Thiel Owners

Hi guys. I'm not up to speed with which amps do and don't, but I do have lots of general Thiel experience. I have noted that current delivery capability is very important, especially in large rooms or loud volumes. If an amp doesn't double its 8 ohm current into 4 ohms, it will probably sound anemic, and its oomph into 2 ohms is relevant also. Ayre amps are good fits. Jim used big Krells. Of related importance is that cabling is more critical on Thiels than usual. One reason is that phase coherence is easily scrambled by cables. On most speakers that doesn't matter because they are already scrambling phase, but the degradation is significant on Thiels. Also, Thiels very low impedance load requires more current through the cable, exacerbating its anomalies. I highly recommend very short runs of speaker cable. Jim used Goertz flat cable in 3.7 development.

Prof, I love it. Which model number? I need some long runs for house sound with CS 2 2s. 

Thieliste, I have used many previous Pass amps and think very highly of them. (I don't know the INT-250). Nelson is a superb designer with values commensurate with driving a Thiel. The simple fact that it is doubling down to 2 ohms speaks clearly to the factors involved in driving a Thiel elegantly.

Because Thiel speakers present a sub-4ohm across most of their spectrum, dropping to 3-ish over extended ranges, the behavior of the amp from 4 to 2 ohms tells more than 8 to 4 does.

naimfan - ordinary screws will work. Those Thiel screws are plated with black phosphate. Rob at Coherent Source Service should have them.

Hey wire nuts - I've gone down a rabbit hole with Steven Hill about wire configurations, pros and cons and theoretical considerations that are way above my pay grade. I'm testing samples and configurations that he sends.

Also, I have developed some observations about the normal vs parallel runs using my OCOS and Morrow wire. I'll chime in after Beetle shares his experience with us.
And - someone here reported his 2.7s as having straight (non twisted) internal wire. I checked with Rob who said that all the 2.7s (all classic Thiels, period) had the normal Thiel wire. So those 2.7s were someone's hotrod. Back to the lab.

Rob - try me at tom thiel at sign worldpath dot net.
Tom

Robert - oops, I forgot where you got them. Nonetheless, they are unique, and were modded after New Thiel went to Nashville. I can speculate that New Thiel was experimenting with different internal wiring. I know they were experimenting with ultra-caps, etc. I wonder if your 2.7s have better than stock XO parts.

Dennis Crossen, the development engineer, was quite experienced and talented. I suspect that whatever you have has merit. Considering my present exploration of wire configurations, I certainly would love to learn how your 2.7s compare to stock. Can you arrange a comparison and give us a report?

Robert - you might be right about assembly. Rob G. said that all 2.7s were built in Lexington, but he was not in the Tennessee loop after they moved. The brands you state are Classic Thiel brands, except Topmay, which I don't recognize (but I'm not in the loop either.) The inductors are most probably Thiel inductors. Thiel bought quantity high purity wire and made small inductors in house and farmed larger values out to custom winders including ERSE who bought long-term supplier AcoustaCoil.
Can you please describe your internal wiring scheme? I'm curious what they tried, and also if they made any component value adjustments to compensate. Good pix of your XOs and/or values for me to compare with the stock schematic would be helpful. But, only if you are diving in anyhow.

Hi Guys, As Beetle mentioned, I have a system for identifying cabinet misbehavior that permits real-time experimentation and amelioration. I use Chladni Patterns in my acoustic guitar design, I love the method, and will summarize here as applied to enclosures. On each panel in sequence, sprinkle glitter (etc), drive the cabinet with variable sine wave, stop when glitter bounces. The physical resonance patterns are mapped via glitter bouncing off the moving anti-nodal areas and accumulating in the still nodes. A seismic map! Take a photo. Increase signal frequency and repeat. Do all the panels.
One neat part is that correlations are easy to make on all panels, allowing a 3-D visualization and suggesting direct intervention strategies. This method is low-tech and enormously informative.
I have identified some modes on my CS2 2s, which are easily addressed. We can quiet down all the cabinets.

On other fronts, good progress is being made. Parts are on order for Beetle's 2.4SEs and my PowerPoint 1.2s. 

Prof, I agree with you. The cabinets are extraordinary in both anti-diffraction and inertness. I regard Thiel's cabinet engineering as state of the art. The approach I am taking is to investigate all areas where additional input might make improvements. Various test reports, such as John Atkinson's thorough work in Stereophile, show various resonant anomalies which I may be able to minimize with minimal cost.

The drivers are fixed and excellent. The cabinets may benefit from small tweaks. The crossovers are the main focus. New improved parts quality throughout and optional outboarding are the main impetus. We are experimenting with 4-nines foil coils in the signal-path feeds and up-sized 4-nines round wire in the shunts, Mills MRA-12 audio resistors throughout and returning to point to point layouts. This undertaking is a labor of love for me and in no way meant to disparage the existing products.

Jay, the cabinets are very quiet, but some issues are present. The 3.7 sings a little where the side meets the cap, and so forth. I'm just saying that we can make incremental improvements beyond the original performance parameters. Such on-going improvements happen regularly in many high-end product lines. We intend to keep Thiel alive in that way.

Rob - JAFant or others here know how to create a virtual system page for all to share. I'll PM you to try to connect behind the curtain.

Indeed, the inboard and outboard options are different beasts. Space / layout constrains parts selection inboard. Making progress every day.

Regarding spikes . . . my knowledge is rather primitive, since all my work was long ago, before any commercial products were available, so there is plenty I don't know. I'll tell what I know. We found hearable and measureable time-domain slurring caused by recoil-swaying of the "unanchored" speaker cabinet. The woofer moves the cabinet in amounts which are very significant to tweeter frequencies, especially their transient attack/timing/phase behavior. Over many years' experience I found the presentation to be more focused with spikes. And transient tests measure more cleanly when spiked. As usual, there are other considerations. Speakers on carpet usually sound smoother, mellower . . . more polite, "nicer". I judge that mellowness to be caused by subtraction of transient detail. And another thing: direct coupling to a wooden floor can cause coupling resonances in the under-structure, euphonic-harmonic and/or dissonant, which are not stimulated with the insulating carpet or isolation-type feet. Another note is that spikes that are not locked down can absorb energy via motion losses between the threads.

There are so many particulars and mitigating circumstances that I hesitate to comment. But you asked, and my comment is that rock-solid stability at the micron scale aids the speaker in its job of transient replication.

Fitter and all, the way the upgrade project is shaping up is to target the most mature iterations of separate-driver classic systems first (with exceptions.) There are many thousands of pairs out there approaching eventual electrolytic capacitor problems. The targets are PowerPoint 1.2 (because I use them in my recording /documentation work and think they have serious professional potential) CS1.6 (no expressed interest), CS2 2 (because I use my pair in my work and it is very special) and CS3.6 for the same reasons. These models use Thiel-developed drivers with proprietary and patented technologies, mature enclosure designs and considerable performance upgrade headroom. The CS5(i) is not on the list. Its fundamental upgrade option is to replace the huge bucket-brigade delay lines with a re-engineered baffle to place the drivers in correct offsets for time-coincidence and elimination of dozens of capacitors and resistors in the signal path. I have ideas and skills, but insufficient life-force to take on that project. Similarly the CS3.5. The CS x.7s will come later. They do have upgrade potential, but are stable in their present lives for now. CS2.4 is in the first tier because of broad interest, significant potential and beetlemania's willingness to collaborate.

All models will have an inboard and outboard option. Inboard space constraints limit some capacitor and coil choices, but all electrolytics will be replaced by high-functioning propylene caps, and feed coils replaced with 4-9s foil coils. Outboard applies same strategies with some larger, higher-voltage feed caps and larger gauge coils and optimized layouts. The cap choices settled on ClarityCap and Multicap RTX bypasses, after exhaustive research and correlation with Thiel history and values. All parts are ordered for first-round samples for PP inboard, PP outboard, CS2 2 inboard, and CS2.4 outboard. We'll be making music this summer.

I will be using an unconventional evaluation approach, rather than MLSSA and related lab development and measurement tools, which I don't have, but would love to find a collaborator for. I use Metric Halo SpectraFoo, a pro-audio tool for tuning rooms, performances and recording, mixing and mastering music, along with some precision meters and scopes. Our benchmark method will be the system I use for my professional instrument and music making. We'll record live in my studio direct to hard disc at 24x192kHz for playback in the same recording space through the same signal chain, through speakers under test. We have live thru playback (music or technical content) with very few unknowns and considerable control and documentation of bass, peak and mix levels. The results going in and coming out are all sampled and analyzed via technical measurements and simultaneous listening. Someone could build a speaker company on this methodology . . . 

You might notice that this scenario looks like more than an after-work undertaking. It is and I don't know how that will all settle. But I am enjoying this challenge and have high hopes for outcomes. Stay tuned.

The x.7 drivers are indeed breakthrough design. That said, the 2.4 drivers are very mature and made and tested in Lexington with significant, advanced technologies. The 2.7 cabinet is stiffer, but the 2.4 will come close in vibrational performance, especially after tweaking. Regarding passive component quality, the upgrades will leap-frog the stock x.7 series into pretty rarified territory. We can expect some veils to be lifted by applying technologies that were out of reach for the stock product line.

FYI: I have contacted the world leader in audio cryogenic immersion. I am previously familiar with what can go right and wrong in that process, and how to stay out of trouble. I will be comparing stock and cryo-treated crossovers to evaluate efficacy. 

Just for clarity, I think the CS3.5 is in many ways Jim's pure vision, especially the bass equalizer. I believe it to be a strong contender for an update / upgrade and I would be interested in addressing the crossover. However, I lack the chops and time to upgrade the equalizer, which would be required for system synergy, since in terms of sonic transparency, the EQ is already a performance limit. So many good ideas, so little time.

Beetle, I haven't yet looked at the 3.5, but will look into it. The equalizer could indeed be ignored in an upgrade. The 3.5 has 6-nines coils and hookup wire and custom 1uF styrene bypasses. We could renew the 'lytics with best of class 'lytics, especially in shunts. But any 'lytics in Signal Path would benefit from our newly developing custom CSA-160 volt propylene cap. Mills resistors. Possible SP feed coil upgrade. Layout can be addressed by building a new midrange board to allow space on the main board for woofer and tweeter upgrade bulk. I'm settling to a system. I "did" the 3.6 today with our first-round assumptions which need confirmation with first-pair trials of 2.4 and PPs.

I can offer a general answer. For perspective, the difference between 33 and 35 Hz is barely academic, virtually identical. Thinking in half-octaves (15 Hz at 30 Hz) is more germane to performance class than is a few cycles. Low E1 on a bass is 40 Hz.

Differences are accounted for by technical particulars. Bass alignment is very specific to cabinet volume, rigidity, driver position, diaphragm mass, maximum excursion, suspension characteristics and motor factors. An alignment is optimized considering these many factors among others. 

As alluded to above via driver size, air-coupling is a factor in the experience of bass authority. The 3.7 10" woofer has 50% greater area to drive its wave-front, reducing compressive non-linearities of air. Bigger makes cleaner. The 3.7 cabinet is also more inert and its baffle is less compressive for attack transient integrity. Regarding prof's thought about bass voicing, the 3.7 was Jim's work, the 2.7 incorporated various outside engineering input. Jim tuned his bass alignment to Q.707. Many designers give a little slop for "bloom, bigness, warmth", etc. I heard the two speakers at Thiel at finalization and heard the 2.7 bass as tuned a little looser. Most designers try to second-guess popularity, expectations and so forth. Jim was fairly immune to those ways.

Also note that image is highly dependent on cabinet and driver edge effects. Prof mentioned the driver part. The 3.7 cabinet design (love it or not) is very highly functional regarding diffraction, even though the 2.7 might qualify for world-class, it is not as good. 

A factor that contributes to "great bass" is that of articulation which, in a first-order design, includes the low end of the mid-range driver. Jim's 3.7 XO treatment is much more sophisticated and uses better parts than the 2.7 due to budget constraints and designer choices. The 2.7 midrange is fed through a 400uF electrolytic cap, albeit with a PP and styrene bypass in parallel. The 3.7 uses a bank of 75uF polypropylene caps with a styrene bypass. Multiple smaller PP caps provide faster reactions and less distortion than a large electrolytic.

Unsound, the 3.5 equalizer addresses only the bass with a simple, shaped boost centered at 22Hz or 40Hz depending on selection. Our reference set-up during development was bi-amped with identical amplifiers and 4 identical wire runs providing no EQ pollution into the midrange-tweeter circuit. (We were subsequently amazed by how many ways users could screw things up with varying amp and cable configurations working at cross purposes. So the bi-wiring option went away.) The EQ circuitry is elegant enough, but the budget required utilitarian execution (and use of generally inferior interconnects) adding some grain and haze to the signal. Jim considered the ported solution (O2, O4, CS2) to be inferior to sealed-box bass and only grudgingly accepted the market necessity of the passive radiator rather than the equalized bass in aspiring products. We aired the possibility of an EQ for the CS5 (the CS5 followed the equalized 3.5) and we talked about a follow-up super edition with an equalized bass. But Thiel was a one-man development lab experiencing high growth, and there was not time to explore such niceties.

One intriguing reincarnation for a CS5 Super would be to add balanced equalizers to the CS5 bass driving a separate bass input. That bass section has three  woofers in two configurations loaded by two sub-enclosures. All bass frequencies up to 500 or so are covered by that subsystem.     

Unsound et al, Thiel's electronics were all developed and built in-house. Jim was a circuit guy before we took up loudspeakers. His first patent was a sweet phono head amp circuit which we built and shipped for Monster Cable to market. The various equalizers suffered from an identity crisis, being seen as such a valuable and viable solution to us, but being considered a pariah in the marketplace, in no small part via Bose's low-grade application of the idea, and therefore resented by many dealers and consumers as somehow unworthy, therefore limiting the budget to fund exemplary execution.

Regarding the CS5 bass drivers, remember the mid 80s were the Dark Ages. We were designing our drivers and Vifa was making some, such as the CS5 proprietary tweeter. Otherwise, we identified some driver manufacturers to customize appropriate drivers. Those CS5 Kevlar units incorporate Thiel magnet structures and copper shunt rings and long excursions. Those 8" Kevlar cones were extraordinary, much better than their 10" stable-mate, and crossed over superbly to the 5" low midrange. With their extremely long excursions, they moved lots of air. The CS5 bass alignment is different than it looks. The bottom and third subwoofer have mass-loaded cones to lower their natural resonance plus damp the upper breakup mode. The way they straddle the center woofer allows them to create a larger-than-obvious radiation pattern for the 20 to 50 Hz (plus first-order roll out) range. The center woofer is lighter and covers 50 to 400 Hz loaded by its own sub enclosure. Let's just say that drivers are chosen for many interacting reasons. These did the job very well.

As I mentioned, the CS5's Achilles Heel is the low impedance bass load and the delay lines required for proper time-alignment of so many drivers. As I mentioned some time ago, that speaker would do well via incorporating an SS2 powered bass driven through a custom Thiel external passive crossover for the two subwoofers. Above that we might drive the upper woofers via an equalizer and handle the top end with my imaginary driver that mounts a 3.7 coax within a 6.5" wavy cone triax. Let's make all those drivers with carbon fiber diaphragms while we're at it.

Jim's low-impedance choice is indeed a quandary. Even with a 4 to 5 ohm minimum, the voltage sensitivity would not have been too low. If Thiel were starting over today, I think we would settle on higher impedance not only to spare amplifiers, but to reduce cable interactions. Such speculation distinguishes imagination from history. 

Rob, the passive parts quality will leap-frog present Thiel crossovers, plus we'll address other subtle issues.
Thielrules, since you have gear, give it a try. You can see the boost curve in John Atkinson's Stereophile review. That boost is designed to not overload the woofer on normal music content. Be careful with movies or other bombast.I among others here would be interested in what you learn.

Harry, the CS3.5 was introduced in 1988, but its lineage goes straight back to the O3 in 1978, updated to the O3a (equalized) in 1979, O3b and then CS3 in 1983 and the 3.5 in 1988, numbered such as the 5th incarnation of the fundamental "3" design. They all share the same design goal of full-range, accurate in all spheres 10"x 3-way reference monitors. One core value, early on, was sensitivity / efficiency. The O1 and O2 came in at 93+dB @1w/1m. The hard facts of physics relegated efficiency to a lower rung on the survival ladder, which haunted the brand furthermore.

A simple passive part update for the stereo 3.5 EQ might be feasible, depending on how much space is available in the chassis.

Wayne, from the beginning, before the O1 in 1976, we experimented with many wave-form paradigms before settling on the wide-dispersion point source that became synonymous with Thiel. A highly intriguing form was the sound-field created by two back-to-back speakers forming a bi-pole radiator. I have run a pair of CS1.5s backed against my CS2 2s with wonderful results. The back-firing pair can get by with less evolved amplification.

The designer's job becomes extremely easier by using 4th order Linkwitz Reilly crossovers which give you everything except phase coherence at the XO point. They introduce 360° / 1 full cycle, which is considered by some, including someone recently posting on this forum, to be zero. The full-cycle time error can be corrected with digital signal processing, if one is in the digital domain. The "new Thiel" pulled a AD-correction-DA jujitsu. It works except for that last truth of ultimate rightness of doing it right in the analog domain. I suspect the Magico to be doing in some fashion what I described, but I don't know anything about it.

Here's a little wood trip for those sort of folks. There is a wood in Bolivia of two principal species, Machaerium schleroxylon or acutifolium, the former upland and the latter lowland. Since mid twentieth century the Germans have exported it to Europe as Santos Rosewood, named for a port in Southeast Brazil as its "source". Lots of intrigue there, but not today. The Danish "Rosewood" finish has been this wood since the beginning. Not a Dalbergia (rosewood species) it acts and looks a lot like rosewood. In the 1980s I developed a direct Bolivian source including a substantial plantation project with the Chiquitano Natives. They distinguished between the upland and lowland types and I applied the names "Amberwood" to the lighter colored, more contrasty upland type and "Morado" to the darker, more purple, homogenous lowland type. Note that they are both photo-sensitive and lighten toward amber with exposure to ultraviolet. The photo tagged above seems to have a fairly dark stain, which looks like the "Dark Cherry" color, but Morado and Cherry are structurally different. Amberwood generally has no stain.

Jab, your lumber dealer will call the wood Pau Ferro, Bolivian Rosewood or Santos Rosewood. You may find some light colored boards, but the upland variety is more rare. You can bleach the darker wood to match your Amberwood.

Prof, Morado can be quite dark naturally. Thiel generally stained to match a standard sample, so some cabinets may be natural or some stained at varying levels of coverage to match the target.

Pau Ferro (Amberwood / Morado) is an excellent-sounding wood. My main business is supplying tonewood to guitarmakers. We introduced Morado to that world in the mid 90s and it is considered by many to be the sonically-preferred fingerboard stock, as well as highly regarded for acoustic guitar body wood. Like Maple, it has virtually no open pores and is a joy to finish. Be careful, about 15% of people exhibit a poison-ivy-like reaction to the dust. Wear gloves and a mask when working it.

Thielrules, your 3.5 performance is not normal. Perhaps Rob might supply guidance.

As you might surmise, I have been looking closely at the classic models to determine where to best expend resources for upgrades. A sad truth is that the 3.5 is old. Those drivers were modified commercial units and are no longer available. The tweeters and midranges are not rebuildable, but I believe Rob can rebuild the woofers. The 3.5 dates from 1988, so early serial numbers are approaching end-of-service-life for the electrolytic caps in the EQ and XO. Those are replaceable.

After considerable deliberation I have decided to not develop an upgrade kit for the 3.5. There are too many disqualifying factors recommending against spending significant money there. However, Rob at Coherent Source Service can help you keep them in service in stock or enhanced stock form.

Thank you for the reminder. Such a complex puzzle, for something as simple as connecting stuff.

Batman, please remember that I was not involved with THIEL since the mid 1990s, so my observations reflect general approaches rather than specific product knowledge. Jim's approach was very analytical and different from many audio engineers. Jim chose to make the speaker's job that of translating its input signal into the room. A corollary is that preceding links in the chain were also required to do their part. A signal chain must be clean, which is a tall order. Most designers of my experience hedge their bets toward forgiveness. Jim went for truth.  When it came to amps and cable, Jim would measure and listen and determine how well a link, cable in this case, was performing its job transparently. I judge that Jim chose that Goertz wire because he determined that it did a better job of being wire and that silver, being a better conductor did it better, both measurably and aurally . . . and so forth. This determination is different from whether he preferred the sound of silver to copper, flat or round but rather that one was a better wire having less distortions. The lab amp lived under the measuring tower with cable runs of about a meter and the listening room had several highly regarded amps with short cable runs.

FYI, Gary Dayton, Thiel marketing and Jim's lab assistant, is working for Bryston and should have deep knowledge of Thiel-Bryston synergies. I imagine he can be contacted through Bryston or [email protected]

Searching is a sweet part of finding.

Since I was behind the scenes, it is normal for folks to be confused. As Beetle says, I was part of the mix from the beginning. Before co-founding Thiel Audio in the mid 70s with Jim and Kathy plus two more indispensible players, I had been a singer-songwriter-guitarist at the college coffeehouse level, had done some recording and had a producing art-craft studio with Walter Kling, another of the founding members. We made various artifacts of our own design - I made about 500 stringed instruments by the time Thiel Audio usurped my life. Fred Collopy, a college friend of mine joined us as business manager and Kathy stepped up for marketing and we decided to give speakers a try.  The Strata-gee.com interview from earlier this year scratches the surface pretty well.

Regarding Jim's records; they were detailed and extensive. They are missing in action through the change-of-ownership. I hope to recover them, or access to them. I agree, they are foundational for any real progress going forward.

Jay, my songs included Dylan, Leonard Cohen and many of the more insight and art driven folk fare of the time with a focus on my own music and that of our folk trio built around a female vocalist-arranger. We applied the seriousness of youth without any thought of making a living at it. A huge lesson for me was the radical difference between any recording and the live event and that commercially produced music could not be counted on as a reference. Record-making is a deep jungle.

In the long run, I believe the 2.7 (and 1.7) will become collectors' items. They were made in Lexington after Jim's death and before the sale, using old-Thiel type parts (high purity inductors, PP caps with PS bypasses, etc. on point to point boards. The 2.7 uses the 3.7 coax with a smaller woofer for a hypothetically "easier" crossover than the 3.7. I suspect that further tweaking, such as replacing the big electrolytics with PPs might make the 2.7 one for history.

Dan, I have one SS1 and two SS2s (one working) in my studio. For my purposes of verity to the source, I've not heard anything close.

Jay, thanks for the Classe lead. I hope to identify and develop a solid electronics repair solution for EQs, SmartSubs, etc. that fall beyond Rob's grasp. Please keep that in mind for all of us.

Indeed, I came of age in the mid 60s and identified with the counter-culture and the folk resurgence, but from an unusual perspective. I avoided the drug thing and much of the dismissal of tradition. After high school I became a Brother in a teaching-social justice oriented religious order and continued studies in philosophy / social justice / human development until 1974 when I exited a PhD program to open my Conceptions Studio enterprise. My anti Vietnam War work and orientation underlay that time.
I formed a commune of sorts and bought land in the Cumberland Plateau of Kentucky to go "back to the land" and live outside the norms of mainstream society.  Thiel Audio was one of our community-cohesive work ideas that took root and "made it" to the exclusion of everything else. Have you ever ridden a wild horse?    

Ron, I likewise have noticed that except for the likes of us here, Thiel seems rather passé. I expect that attitude to change as we gradually and by degrees create new possibilities for these classic models via upgrade options. A quiet Resurrection is what I envision. I am investing considerable time in the ReXO project.  A full 2.7 XO re-envision is a way down the road, but to the extent it might interest you, there are some upgrades that you could perform yourself or with a handy friend. I guess my point is that rather than taking a beating in a soft market, you might change the terms of engagement.

Dan, I need my SmartSubs for my emerging guitar-making-research-recording-evaluation studio. One SS2 is being repaired. I'll report on the level of success.

350wpc is not too much. Too much really doesn't exist in my book, although there are small-amp aficionados who specifically want the wet sound of overdriven tubes. The amp you listed above @ 250-8 x 375-4 I judge to be inadequate (without ever hearing it.) If an amp doesn't double its output from 8 to 4, it is current limited and therefore producing distortion when driven hard. Thiel speakers drive hard, dropping below 3 ohms over extended ranges. That amp is likely to sound ragged / dry / conjested plus it could blow up drivers at high volume. Look for an amp that doubles from 8 to 4 and makes at least 1.5x its 4 ohm rating at 2.

Dan, if the MacIntosh amps deliver the same value into the three loads, the design is power-supply-current limited and will not do a good job. I have no current (ugh) knowledge of your best choices. Members here are quite well informed about best offerings. I have used Krell, Pass/Threshold, Ayre, Levinson, Conrad-Johnson and old (DR6 & 9) Classe. I have never heard a class D amp that I liked with Thiel. Whoever knows something definitive please speak up.

Jon, your experience coincides with mine and many, many more in my experience. The fact that the amp doesn't double into 4 ohms says that it is out of power as current demand increases. Thiel speakers have low impedance for high current appetite , especially in the bass. So, doubling your available current in the bass is germane. Your blown midrange was almost certainly caused by distortion which heated the midrange voice coil beyond its limits.

As an aside, the only driver failures I have seen have been burned voice coils, caused by overheating, caused by distortion. In all the years of heavy-duty, overnight, stress and show conditions, I am not aware of ever "blowing" a Thiel driver. We had big amps with high current delivery power supplies.

Thosb, As Beetle mentioned, the 2.3 is not a front line candidate because more interested folks have 2.4s which have more evolved drivers. But 2.3s have potential and everything we are learning on other models applies to the 2.3. If you are so inclined, we could help you create a pair of well hot-rodded 2.3s that will definitely address the "tizziness / brightness" to a significant degree. 

Be aware that your room has a significant issue of the length being close to double the width. Your standing waves and timing reflections are likely the significant limitation to performance. Diffusion in the corners will help. If possible, consider orienting on an axis not parallel with any wall to spread out early reflections.

Welcome.

Harry, I have looked long and hard at which models to hot-rod. My choices center on Thiel-developed drivers and a long life horizon going forward. Although the 3.5 was a defining product in its time, it doesn't check enough boxes. The drivers are modifications of commercial units with little of Thiel's proprietary magnetics or cone materials and geometry. Those drivers are becoming obsolete. Same can be said for the equalizer. Redevelopment of a product is a substantial undertaking requiring custom component development and re-voicing via systematic measurement and listening, like when a new product is developed. I expect the results to be worthwhile.

The first-tier models are the PowerPoint1.2, CS1.6, CS2 2, 2.4 and 3.6. Next round includes the 2.7, 3.7, CS6 and 7.2 - in order of expressed interest and personal time availability. First tier hopes to debut before year's end. 

I agree with your "masking" comment, which can be attributed to various factors. Among them is the presence of electrolytic capacitors in the signal path which are at or nearing their end of life expectancy. Since you love the speakers and if you or someone you know is capable, I believe you can make a great improvement yourself by removing the crossovers and replacing the electrolytic caps and resistors with like-values. (Your coils, wire and layout are excellent.) If that prospect interests you, I can point you to brands and sources which have emerged from our upgrade project research. You are welcome to send a PM to explore this avenue.

Harry, I neglected to mention that Rob Gillum of Coherent Source Service is capable of doing the job if you prefer that route.

Regarding impedance, it is what it is. The designer chooses the nominal impedance for fundamental reasons and then optimizes all parameters around those dirver particulars. In particular, Jim needed low inductance drivers for high-end extension and low distortion, which is determined principally by voice coil turn count. Jim believed that if amp designers took their work as seriously as he took his, they would make amps that did their job of doubling their power into halved impedance loads to 2 ohms.

I have expressed a differing opinion here previously. IF the nominal impedance could have been 6 with minimum 4 ohms, the voltage sensitivity would be lower, but more amps could drive the load well. But, that's a big IF, because the laws of physics can not be suspended. So, we have 4 ohm speakers to contend with and the amplifier choice is of fundamental relevance to system performance. 

For the Record: Audio Consultants in the Chicago area was indeed as good a dealer as a manufacturer could ask for. Count on them for straight dealing. In those days of knowledgeable brick and mortar dealers, such questions of system pitfalls and synergy were a big part of the dealers' function. Nobody would be sold an anemic amp with a Thiel speaker and dissatisfaction and driver failures were a rare occurrence. 

Stspur - thanks for your report. I am interested in further sonic developments through time. Let us know what you hear.

Thiel has used the same internal wire since the 03 in 1978, before the time that wire became a component. We identified a high purity, long crystal, oxygen-free polished wire in teflon jacket, developed for the aerospace avionics industry. That wire, twisted at 2.5 per inch does the job in separate runs from input terminals to XO and XO to each driver, with no shared grounds, etc. The wire is a known component of the XO network, so all its factors are measured and listened to as system characteristics.

When ITT / Florida stopped making that wire, StraightWire developed an equally excellent source and Thiel used that exclusively going forward.

This internal wire is a different animal from amp to speaker cable runs which must be engineered without knowing source or load or ambient impedance, capacitance, length of run and so forth.

I have no actual knowledge of Goertz except that Jim used it in the lab as a reference quality cable. The industry works in many ways and each company develops its own value-style. Thiel valued long-term mutually sustaining relationships built on product, company and dealer synergies. In such relationships, the players share feedback and experience for mutual benefit. Thiel and SW worked this way for many decades. So it is no surprise they used each others' products or co-exhibited at shows.

I can't make product recommendations, because my personal experience is thin and very dated.

GS - no definitive answer. Check with Rob G at Coherent Source Service. Photos may be required. Generally speaking, the high mass of that baffle precludes much movement in use. Shipping ? If mine, I would consider an expansive hydraulic grout or cementatious epoxy repair.

Indeed wire is a squirrelly stew. Considerable knowledge exists, much of it in the realm of very high tech-space research-at the fringe of measurement. High end audio "found" wire as a new frontier in the 1970s and some very good work has been done since. There also seems to be considerable borrowing of second-hand knowledge, repackaged with huge mark-ups attached - commonly called snake oil. But most oil has no snake in it and much wire technology is indeed real and crucial for ultimate performance. To the material dynamics of bare wire, add reactance of coatings and influence of geometry, and you see that complexity increases exponentially. We know much less than we don't know.

Exploration of these frontiers adds to the richness of our journey.

Presently I am comparing coil configurations as fed from the circumference or the core. The oscillation of the electromagnetic fields behaves differently in each layout. Opposing opinions exist among intelligent and qualified observers. Most say "it doesn't matter", which I interpret as "I don't wanna go there." Who knows where this road goes?

I second jonandfamily's request for direct experience with the soft dome for the Thiel UltraTweeter.

Unsound - the SCSs seem fairly invisible here. My slight experience might help a little. The SCS is essentially a grown-up O2 - 6.5" ported two-way with polarity correct second order crossovers. They produce a decent step response and frequency response and can handle more power more easily than first order XOs. The SCS designation began with Thiel's first coincident / concentric driver in 1989. The specs and general performance of any generation will be similar with more modern iterations having more serious driver technologies and improved port loading, with more sophisticated performance.
I have 3 pair of O2s and a single SCS4. Side by side, the SCS4 has greater detail and nuance, but its overall  musical impression is fairly similar to the O2 from 1976. As models progressed, the ports migrated from single to dual to rounded outlet to flared horn outlet which gradually reduced port air velocity and chuffing, creating better room coupling.

The ear-brain does an interesting thing with the fundamental frequencies. Any distortion there serves to slur the upper harmonics - the auditory brain interprets all sound from the bottom up. I am working with a very exciting technology, modeled on shark skin, that organizes the wave launch from the port, thereby cleaning up the entire spectrum to an unexpected degree. Not ready to show, but will upgrade any ported speaker, including Thiel O2, O4, SCS (all), CS1, 1.2, 1.6 and CS2. Stay tuned.

FYI: I picked up a single SCS4 for a couple hundred bucks and looking for another one to make a pair.