And regarding your replacement midrange drivers. Rob's replacement is a CS2.2 mid replacement and it will work. And none have failed in the field. However, be aware that its power handling spec is far less than the 3.5 mid. High power levels will damage it.
The good things about that driver include its physical drop-in aspect and its suitable performance. At moderate levels it will sound good.
|
JA - the upper crosspoint is the same 3kHz with very similar tweeter low rolloffs. The principal problem comes from the lower crosspoints: CS2.2 @ 800Hz, and the CS3.5 a full octive lower at 400Hz. The driver is asked to move a lot more air with longer excursions at lower frequencies. But, as I said, there have been no failures and it works well until we get an optimized solution.
|
Woody - the problem is that first order rolloffs engage the driver for 7+ octaves and Jim's designs control any anomalies via circuitry within that broad range. So, you can find many similar drivers, but re-engineered crossover circuits are required to replicate the original close-tolerance performance.
|
As a quirk of history, the lower midrange boost referenced by unsound caused a bit of weirdness that lasted the life of the company. In order for the woofer/midrange to retain their coherence both drivers had to be fed the same equalized/boosted signal. In other words the EQ couldn't simply boost the woofer because then the crossover slopes would be incorrect and not add properly. (Point of subtlety: the lower ranges of the tweeter also got boosted - it was still contributing below 200 Hz.) Note that the CS3 was bi-wireable with the woofer and midrange on one set of terminals and the tweeter on the other. The ramifications were quite complex depending on what different wire and/or amps, EQ or no EQ, etc. Jim didn't like cans of worms and decided that the dealer / customer / marketplace had too many opportunities to screw up his design intent, which led to banning double inputs on future products.
The decision was people / marketplace driven rather than performance driven, even though I often hear that Jim was against bi-wiring for performance reasons. (But only if the customer screwed it up.) Our development prototypes were often bi-wired to remove the intermodulation and other cable distortions from the speaker development decisions. It's always a can of worms.
|
Rob - I can recommend the SonicCap from Sonic Craft that still has the old aspect ratio. Anything Sonic Craft carries has been carefully considered and chosen.
|
To expand on unsound’s answer - the 3 woofers include the upper woofer centered between the two subwoofers. The subwoofer cones of the CS5 were mass-loaded and damped with rubber. Those drivers were modifications of stock ScanSpeak woofers. In his usual fashion, Jim improved the motors of those woofers as well as replaced the cones. The new subwoofers are mass-loaded with a center MDF puck. The upgrade was expensive and popular. Rob says practically all CS5 owners upgraded to the 5i (improved). I don’t think there were any crossover changes, and I don’t think think the improved woofers are available. I am presently conversing with ScanSpeak about replacement drivers for classic models including the CS5, but the fruit of these discussions is in the future.
|
A thought proposal: Multi-amp CS5s.
Over the years I’ve heard comments regarding CS5 performance that sound a lot like amplifier misbehavior than speaker limitations. My experience with the CS5, with the Krell FPB-600 and other big guns, doesn’t include flabby bass, anemic transients and brittle highs. (Note, I think the low impedance load is brutal and inexcusable . . . but) great sound can be achieved with great amps. CS5 owners please share your relevant experience.
My thought proposal is to wire a pair of CS5s for multi-amp inputs.
Configuration A: bi-amp. One channel on subs and upper woofers. Other channel on top 3 drivers.Configuration B: tri-amp. One channel on subs, other on upper woofer. Third amp (one stereo channel) on mids-tweeter.Three amps required. All run in stereo due to low impedance load.This idea is inspired by the excellent performance of the Benchmark AHB-2 and its relative affordability at $3K / each. The CS5 internal access is feasible because each driver circuit is routed independently to the binding posts, as are all Thiels.
My interest springs from my love of sealed bass performance and a product which didn't meet its market potential; only 500 pairs were built. My homework is to learn more about the present bass limitations for developing replacement woofers with ScanSpeak. Any and all feedback is welcome.
|
All - it is likely that those serial numbers will have classic, Lexington built, point to point crossovers - which I consider superior to later Asian iterations. TMR would very likely verify my presumption of PtoP for you. The 2.4 is high on our list of upgrade attention.
|
Unsound - the CS5's low impedance is restricted to the very low frequencies. It is about 2 ohms from 20 to 50 and drops to about 1.6 at 10Hz. But there is little power required down there- the low rolloff is to minimize phase shift at the deep fundamental musical foundation. John Siau (Benchmark designer) says no problem for the AHB-2 in stereo mode.
|
Brett - your point is why Jim refused to endorse bi / multi-wiring. Indeed everything can be screwed up and he firmly believed in Murphy's Law. Indeed we would abide by reason, using identical amps and speaker cable runs, and actually keep phase integrity somewhat better because intermodulation and other cable propagation artifacts would be minimized via multiple sources.
|
sdl4 - not Grimm, but yes Mogami and Benchmark. The pro brands I have tried perform adequately, but always come up with less ultimate resolution and 'you are thereness' against my references, which include some upper end Straightwire and Morrow.
My previous setup as shown in my virtual system put all the gear including speakers at the same end of the room. Next time I'll go for even shorter runs of both interconnects and speaker cable. As shorter is better.
|
Hey all - I would LOVE to hear some answers to Brett's query about "best" Thiel speakers, especially your reasons for your choice(s). Some of you are quite experienced up and down the line and your sharing would help my perspective.
|
Sdl4 - pro / high-end contains very valid concerns. I find it fascinating how people limit their bandwidths when approaching complex subjects. Benchmark makes undeniably great products and John Siau is an obviously world-class designer. I broached the subject of wire with him and his response was that the star quad configuration cancels all relevant distortions and is therefore as good as it gets. (my simplification.) I brought up "teflon" vs polypropylene dielectrics, and metal purity, crystal structure and surface polish. He dismissed those elements as irrelevant. My personal history proved those elements to be relevant and audible. People admit what they can understand and effective people tend to reject what they can't cope with. You might be well served to get some Benchmark and Morrow (on sale) for direct comparison. Please report back.
|
@beetlemania
re: " Curious if you have an opinion or knowledge about how to pair "main caps" if the full capacitance is not available in a single cap. Is it best to optimize balance between caps (eg, 7+7=14) or is there some degree of leeway (eg, 10+3.9+0.1)?"
My opinion is based on Thiel experience and related observations - both approaches are valid, but will produce slightly different sonic signatures. I believe the cascading bypass format compensates for deficiencies in the large cap which produces dielectric cycle anomalies (charge lag, erratic discharge, Effective Series Resistance irregularities relative to power levels, etc.), along with possible smoothing of the signal in general. A small cap is inherently less reactive, plus a higher quality small cap is more affordable considering budget constraints. As you have previously mentioned, Thiel’s 1uF bypass cap might sometimes introduce less-than-best characteristic balance when bypassing at greater than 10% of the base cap value. We arrived at our format because we developed that great 1uF tin foil / styrene film cap for the CS5, bought 6-figures of them, and put them to work wherever they made an improvement (that we could afford.) Our main (base) cap was a single cap unless its value exceeded 100uF, where the deleterious secondary effects broke through. In those cases, we used single value multiples. You’ll see ganged 100uF caps (often electrolytic) bypassed by a good polypropylene, bypassed by the great (yellow) 1uF tin/styrene. 100uF was the maximum cap value we identified from listening, and same-value was identified as more correct than cascaded values.
Note this scheme runs contrary to common practice. Note the CS2.7 (outsource engineered after Jim’s death) uses a 400uF lytic in the midrange feed (400E/15PP/1S) and other large, cascaded lytics in that shunt circuit (330E/220E/15PP.) That scheme differs from Jim’s. Note that people who prefer the 2.7 sometimes note they like the "smoothness" or "refinement" of that model which may be attributable to this cascaded approach. Note also, that Vandersteen employs such a cascaded-value, multi type approach to cap bundles. I have wondered if that difference might be a significant contributor to the obvious character difference between Thiel and VdS. The two brands share very similar philosophies and solutions down to wire configuration, but exhibit very different personalities. Jim’s engineering approach was always to identify the solution that addressed the hard-core principles most directly, optimize the particulars and confirm by ear that the most accurate solution had been found. Little to no slack was cut to make harsh recordings sound smoother or mitigate other signal chain problems.
In summary, I would choose the 7+7 option as truer to Jim’s approach. There is, of course, leeway and your other string presents the option of using a CMR (PUR+) as the 0.1uF value without breaking the bank.
If you climb back into your 2.4s and try these options, let us know what you learn. No end to the fun.
|
sdl4 - thanks for your cable feedback. Indeed star quad's major claim to fame is rejection of incoming noise - and that is not as big a concern for hi fi as for pro environments. BTW: I use Vovox unshielded for my recording mic feeds and love it. Please keep us apprised of your studies. I among many want to know a cost-effective wire solution.
|
ydjames - regarding your 2.2s. They will be viable for a long time. The Thiel ultratweeter has a drop in moving system that is fiber rather than aluminum. I'll be evaluating any XO changes that might be needed. The midrange has a modern ScanSpeak drop in replacement. I'll be tweaking that also. The woofer is all-Thiel, being the first Thiel driver after the CS5i woofers to utilize the decked out motor, plus the double cone geometry. Those are bomb-proof, plus Rob has hundreds on hand and lots of moving system rebuild kits.
The 2.2 has been my professional monitor since 1990 for artist evaluations of mixes and mastes. I know it and love it. I have developed a simple cabinet brace between the woofer and passive radiator to quiet that 300Hz resonance, for a more solid upper bass / lower midrange. And the same ClarityCap x Mills upgrades that Beetlemania did on his 2.4s all apply to the 2.2s. And there's more. I've done more work on the 2.2 than all other models combined. When you're ready to hotrod them, I hope to be ready to make you happy. Meanwhile, Rob can handle any maintenance issues.
|
Brettmcee - it seems we probably won’t get anyone’s list of top Thiels, in great part because that conversation already peppers these two hundred plus pages. Preference ranking is also a very personal thing and includes personal values and constraints. I can, however, supply some relevant history and perspective that may interest you.
Point one is that each revision of each model represents a step ahead in Jim’s pursuit of the same goals stated for that model at its inception. New drivers are a core element of each iteration. The 5i is merely "improved" because the top 3 drivers remained the same and the woofers are modifications, not ground-up designs. But in Jim’s lexicon a later model is always better.
Point two is that model progressions developed over time rather than some foreshadowed master-plan. The original series 1,2,3,4 displayed Jim’s linearity. The next model would get the next number, not necessarily bigger or more expensive, just chronology. (I tried to make the leading "O" stand for Original, but to Jim it was merely a leading zero to note he planned on more than 9 products. The CS designation changed that. The CS3 would have been the 03b but for Peter Moncrief of the International Audio Review who dubbed it a "Coherent Source" (which of course the 03 and 04 had already been.) So the floor-standing Coherent Sources developed their decimal iterations, the higher the number after the decimal, the more recent and more well developed. Some listeners prefer some older models due to more forgiving attributes in those older models. Note that impedances began around 6 ohms and fell consistently over time to more like half that value.
Point three is that market realities intruded into Jim’s natural rhythm and wishes.
We had developed a fairly recognizable lineup with the CS3 as the rather bold 10" 3-way, the CS2 as its little sister for smaller rooms and lower price, and the CS1 as the little two-way that could hold its own at half the price. The next product in the works was the CS4, which would have been a 4-way, and in hindsight possibly another breakout product like the CS3 had been. Jim was working on the 12" woofer which eventually found its way into the CS7, crossed to a 6-7" lower midrange and a CS2-type small upper midrange, and the 1" ultratweeter that landed in the CS5. The late 80s were high times and many companies were unveiling statement products at 5-figure prices. Add to that, our most enthusiastic high-end market was Japan and the Asians who followed its lead. Japan (and others) wanted more speaker than the CS4, plus in Japanese culture the number 4 is associated with death and myriad superstitious troubles - a model 4 would not fly in Japan. So the 5-way CS5 was developed, leaving the door open for the CS4, which never materialized. In my opinion the CS5's technical difficulties could have been left behind by a CS5.2; but that never materialized.
The 5 was followed by the 2.2 which I consider an in-character small masterwork. Then the SCS introduced the coax to greatly solve the ear-position problem of a portable speaker. It retains its port and its second order crossover from the 02. Fair enough. But the following CS3.6 changes the core DNA of the model 3 to reflex bass and the model 6 and 7 follow suit while side-stepping both bass equalization and sealed bass. They don’t fit quite as neatly into the tacit philosophical matrix. There are many factors involved in my take on the subject. One of those is my departure in 1995, having established our production capability and needing to leave a very hot hotseat of directing manufacturing operations for a production-limited company growing at a self-defined 30%/ year with marketing pressure for greater growth. Add also, Thiel’s entry into "home theater" requiring the development of the feature-laden subwoofers, and the necessary transfer of much of that HT manufacturing to China. The company’s evolving DNA no longer fit my personal sensibilities strongly enough to ride the wild beast.
Let’s address that pesky model 3 bass alignment. Believe it or not, a major component of Thiel’s musical reproduction philosophy rested on solid bass execution from which the harmonic structure of the music arises meaningfully. A sealed bass system pulls off that feat far more deftly than a reflex system in which the 4th order bass alignment places the lowest bass harmonics a full cycle behind the higher harmonics - effectively putting the deep bass 10 or more feet behind the rest of the music. In the day it was called "slow bass". Jim’s equalized bass solved the problem elegantly, but faced significant market resistance. His CS5 non EQ solution required progressively dropping impedance in the deep bass. However, I believe the problems of driving that load could have been mitigated via higher impedance across the board. Jim was insistent that higher impedances "wasted amplifier power". And his word was gospel. A very interesting aspect of Jim’s thinking is his reduction to simplicity. He reasoned that if "some" amplifier(s) could drive the CS5, then any that couldn’t were defective." He always reverted to some version of: "they have an amp problem, not a speaker problem" when addressing difficulties driving the punishing load.
Cutting to the chase, the CS5 failed to meet our market expectations and the CS3.5 EQ was similarly problematic. I believed that we could improve both to surmount their limitations and carve a clear market niche. That wasn’t my call to make, as Kathy as Marketing Director decided that those considerations were too esoteric for a meaningfully large market and that reflex bass was "good enough". Jim set about executing reflex bass very well indeed; but the lagging time alignment is baked into the physics of the beast.
I detect a more avid fan base for the early Thiels where sealed bass is part of the DNA of the upper models and reflex bass is a necessary accommodation for cost-reduction in the lower models. I recognize that my perspective is colored by my own lived experience, and that there are many brands priced up to $Six Figures that rely on reflex bass. Just my take reflects my values.
So, my list of favorites would include the older models, even though their execution is not up to the standards that evolved as Jim learned more about driver improvements and other elements of his art. I have heard that the 7.2 was his favorite all-in (although other stories are out there.) The breakthroughs of the radial wave and star plane drivers of the 3.7 would have opened a doors into a new level of performance and possibilities. But the story doesn’t necessarily unfold as we might wish.
|
sdl4 - I have not used the Vovox for interconnects - just because I never felt the need. I'm happy with my Morrows. My opinion is that unshielded sounds better than shielded; so only use shielded when there is a need (evidenced by shielded sounding better.) I'm very interested in what you learn about Vovox as interconnects.
|
thoft - among your short list, the Dunlavy is phase/time coherent like Vandersteens and Thiels. The others are not.
brett - I don't know how many of 'us' have ventured into DSP / active, etc.
"That bass affect" is present in all reflex bass alignments. The 3.6 passive is in proper polarity, pumping out when the woofer pumps out, but it is a whole cycle out of phase - 360° phase shift, just like other reflex systems. That same 360° phase shift is present in most "modern" crossovers which use 4th order filters on all drivers. The ear-brain compensates or learns the phase lag. But Thiel fans tend to appreciate the lack of that phase shift and resulting naturalness of zero phase shift through the crossovers.
Among Thiel speakers, those with sealed bass (like your Kappa 8s) include the model 01, 03, 03a, CS3, CS3.5, and CS5 plus the PowerPoint / PowerPlane with the 6.5" PowerDriver. The others have either ports or passive radiators.
|
thoft - I’ll pick up this SCS4t query. It is a regular SCS4 on a built-in stand which is acoustically non-functional. The stand provides no bass boost or any other function than to elevate the speaker. I have an SCS4 here - the final iteration of the 02-SCS series. It uses the coax PowerDriver that inhabits all the 6.5" HT models. It is a well developed driver with all the Thiel special sauces. The twin ports seem to be some sort of tractrix or similar advanced room coupling form, but they present a rather convoluted baffle plane for the waveform launch. Perhaps the unbraced cabinet might have some resonances. All that said, I don’t warm up to the SCS4; it presents a somewhat hollow and edgy sound compared to many other Thiel speakers. In my experience (far from all Thiel speakers) I would place it as the weakest entry for dedicated music listening; but it's probably a pretty good Home Theater contender.
As an aside, I use one (or two) SCS4s on stands for live sound reinforcement in a small venue where it shreds the regular contenders. Try to get an audition against some known speaker of your choice.
|
unsound - thank you as always for your comments on multi-amping the CS5. You are correct that this cure might be worse than the disease. I'll note that your "simple" solution of using a suitable amplifier is also Jim's solution. And it works well and beautifully. I wish that John Atkinson and Larry Archibald had applied your solution. They got best results with the Krells of the day, but left the (deserved) impression that most amps couldn't successfully drive the load.To my multi-AHB-2 solution: I suppose that's a personal proposition. I agree with you noting all those associated expenses and complications. And I admit that my intrigue may be closer to a lab experiment than a market solution. I have 2: AHB-2s and want 2 more for live site work . . . (creeping justification strikes insidiously.)I don't know how the amp might drive that low end, since it is so nicely resistive. The clipping LEDs are sensitive enough to use as lab tools.
To your comment about re-sculpting the baffles: What if? The MB upper midrange dome presents the problem of being 3/4" too far forward. A driver built on the radial wave / star-plane diaphragm constructs might outperform the MB, plus place its acoustic center closer to proper placement.
Gotta go now.
|
Beetle and Duramax - I'd love to see pix or descriptions. I can supply the general MO for that time. The 2.4 was introduced in 2003, before Jim got sick. All hi-fi products were still being made in Lexington with components selected, developed and upgraded through the years. So, #005/006 would have Solen caps, ERSE 5-9s coils and resistors, teflon jacketed wire and masonite boards. In other words, what we consider "real Thiels". My notes say the 2.4 XO production went to FST around #220-230.
Also, the first (100 or so) units were treated as shake-downs with every unit going through an R&D type course of measurements and listening cross-checks in order to set the production measurement limits. Every Thiel speaker was tested in an anechoic tower with strict pass/fail limits which were developed via this test/listen process at the beginning of each product's life cycle. So, duramax's
CS2.4s would be the equivalent of "blueprinted" in the racing world.
Duramax - if you ever upgrade / hotrod those 2.4s, I would love to get the original crossovers as my laboratory reference for the model. I have a pair of late / FST (3729/30) 2.4s here on loan. Wouldn't it be interesting to compare early vs late?
|
|
Max et al - #1&2 were usually kept in Thiel's music room and 3&4 at the farmhouse or insider rotation. You may very well have the earliest 2.4s out there in the world.
|
Max - what serial numbers are your straight 2.4s?
|
sdecker - thanks for the input. My 220-230 estimate was from Rob’s memory. I’ll replace that estimate with your 611-12 old-style 2.4s.
Now, a thought about Lexington vs FST coils. It seems that one of the extenuating circumstances around taking XO production to FST was that our original coil wire became unavailable. As a reminder: we had "found" an aerospace communication wire in 1978 while developing the 03. It was 6-9s oxygen free, long crystal, slow drawn, polished wire. Its use in our coils was transformative. (Longer story embedded in this forum’s prior history.) Over the years, that wire was "degraded" to CDA101 (5-9s) with less stringent production values. Then CDA102 (4-9s), still a league above "normal" CDA110 electrolytic tough pitch magnet wire used in most applications and by most hi-fi manufacturers. ERSE tells me that in the mid-2000s those high conductivity, high purity wires became "unavailable" and were replaced by CDA110-special, which meets only the 99.9% / 3-9s spec, with special attention paid to extrusion, surface quality, etc.
Bottom line is that Lexington boards have at least CDA102, and older boards have CDA101 coils and they are wound to optimum aspect ratio, tensile pull and bonding formulation. These Lexington coils are world class. Coils on ’glass’ boards from FST are best presently available, CDA110 special, but certainly lesser than the Lex / Masonite board coils.
|
Max - I use Delignit / panzerholtz in piano restoration work - and love it.The earliest Thiels from 1975 to 1985, up to the CS3.5 used various products. The 01 and 02 were medium density particle board. The 03 and 03a were Baltic Birch (actually the higher quality FinPly). The CS2, 3 and 3.5 were high density particle board walls with high density particle board baffles. The model 2 and 3 had internal spruce struts inside the cabinets. I consider all these models to be of superior materials and outcome than the later MDF cabinets.
MDF was chosen for its consistency, universality of sourcing, flatness for laminating and gluing, and machinibility. With well engineered internal shelf braces, it was adequate for the job. Note that Thiel was producing extremely complex cabinets in a production environment with no room for anything to go wrong, and needing a universal material for feasible stock management. 1" MDF filled that bill. We were never a boutique, cost no object shop. Our finished products sold for sometimes a small fraction of what some manufactures lavished on their cabinets alone.
There are no CAD / manufacturing drawings available; but that may change in the future. I believe a more cost-effective retrospective approach might be to experiment with Panzerholtz, etc. as bracing material to quiet any cabinet wall resonances that can be identified. I am adding strategic long-grain spruce cross-braces to settle the enclosures, and have chosen torrified bamboo for bracing as a very high performance / cost effective candidate.
In the mid to late 80s, I experimented with various Corian-type products as well as honeycomb substrates with technical skins, curved panels and so forth. So much cost was added that we stayed away from them. We went to cast baffles machined with abrasive diamond tooling as an effective solution to affordable extreme rigidity It would have been very easy to double our manufacturing costs with multipliers applied for retail prices, and stray from our intended niche of phenomenal high value. Cost effectiveness was one of Thiel's evaluation lenses.
By the way, the CS3.7 / 2.7 returned to one of those 1980 experiments. The curved side panels are 17 ply birch laminates, using the material and its geometry for added rigidity and waveform propagation.
|
Max - I had never considered that Thiel's weak link was its cabinets. Let's put that idea on the back burner for now.
Regarding Spruce as a cabinet material - that has merit. Spruce is an extraordinary wood, landing it as the soundboard of choice for many musical instruments. My primary business after Thiel Audio has been producing tonewood for high-end instruments. Red Spruce (the holy grail of tonewoods) is my specialty. As a grown material it's hard to beat for stiffness to weight and so forth. But as a cabinet panel, it doesn't measure up all that well. It is a naturally resonant material, which is what you want to avoid in an enclosure. It is highly directional, being 7-10 times as stiff along the grain as across. Its micro-structure encourages transmission of sound through its thickness, another thing we're trying to avoid. And even though it is relatively stable for as a lumber species, it still expands and contracts enough with humidity changes that the cabinet would change characteristics through the year and eventually crack. So, all in, the idea doesn't check my boxes.
Among the materials science aspects of Thiel's cabinet walls, I chose a Canadian MDF made entirely of Spruce fiber, for the reasons you cite. We then laminated veneer on both sides with a structural adhesive, creating an I-beam effect with about double the stiffness of an unlaminated panel and more tha 1.5x that of an exterior-only veneered panel.
Dr. Toole and his proteges at the Canadian Research Institute looked at cabinet material candidates regarding stiffness to weight, stability and damping, and landed on MDF as the closest to ideal. PSB, et al made a case of MDF being "better" than most of the exotics out there. I wouldn't go that far. But putting cost into the equation tilts things significantly toward MDF being a pretty good substrate.
All that said, my present design work uses other materials to augment and upgrade the basic MDF cabinet.
|
Max - I consider Delignit to be a wonderful material and would be interested in what you know about it as a bracing material.
I have had some success adding cross struts of bamboo (in the league of Delignit) to the MDF shelf braces, which keep the struts from vibrational deformation - a nice pairing of attributes.
|
Rob - this wire issue deserves more fleshing out.
CDA101 is 99.999%CDA102 is 99.99%CDA110 is 99.9%There are other requirements for each gradeAnnealed Copper is less conductive than these grades, which are all Electrolytic Tough Pitch-at the pinnacle of conductivity for hardness.
Regarding ERSE's CDA101. As you know, I've been playing detective to study what would have been obvious at Thiel Audio, which I left in 1995. I have good inside sources, most of whom either never knew, don't remember or don't have access to relevant records.
What I know: AcoustaCoil (later absorbed by ERSE) produced all our CDA101 coils since 1978. State of the art. When trying to get a definitive statement from them regarding their website claim of CDA101 - I didn't get a straight answer. I got diverted to Barbara in sales (who seems to run the show) who went to engineering who came back with the answer a couple weeks later of "ETP / CDA110-special. Now, let me float some conjectures.
ERSE along with ScanSpeak and most everyone else makes most of their world goods in China with FST being a prime source. My conjecture is that the Chinese wire plants are using CDA110-special and that is why the "answer" to my query came back that way. However, I also know that ERSE makes their custom coils for the American Market at their Cleveland plant. My conjecture is that those coils are indeed CDA101 as stated on their website and consistent with our long history with them.
I can't prove that and really don't have the will to drill down - there's only so much time. But for discussion, let's suppose that's the case.We know that the 3.7 XOs went to FST some time around #500 (no hard evidence yet.) We know the resistors and caps were CYC clones of Thiel's ERSE and Solen originals. We know that Jim "approved" them. We also know that direct comparison of those components in the 2.4 illustrated the original Thiel parts being "better", as discussed previously in this thread. We know the FST coils look decidedly inferior to ERSE coils. I have not yet made a direct sonic comparison. Beetlemania may have directly compared coils to coils - I don't remember.
We also know that all CS2.7s were made in Lexington with original parts including ERSE coils and resistors, Solen caps and Masonite boards. It is possible that the reported fluidity and grace of the 2.7 might be in part due to that factor of All Lexington build practice.
To the other point of foil and/or stranded - I have used some ERSE foil coils, but not directly compared to (known to be CDA101) ERSE coils in the 2.2. I have not used stranded. Regarding both foil and stranded, they make a difference in resistance and capacative reactance - so those changes would have to be compensated in the XO. I do not yet have the chops to make those compensations, so I haven't gone there in my development work.
That's about it for now. TT
|
Jazzman - as a point of record, the 2.2 never suffered any production changes. This may be unique in Thiel's history. The drivers remained extraordinarily stable throughout its 7 year x 4000 pair run and no crossover changes were deemed necessary to fine tune the design.
TomTweak - if those 8 gauge coils are for an existing design, their effective series resistance will be much lower than smaller gauge wire. You may need to compensate to keep the crosspoint and level as it was designed.
|
Rob - the new ClarityCap "purity and purity plus" caps do not yet exist. You may PM me regarding additional details, but don't look for them in the marketplace in 2021.
|
TomTweak - these speakers seem chock full of exotic solutions. Have they been submitted for review, or do you have any white papers or third-party commentary?
|
gs5556- yes the passive screws lock the radiator to the baffle, and must be tight. And yes, they do come loose with vibration. You might try some removable Locktite, etc. or snug them every month or curse them, but that probably won't help.
|
Adding to Beetle’s excellent summary, here’s a cheapskate trick for the Thiel resistors that you might keep.
Turns out, a big part of those resistors’ ’sonic problem’ is the thermodynamics due to their mounting. They are non-inductive wire-wound with current flowing around their coil circumference, dissipating into the ceramic body. The bottom of that body is glued to the panel, creating a large differential in wire heat distribution. Solution is to mount that same resistor on an edge rather than a flat. Put the edge on Mortite / BluTac pads for even better results. Cheap fix, much of the improvement of Mills, suitable for woofer and midrange. Spend your savings on a Path, etc. for the 12Z tweeter feed.
|
Some of the power resistors can benefit from 25 watt versions, thus 2x12watt. This resistor thing was a short suit for Jim (in my opinion), since the compact layouts can only afford so much real estate. Modeling software can ID those candidates; so can pink noise and a thermometer (or fingers.)
Beetle - I do remember the aluminum cooling bar mis-step. Moving air and greater distance work better.
|
Beetle - I looked extensively at thermal dynamics as a source of 'aural congestion', drifting crosspoints, etc. That's where the new layouts came from, which have all the resistors separated from caps as well as positioned over ventilation holes in the board for natural convection, and moved either to the exterior of the cabinet or to separate XO enclosures. The hookup wire(s) also mount to pin-type heat exchangers close to each driver to drain voicecoil heat directly through the lead wires. Taken together the thermal management significantly improves dynamic range and 'sonic orderliness' during high-power use. The small-signal Thiel reputation got bigger.
My investigation began by remembering 'warranty claims' for melted caps and charred XO boards in the old days. In my power tests, some resistors got too hot to touch. Series coils also get hot - I put them on 3 rubber feet for all-round radiating and convection. My tool of choice is a non-contact infra-red thermometer to identify small scale/ local differentials. That led to some heat-sinks being added in-line on the XO board for dissipation before entering caps. I don't remember at what stage all that investigation stood when we hot-rodded your 2.4s. There's always more.
|
Guys - I'll have some things to say about Thiel's take on positioning.Meanwhile, is there a physicist or other knowledgeable person here to address sdecker's query about 'interesting' bass alignment? It's important, and veiled in obscurity, and reflex bass is the industry standard. Any comments?
|
Tweak - in a first order alignment, neither driver rotates more than 90° phase shift, and the overlapping drivers are in opposite directions - they sum to zero. But that's only if each driver, in its enclosure rolls off at 6dB / octave. All that 'extra stuff' in Jim's crossovers is to counteract any anomalies in the entire system such that they do just that. Then they cancel each other and sum to zero phase shift. Since the world is never perfect, plus/minus 10% is achieved, except for the CS5 @ 5%. Convention allows these claims for 100Hz and up. We'll talk about bass later.
|
sdecker - I am not fluent enough in the art to answer much about the nuances of Thiel's phase coherence. I will point out that your "completely time and phase coherent" is not something that Thiel claimed. The applicable term of art is "minimum phase transducer", which has a complex set of definitions revolving around the least phase error possible given the particulars of the design parameters. To a prior question of whether this is marketing-speak, I say it is not. Jim was extremely careful to not claim anything beyond the real, hard facts. I must demure to more knowledgeable folks regarding bass reflex phase error. Please let us know what you learn from your studies, The Asylum, etc.
Here is something I do know. If the bass extends below the program limit, much of the phase shift problem evaporates. All Thiel models push that lower limit quite deep. Note also that Thiel's bass tunings typically produce low and controlled impedance peaks with phase plot following closely and well, indicating significant attention paid to coherence at all frequencies.
As you know, I am a fan of sealed bass, and I have shared here how disappointed Jim was that his equalized bass solution did not meet better market acceptance. I speculate that a more purist EQ execution, along with true balanced design, and today's better interconnects - that an equalized sealed-box bass might be viable.
Just for fun, you all might enjoy reading about Thiel's first product the 01. It is an equalized 10" 2-way with bass to 30Hz with third order slopes which are phase correct in many ways, excluding polar dispersion, in a 1.2 cubic foot enclosure at 94dB efficiency. That product got us to Germany by 1978 and via Europe got the attention of advante-garde American dealers. I have seen no evidence of audiophile interest in the 01, but it most clearly represents Jim's initial design impulses and orientation.
|
magnet7 - yes, that 8uF tweeter blocking cap is the best bang for the buck. Psychoacoustically, its function range in the midrange and treble is where our hearing is most sensitive.
|
sdecker - I appreciate your feedback. Remember that I did not keep up with Thiel for its last 15 years. I agree "completely time and phase coherent" is a marketing, rather than an engineering term, so it would have come through Kathy's channel. Fair point.
On the other hand, excepting that stuff in the bass, those speakers are phase coherent.
What I mean by the "phase shift problem evaporating" is not clear to me either. My measurements show group delay remaining constant through that 33 to 100Hz range, and I don't have a handle on what frequency the full rotation occurs to cause the lower fundamental to be be behind the upper range. I'm not being coy, I'm admitting my limited knowledge of the particulars. I do believe the effect is audible, if we compared your 2.4 to a CS5. So much to learn, so little time. Please educate us if you learn more from the Asylum or other study.
|
thielrules - as I see it, the output of the port is a full cycle behind the output of the upper drivers, with 4th order (24dB/octave) slopes induced by the physical / mechanical interaction of the resonating port (no electronics involved.) Since I think that 4th order slopes maintain phase linearity, the only non-coherence would be in the time domain, where the frequencies covered by the port would emanate one cycle behind (but in phase). My calculation says that lowest tone centered on 38Hz (port peak output) would sound like it comes from 30' behind the speaker.Above 50Hz (the -3dB/half power crosspoint) the output becomes dominated by the woofer which is in-plane and therefore in-time with the upper signal.
The lowest note of a bass is about 40Hz, and most of its information is above that 50Hz crosspoint. So, I'm not surprised if many people can't distinguish between ported and sealed bass. Also, your DSP solution that further minimizes slight phase discrepancies in Jim's original analog filters, does introduce a digital conversion as well as some pre-ringing (which you have minimized.) I am one of those folks who is sensitive to pre-ringing, so I might choose different trade-offs than you do.
Remember that most modern speakers admit these 360° / full rotation phase shifts at every crossover point, where there is lots of musical information. The common call is that "it can't be heard". I suspect most Thiel fans prefer it not be there, even if their ear-brain is deemed to be able to ignore it.
|
sdl4 - stock Blue Jeans
This cable was sent to me by Iconoclast for evaluation as their best offering from the BAV line, short of Iconoclast. I don't believe it shows up in their stock offerings with RCA terminations. The cable shows up in BJ's SDI section, but the connectors look like their stock audio RCA's which are Canare RCAPs.
|
thielrules - I understand the trade-off between linear and minimum plusses and minuses. Regarding the DSP solution, that's a realm that is morphing in real time. Even at the very high end, digitization still brings its own set of potential problems. Digital circuitry is noisy and requires careful isolation from analog. Assuming you can manage all that and get stellar results, there is the matter of different skill sets. Of course in Thiel Audio's day, the realm didn't exist, or was still in its infancy and adolescence. Jim developed his tricks and his niche firmly in the analog sphere, and nurtured and grew those skills over his whole career. I am committed to consolidating that legacy for its stability into the future.
I am pleased that you are applying digital solutions to the design, but such an undertaking beyond DIY, would take a concerted effort to make it so.
|
Jafant- hands-on audio is on hold till I find new workspace. I hope to hit Virginia within a month. No Classe report yet from Bill. Audio care goes slower than medical. |
Tony - I am not surprised you heard improvements with your wood floor. That is consistent with my long-term experience with hi fi and live sound. One problem with carpet is its frequency-variability; different frequency bands are absorbed or not in non-linear ways. For the record, I have gotten best carpet results, including at Thiel's Music Room in Lexington, with wool carpet over commercial hair-pad underlayment. That's when carpet is necessary, such as when covering concrete where wood is not an option. In live performance, ensembles and/or the piano is often placed on carpet to quiet the reflections for the player(s). It does that, but not linearly with frequency; plus the effect in the audience is often an unpleasant harmonic unpredictability.
In my personal experience, it doesn't get any better than a wood floor. The record-making process tends to assume a wood floor for playback, as does the underlying theoretical physics. Jim's designs all employ a 2dB shelf below 200Hz in the anechoic response to account for the presence of a floor - converting the full-sphere anechoic response to the half-sphere in-room response. Some brands ignore this design element. Some users call Jim's designs "lean", etc. That "leanness" is exacerbated by carpet, sucking out some areas of response.
Congratulations on your wood floor. Music heals some aches and pains.
|
Sutherland Felt - manufacturer. F11 absorbs audio frequencies. 1/8" is where I landed with multiple layers behind midrange driver.
|
jafant - the pleasure is mine. Good work here, guy.
magnet7 - my latest schematic (22dec'87) shows an 8uF (not 8.2) which had replaced the original 7uF spec.
|
Tom - highly congested territory with more questions than answers. I have been working with turbulent / laminar flow for a couple of years now. I suggest you experiment with an open mind, since there is much more to the area than first appears. As a point of Thiel history, these multi-faceted problems of ports led Jim to migrate to passive radiators rather than ports. Note the model 1 which leap-frogged from ports to radiators, landing on that long, slotted low-turbulence port on the 1.6.
There are too many issues to address here, but note that taking the port outboard exacerbates the problems of coupling the port’s output to the surrounding air. The baffle flush with the port end serves to support that transition. Moving your port outboard may reduce some internal turbulence, but at the expense of increased outboard turbulence.
One thing I’ve learned is that the surface turbulences are audible, that they affect the full-range sound, not just the bass tuning frequencies, and that the particulars of the baffle surfaces surrounding the port have audible effects.
I don’t know how that coin-toss of internal vs external would land. Please keep us posted about anything you learn.
|
