Thiel Owners

anzen - lots of interacting variables and solution visions. Among those is de-coupling which Thiel rejected. Thiel’s vision landed on maximizing the connection between the driver and cabinet to reduce recoil, increase effective mass and stiffness to tighten the onset transient wave-form. Early-on we used a thin gasket-paper interface. As we developed precision CNC machining we eliminated that gasket by using an O-ring at the ID of the mounting rabbet as an air seal. The mechanical joint was driver to baffle with no interface. It seems that later (after TT) products migrated to a very thin foam interface that essentially collapses under mechanical load while forming an air seal. In all cases Thiel attempted to unify the driver with the baffle for maximum rigidity.

The opposite approach of ’floating’ the driver with its claims of resonance isolation may have merit, but I never found it in total system analysis. The decoupled driver vibrates more and longer than the coupled one and its recoil is far greater - producing transient slur. I should note that such slur is without much consequence when a speaker’s onset transient is already compromised via non-coherence. But a coherent waveform begs the brain to figure out what’s wrong with the slurred leading edge, and that 'work' decreases musical intimacy.

Another problem with de-coupled drivers is keeping them tight. A vibration-absorbing gasket permits relative motion between the basket and baffle. That motion induces loosening of the fasteners - a very bad deal.

So, Thiel unitized (as much as we could) the vibro-mechanical system of drivers and cabinet. The cabinet-floor interface is another very complex equation to solve. Stand by for future reports on that front.

A cable note worth mentioning (again?) regards binding posts. Going back to the 2002 CS1.6, Thiel used gold plated brass binding posts with a big knurled knob. I could speculate about how that came to be, but I won't. I will say you can achieve considerable sonic improvement by removing the knob and using banana plugs.

The slightly opaque, clacky quality, especially noticeable on the high end will vanish. Easy, inexpensive, worthwhile.

A complication is the center hole being slightly undersized so all males plugs do not fit. I re-bore with a 4mm / 5/32" drill. I use the WBT-style angled, stackable, expanding male ends, and make contact in DeOxIT contact grease.

Much is often made of the large surface, high pressure connection of spades. My exploration led to examining field propagation along the wire run and its scrambling at a bulky binding post. A low bulk, straight connection does everything right, and an expanding plug makes good contact. Try it.

When you say 'nude termination' I take it to mean the cable is self-terminated with no additional hardware. Good move. But now to the speaker input terminals. Whenever there is a cross-sectional excess of conductor in the signal path, eddy currents are induced along with other forms of non-linear flow. Those induced reactive fields present erratic impedance to the input signal. I suggest you put some straight or 45° banana plugs on the ends of your SP7s and fit them to the terminal center holes. Set the nut aside. (The actual post is OK.) Compare. Report.

The clones I described above are gold plated brass from parts express. Genuine WBT multiplies the price. 

We first discovered eddy currents in the steel baskets of the O series. When we auditioned the phase coherent version we heard weirdness, but the phase-correct / time incorrect third order slopes 'erased' the weirdness. We thought it was mechanical / vibrational, but all the epoxy we could pile on the basket didn't help. It turned out to be magneto electrical eddie. Aluminum baskets cured it. Later we tried aluminum voice coil bobbins for their thermal conduction. Eddies were purely electrical, but problematic nonetheless.

This wavelaunch stuff I've been working with lately is a manifestation of the same effects in the realm of surface propagation and boundary layer turbulence. And, of course, the binding posts. Jeff Roland first brought that problem to my attention. They had put huge output posts on their amps seeking better performance and a marketing ploy. They failed sonically. 

Yep, same file cabinet.

unsound - no doubt you're right. Jim gravitated to the Geortz from StraightWire. I think it was Tony Cordesman or John Atkinson who borrowed those Goertz for their reviews and considered them an improvement.

I haven't gone there myself because I prefer to stay in the middle lane with ancillary products that are likely to be in widespread use among Thiel owners and fans. However, I would love to audition some flat-ribbon cables for comparison.

My recent wire excursions have me cross-checking driving speakers with and without crossover networks. The SCS4 is quite well behaved with no XO at all. And, guess what, cables interact differently with bare drivers vs networks.

How would you characterize 'their very obvious differences'?

Your description of the Maestro sounds very sophisticated. I don't see it in their present offerings. Steve has told me that they no longer use individually coated wires due to centering difficulty, expense, etc. Their present core technology is to compress the bare copper wires to maximize and unify contact, then insulating those groupings and applying their various cable geometries. My StraightWire is way down the line at Octave which utilizes their core compression technology, their insulation mix, geometry, etc. He characterized the Octave II (now III) as most of their available performance before climbing the cost ladder of 'designer sound'.

I don't remember which SW cable we had at Thiel when I left. I do know that the music room was outfitted with the Kimber Black Pearl when I auditioned the 3.7 vs nearly finalized 2.7 in 2012. Those cables are astounding - at $5 figures.

Kimber discontinued the Black Pearl due to time consumption and production requirements that couldn't be sustained at any asking price. Ray had gifted a first-run pair to us around 1980 and later traded them for a last-run finalized version. He, of course, used our speakers in his development work.

I was honored and pleased to evaluate the finalized CS2.7 even though I was not involved in its development. I saw the careful and sustained work of Team Thiel to bring that product to market in Jim's honor.

Rob says 1500 pair.

The first brigade of New Thiel had some good talent. And the new owners were trying out building on Jim's legacy. In particular Steve DeFuria had been an early Thiel retailer and long-time inner circle supporter. Steve became the first national sales manager and was able to re-populate the dealer base by decommissioning Crutchfield and other sincere promises. Those 1500 pairs were made entirely at Thiel Lexington with PtoP boards and best historic practices. They even got the (new at that time) Klippel Quality Control system, but I don't think they learned its ropes before changing horses with the next management team. Five new teams in five years to try to please a very difficult principal owner. $ten million spent to scuttle Jim's legacy.

Yes, Rob definitely had a role. Home Team Thiel managed the executive design and manufacturing engineering and execution. Principal players were Kathy Gornik, Dawn, Rob, Gary Dayton and probably some others. Thiel in 2012 was a skeleton crew and all hands believed that bringing this post-Jim product to market was essential to any kind of future. The drivers already existed in the 3.7 coax and 2.4 woofer. Technical crossover engineering was done by Warkwyn in Toronto, which resembled pulling teeth to get an authentic ’Thiel’ product from an outside firm with ideas of their own. Indeed the CS2.7 was part of what attracted a new buyer . . . which could have carried on the company, but was not a good fit and things went south. But the product exists as a testament to Home Team Thiel’s dogged determination. (Also, there is some low-hanging fruit for improvement.)

Rob - I have likewise not heard of any high 2.7 serial numbers. Rob gave me the 1500 pair figure and I take it on his word. I do know that by that time Thiel’s export markets were stronger than domestic. We’ll see what surfaces as time goes on.

Yes, that 400uF electrolytic cap in the midrange feed is not Jim Thiel - never done. As a fly on the wall, I gathered that there were ongoing struggles between Home Team Thiel and Warkwyn who ’knew best’. New ’finalized’ samples would come from Toronto with persisting failures to honor such requests. When their tab hit $250K, Kathy pulled the plug and began production. New Thiel didn’t deem it worth the trouble to make future changes.

Regarding 3.7 vs 2.7, I see almost every value changed for the 2.7. Circuitry layouts are similar in that the same basic resonances are being compensated. But there are significant differences and no shared component values between the three models' crossovers.

I’ve heard the models 3.7 and 2.7 only once at Thiel’s music room September 2012. I was overjoyed by the performance of both, especially the 3.7. What a Swan Song.

anzen - are you posting on Micah Sheveloff's Thiel Legacy site or somewhere different? I'd like to see your layout before responding.

Biannuzzi22 27june23

"I am seeing a lot of chatter about upgrading the CS2.7 speakers. Can anyone send me instructions on how to do this? In particular, it looks like using the Clarity Purity 1uf 800v capacitors is a good upgrade? Where are these located and how would I replace what is currently on the crossover?"

Crossover upgrades are, at this time, conversational rather than commercial. I hope to eventually have plans and kits, but not yet. The CS2.7 is the final Jim Thiel-esque product, developed after his death, but by the original in-house team with outsourced engineering help, and built in Lexington with traditional classic methods. It qualifies in my thinking as genuine and worthy of upgrade.

Clarity’s Purity800-S is their hands-down winner. Its internal circuitry is more sophisticated than the other PUR caps, plus the thick copper end caps and silver wire refinements. All that said, I find that replacing the stock Thiel styrene/tin - film/foil 1uF ’yellow’ cap with the PUR800 1uF is a lot of money spent on marginal improvement. I suggest you rather might consider unifying the split/bypass cap into a single PUR800-S where most appropriate (eliminating the bypass.) Your best prospect is the tweeter series feed which is now comprised of a 15uF PP and the 1uF yellow bypass. Replace them both with a single 16uF PUR800-S. (That leaves potential for a more sophisticated bypass, in the works, at a later time.)

The weakest link in the 2.7XO is the 400uF electrolytic cap in the midrange feed. I am working on an all-film solution which is not yet ready. My first focus is on early products to keep them up and running. New midranges and tweeters will bring later Thiel driver technologies and longevity all the way to the early designs. As various caps are finalized I’ll announce here. Stay tuned – a lot is presently coming together - I anticipate real progress this year.

Rob - good to see you making steady progress, and eagerly await your report on that parallel tweeter cap.

Regarding coils - bigger is not better (despite conventional opinion.) A significant consideration is the saturation frequency rf wire size. 18 gauge saturates at 17kHz above which turbulence results in distortion, heat and related instability. I consider 18 gauge the best compromise for woofers and OK for midrange with smaller being better and more so for the tweeter. That 20 gauge midrange feed coil is not an accident or compromise. Note the trimming resistors that follow it, so less coil resistance would not be an advantage. Au contraire. The dissipation factor of a coil is far greater than a resistor.

The 20 gauge tweeter feed coil was probably the best compromise between its required inductance and a resistance value that matches an available resistor value. I know that Jim would have preferred a 22 or higher gauge there if practical.

Also note that Thiel inductors were best-of-form from 1978’s model 03 onward. The 2.7 ERSE inductors directly trace their lineage all the way back to 1978 when Thiel introduced ’good wire’ to Acousta-Coil for the 03. At that time what is now CDA101 (highest grade) wire was an aerospace-only product developed for the Jupiter Deep Space Probe by ITT. The rest is history which I recounted in these pages years ago. I don’t think you’ll find better wire or winding, and thicker gauge would degrade performance - and make Jim squirm.

igmaty - the CS2.3 was introduced december 1997 and had a minor XO tweak November 1998. Thiel’s serial numbers always began with #1 and 1/2 were the first pair, etc. My notes say about 2750 pairs were made, so numbers would top out about 5500. The coax was updated some time in the product life which necessitated adding an 8.5 ohm resistor and 3uF cap. That change came February 2001. Try Rob Gillum about the particulars of how to identify if you have early or late 2.3s. Once you get oriented, there is substantial / qualitative sonic improvement to be had via relatively simple crossover component substitutions.

First get them up to speed and have fun listening.

The early 3.7 coax had variable tweeter amplitude which required individual trimming resistors. The problem was solved and the ‘late’ drivers were painted black to differentiate between n inventory. Blacks are indeed identical to ‘correct’ chrome rings. 

Guys - about those 2.7/3.7 coaxes.

The black ones are the original chrome drivers that have been painted black. They are chrome under the black paint. They were painted to differentiate them from the early drivers which had erratic tweeter output levels which required trimming resistors to bring them into spec. When the tweeter level was solved, the 'finalized' drivers were painted.

If your chrome driver has no trim resistor, it is the same as the black ones. If it has one, then you might consider replacing that resistor with a higher quality resistor of the same value.

Of course it is possible that your old driver is not performing optimally, in which case having it rebuilt or replaced would upgrade performance. If Rob were to install a 'new' tweeter needing no trim, I bet you could get him to paint the frame black.

Pull the driver and if the wires go directly to the terminals, it is 'new'. If there is an electronic component on one of the terminals it is 'old'.

Rob hasn't mentioned any serial number record, etc. Many records were 'lost' under New Thiel.

JA - concerning the workspace. No room for multiple speakers, so no longer a 'hotrod garage' - more like a lab/studio. I've spent half-days working here since April with real progress.

One surprise is that capacitor (and all components to a lesser degree) directionality is far more important than I thought. Thiel didn't consider this factor in production. I find directionality more efficacious than higher-cost parts. My present method of determination and optimization is very cumbersome. I'm working toward an easier method. Considerable upgrade potential with your same parts!

I wholeheartedly endorse replacing your stock Thiel resistors with Mills MRA12s. Short money, real improvement.

Wire is a rabbit-warren way deeper than expected. Serious work in that department for two years. I (right now) have a solution that is very satisfying. It's another matter to get it manufactured . . . another story still being written. My present sample development method may amuse you. Imagine a bracket near the ceiling to which stuff can be clamped. Imagine a miniature 'diving board' clamped to it with some array of holes in a circle to accept wires / fiber cords, at various spacings and radii. Imagine a similar set of holes in a hand-held puck. Imagine a center non-conductive cord fastened to the ceiling bracket and stretched tightly to a turnbuckle on the floor - a vertical tight-rope. That core and/or other separating cords are fixed to the diving board and through the puck which is hand-twisted to determine the cable lay and packing density. The various wire and cord components are typically cut 44" long which wind up to a meter long. Those 1 meter setups are installed between a pair of RCA plugs with screw terminations. Placed between my Benchmark DAC3 and HPA4, their attributes can be readily heard and their FuzzMeasure sweeps show differences that correlate with listening. The test method is time-consuming and tedious - making a meter of cable takes an hour or two. But in a half-day session I can make and document an idea. Rinse and repeat. Winners are long enough to cut down to woofer and tweeter runs (two different cables) in a pair of SCS4s.

A significant part of my work is less-technical. Suitable wire comes from many approaches. Partners here who have upgraded their speakers have (I believe) always chosen to upgrade from the stock Thiel 18-2 solid in teflon. I have tested many/most of the wires they used. My determination is that their differences from stock gain something and lose something and change the fundamental characteristic sound signature - which I am unwilling to do. I must say that the cables out there are works of art, intellectual - technical - musical. If anyone tells you 'wire doesn't matter', be suspicious of whatever else they may say.

My work has surveyed a broad range of approaches and I owe a debt of gratitude to the manufacturers who have provided samples, knowledge, and special insight. My approach has been from the perspective of Applied Physics rather than engineering and I thank my friend, neighbor and MIT operative for references and insights from beyond the usual perspectives. This has been a very long, deep journey - the deepest dive of my lifetime. It wouldn't have happened without the Thiel community, which I have accessed here on this forum. 

There is more work to do, but this pair of SCS4s, in the works for around a year, is about ready to head to California for Fluid Dynamics wavelaunch technologies. That rabbit-hole is another world of its own.

Off to the lab/studio -

jonandfamily - about the CS1.2

I have a crossover layout. Feel free to PM me .

Tom

pablohoney - my time at Thiel ended in 1995. I don't know what changed and what stayed the same. Rob Gillum at CoherentSourceService would know. But Rob is very busy and doesn't necessarily respond to non-transactional queries. I'll post a finishing narrative here shortly.

Pablohoney

Birdseye Maple finish

Wood finishing is an art of its own. Thiel took its cabinetry and finishing quite seriously, far beyond normal utilitarian protection of the wood. In Thiel's first two decades under my direction, we built cabinets that were fully mirror images of each other. Our wood stable included dozens of species – basically everything that anyone would request.

Our approach was to let the woods' natural color and grain shine through, with mild exceptions. Those included toning unusual flitches of a species toward the norm as well as producing the 'expected shade' for others such as Mahogany and Cherry which have a long history of particular darker-than-natural shades which approximate how those woods age through time when exposed to natural ultraviolet rays. But all in, let's call Thiel's finishes 'natural'.

The colorants we used were dyes rather than pigments. Dyes are very small molecules of color for shifting tone without obfuscation, whereas pigments are quite large – resulting in a more opaque covering effect. I note that later finishes such as the signature Red Birdseye on the 2.4SE, and the very red Morado on my CS1.6s, as well as the 'Dark Cherry' on the SCS4s are all pigments. They obscure the underlying grain and are easier to apply more consistently than dye stains.

Back to natural finishes. An important aspect of a natural finish is that it enhances the natural characteristics of the underlying wood. A good natural finish does not visually sit on the surface. Many water-based and other plastic finishes appear as a separate film over the wood. They have very large molecular structures compared to the underlying wood and their refractive index (how the incoming light is bent) is quite different from wood. Nitrocellulose lacquer and shellac, exhibit refractive indexes closest to wood. Our production finish was a pre-catalyzed nitrocellulose lacquer with its first-coat sealer having additives for ease of sanding. This finish handles and acts like lacquer, but continues to get stronger after solvent evaporation allows molecular cross-linking. In our custom shop, we often used shellac as the sealer for its subtle range of coloration and old-school visual appeal. Our industrial finishes were custom blended with safer solvents and premium working and durability qualities. But a close working approximation on the open market would be Sherwin-Williams Sherwood Catalyzed (precat) Lacquer over its branded sealer or shellac.

On to the particulars of Birdseye Maple. Birdseye is an anomaly which acts and looks differently when viewed from the tree's exterior or interior. Therefore if your panels contain mirror-image bookmatches, you will have 'innie' craters on one half and 'outie' craters on the other half. To avoid large reflective differences that can make one half look darker or lighter than the other, good practice is to apply a penetrating and a wash coat to the panels to plug the tubular pores in the wood. The first penetrating coat is sealer cut to 33% viscosity and the second wash coat is cut to 66%. The third full viscosity sealer coat is sanded level to accept the topcoats. If you are darkening the maple, it is best to use clear first and second penetrating coats and add color to the third full viscosity sealer as well as early topcoats if needed. Staining bare Maple is risky due to unpredictable color striking over differing micro-grain structures. The final topcoat should best be clear. Our production finish was 1 sealer and 2 topcoats, unless more was needed for color or grain needs. Note that more thinner coats produces less orange-peel than fewer thicker coats. You don't want orange-peel.

A pro tip learned from my musical instrument making is that you can level a final finish without the laborious rub-out process like you see on high-end guitars, etc. Allow your final topcoat to flash off dry to the touch (5± minutes depending on temperature). Then briskly and firmly rub the finish with a soft cotton tee-shirt, etc. The finish is soft enough to lay down under this pressure for a lovely soft, level gloss. You can soften the gloss further by using regular builders' fiberglass first, and following with the cotton knit if desired. Practice on scrap. If you are making a showpiece plinth, I suspect you might like what you get for the extra effort. Such showpieces often get a coat of carnauba wax for a soft feel and look. Thiel did not wax its finishes, but some users did.

I've tried to choke the fire-hose enough to not overwhelm while supplying enough content to chew on. Let us see your results when they come.

pablohoney -

As fate would have it, our original ’O’ series used Danish Oil as a first coat to seal and warm the tone. That went out with the 1983 CS3.

Our sealer or shellac are clear, which in shellac land is called ’white’, unless you want some color to match your aged veneers, etc. Then there are many shellac shades available beginning with ’blond’, all the way to ’dark amber’ and ’orange’. Begin with the thinned clear wash coats like I described. Just wade in.

I advise against polyurethane as that chemical family has much larger molecules and different light refractive index that I disparaged above. Try for lacquer to get a ’Thiel’ look.

Bill exemplifies a dying breed - someone who knows musical technologies with fluency coupled to the requirements of the musical arts. He will be missed, and I don't know if he'll be replaced.

Bill has had the Thiel CS3 / 3.5 equalizer in his backlog, but alas, has run out of time to address it. But he did do his last Oppo 105 Music Technologies upgrade for me. It is wonderful and serves as my CD and SACD source at considerably higher level than either the stock Sony 9000ES or SCD-1 from Thiel Audio's sale.

I had met Bill in the way-back days when he was the manufacturing director of Conrad Johnson in their first two decades. But we lost touch on my hiatus from audio land. I really appreciate you recommending him to me; we have had a few years of enjoyable, productive interactions. I wish it could go on - All things must pass.

I highly recommend anything with Bill's signature on it.

FWIW: These serial numbers place this pair in Lexington-Thiel production with USA AcoustaCoil coils, best-spec wire, and Solen polypropylene caps, on point to point masonite boards. We have seen that later Chinese production had lower spec parts including polyester caps on PC boards and variable coil quality. If you want a pair, this one would be a good one to get.

solobone22 - isn't it lovely when you dial in your speakers to the room?

I've never used any sub except Thiel's with the matching Passive XO. I learned a few things. The 'normal' input engages the room boundary circuitry, whereas (I think) the LFE does not. That boundary interface is magic. The bottom end of the 2 series is 4th order reflex, which is optimal for matching a 4th order low-pass / high cut on the sub. The 2.7's -3dB point is stated as 23Hz (can that be true?).

I'm speculating here, but since there is so little musical energy below 30Hz, I wonder how it might work with the 2.7 running free and the REL selected for as low as you can, even if above the speaker low limit.

I do hope that folks with greater sub knowledge and experience check in. You can use REW to graph the actual system output from 0 Hz. So at least you won't get lost.

catalysis - I totally agree about Benchmark. I run a full BM setup and couldn't be happier. Some folks quibble about BM's lack of lower harmonic distortion. BM intentionally practically eliminates it, while many users like the added 'warmth' and 'fullness' that the HD adds to the signal. I agree that Jim would have loved this gear. It brings truth without costing a fortune.

To your question about progress and kits. The question is valid and the answer is complex. I am working every day on upgrades, and making real progress. A somewhat unexpected twist (which should have been obvious) is that Jim's mastery of value engineering was phenomenal. It takes quite a bit of money to break through to a higher performance/cost level, which turns the idea of upgrading on its head.

Several folks here have upgraded their speakers, and all have been satisfied . . . at considerable cost - which was OEM special buy with no mark-up. Looking at real costs with overhead, the cost/performance equation gets murky. Other interesting developments include that every (?) rebuilder chose to use 'better' internal wire - and were pleased. I have tested all of those wires and plenty more. I am not satisfied with any of them as preserving the unaltered, pure signal. So, i went down a very deep and wide rabbit hole. The result is a proprietary wiring harness that breaks new ground and settles some subtle but persistent sonic issues.

There are other solutions being brought to bear. Analog layout, field effects, thermal management and wavelaunch propagation are all being advanced. All in, I am achieving up to 20dB lower harmonic distortion and other background 'noise' (collectively referred to as 'computational irreducibles') caused by dynamic interactions in the complex speaker system. There really is too much to itemize and summarize, and much of the work is proprietary. I can assure you that we will achieve significant performance improvements while changing nothing of Jim's work, and in fact extending his approach and methods toward qualitatively greater fidelity. Much like the Benhmark gear sounds so good by getting it so right.

The vision has shifted from 'upgrade kits' toward development of novel technologies which do 'it' better for less. The goal is to offer these technologies and finance driver redevelopment and improvements - for a very long life horizon for these heirloom speakers. This autumn I plan to exhibit these technologies to a group of invited guests in a purpose-built playback room, and court more team formation.

FWIW: I'm picking up a pair of CS5i on this trip. I now have 5 pairs of CS3/3.5s. I'm concentrating on early / older models because their drivers are unavailable and aging. Their replacements will be a new iteration of the UltraTweeter (used in the CS5, 3.6 and 2.2) along with a new iteration of the CS3.6 midrange to retrofit into the CS3/3.5. We have a good working relationship with ScanSpeak which has all the working drawings from our development partnership with them. The equalizer has been expertly and impressively upgraded, but the finalization of that job was on Bill Thalmann's plate as he succumbed to illness. We're scouting another electronics designer for that project. The workhorse model shared with my collaborators is the SCS4. If any of you want a Small Coherent Bookshelf / Monitor, that will be the first speaker to be available at various upgrade levels. I am truly encouraged by our improvements.

I recognize that you all expected more progress sooner. I'm playing the hand I'm dealt the best I can. Progress is slow, but I am encouraged by the results to date.

jafant - I think so. There are lots of elements in place. I'll post an update summarizing what's been learned and where we're going. Our collaborative workhorse has become the SCS4. So that'll be the first out of the gate. If you want an extremely good bookshelf speaker, or one that's been modified for stand mounting, get yourself a pair of these critters. I don't think there's another phase/time coherent small monitor out there.

More to come.

pablohoney - nice rack.

Regarding color, Maple does darken with time. Exposure to UV is the key. There are UV lamps. Sunlight has tons more than any bulb. Good work.

We're in the same place with 7.2s. I bought a pair awhile back, but haven't fired it up yet (lots of reasons). I did hear a pair at Rob's a few years ago. Extremely good. Also, the company line is that Jim considered the 3.7 to be 'better' and therefore discontinued the 7.2, etc. In fact the 3.7 breaks new ground and is the statement of his art. However, the 7.2 is not a lesser speaker. Those speakers were phased out to simplify the operation to give it some chance of surviving his death. I'll be interested in hearing your comparison between your 3.6 and 7.2 when that happens.

thoft - about drivers. I'll prepare a more detailed list but here's a general arc.

Early (0 series) were all essentially stock drivers with small choices such as dust cap, surround type, etc.

The 1983 CS3 had a customized Vifa woofer. Very interactive development, but stock parts. Through the 80s that process became more and more custom. The 1990 CS2.2 had all custom drivers, co-developed with and made by Vifa.

Early 90s we developed our own capability which was used for rapid prototyping. Vifa made them from parts and technologies we developed. Those Vifa catalog numbers can be bought and/or easily cross-referenced.

Our needs outgrew Vifa's (and anybody else's) willingness to produce our eccentric designs. 1995 onward we began making our own drivers in-house. CS6 woofer was (I think) the first all-thiel designed and manufactured driver. All the coaxes (after the 1990 SCS) were in-house Thiel design and manufacturing.

In the 2000s, some of those in-house coaxes were made by FST / China. Same Thiel design, off-shore source. (SCS, MCS and PowerPoint coaxes). The '.7' star plane and radial wave drivers were designed and prototyped in Lexington, but (as far as I know) all made by FST.

That's the broad-brush overview. I'll compile and follow up with a list if you wish.

JA - my news may sound like a skipping record, but from here it sounds good. The lab has only the SCS4s as the test mule due to tight space. Other models are in storage. I'm picking up a  CS5i pair this month. The SCS4s have focused my efforts and increased productivity. The big, general lesson has been Jim's brilliance as a value engineer (in addition to his inventiveness.) A lot of funds can be spent on Thiel speaker upgrades to get very little musical upgrade and/or even sonic downgrade in some aspects. My task is to keep every ounce of Jim's solutions while adding qualitative upgrades at defensible cost. Real progress is being made.

In particular, I have developed a hookup wire solution that keeps the solid simplicity of Thiel's traditional solid 18-2 in Teflon while further augmenting performance. This wire gig has been a very deep dive and the result adds high end air and ease as well as full range increase in clarity and dimensionality. Presently looking for a manufacturer willing to cope with its particulars.

I'm filling my prototype iteration needs via hand-spinning with a shop-built contraption. Starting to hot-rod / repurpose an old-school spinning wheel to make first generation samples. This wire is also being tested in inductors, which may add value without the down-sides of foil coils. Stay tuned.

Lots of capacitor round-robin has led to our own custom caps by Reliable. Trials have been definitive in favor of their 'segmented' (Multi-Cap) technology. The segment ratios represent new work. First small run for 4 pairs of SCS4s is expected next week for beta review by collaborators.

The surface wavelaunch and port airflow technology keeps getting better. I'll get next generation samples this coming week for trials.

Some cabinet stiffening schemes are getting mature. Both internal and external treatments will diminish chatter and raise vibrational frequencies to substantially reduce their impact.

Later this month I'm looking forward to visiting a collaborator who has built a purpose-built listening room with my design input. We'll be tuning it and making future plans before the end of the year.

It's been very slow going, but a corner has been turned and real goods and products are taking form.

Here's a readout of the CS2 tribute.

CS2 Chapter in Thiel History

The CS2 began life in 1984 after the game-changing introduction of the CS3 in 1983.

The CS3 was the 4^th iteration of the model 3 – equalized sealed 10” 3-way with bass response to 20Hz. It demanded a fairly large room, a very robust woofer, and a midrange to cover 7+ octaves including in and out ramps. The model 3 spent significant budget on that very capable woofer and the active equalizer to take it so low. Its right price was considerably higher than Jim wanted to charge, and many of its virtues were not needed by many listeners.

The model 4 filled smaller spaces with a bass limit in the mid 40s from a ported 6.5” two-way floor-stander. That format became the CS1 series.

Thiel needed something for smaller rooms with less demanding bass at significantly lower cost than the CS3 while providing better performance than the 04/CS1. The CS2 was born from that need. From the beginning its identity included trimming costs without sacrificing performance beyond bass extension. In fact, its midrange could be cleaner since its crosspoint came in at 800Hz rather than the model 3's 400Hz.

By this time Thiel had established a strong working relationship with Vifa, who co-developed a 3.5” full-range driver for our CS2 midrange needs. Vifa and Jim co-developed the woofer with some of his emerging motor geometries and techniques, even though it was still a conventional overhung motor design. A reflex woofer costs about half of a boosted sealed woofer because its low-frequency linear excursion requirement stops at the port tuning.

In 1984 all cabinet work was still conventional tablesaw work along with our newly acquired inverted router. The CS3 baffle was being sculpted with hand tools, at considerable cost. We developed the routable CS2 grille board as a wave-guide and diffraction control mechanism with considerable success, at very low cost compared to the CS3 baffle. Later when we bonded the grille to the baffle with rubber tape, it became even more effective. The port cost nearly nothing compared to the $ multi-hundred equalizer, which audiophiles wanted to be more transparent (requiring higher cost.) The CS2 load was a very resistive 6 ohms minimum, and moderate 87dB sensitivity, making it extraordinarily easy to drive. We focused our collective energies into cost-effectively producing this low-cost / nearly full range, coherent source for smaller spaces. It was our first real hit.

The introductory price was $1350/ pair against the CS3's (insufficient) $1950. It met its market and sold consistently well. Its 1991 replacement CS2.2 was driven by our developing CNC capability for a more sophisticated cabinet to support new driver technologies first developed for the 1988 CS5.

The CS2 served as a sophisticated, elegant entry-level speaker for a broad audience. It sold about 7500 pairs over its 7 years, the most of any Thiel model.

jchussey - the CS3 EQ is approximately the same as the CS3.5. You could substitute them. The CS3 EQ starts a little higher at about 100Hz rather than 80Hz for the 3.5 and maxes out a little higher at about 26Hz rather than 20Hz for the 3.5.

Look at the CS3 Stereophile review to see the curve. The response without the EQ will be the mirror image of that boost curve drawn below the 0 axis.

The CS3 and 3.5 are identical in intent. The tweeters are the same, the woofers are the same except the 3.5 cone is coated for a smoother response. Rob has that coating that you could apply. The midrange is the big difference. The original SEAS wasn't up to the task and overloaded in the lower mid / upper bass and sometimes failed. The ScanSpeak 3.5 mid was far more robust and broader range. Note the sculpted baffle edges of your CS3s are true conical sections, which are a bit 'better' than the single radius curves of the 3.5. 

I am treating the 2 models the same, they will get the same new tweeter and midrange built on the CS3.6. And the EQ will be substantially better. It is prototyped and sounding good.

JA - congratulations on this forum and on the likes of unsound and pops chiming in after all this time. 2024 is bringing some tangible fruition. Long, slow trudge, with so many lessons learned and alliances made. I thank all of you who have taken on your various tasks and contributed your knowledge and shared your outcomes.

My best wishes to all. Tom

@beetlemaina - good to see you. Your nod to Cardas led to their inclusion in this winding journey, where I ended with their 'chassis wire', which I consider a masterpiece of innovative engineering. Straightwire, Iconoclast, Morrow, Kimber, Fishman and MIT physics have also contributed knowledge, expertise, samples and guidance.

Interesting that all of you beta rebuilders have replaced Thiel stock wire, all being pleased. I agree with your perceived need. It would take a book to sift through the particulars; but suffice to say that my solution contains multiple elements of new art and addresses the shortcomings of Thiel's wire without sacrificing an iota of its intention and successes. I'm sorry that I can't expound.

My guiding lights include correctness and cost-effectiveness, which doesn't translate as rejecting better if too expensive. My cost-effective solution is better than anything I've experienced for my purposes of correct transmission of audio signal. Yes, I'm pleased.

One of my development input streams is my recording. I record acoustic music in real time and space, very simply and directly. A pair of mics flat to 50kHz on a separating disc of Golden Proportion and non-diffractive / non-reflective surfaces. The 8' cable pair feeds a SoundDevices recorder at 32bit/192kHz without any protection, manipulation, or post processing. The SSD recording is the mic feed. Comparing and evaluating recordings with various cables (the only cables in the chain) has been extremely helpful in an extremely complex maze. I am confident that you will at least enjoy and I hope support my cables.

After the better part of a year of rejections, I have found an american cable manufacturer willing to take on the particulars required. We're getting started now and will have beta samples within the foreseeable future (God willing and the creek don't rise.)  We'll announce here. 

@unsound - I find it more cognitively dissonant in the used / classic speaker market than it was in the new / heyday speaker market for a speaker to require extremely specialized and expensive amplification in order to work, even reasonably well. Of course there are users with relatively small rooms who listen at relatively moderate amplitude and who by good research or luck have a moderately priced amp that can drive later-career Thiels. But there are more who don't and plenty of driver damage done by less-than-stable amps. Yes there were real problems winding underhung voice coils with enough turns to raise driver impedance. But there was also a look-the-other way attitude about less than ideal amps. Note that Jim's recommended amps were always priced at a substantial $ multiplier of a pair of Thiel speakers. 

kheine - Yes.

There's a long story about the bottom inputs, but we can summarize that my design goal of furniture elegance without electronic clutter, has been less than popular.

Here are some considerations for your move that include the fruit of my recent redevelopment work. I believe that the 'wire as waveguide' construct is more functional than 'wire as electron conductor'. Sound quality will be improved more by managing propagation fields than by making lower resistance connections. I now land firmly on banana plugs rather than spades or bare wire. I also remove any excess conductive mass in the signal propagation path to reduce eddy currents. Therefore I judge your 3.6 plastic capped binding posts as sonically superior to the later large metal lugs and knobs. But greater improvement can be made by streamlining the signal path further by repositioning the inputs away from the electrically conductive input panel and using a straight-line, a low mass terminal system directly through the cabinet (bottom or) back. There is something to be said for positioning them at the back edge of the bottom, facing down. (I suspect you would rather not.)

I highly recommend the Electra connectors from GR Research. In fact, as a stand-alone improvement I suspect their ise will bring the largest SQ improvement / cost of anything you can do to your speakers. You would mount the pair at the bottom / back of the cabinet, right through the MDF panel (bottom or back). You can gain sufficient access through the woofer and the existing input panel.

Let us hear about your outcome.

I glad you get it and like it. That was the idea, but we're in the minority.    FWIW, I have done installations where the cable was routed under the floor and through a  drilled a hole through the floor  - voila - wireless speaker.

biannuzzi22 - I'm not a 2.3 expert, but I might shed some historic light around it.

The CS2.3 was the first-generation Thiel-designed coax. The 1990 SCS used a stock ScanSpeak 6.5" coax which solved Thiel's fundamental problem of working properly only for a seated listener 8' or more away. The coax upper driver was then applied to higher priced Thiels. By 1996 enough experience had been gained to produce a cost-effective coax. The model 2 was the chosen platform because its use in rooms smaller than for the model 3 and up made the model 2 a good candidate. Note, the model 1 never got a coax for cost reasons. The brainstorm was to create a mechanical crossover between the voice coil and midrange cone based on resilience/ viscosity to decouple the midrange cone from the directly driven tweeter dome - in just the right way. The product could only be created because we were prototyping all our drivers in-house by that time. It took more than a year of iteration to get it right. It worked, but with shortcomings that Jim addressed in the subsequent 2003 CS2.4. The 2.4 is considered the sweet spot of that design.

The bucking magnet was added in 2001, #4567 to manage the home theater use near televisions. The 2.3 coax was less expensive to produce than the 2.4, so it continued its career in the PCS and MCS models.

Jim continued work on the wrinkles of the mechanical coupling and had hoped to use an improved version in the CS7.3. A major element of that driver is the obviation of any electronic crossover between the midrange and tweeter. The joint crossover is devised to drive one electrical motor with compound mechanical elements covering the combined range.

Note: both the woofer and coax crossover sections are quite different between the models implying significant differences between the drivers. They are not interchangeable between the models. Your 2.3s are what they are. Any differences from the 2.4 would be speculation on my part. I have 2.4s and can't believe the wonderful sound quality for the price. That niche was in the DNA of the model 2 from the start.

JA - can you post a link to Abe's tribute. I can't find his tribute to Bill.

jafant - yes, we can read it. Thank you.

It's always sobering when a friend, or associate of any kind, vanishes. I feel fortunate that you re-introduced me to Bill 5 or so years ago. Bill and I knew each other as co-exhibitors at audio shows, but otherwise had no involvement. When you recommended him to upgrade my classic Classé gear it was a great fit. That led to Bill becoming part of the nascent Thiel Renaissance dream. Bill worked on the CS3-3.5 equalizer upgrade, specifically conversion to two separate mono fully balanced units powered by a single outboard power supply. Bill's declining health put that project way on his back burner and I'll be picking up the pieces on my next trip to Virginia.

Bill was a very special person on many fronts. Regarding audio, he would bring his accumulated big bag of special insight and experience to bear on whatever the project at hand. The results were always good. I am pleased to have his final Oppo 105 upgrade. It plays any/all disc formats in elegant form.

Parkinson's and all degenerative diseases seem especially cruel. Bill faced his journey with characteristic upbeat good will and good work. We'll miss him.

vair68robert - I susepect it's they guy who makes aftermarket 'fixes' for various speakers. when I spoke with him a few years ago he said his units 'fixed known abberations in the Thiel's upper midrange'. Rob Gillum bought one to measure and said he was reducing amplitude across the upper midrange and introducing phase-shift. Parts quality was significantly below Thiel's standards. These units are 'pre-crossovers', not replacements. They pre-shape the signal before it goes to the internal crossover.

Again I'm speculating what's up, could be something else.

Some thoughts related to CS3.7 ’brightness’.

Brightness, or excessive upper midrange energy is a common complaint about Thiel speakers. Knocking those frequencies down will reduce the perceived problem, but I’d like to add some perspective to that solution.

I know intimately how Jim developed and voiced his speakers. He compared and correlated ear, impulses, sweeps and pink noise using floor-standing, aerially suspended, and ground-plane techniques. He achieved his goal of less than 1/10th dB / octave-averaged power of differences. I believe Thiel speakers are flat. That judgement runs contrary to many user’s experience. Following are some thoughts about that discrepancy.

Jim’s design listening position is off-axis via speakers pointing straight ahead. Lots of irregularities are reduced or eliminated at the resultant 25-30° off axis listening position. Note that most listeners, nearly universally, toe the speakers in perhaps half that amount. They then get more high frequency amplitude and more discrepancy between drivers than a properly (as designed) positioned listener.

So why do most listeners prefer listening closer to on-axis than designed? One reason might be that Thiel speakers have unusually wide dispersion patterns. That is by design - approximating how a real sound source would radiate into the listening environment. Look at the off-axis plots of Thiel speakers and you’ll see only a few dB of high frequency droop with very similar response curves to the on axis response. That wide dispersion provides better in-room power response, better stereo imaging and better phase coherence. But it comes at the cost of more side-wall reflected energy than most speakers. I suggest trying straight-ahead orientation while adding sonic absorbancy at the side wall first reflection point. My experience is that the reflection is causing the perceived problem and that absorbing the reflection presents a better solution than reducing the amplitude of the offending frequencies.

I have only heard 3.7s once, in 2012 in Thiel’s listening room, powered by the Krell FPB-600. That room measures 14’ high x 22’ wide x 36’ long, with 1.5" thick wall panels covering perhaps 1/4 - 1/3 of the wall surfaces and wall to wall wool carpet on hair pad. Room is of medium vibrancy, not dead, not live. The 3.6s sounded magnificent without a trace of brightness.

Not everybody has such a room. My point is that rooms make tons of difference which are best addressed via room / acoustic solutions rather than loudspeaker modifications.

I have been wrestling with other causes of the persistent complaints of high-frequency less than best-ness. I have identified (and previously mentioned) wavefront propagation anomalies that when mitigated, control a sonic instability perceived mostly in the high frequencies. The propagation turbulence is actually full-spectrum and my newly emerging acoustic wavefront propagation solution produces a full bandwidth uniquely settled and natural sound.

Another situation is that Jim chose aluminum driver diaphragms for lots of good reasons. However the tweeters breakup harshly around 25kHz. Even though that is above (generally accepted) hearing limits, that breakup mode couples with some wire anomalies to produce less than best transparency in the highs. Thiel’s solid18-2 in teflon from the 1978 model 03 was reconfirmed for the 1988 CS5 and again for the 2006 CS3.7. I like it a lot compared with other contenders. But there is a problem with skin-depth saturation beginning at 17kHz which interacts with the instability of that tweeter breakup mode. My new cable uses two smaller gauge conductors per leg to raise the saturation frequency to 42kHz. The chronic Thiel high frequency blockiness dissolves into grace and air. My assessment is that any perceived high frequency ’brightness’ will vanish without reducing amplitude via a series resistor.

I realize that this new cable solution has been a long time in the pipeline. It's now in its fourth generation prototype, and the design is settled. My 3rd generation hand loom is being built which allows small scale prototypes for beta co-developers and critical feedback. I believe it will be worth the substantial wait.