Thiel Owners

Rob Gillum at Coherent Source Service might know. And some ebay, etc. resellers sometimes salvage old gear. Also, Madisound was still supplying some parts in the mid 80s and may have a clue. It was probably made (I don't remember for certain) by Vifa. But a stock driver will not work. The motor was custom designed and everything else was specified by Thiel. Normal woofers don't have the smooth high-end rollout needed for first order slopes.

Guys - I just snagged a beater pair of MCS1s for the hot rod garage.
Tom

Subwoofers are inherently expensive, especially at Vandy / Thiel levels of execution.

Hypothesis: The ear-brain builds its sonic interpretation from the bottom up. It infers harmonic structure and actually creates a 'phantom fundamental' if none is present in the reproduced waveform. My opinion is that the listener can be better off creating his own phantom fundamental than interpolating a scrambled waveform.

But the choices are never that simple. It's always some mixed bag.
Enter amp interaction. As mentioned above, the amp can be drained and strained by deep bass demands, especially into Thielesque low impedance loads, especially as in the CS5. An obvious remedy is bi-amping' where one channel drives the bass and the other the highs.

Thiel stepped away from the bi solution for various reasons, among which was over-all system cost-effectiveness. The customer buys 2 amps. Another was amp and/or cable mix and match - there are many ways to screw up the sound and the almost-adept-enough dealer / audiophile is skilled at doing so. We determined that the straight-forward, vastly more safe, simple and less expensive route of single inputs was the best way forward, especially given Thiel's high performance per price niche.

At this time I am exploring various bi options while providing a high-quality single input option of jumper straps of solid high purity copper with conductive grease.

At one show before the advent of surround, we positioned 2 pairs of 3.5s in the normal stereo position - back to back. There was about 5' between the back-firing pair and the wall and plenty of side wall space as well. The effect is magnificent. The bass wave forms a quasi spherical wave front and the transition to upper frequency in-fill is seamless. If you like bi-polar presentation, try it. That's if you have a spare pair of 3.5s, or your main speaker of choice, available.

Gary - you must find a 7.2 woofer solution because all those owners including yourself can't be left stranded. Good luck with Rob.

The 7.2 is on my wish list of eventual upgrades. But the only information I have is the owners manual and product reviews. I would appreciate any information anyone might have.

Thanks, Tom

I know that there is some polarity jumble out there. In the 1980s early CD period, the BBC did a study that determined "most people" preferred aggressive sounds (trumpet, drums, etc.) to be in reverse polarity so that the cone moves inward on its initial transient attack.

In today's world of large track counts and greater reliance on recording technologies, it seems that more records are made in proper polarity and don't require any diddling.   

esprits4s - mass always tries to sag, and these speakers are about 30 years old. Before incurring expense, I suggest turning all 6 woofers top for bottom. The offending woofer might take some time to re-center in its gap, but it could fix itself.

Rob - if boredom continues, find a pink noise source - equal energy per octave - and experiment with ear height at 8' to 10' from one speaker. Find the height where it just sounds right and deteriorates if you move up or down.

If your listening ear height is different, you can brush up on your trig and compensate the spikes. Have fun and let us know how it goes.

Sam - my cap of choice is ClarityCap CSA with potential CMR bypasses. My resistor choice is Mills MRA-12 which is nearly best of form for reasonable cost. Our work is still in process.

Main benefits from upgrade include higher resolution, micro-dynamics and transparancy, which are already Thiel's strengths. I have not yet addressed the 2.7, so have no firm advice except that replacing all the sandcast resistors with MRAs will increase delicacy and refinement.

Andy - If you are replacing a few parts and they are accessible, you can snip the lead wires and twist your component leads to those stubs - and then solder.

Remember that if you are using lead solder to be very careful with smoke extraction - lead accumulates and causes brain and nerve damage. Thiel solder is 96-4 / silver-tin, but it is much harder to work than lead solder such as Cardas eutectic. You decide, but do be careful.

Sam, your experience is consistent with mine during 50 years with myriad amps and signal chains. More power, especially current, works better. One insight is that current lag induces phase lag. In most speakers phase lag is not audible, since phase is already compromised and the ear-brain is reconstructing - inferring the wave-form. But minimum phase transducers eliminate that fore brain engagement (because the wave-form is not phase compromised), and therefore any phase anomalies are much more apparent. Similarly holco's GD upgrade is more evident on Thiel (or other such as V'steen) minimum phase speakers.

I'm speaking from hard experience. Upstream transparency was the single most critical aspect of success or failure of Thiel auditions. The speaker always gets blamed for any musical failing. But if you think about it, the speaker is critically translating its input signal including phase response, which is why it can sound so good when the signal is right.

Sam - sorry for the dense language. The insight that is somewhat obscure is on the listener-cognition side of the equation. Phase coherence lifts a veil that allows the listener to believe the sound is real, and therefore the sound is subjected to a much higher level of scrutiny. So the garbage sounds worse on Thiels. 

Andy - I can share information, but can't send drawings until Thiel Audio's bankruptcy is settled.

You are correct; the 2.7 and 3.7 have 3 separate drivers with electrical crossovers, unlike the 2.4 which is electrically 2-way plus the mechanical mid-tweeter xo.

There is no time compensation in those crossovers. The coincident alignment is achieved geometrically via the position of the coaxial midrange and tweeter.

If I ever get my recapped Classe's back from the shop, I'll report. DR6 pre and pair of DR9 power that I've had since  the late 80s. 

And it may interest you that Dave was Thiel's national sales manager from the late 80s to 90s. He moved his family to Lexington to help Thiel navigate their rapid growth into compatible and excellent dealers that he knew from ARC and Magnepan. He eventually went back home to ARC, and everyone was better for the experience.

Happy New Year everyone. This forum has been a focusing inspiration for me taking on this upgrade project. This project has served to reorient my work life more toward music, which has always been important to me, but has often take a back-seat to other initiatives. So thank you.

Beetle and I are making progress and should have something to report pretty soon.

Sqmlaw - the DR8 and 9 are quite similar except for the power output. If you like the amps and if they don't need recapping, you are better off than most music lovers. Power requirement rises dramatically for large rooms and loud levels. If it sounds good to you, you may be within the requirements of your room and levels. Enjoy.

I would call a 20' x 20' room large. Square can cause standing wave problems, but openings on 3 sides helps that a lot. No real delineations apply for "large" or "small"; the 'how much power is enough' is an ongoing unknown. But swapping amps is a big deal; but reports are consistent that more is better, all else equal, which it never is . . .  As you all know, Thiel speakers are low impedance, requiring lots of current. However, that low impedance is gracefully resistive, which minimizes the problems. Classé DR amps are true current source output stages, so they match Thiel very well.  All that said, bridged /mono configuration is less friendly to low impedance loads than normal mode. So, it's a soup, and I suspect your soup tastes pretty good.

Additionally I am exploring bi-amping and believe that we can get better performance from those two amps via split normal mode than from bridged mono mode. Since most solid state amps can drive half the impedance in normal mode relative to bridged mono, and since Thiel loads are low impedance, my hypothesis, my experience in general and with the CS3 and house 3.5s suggests that (vertically) splitting a normal-mode amp between bass and upper registers delivers the best of all performance. You have two appropriate amps. In your shoes, I would explore vertical bi-amping rather than swapping those amps for bigger ones.

Beetlemania will soon test that hypothesis with his bi-ampable hot-rod 2.4s. Also, I will be outfitting my PPs, 2.2 and 3.6s for comparative experiments as time allows, since I have 4 channels of Classé DR9 in my rig.

Pops - What is the power rating of the Macs? Do you use the 4 ohm tap?

The CA200 is rated at 200 into 8 ohms and 500 into 2 ohms, so it is current limited, which interests me. Have you used it in both stereo and bridged mono? 

What is your ceiling height on that great room?

Room sounds really nice. That smallest floor to ceiling dimension is a significant limiting factor to normal playback rooms. Good work with your ceiling treatment.

I'll post when I have some feedback regarding bridged mono vs vertical bi-amping performance.

I am likewise eager to hear people's experiences. Please cite your amp particulars; this performance would be related to low-impedance behavior.

2.2 and 3.6 have same tweeter.

Barry - Your speakers can probably be serviced by long-time Thiel service manager Rob Gillum now of Coherent Source Service in Lexington.

The O3a is second generation O3 which put Thiel on the map. The A generation dealt with baffle edge diffraction via wool felt absorption - very effective but somewhat homely-looking. Your cabinets are made of Finland-Ply (high-end Baltic Birch), similar to the latest CS3.7 and 2.7s curved ply panels. You also got the first dose of styrene ultra-bypass caps. With the equalizer, you can get sub 30Hz sealed box bass from that heavy-duty custom Eminance woofer. The O3b added some driver refinements and more precise time alignment. The CS3 incorporated a curved-edge baffle for geometric rather than absorptive diffraction control and an all new driver complement. The CS3.5 (5th generation Model 3) still stands an an epitome the design family. If parts were readily available, which they are not, I would like to hot-rod the CS3.5 as a tribute to its place in the development of high-end audio. 

The point is that each successive model generation applied new ideas, techniques and technologies to stand on the shoulders of its predecessors. The O3a is a true minimum phase, minimum diffraction, extended range Thiel speaker. It is probable Rob can keep them on the road and their performance will embarrass many contemporary speakers.

Barry - to your question, each model (3) keeps its overall character and intent. The differences are in refinement of execution of that intent.

I have some thoughts on The Listening Room. Don Hoatson was an early Thiel adopter, a pioneer from his opening in 1978. He was going strong with Thiel when I left in 1995 when I lost touch. His rooms were indeed great sounding due to their dimensions, construction, and that he used commercial horse-hair pad under the wool carpets to take the bottom octave absorption about an octave below any of the foam underpadding  products. Don died in 2016. 

This morning I spoke with his grandson, Michael, who grew up in Don's business and listening rooms. Mike has taken over and moved the business to Chestertown on Maryland's Eastern Shore where he has been open one year, is growing and expects to honor everything his grandfather taught him. He is staking his future on customers like you guys.

Don was an electrical engineer as is his son Dave, Michael's father.

Guys, can any of you help me with some basic information re the CS1.6?  I just bought a pair for the hot rod garage, but won't see them for a couple of weeks - I've never seen a pair. I need to know the dimensions of the slot / port. How wide is the slot and how deep from the inside / back to the start of the flare? Is the inside edge eased / rounded in any way? 

Any information would be helpful in developing some ideas around upgrading port performance. Thank you in advance.

Fan - Very helpful in deconstructing some performance differences between the round and slot reflex system. When we're finished with the 1.6, I believe we'll have something quite worth auditioning.

I vote for cables as short as you can make them. And then however they're made doesn't matter as much.

Andy - internal wire is definitely a factor. It carries a slightly different set of requirements from external cable, which must cope with many unknowns of run length, electromagnetic environment, speaker impedance fluctuation, and more. The internal wire is engineerable to the known requirements of the speaker where it is installed.

Thiel wire is very good. In fact, Thiel "found" wire as a design variable in the development of the O3 in 1978 and introduced aerospace wire to the industry. All models of Thiel wire is, and has always been since that time, 99.9999% pure, low oxygen, long crystal copper developed originally by ITT for the space project, and since then cloned by many makers around the world and now improved by continuous casting and other advanced technologies. I'm guessing that Thiel wire trumps all but some of the highest cost and/or DIY speakers out there. 

Although various configurations are appropriate for hookup wire, Jim chose 18 gauge solid in teflon @ 2.5 twists per inch as a well considered judgement of maximum versatility. Economy of scale was a major concern - good wire is very expensive and big order quantities reduced cost, so all hookup wire is the same, and therefore might stand upgrading via specific-use engineering. More on that in a minute. Beyond hookup wire, the same quality criteria apply to coils. The signal meets hundreds of feet of series and shunt coils inside the speaker. Thiel uses 6-9s copper coils, tightly wound and oven baked. Great coils. In Thiel's and  third party testing, our wire and coils are best-in-world performers.

However, it's never that simple. In the present XO upgrade project I have learned that later speakers including 2.4's were supplied with Chinese-made crossovers. The copper (and other parts) is supposedly a clone of Thiel Copper, but it is not certified and my research causes me to doubt the claim. The coils are obviously lower grade manufacture as learned from Beetlemania's work on his 2.4 upgrade project. The resistances and topologies of the wire and coils are all good, in fact the coil values have been tweaked to account for inductive changes with Printed Circuit Board mounting. The XO performs properly, but sound quality probably suffers a little, as ascertained by Thiel insiders.

My plan is to evaluate the source of each XO encountered to decide whether to keep or replace particular parts. Beetle's late 2.4s have every element replaced. My 2.2s keep all coils and French-film PolyPropylene caps and German Styrene x Tin foil bypasses. My PowerPoints include a Lex and a China pair. Lex are better and will retain some parts. No China parts are being reused. It's a riddle which I am systematically solving.

Another element is wire gauge. Lower resistance is the primary reason for larger wire. However, the overall circuit resistance from input terminal to driver lug is measured and accounted for in the XO, so resistance considerations become nearly insignificant. Our upgrades are increasing some (woofer) hookup wire diameters for greater ion flow optimization. Coil gauge is only changed where required for foil cap upgrade availability. Contrary to some opinions, the small gauge shaping coils are superior to larger gauges, because they can be tuned to their parallel resistor for balanced resonant circuit performance.

I am accounting for wire and coil resistance changes, which may sometimes require a resistor value tweak for proper system performance. Project process is being made. We'll have more to report when the Thiel Audio bankruptcy settles.

Beetle - that is correct, the twisted pair wire I sent you was sourced from StraightWire, which is an industrial as well as OEM and consumer source. As you mentioned, we will be comparing that wire in various gauges against other contenders when that time comes.

Sgmlaw - indeed electrolytics are an issue; they fail over time, drastically and sometimes catastrophically at the end, but also by decrease of value over time, as you say. ELs show up in 6 figure speakers, it's expensive to avoid them.

So, a 30 year old speaker XO has almost surely drifted out of spec, as well as in some level of danger of failure. Rob says he has never seen a failed Thiel cap so far. Thiel used VersaTronix and ERSE, both in their long-life versions with a predicted life of 30 to 50 years. All are bypassed in some way, which extends main cap life more so.

All that said, I am replacing ALL electrolytics to create an effectively "permanent" XO, comparing cost / performance at this time. The likely first-tier upgrade will use ERSE metalized polypropylenes @ highest voltage that will fit, bypassed or not depending on function. Next-tier upgrades will be to a custom ClarityCap CSA - expensive, but stunningly high performing - again, bypassed or not with Cornell Dubilier PPs or RTX styrene film and tin foil, depending.

Early indications sound like we'll be playing a league or two up from original Thiel. The view is getting clearer. 

All - regarding upgrade pricing - there will be multiple plateaus for each model which are self-defining as the project proceeds. You know it is possible to spend more on crossover parts than the entire original speaker system parts cost. I'm not going there. In other words, I will not be putting Duelunds and Paths in a 2.4, even if an individual DIY owner might go that far.

My work is to determine the hard limits of format, cabinet and drivers, and then determine an ultimate cost-feasible implementation of that format considering traditional Thiel values. Beetlemania's CS2.4s are at or close to the highest imagined implementation tier. We must learn whether its $4 figure cost will be sonically justified.

There seems to be a middle plateau upgrade at half the parts cost of that. Brands have been chosen, but much remains to be decided.

The first tier upgrade will be a step up from stock in that sand-cast resistors will be replaced with Mills MRAs and all electrolytics will be replaced with polypropylenes. ERSE MPX caps are an outstanding bargain for that use. In this and all tiers, any questionable wire & coils will be replaced with 6-9s as in Lexington Thiels.

I would like to address design style. Many designers take many approaches to this subtle work - a high level of art is involved. I am honoring Jim's approach of reason-based, experimentally and musically verified choice. In Jim's approach, every element or change or choice must BOTH improve the sound and align more closely with the measured technical ideal. That might seem obvious, but it is far from common. Many designers mix and match various euphonic anomalies to arrive at a pleasant end result. A corollary of Jim's approach is that each component must be pure. A capacitor is a capacitor, etc. Again, that approach is far from the norm.

An example of Jim's straightforward style is wire. Wire is enormously complex in how it behaves and how it sounds. With the help of Ted and GE aerospace avionics, Jim decided on his wire configuration. That 6-9s, polished surface, teflon jacket twisted pair sounds great and acts predictably like an ideal, engineerable wire. Over the years he investigated other types and brands, but kept our wire because it does what good wire should do and doesn't introduce spurious anomalies (which some folks might prefer.)

Holco- your wire idea is fine. Its identical geometry and jacket will not alter the measured system functioning. Your diameters are similar to my calculations for upgrades. Silver is a better conductor than copper, but coming at a hefty cost. Silver also imparts a different sonic signature, which some people like. I am very interested in your feedback should you choose to make this investment.

From my perspective that departure, both in value engineering and sonic signature, rules it out of consideration. We could make a sonic contributions list, and near the top would be inductors/coils. The short silver hookup runs attach to hundreds of feet of coil conductors. My reason-based choice dictates Thiel copper, same as the coils. My further investigations for hookup wire include litzed, graduated, multi-strand high purity copper - because theoretical considerations suggest its possible superiority. It would also have to improve sonics, and fall in my affordability value plateau structure. 

Right now I have 4 wire configurations awaiting evaluation and testing. Tier One upgrade keeps traditional Thiel wire because it withstood decades of testing and comparison. Other tiers may or may not change. Further work. Input welcome.

Holco - I am very interested in your outcome reports. I have never experimented with such configurations, but considerable dynamic electromagnetic fields are at work in the speaker, beyond my understanding of their inner relationships.

I am unclear what elements would be grounded in the speaker, since the + and - leads are both active in the AC / signal feed. Are you envisioning some sort of additional shield in the cabinet? 

So, it seems you may be grounding the driver frames?
Thiel drivers since CS2.2 in 1990 use Faraday networks / shorting rings in the voice coil / pole piece structure which should cancel stray fields. But, stray residual eddy currents are possibly in play.  Please report your findings.

Rules - I appreciate and concur with your concerns. You would be pleased to know the level of verification we employed at Thiel. Serious science.

But there is this other dimension. In my acoustic guitar design and archival recording I have identified an accumulation factor. Recording engineers and piano techs and other technical artists also experience this phenomenon. When on a particular path of exploration, guided by both cognition and intuition, there are many choices which are not provable or even discernible. But a conglomerate effect becomes identifiable / meaningful over time. There are so many subtle factors contributing to the overall result, that each of them could be ignored or over-ruled, but they can matter in their aggregate.

Andy - you're on it. My study over the decades has taught me that listening is far more active and synthetic than we would assume. Auditory input is pretty sketchy in that sound pressure moving the auditory cilia must be fundamentally interpreted for meaning. That interpretation occurs in many parts of the brain and is associated with many different functions of memory, emotion and cognitive processing. We're making up most of what we hear.

Sound can be broken into two processing categories much like light acts as photon particles and waves. The wave aspect of sound relates to the frequency domain of pitch and timbre. The particle aspect is the time domain. The gating mechanisms you reference have more to do with time than frequency. Temporal propagation occurs in real, interpretable space. Any sound, such as a finger snap, arrives at the listening pair of ears with time information that allows us to know what it is as well as where it comes from, including its reflective and absorptive environment. As you allude, the processing power of the auditory brain would be overloaded without organizing mechanisms. One such mechanism is the time threshold, generally considered about 5 milliseconds. Components within that 5ms envelope are conflated into the original sound, while those arriving afterward are treated as reflections/echos. Of note is that those sub 5ms components are perceived as slurred or de-focused when the various frequencies of the arrival transient cannot be combined into a sensible single event. A real acoustic sound source (the finger snap) arrives with all (frequencies of) transients intact and its reflections off the nearby boundaries also intact. The analytical fore-brain figures out / decides the nature of the source (the snap) and the particulars of the walls of the room. And we are very good at it, being necessary for survival.

Trouble comes when aspects of the transient event have been compromised by the reproduction process. Though there are many opportunities to compromise this transient information, the most pervasive is that scrambling introduced by non time/phase coherent loudspeakers where various frequency bands arrive at the ear at different times than a real-life intact signal would. In that case, the auditory brain must analyze the sonic elements and synthesize an opinion of its nature (finger snap). It must also repeat that analysis for each reflection. Those additional layers of decoding are processor-intensive and serve to distance the whole listener from the heard experience. One fine twist is that the more sophisticated the listener, the more he tolerates / succeeds at the cognitive process of figuring out what is being heard. Therefore I trust the aural impressions of non-sophisticated listeners simply because they are in closer contact with the whole, natural auditory experience, whereas the sophisticated listener can "overlook" the deficiencies of a temporally inaccurate sound because his skill enables him to "hear" it despite its shortcomings. Teenage girls are my first choice for test listeners.

Sgmlaw - We were aware of the Bell Labs research, which concurred with Jon Dahlquist's exclusive use of 18 solid in the DQ10. We compared 18 solid with 18 stranded and the solid won hands-down. Cousin Ted concurred and we went with 18.

In my present work, I am considering ganged runs of 18 solid for mid and/or woofer where the larger cross-section would increase instantaneous ampacity. Also in the mix is a various strand gauge litzed solid bundle with 15 gauge aggregate size.

Straight wire does have effects beyond resistance. I am keeping my layouts linear around a central wire corridor to minimize antenna effects and stray field losses. I have various wires to compare - in the mi are single, double or triple twisted 18s depending on use.

To clarify, in the last paragraph of my previous post I was referring to "wire", not the brand StraightWire. I meant to say that all wire has many effects and artifacts beyond resistance, capacitance and inductance. Most of those spurious artifacts exhibit in the time domain with frequency related phase shifts and distortions.

That said, the brand StraightWire was Thiel's long term wire source. My summer of 1978 was spent researching wire, which led to the ITT aerospace 18/2 twisted in teflon and same-source coil wire. As time went by, that technology and non-military market was transferred to another Florida company, and StraightWire became their distributor. Thiel became StraightWire's customer sometime in the mid 80s. One of the partners spun off to form WireWorld and Thiel sometimes used their cable products at shows along with various other brands. But the industrial wire from the factory continued to be supplied to a broad market by StraightWire. The wire I got from Rob earlier this year was that same StraightWire. The wire in the Chinese Thiel crossovers looks the same, but I have no way to verify its particulars.

Andy - my first-choice idea for multiple runs is to twist the bunch (2 or 3) 18 ga varnished coil wires having essentially no dialectric. That wire-twine is insulated with a cotton sleeve (very low dielectric absorption) if physically necessary - and/or physically kept from contacting its opposite polarity mate or other conductor. Driver runs will require spacers so that the ± twine-pair can be twisted to reduce capacitance. I am looking at teflon donuts for that purpose.

Much of the insulation requirement for speaker cables and interconnects is due to handling, being stepped on and so forth. Inside a speaker cabinet is a safer, more predictable environment.

I share the surprise of learning of Chinese sourced components. I am gradually making sense of the history and can speak more readily after the Thiel bankruptcy closes.

I will say that Jim's decisions were careful and considered. I know that FST, the Chinese manufacturing source, was chosen for Thiel's Home Theater product components in the mid 90s and that Jim worked with them in his last years on the 3.7 drivers, which could not have been developed any other way.

So, even though I and many others judge Chinese components less worthy, the facts may not support that opinion. However, my own research and observations lead me to return to Thiel's traditional use of best-of-form sources which are generally European, Japanese and/or American. So much subtle sophistication goes into creating the very best of anything, that I don't trust even an ISO 9001 spec unless I verify the results for myself.

I have learned that the CYC caps in the 3.7 use a Chinese film claimed to be equivalent to the best. Solen and most other very sophisticated brands use it also. I believe from trustworthy input that that film underperforms the best-in-world Japanese and Danish film, from which Thiel's prior custom caps were wound. I am returning to known best of form suppliers even if the Chinese suppliers might be as good. I would have to prove it for myself.

In the case of wire, Beetlemania's 2.4 coils were from the Chinese source with same claims as above. I can't compare wire alloy or manufacture methods and specs, but I can say B's 2.4 coils were visibly inferior in wind quality. Thiel's known ERSE and Jantzen coils are my choice. And so forth and so on. We only know what we know. We don't know what rigor or trade-offs were applied to late Thiel products, especially after Jim's death, but even before that. I saw a photo of Jim showing off a CS3.7 crossover with all CYC caps.

For my part, I plan to create iterations that best serve the musical capabilities of the design intent.

Beetle - thanks also from me for your steady, careful implementation, especially when backing out of blind allies.

You all may be interested to know that the 03, 03a/b, CS3, CS3.5 and CS2 had styrene micro bypass caps (0.01 and 0.015uF), similar to (but not as good as) the RTX bypasses. They were dropped @ CS2.2 and CS3.6 and onward. Those earlier models also used the French polypropylene Solens @ 400 volt and German polystyrene x tin foil 1uF, both best in world at the time. Those companies ceased making film, and replacement caps gradually became more ordinary. I suspect that some of the love for the CS3.5 might be attributable to those pristine caps and double bypasses.

In this project we are using best of form caps that exceed anything Thiel had ever used, including those described above.  Thanks again to Beetle for making various controlled comparisons. I will be making more.

sgmlaw - your Hafler story represents the approach of a wide range of designers, and practically the entire DIY community. I will offer a different perspective which included Nelson Pass, John Dunlavy, Benchmark's John Siau and Jim Thiel to name a few.

That Thiel approach says that in a situation where an inferior part creates a better sounding outcome - that is evidence of another problem. Something else is going on which has not yet been identified.

Your Hafler experience happened routinely in Thiel product development. As a matter of process, we took that outccome back to the lab to figure out what was happening. Usually that "fix" was masking something much like dither noise masks or randomizes artifacts of digitization. Our MO was that a solution had to both sound better and measure better or our work wasn't finished.

I recognize that I am speaking anathema to many or even most of you. But I offer it as a signpost along the Thiel road.

I have commented before, but an additional comment might be in order.That is that terminology often obscures the discussion. Some of that obfuscation is purposeful, some is accidental. It may help to insert the word ’polarity’ when it fits. Polarity, here, is the direction the driver moves when fed a signal. A 6-volt dc (battery) signal is appropriate. With plus to plus, the driver(s) should move out into the room. Coherent speakers do that, most don’t. A Wilson, KEF, etc. will leap-frog such that the woofer comes out, midrange goes in, tweeter comes out, etc. Those speakers are polarity-incorrect. Conversely, when all drivers come out, many companies call that ’phase correct’, despite any other phase or timing anomalies. I would call that ’all out’ condition ’polarity correct’. But most designers and critics consider either behavior as OK, whether all drivers move out, or leapfrog up the array, irrespective of many phase and time anomalies. That’s because the ear-brain can reassemble the intended wave-form, and do it quite well. I believe that many listeners, including pros, actually enjoy the mental gymnastics required to reassemble the waveforms. Consider that the BBC ruled that if an individual person can distinguish positive or negative air pressure (polarity inversion), the preference-judgement is personal. Their ’research’ demonstrated that a majority of their subjects preferred negative polarity, which would make a drum hit (for instance) suck rather than blow. The leading edge would be a vacuum whereas it was a pressure wave from the real drum. The BBC deemed that the negative pressure attack was more polite and acceptable. - preferred by more (British) listeners.

In that light, the typical designer generally strives for ’listenability’, that polite, acceptable presentation, which is quite often not what the microphone ’heard’ or the recording stream produced. (Deeper discussion deferred that many ’modern’ recordings invert various polarities to ’fill the mix’.) But I am addressing the speaker reproducing its input signal. Wilson (as example) inverts polarity at each driver exchange (crossover point.) Good engineering executes the hand-off between drivers with smooth phase transitions. The absence of abrupt glitches gives the ear-brain no hard evidence of trouble. And most design styles, companies and critics call that victory.

In contrast Jim Thiel, Richard Vandersteen (and a few other oddballs) chose to preserve the phase-time information intact. Many of you guys appreciate that, most people do not. When phase is kept intact AND the drivers are aligned so that their leading edge transients all reach the ear simultaneously (time-alignment), we call that Coherence. I notice that today the C word usually means ’smooth phase transitions’. rather than our assignment of ’integrated waveform’ period. First order roll-offs (including electrical and acoustic elements) sum to produce zero phase shift. (One driver leads by the same amount that the other lags, such that at the design listening distance, they sum to produce no shift.) Add physically equidistant sound sources, and you get an actual representation of the input signal representing the recorded sound, with no need for the brain to descramble the phase and time information. Thiel and Vandersteen decided that goal of authenticity was worth all the difficulty of making it right. The industry at large does not consider that element of fidelity to be important, or important enough to warrant its difficulties.

I suggest pulling out ’phase’, ’polarity’ and ’time’ in trying to understand the landscape. Richard Hardesty’s journal has been cited here. I consider his clarity and teaching style to be stellar.

JA - I suspect you haven't gotten satisfaction re your impedance / phase query. Technical study and understanding underlies amp loads. Here's a superficial sketch: Impedance (AC resistance) can be mechanical, electrical, inertial, etc. Voltage is the driving force. Current flows during motion / work. Phase shift occurs when current (work) lags or leads voltage (motive force). That situation exacerbates amplifier difficulties. Changing impedance and phase angles cause amp distress which is especially troublesome when reaching for max current delivery ie low impedance load.

You are correct that those top-tier, and probably all, designers know how their amp responds to 2 or 1 ohm loads under various phase angles. If that performance is exceptional or even good, they publish it. If not, they suggest its un-importance by their neglect.

Test specs can be misleading on many fronts. User success with a design proves its worth.

JA - how do we access desmond's project. Or, can you get him to report here?

Andy - thanks for the clarification. The C is for Copper in CSA and the CSA is reported to be a big step better than the SA. One day I will directly compare the SAs from the 2.4SE with the new CSAs.

"I have used the Clarity CSA (the older version, not the latest one with their newer Copper Technology), Mundorf Supreme cap, and Jantzen Z-Silver."

Andy - would you please clarify which ClarityCap you are referencing? The CSA has the copper ends.

Thanks, Tom

Through all of Jim's final listening / troubleshooting / optimizing phases, the parts were mounted on the actual crossover boards with all input and output leads attached to real drivers in their real cabinets. Therefore any leakage, environmental changes, etc. would be accounted for in the final as-built design. Our changes affect his outcomes and we will be reassessing actual measured performance and tweaking values as necessary.

I expect our expanded layouts to exhibit less crosstalk as well as greater thermal stability. But our target slopes and voicing will be the same as Jim's