subwoofers, I had some rather specific targets, given that I had no desire to build another set of LF cabinets.

As mentioned for the original Isiris, and the Aurasound NS12-513A used in them, I kind of have this thing for woofers that straddle the divide between conventional woofers built for gorilla coffins and medium throw subwoofers.

What does that mean for a 12" woofer, in practical terms?

• Sensitivity target for an 8 ohm driver of 86-87 dB
• MMS in the range of 150g to 200g. This has a strong bearing on the T/S parameters achievable and the final enclosure size and Fb.
• Xmax of 10mm minimum for underhung linear motor, or 12-14mm for overhung motor, so that before BL droop there's soemething like a fairly linear 6mm range.
• Qes and Qms resulting in a Qts in the range of 0.38 to 0.45
• Le between say, 0.8 and 2 mH- no higher
• Sufficient copper in the gap to keep midrange distortion down, reduce inductivity modulation, and have monotonically decreasing distortion up to 1kHz (I.E., for constant SPL drive)
• Fs of about 20Hz, and Fb for a pair in a 110L sealed enclosure of about 30Hz. Would consider a reflex alignment, if similar Fb tuning was feasible given driver parameters.
• Fairly smooth response to at least 1kHz and first breakup mode at 2kHz, to minimize resonance amplification of distortion products.

I suppose while I'm at it, I might as well toss in a winning lottery ticket, too. But then, that would have eliminated the RSS315HFA-8. But that quip does highlight the degree to which that combination of characteristics is quite hard to find. I looked at every 12" woofer/subwoofer driver on PE, Madisound, and Solen. Many, if not most of them, are far more expensive than the RSS315HFA-8. BTW, there are a couple of RSS315HFA-8 in stock at Solen, the last time I looked.

And just for fun, here is the blurb for the RSS315HFA-8 and you might guess why it tipped me off they had the right things in mind...

The Dayton Audio Reference Series subwoofers take the quality and performance of the much acclaimed RS line into the subwoofer realm. As with the rest of the Reference Series, the main focus of the RSS315HFA-8 12″ driver is ultra-low-distortion. Second, third, and intermodulation distortion are kept extremely low by utilizing three magnetic short-circuit paths within the motor structure. A durable aramid fiber spider and extensive venting throughout provide quiet and uniform excursion. The black anodized cone, rubber surround, and custom basket give this driver a unique, high-end look. Choose this High Fidelity (HF) version for home speaker systems when your priority is clean, low bass and you are not restricted by cabinet size.

Now, the irony for that copy is in the last line, but then, they are comparing to subwoofer drivers. When you look at typical 12" woofers, you see MMS in the range of 100g or less, Qts in the range of 0.27-0.32, sensitivity often around 90-92 dB, and you need a gorilla coffin for a single driver to get good LF extension, due to Hoffmins Iron law and how the parameters interact- low MMS and high BL product gives you high sensitivity and low Qts, and you can't get bass extension within a sealed system- you must have ported or PR approach, and relatively large box. Or, shall I say, LARGER box.

[Finleyville]

• I appreciate your inputs about costs and wanting to do a sort of "final solution"
• I looked at a LOT of drivers before picking these, and while cost did have some bearing, I was still looking closely at performance.

Yeah, my "Final Solution" quote was not about the drivers selected. The performance graphs you provided looked great to me for their intended use. I was more interested in the best upgrades within the crossover circuit. The 80%/20% rule sounds like a great starting place. I have done zero experimentation with crossover parts/upgrades and would need my hand held moving down that rabbit hole.

Talking about crossover designs, whether you believe in it or not, do you foresee any problems if I decided to separate the woofer section from the T/M section so I could bi-wire the cabinets? In my limited testing, I have always found some small increase in performance going this route.


As far as perhaps saving some money in the build and instead placing those funds in better equipment, my current front-end consists of items that would need a huge jump in money to achieve any worthwhile improvements. An extra $1000 for an "ultimate" crossover would not be better spent anywhere else. Well, except for room treatments. But those improvements I do not mind stretching out over some time anyway.

[JonMarsh]

With the modular external crossover board approach I use for this project, bi-wiring is very straight forward and easy. And it's a good way to lighten up the load on what on might term intermediate level amplifiers. This is especially true if the amplifiers you use don't have massive capacitor banks.

What I found back in the 70's, experimenting with Dynaco Stereo 400 modules, is that for them to "sound" the best, they needed about 100,000 uF of storage in the power supply (for electrolytics, with low ESR) and say at least 100uF of film capacitors to cover the top end. These were experiments I was doing when I was a co-owner in a high end shop, The Boulder Sound Gallery. The commercial electronics we sold included Mark Levinson, Stax (class A models) and Luxman.

It drove my partner and the sales people nuts when I brought in my black box, and we hooked up the Ampex ATR100 with a Mark Levinson preamp and our best B&W speakers and played high dynamic range material (our own live recordings, plus carefully made needle drops from virgin vinyl with a Rabco tone arm and Denon moving coil cartridge. At louder playback levels where the big 200W/channel Luxman and Stax DA150 sounded strained, the modified Dynaco (original power transformer, just one) was effortless and serene.

I'm sure this is something Frank Van Alstine discovered later and was the basis for his modified Dyna builds.

But splitting the load up by frequency bands for amplifiers with more mundane power supplies should accomplish much the same thing.

BTW, some of the HTG experimenters have determined that this phenomena also holds for Class D amplifiers such as the Hypex modules.

[tktran]

Talking about crossover designs, whether you believe in it or not, do you foresee any problems if I decided to separate the woofer section from the T/M section so I could bi-wire the cabinets? In my limited testing, I have always found some small increase in performance going this route.

If I may contribute, but try not to derail this great thread by Jon, I will say that in my experience separating the woofer section from the M/T has pros and cons. A big positive is the lighter cabinets, which makes much more manageable to carry. Though in this case, probably not much lighter, because of the dual RSS315HF8A are heavy.

The con is that you'd have to be careful with changes in the cabinets vibration modes. I use sorbothane feet, but I haven't done extensive experiments to measure any differences. They sound great to me, and makes my back feel better.

[JonMarsh]

I’m guessing your MTM section is some type of cardioid mid bass/mid range construction; that, or you just put on those recessed grill sections to drive people like me crazy trying to figure out what’s going on!

I have something faintly similar in mind, but with some Dutch & Dutch 8C ideas thrown in for good measure.

[tktran]

This is John Kreskovsky's NaO II design; which is a dipole MTM, with a selectable cardioid or sealed woofer system.


When built it, I turned it a 2 box system for ease of transport, and opted for the sealed box only design. Everything is built to John's spec, except the crazy wings that I put on because I wasn't ready for the open baffle look, at that time.

Apart from the separate woofer bin, the other thing that NaO II has is a woofer system that has a LP @LR4 120Hz, which is same spec as the LFE channel in home theatre. This allowed me to build separate (sub)woofer cabinets, and move the woofers around for smoother in-room bass response.

So there's a total of 2 woofers per speaker (not pictured). In the picture you see only 1 sealed woofer for the MTM dipole. The other woofer is placed not below the mid panel but elsewhere in the room.

The change is not sanctioned by JohnK, but nevertheless I've learnt a lot during the process of implementing the changes, the tradeoffs, pros/cons etc.

[JonMarsh]

OK, that brings back memories of some of JohnK's designs, including the NaO, though I think I'm not familiar with the NaO II.

Now, what's interesting about what you're doing, is that it is somewhat akin to what the Dutch & Dutch 8C does, with a cardioid upper bass/midrange, and a pressure zone loaded woofer system (rear firing; has some flexibility about how far from the rear wall), but then the 8C uses a tweeter with waveguide on the top end.

However, they do a lot of other "modern things" including wireless data transmission, syncing the stereo channels for two speakers wirelessly, DSP and built in amps.

From my perspective, that gives a lot of places were things could go wrong, or not quite as "right" as you'd like. And they've had issues with production units.

But acoustically it's an interesting design to build off of.

https://dutchdutch.com

https://hifiplus.com/articles/dutch-dutch-8c-stand-mount-loudspeaker/

https://www.audiosciencereview.com/forum/index.php?threads/soundstage’s-review-of-dutch-dutch’s-8c.7308/page-2

[tktran]

Yes, I'm well aware of the 8C. I actually auditioned them in November at a high-end hi-fi dealer. They are ok, but have all of the limitations you'd expect of a small speaker. Erin Hardison has measured them with his Klippel Near Field Scanner:

https://www.erinsaudiocorner.com/loudspeakers/dutch_dutch_8c/

They were no match for the Revel Performa F338Be, which uses a triplet of 8" woofers, a 5" dedicated midrange and 1" tweeter in a ovoid waveguide, which I also auditioned on the same day. The KEF Blade 2 was in the same room as the F338Be, which made for an interesting comparison, but the dealer was careful not to put the 8C in the same room as the 338Be/Blade 2.


But this is both a larger and more expensive speaker, so it's not a fair contest. But it did highlight, to me, the merits of using three 8" woofers in a larger box, unlike 8C's 2x8inch "subwoofers" crammed into a small box.

This idea that distortion doesn't matter, predominantly driven by one academic in his quest to gain wider acceptance for his own distortion metric, IMHO, is a shame, and in my view counterproductive. Yes THD is not perfect, but let's not throw out the baby with the bathwater. Non-linear distortion is important, and this view is shared by other engineers and people trained in the physical and biological sciences, but also people who may not be formally trained but have have done a LOT of testing and empirical "research", even though they may not be published in the AES or other journals for peer review. Many people are publishing on their own websites, without any need for publication, getting grant approvals, or otherwise furthering their career. eg. our dear John K, but also the late SL, John Krutke, Mark Krawiec, Monte Kay etc

As Monte says, how do you know low non-linear distortion doesn't matter, if you've never truly heard it?
Here he uses TWENTY FOUR Dayton Audio RSS390HF woofers:

http://www.mfk-projects.com/Home_Theatre/theatre_woofer.html

And furthermore, how can you say that distortion doesn't matter, when all you've measured is THD (which is mainly influenced by HD2/HD3).

As recently as September 2021, Sean Olive said that his studies is "limited to one room" and "limited to the accuracy of the subjective measurements" and that more research needs to be done. The "model doesn't include non-linear distortion and directivity" and "doesn't include non-linear distortion (again), spatial attributes or max SPL" . Sean, you are also my audio hero. This is a true scientist. Research begets more research. So what we (thought we) knew in 1990 in the era of Consumer Reports, is not the same as what we know now in 2020.

His comments are here, again with Erin, this time on Youtube:

https://www.youtube.com/watch?v=IEtYH03pfOI&t=4014s

I too, share his frustration, and Sean talks about how easy it is to design a small speaker that sounds ok at low levels, but then drops out above a certain SPL. So just because a $100 DSP enabled small Bluetooth speaker can have essentially the same frequency response as a Revel Salon 2, there's nothing in the CTA2034 specification that can convince people to buy a $20,000 speaker that's the size of a fridge... other than it being nice piece of furniture.

I shudder to think of the day when we'll have have ten 1" "subwoofers" with 2" x-max crammed into a cabinet the size of a glass of milk, and consumers think that it is a good speaker...

Luckily we still have DIY...

[JonMarsh]

As one of my former colleagues would be wont to say, "I'm in violent agreement with you!"

This is the money quote, for me...

This idea that distortion doesn't matter, predominantly driven by one academic in his quest to gain wider acceptance for his own distortion metric, IMHO, is a shame, and in my view counterproductive. Yes THD is not perfect, but let's not throw out the baby with the bathwater. Non-linear distortion is important, and this view is shared other engineers or people trained in the physical and biological sciences, but also people who are have done a LOT of testing and "research", even though they may not be published in the AES or other journal for peer review. Many people are publishing on their own websites, without any need publications, getting grant approvals, or otherwise furthering their career. eg. our dear John K, but also SL, John Krutke, Mark Krawiec, Monte Kay etc

As Monte says, how do you know low non-linear distortion doesn't matter, if you've never truly heard it?
Here is uses TWENTY FOUR Dayton Audio RSS390 HF

http://www.mfk-projects.com/Home_The...re_woofer.html

And furthermore, how can you say that distortion doesn't matter, when all you've measured is THD (which mainly influenced by HD2/HD3).

I think the folks at PuriFi are on the right path, when they have such a focus on IM.

Even back in the 70's, if someone thought distortion didn't matter, we just took them into the high end room and played the Levinson HQD system. That was usually the equivalent of a transcendental experience back then... even though we "cheated" and used Janis subs instead of Hartley.

[JonMarsh]

I keep checking back every month on the PuriFi website, and now the 4 ohm Purifi Midrange is listed, data sheet available, and being offered in sample quantities on the web shop, with a lead time of 4-6 weeks.

So these are now real for normal folks, too.

:T


~Jon

[Finleyville]

With the modular external crossover board approach I use for this project, bi-wiring is very straight forward and easy. And it's a good way to lighten up the load on what on might term intermediate level amplifiers. This is especially true if the amplifiers you use don't have massive capacitor banks.

What I found back in the 70's, experimenting with Dynaco Stereo 400 modules, is that for them to "sound" the best, they needed about 100,000 uF of storage in the power supply (for electrolytics, with low ESR) and say at least 100uF of film capacitors to cover the top end.

I think I have that covered. Each one of my mono amps has 120,000 uF of storage. I am glad I can now use two pairs of binding posts into the back of the cabinet!

If I may contribute, but try not to derail this great thread by Jon, I will say that in my experience separating the woofer section from the M/T has pros and cons. A big positive is the lighter cabinets, which makes much more manageable to carry. Though in this case, probably not much lighter, because of the dual RSS315HF8A are heavy.

The con is that you'd have to be careful with changes in the cabinets vibration modes. I use sorbothane feet, but I haven't done extensive experiments to measure any differences. They sound great to me, and makes my back feel better.

You might have misunderstood me tktran. I was not looking to separate the top and bottom of the enclosure itself. I will build the cabinets as one piece if that is what the final design calls for. I was asking about the separation of the M/T xover section from the bass xover section only.

[tktran]

You might have misunderstood me tktran. I was not looking to separate the top and bottom of the enclosure itself. I will build the cabinets as one piece if that is what the final design calls for. I was asking about the separation of the M/T xover section from the bass xover section only.

Of well then that’s an easier answer. Of course. I think that’s a great idea.

I have not found a benefit of bi-wiring, but that doesn’t mean it doesn’t exist.

Perhaps its because I’ve gone straight for bi-amping.
I confess I’m a bit of an amp collector and have too many mono or stereo amps; enough for a Dolby Atmos system. All of my amps are solid state and 2 ohm stable, and all a minimum 100/200W WPC RMS into 8/4 ohms.

Second, it’s because I choose to design or buy speakers so that don’t drop too far below 4 ohms. 3 ohms is an absolute minimum. Asked me why and it’s got to do with releasing the magic smoke as a young person.

Of course bi or tri or quad binding posts also allows for conversion of a passive speaker to active, or from active to passive at a later date. So that’s nice.

[Bear]

I keep checking back every month on the PuriFi website, and now the 4 ohm Purifi Midrange is listed, data sheet available, and being offered in sample quantities on the web shop, with a lead time of 4-6 weeks.

My browser isn't cooperating with me to find these.

[dwk]

My browser isn't cooperating with me to find these.

https://purifi-audio.com/vare/ptt6-5m0z-nfa-01/ is what Jon is referring to, I think.

I spend far to long looking for the 4 inch midrange, rather than the 4 ohm

[JonMarsh]

OK, ladies and gentleman, here is another tidbit to tantalize you with on this Friday...

We often discuss the pros and cons of different crossover components, but let me say, experimenting with different parts on these test crossover builds has highlighted for me how you can MEASURE the difference between different component technologies- in this case, specifically inductors, and none of them cheap inductors.

These graphs are more evolutionary then reporting on the final outcome, but that I expect to have possibly by the end of next week, as after the last measuring and listening sessions (yesterday!) I finished up some further promising evolution in the design (which including pushing the midrange crossover frequency lower, which unfortunately does raise the cost of admission ) but we'll total up the damage later (though I have a preview for myself, having ordered parts for a new build of the woofer and midrange crossover this morning).

To summarize briefly, I initially I was trying out some very low DCR parts from Jantzen in both the woofer and midrange crossovers... and the distortion plots looked like this for the system:



Now, of course I had already run distortion measurements and SPL measurements on the raw drivers in these cabinets, so to say that this caught my eye is something of an understatement...


This is a similar measurement with a newer version of the midrange crossover, and even more expensive inductors in the woofer crossover:





The woofer roll off is still contributing to distortion in the midrange... but compare the levels at 1kHz or 2kHz.

I already had some new inductors for the woofer section on order, should be here next Tuesday, along with all the other new parts.

You know what they say, slow work takes time!

[tktran]

Very interesting. The official datasheets show that Purifi 6.5" midranges have HD2 and HD3 below -60 and -70dB respectively. So I was a bit surprised by your first distortion measurement, but wasn't sure what to make of it. I wasn't sure whether you were using the 4 or 8 ohm midrange, (not that it matters, according to the datasheet), perhaps you were using an early prototype, or perhaps I hadn't been paying attention and you'd been using a high drive level. eg. NOT 2.83V.

Your graphs only show distortion at dBr, not absolute SPL.
Jon, what is the output level? You hinted that the changes were due to more expensive inductors? What is the difference? Is it air core vs Steel laminate?

The other interesting thing is that the measurements now reveals the weak link- the PAIR of 12” woofers!

Anyway, you're very inspiring:
I'll have to have a try at dual 12" woofered 3 (or 4) way!

[JonMarsh]

Very interesting. The official datasheets show that Purifi 6.5" midranges have HD2 and HD3 below -60 and -70dB respectively. So I was a bit surprised by your first distortion measurement, but wasn't sure what to make of it. I wasn't sure whether you were using the 4 or 8 ohm midrange, (not that it matters, according to the datasheet), perhaps you were using an early prototype, or perhaps I hadn't been paying attention and you'd been using a high drive level. eg. NOT 2.83V.

Your graphs only show distortion at dBr, not absolute SPL.
Jon, what is the output level? You hinted that the changes were due to more expensive inductors? What is the difference? Is it air core vs Steel laminate?

The other interesting thing is that the measurements now reveals the weak link- the PAIR of 12” woofers!

Anyway, you're very inspiring:
I'll have to have a try at dual 12" woofered 3 (or 4) way!

Well, going back and reading that post, I am DEFINITELY perceiving the influence of lack of sleep last night (pondering design issues) and an additional cup of coffee over my normal morning allotment- a little less coherent and factual than I would prefer!

So, let me clarify a little- in the first measurement, there were Jantzen P-Core inductors in both the midrange and woofer sections- nice AWG13 parts that were not ludicrously expensive. Well, it just seems to me that they have some kind of hysteresis going on that contributes to a nicely matching HD2 and HD3 "noise floor" as regards distortion. I call it "noise floor" because this was not hard drive- good ole 2.83VRMS. For some drivers, I'd never notice that kind of level. It's not saturation in a conventional sense, but more a low level nonlinearity.

The second build has only air core inductors in the midrange, but P-Core inductors in the woofer. Pricey, and very low DCR, but obviously still with certain issues.

Earlier this week I ordered some Solen AWG12 Litz air core inductors. Should be here next Tuesday.

But I also did a revamp on the overall design, pushing the lower midrange crossover point down to 250Hz, and this AM starting early was all about finalizing the BOM and doing some more simulation checks and tweaks. If I say something like, "this will be the final version" then of course I'll be jinxing it.

But you know, it just might be.

[augerpro]

So the second plot is air core on the mid, and P-core on the woofer? Looks the mid distortion is fine, so maybe an air core on the woofer would also perform well?

[JonMarsh]

So the second plot is air core on the mid, and P-core on the woofer? Looks the mid distortion is fine, so maybe an air core on the woofer would also perform well?

That's the plan- moving to all air cores.

[tktran]

The on axis looks great; what are the measurements conditions to get 20Hz to 200Hz all in the one graph? Or do you merge or blend the near field with the far field?

how’s it looking at 30/60/90 degrees?

[JonMarsh]

The on axis looks great; what are the measurements conditions to get 20Hz to 200Hz all in the one graph? Or do you merge or blend the near field with the far field?

how’s it looking at 30/60/90 degrees?

That plot is a VituixCAD plot of the simulated response using FRD files from the 15 degree axis generated from Fuzzmeasure using Half-Hamming window measurements and only 1/24th octave smoothing (pretty minimal). These are done with the cabinet well out into the room. No blending near and far field to clean up LF- Hamming employs some variable filtering at the lower frequencies can reduce the information if over done. The data used for the simulations includes room reflections due to the measurement distance and long window time. All the drivers have been measured relatively near field to be sure of any driver issues before proceeding to the crossover development stage.

For example, this dual woofer measurement was made with the test cabinet well out into the room, and some additional acoustic draping so as to reduce LF reflections and room issues- wanted to just look at the summed behavior of the two RSS315HFA drivers.



The raw response of the woofers roll off a bit, but then there is baffle step from the cabinet. This is still a small factor even for a 250 Hz crossover, due to the size of the cabinets.

Fuzzmeasure (from RØDE Test) is an interesting and capable program-having six different operating mode configuration selected when starting a file-

• Studio
• Classic Speaker testing
• Live Sound
• Acoustics
• Advanced Acoustics
• Equipment testing

As far as I know it's still a Mac only program- it uses too many of the advanced API and program functions built into MacOS and not available in Windows.

[JonMarsh]

Testing again soon

OK, in spite of shipping delays with the holidays and vendors not stocking parts, I'm making some progress anyway-

Enough parts came in for the midrange crossover to build the complete new one, with the changes under way modifying the original board just didn't seem reasonable or practical- needed more room, what with the lower crossover point for the midrange.

First, a peak behind the current at the current state of the schematic - I think it's pretty close now.





The target response curve and the impedance curve seem pretty much up to snuff, so if the build matches it, we may be in good shape.








The test midrange board is built, and there's room for tweaks if needed.




I'm still missing inductors for the updated woofer crossover, but may be able to kluge something up.


The tweeter crossover hasn't changed any since the start. So, at the least I can test the midrange and tweeter sections together soon...

[Zvu]

Hi Jon,

When i measured distortion of the inductors, i've found that Janzen C coil doesn't influence the driver in such manner.

Also, ERSE seemed like a good option, measuring even a little better than C coils but costing a lot less. You could try that. Pure air core inductors are the best but, as you certainly are aware of, they get enormous for values over 5mH if you want low DCR.

[JonMarsh]

My plan now is to have a shoot out. I've used the high end Erse parts in the past, and have some on hand, plus the C-Coils.

I just put in an order with Solen for the big bad boys in air core; (AWG10) we'll see how long that takes.

And pricey they are, about as expensive as the woofers I'm using.

We'll have a performance shootout with just the inductors being changed. Could be interesting.

[Zvu]

......................

We'll have a performance shootout with just the inductors being changed. Could be interesting.

I always enjoy a good battle

[tktran]

Me too.

Jon, Yevgeniy of HifiCompass did a experiment with inductors too- you might be curious to check it out, if you haven’t already.

Warning: spoiler alert.

https://hificompass.com/en/projects/experiment/test-couple-iron-core-inductors