Thiel Owners
@unsound Can’t decide whether to trust you or my lying ears. @ronkent probably in the same predicament. But because you’re apparently into numbers, riddle me this. How many watts are needed to get 85 dB from CS2.4 at the listening position 8-9’ away? Also, what do you imagine JA meant when he wrote: “Thiel CS2.4 owners should make sure they have a good 4 ohm–rated amplifier to drive this speaker.” |
@unsound If you’re trying to recreate orchestral climaxes in your room, I suggest you have the wrong speaker rather myself with the wrong amp. Most Wilsons or the JBL K2 9800 will play loud as hell without issue. I hope buyers of the CS2.4 are not expecting to get ear splitting SPLs. And I hope, for the sake of your ears, that you reserve that experience for an actual orchestra a few times per year rather than every night in your room. Yes, 85 dB is a low bar. It’s also close my upper limit of comfort. Unless I’m going to listen to folk or a quartet, I bring ear plugs to most concerts. Even at 90 dB, my Ayre has plenty of headroom, probably <15W with the 88 dB 2.4. It’s laughable to think my amp is sweating. 2 ohm, 4 ohm, doesn’t matter at this level. Meanwhile, go take a look at Soundstages measurements of the 2.4. Speaker distortion probably becomes audible before amp distortion (and the CS2.4 has superbly low distortion). Yes, extra power sounds better, all other factors equal. All other factors are not equal. I’ll take an Ayre, ARC, or Aesthetix over any Krell or Levinson, thank you (hey, I’m an A-hole!). I won’t assume your system is wrong for you. Please grant me the same and ronkent. |
Under the conditions I use my system (seldom above 85dB and probably never above 90), my amp has more than enough power to deliver stress/compression-free music. It sounds superb. Listeners demaning peaks of 105+ should probably look not only for a different amp but a different loudspeaker than the 2.4. YMMV. |
That’s a great, and funny, story, Tom! And a good reminder of the compromises designers must make. Throwing money at the problem helps but even with unlimited budget, consumers will not have a “perfect” system. For myself, I most value transparency, resolution, imaging . . . musical immersion. Those are the very qualities improved with the CS2.4 XO upgrade. And I have assembled a system that addresses those priorities within my particular budget. I have zero interest in recreating SPLs of The Who in my room. |
At first glance, coils are 4-9s, best of form. Significant improvement could be made via cap and resistor upgrades from AXON, electrolytic and sandcasts.The coils are certainly of higher quality construction than those on my FST-sourced 2.4 boards. And the yellow caps appear to be the high quality polystyrene 1 uF that Thiel commonly used. I see one bank of electrolytic caps. Those are probably shunts but I found shunt cap upgrades to be audible. ELs not only drift with age but they are generally lower SQ than film caps. The other caps are probably metalized film type but it’s hard to tell for sure. I see at least one cap with the Solen logo. The resistors appear to be the sandcast wirewounds. I am confident that improved SQ can be had by upgrading the parts quality. If that was my CS6, I would replace all resistors with Mills MRA-12s. I would want to see a schematic before commenting about specific caps but I can say that better caps substantially contribute to improved transparency and resolution. Whether this would fully mitigate JA’s one criticism of the CS6 I have no idea, but it would certainly get you at least part way there. Andy2, curious to know what you see on that board. How did you surmise an all-pass filter? An all-pass filter suppresses no frequencies, right? Seems like such a "filter" would invariably allow the driver to excite its resonant modes. One thing I learned form Tom - and that greatly surprised me - was that Thiel used the same level of parts quality on all models from the 1.X to 7.X. I confidently predict that better parts quality will improve SQ for all models. That said, DIYers should know that it’s not necessarily as simple as replacing OEM parts for higher quality ones with "identical" values. In particular, changes n DC resistance can modify the voicing. |
@andy2 It’s above my pay grade to speculate whether you’re onto something regarding the XO design and JA’s subjective opinion, but I am confident higher parts quality will improve transparency for the CS6. It also seems relevant that Tom has told us Jim T’s CS7.3 was likely to include a mechanical XO for the coax unit albeit perhaps for entirely different reasons. |
Andy, FWIW Brian Damkroger did *not* hear the midrange reticence during his audition: https://www.stereophile.com/content/thiel-cs6-loudspeaker-brian-damkroger-october-2003 My experience with the CS2.4SE included a “glassy” quality in the midrange that was mitigated by replacing the sandcast resistors with Mills MRAs. And the Clarity CSAs (plus better coils) removed veils that I didn’t realize were there (sorry not sorry for the cliche). So, it’s easy to imagine better parts would improve, if not cure, the “reticence” that JA heard. The CS6 is an interesting model that I considered before settling on the 2.4s. Not as big as the massive 7.2 but still with good bass extension and sensitivity on par with the 2.4. That said, Tom Thiel’s comments about the 2.4 coax driver mass leave me wondering if this is the sweetspot (even including the 3.7) for my priorites of resolution and transparency. |
Thanks for the link, holco. I quite like the subjective reviews at hifi-advice. They do an exceptional job of comparing the DUT not only to competing products but also describe how the sound changes with different speakers, even cables. Here is their review of the VX-5 (the amp section of my AX-5 is a less powerful version of the VX-5). https://www.hifi-advice.com/blog/review/analog-reviews/amplifier-reviews/ayre-vx-5-twenty/ |
I think it not improper to point out that JA seems to hear what his measurements tell him. At 50" there is a midrange suck-out and some spikes. Notice that he wonders in print if that is what he is hearing.
Supposedly, Stereophile’s reviewers do not see the measurements until after their subjective evaluation but easy to imagine that JA might deviate from that when he does the evaluation *and* measurements. He does acknowledge, in this case, that the measured suckout is specific to his axis and distance. That said, it seems possible that JA *was* hearing with his measurements. Soundstage measures at 2 m. The CS2.4 has the flattest “listening window” of any speaker in their database (+/-2 dB from 33-20k). http://www.soundstagenetwork.com/measurements/thiel_cs24/ It’s plainly obvious that JA’s 50" distance is inadequate to characterize the frequency response of a Thiel. I suspect the waterfall plot would also be cleaner if taken at 2 m (it's already quite good, especially in the midrange where other, larger drivers continue to have energy after the music has stopped) https://www.stereophile.com/content/thiel-cs24-loudspeaker-measurements |
coincident driver takes the compromise out of nearfield listening. When I have the house to myself for the evening, I pull the coffee table out and sit on a wooden low back chair in front of my couch. I am unable to hear deleterious effects from the coffee table (ie, reflections) but this chair puts my ears probably 3-4” lower and, maybe, 7’ from the speakers. I get great sound from my “normal” position (ears an inch or two below tweeter axis and 8.5-9’ away) but this “nearfield” position is intoxicatingly immediate and engaging. From trials bringing the couch away from the rear wall and closer to the speakers, I think the lower position is more responsible for improved SQ but being closer seems to help, also.
I believe that what is causing the page to stretch horizontally, at least when viewed in Firefox, is the long URL in Unsound’s post dated 4-3-2019 at 12:32 p.m. EDT. Yes, the problem seems to be the ebay link in that post and it shows in Firefox but not Safari. |
The CS3.7 XO pic I’ve seen shows good coils but sandcast resistors and CYC MKT caps. The drivers are probably the best ever from Jim Thiel (ie, some of the best ever from any manufacturer) but the passive parts quality is wanting, IMO. If I had 3.7s I would certainly replace the resistors with Mills MRAs which are cost effective, high reward/low risk. Replacing the caps *might* be more risky but I would probably do that, too. There is a LOT of capacitance on that board, so might have to go outboard if replacing everything including the big electrolytics. I would start with the tweeter and midrange caps. That said, waiting for Tom Thiel’s solution is probably advisable for the less adventuresome. |
The terminal spade is almost certainly rhodium over silver but the strands are copper, according to their website http://www.cardas.com/clear_beyond_sp.php BTW, I hear a more relaxed and liquid presentation with my biwire configuration compared to single cable plus jumper. Also, a bit more dimensional with a tad more clarity. |
@jafant Not sure how to answer that and I haven’t kept careful records other than I had minimum 100-200 hours on new parts before I made comparisons. Most resistors have been run for about one year (3000+ hours?) other than Tom had me replace the 30 ohm with paralleled 60 ohm for improved cooling. And each channel was built independently as I waited for the minimum break in. So, lots of variability. New caps are surely >600 hours by now, some might be past 1000. The last changes I made were the Cardas hookup wire and binding posts. The more recent channel is probably at least 400 hours. I consider everything settled at this point although was not fully confident that 200 hours was adequate for the caps. @sdl4 I am running Cardas Golden Reference. I had Neutral Reference paired with my CS1.6. When I switched to Vandersteens, I briefly used generic speaker wire (16 ga?) to decide whether to splurge for biwiring. I was able to get a good deal on used GRs. Almost sold one pair after I bought the 2.4s. My procrastination paid off! |
@unsound @prof This should soothe the feelings https://www.avsforum.com/double-blind-abx-testing-proves-power-cords-effect-audio/ |
Can’t say I’ve noticed this trait, Andy2. Some songs with the mix in one channel seem to emanate well outside the bounds of the speakers. The CS2.4 does this at least as well as the CS1.6 and Vandy 2Ce Sig II. Just earlier this evening I did a double take as the image was a couple of feet left of the left channel even with my eyes open. I think this is this is more dependent on the recording than other factors, at least with designs that attempt to reduce baffle effects. I continue to be very impressed with SQ after my XO rebuild. If I’m in the hot seat, I can’t read with background music on - too distracting! The only sonic parameter that doesn’t quite match the very best I’ve heard is image density. Perhaps *that* is influenced by the baffle? But the TAD Ref One is among the best I’ve heard in that regard (and other regards!) and I don’t think anyone would accuse that model of having an innocuous baffle. |
@andy2 I can probably find recordings as you describe but not able to compare to anything else in the near future. Any 3.7 owner want to bring your speakers to my room? Interesting hypothesis regarding image density. That sounds plausible. Are your observations consistent in that respect? Yet another example that pretty every design comes with trade-offs. The small midrange diaphragm on the 2.4, as Tom highlighted, is very nimble and lightweight, able to convey a delicacy. It’s been a year or two since I’ve heard a reference level speaker but I'm confident my modded 2.4s are top-shelf in terms of transparency and resolution. But the small drivers also can’t push as much air and that has its own sonic consequences. |
Those SEs are from the same seller mine were from, similar price. They appear to be in slightly better condition but hard to know from the pics. Mine were rougher than described by TMR and the pics didn’t show all the cabinet flaws. The vermillion maple and stainless hardware is stunning. Other than the finish, I can think of only one reason to get the SEs if you intend a Tom Thiel crossover upgrade: the oversized back panel will easily accommodate double binding posts whereas the standard version will fit only one. Of course, you could always drill through the cabinet for the second posts but that’s a less satisfying solution, IMO (see pics on my system page). |
When New Thiel took the reins, Gary held out for Phase Coherence in the new products. He was over-ridden by the New Team, including Steve DeFuria and Bob Brown who had been long-term Thiel supporters and associates. They said that Gary had "drunk the kool-aid". Gary left New Thiel when Mark Mason decided to pursue Canadian-type, non-coherent designs.I've written this before but it's a cryin' shame that New Thiel completely abandoned Jim Thiel's design principles. At that point, why even keep the name? Meanwhile, I'm still happily drunk on the kool-aid of my upgraded 2.4s. Last night, I yet again found myself locking into the music (16/44 Mary Gauthier via Qobuz) despite having my attention elsewhere. |
Regarding the outriggers and spikes, I prefer the sound of my 2.4s with them on. It might simply be, however, that the improvement I hear is more related to having the speakers elevated another 2-3” rather than any coupling to the floor. My floor is tile over slab. The spikes sit on thin wood pieces to protect the Oriental rug I inherited from my parents. Anyhow, I’ve noticed I prefer the sound even more if I sit in a chair lower than my couch and another ~20” closer to the speakers. But this is changing more than one variable at a time, so can’t really say to what degree the lower listening axis makes it sound “better”. Regardless, my modded Thiels sound superb. I hope Tom is able to bring his upgrade kits to market soon! |
Dunlavy and Meadowlark were true time/phase correct speakers (although, in a Stereophile review of a Dunlavy speaker, the designer hedged whether the crossover is truly a first-order filter). Vandersteen is pretty much the last one standing and RV has really achieved something special with his carbon drivers. If he can ever bring that technology to a real world price point . . . Avalon is not time/phase coherent. A top priority was avoiding distortion from out of band break up. The original Avalons (designed by Charles Hansen, Neil Patel, and Jeff Rowland) did this by using steep filters, low XO points, and small midrange drivers. The 3-way Ascent had a 2" titanium midrange. IMO, Jim Thiel was able to control, if not eliminate, out of band ringing by using small midrange (3.5" in my CS2.4 coax) or the wavy diaphragms of the X.7 models despite the slow roll-off filters. None of the megabuck speakers I know of are time/phase aligned (certainly not Magico). That said, some of the best systems I’ve heard had speakers with steep filters. The TAD Ref 1 and Vivid Giya are standouts in my experience. I hope to soon have another listen to some megabuck speakers and compare their sound to my modded 2.4s. The new Stereophile reviewed Dutch and Dutch active speakers that pretty much have time, if not phase, alignment. In fact, all of JA’s measurements were eye-popping. I can’t imagine I’ll ever own active speakers but I would love to hear those! |
@holco Curious why you picked the Rhymthmic. If I ever decided to add subs I would almost certainly go with Vandersteens or used Thiels, so curious to know your rationale. But I don’t feel the need. My old Vandersteens had some audible output into the mid 20 cycles whereas the 2.4s have nothing below 30 yet I’ve only noticed the difference on a single song (organ tones). |
In my wire investigations I have learned that 6-9s wire is nearly extinct and even 5-9s is practically unaffordable. As the world changes toward miniaturization, highly complex component circuitry and wireless data and signal transmission, wire is less critical. It seems that some of the best sources (both audio and aerospace) are using 4-9s with careful casting, drawing, gauging and coating technologies. To your question, Thiel maintained a steadfast commitment to using the best wire availableHi TomI probably already shared this in a PM, but I found the Cardas hookup wire and binding posts to make a really nice improvement in SQ. In fact, I would rank the improvement as tied with the Mills resistors for second biggest improvement after replacing all caps and coils. I hope you are able to audition Cardas as you work on the R upgrades. That said, my experience aligns with with Prof in that $$$ AC cabling does not yield a corresponding improvement in SQ. As always, YMMV. |
We usually had a latest CJ and ARC model in our listening room.I have very limited experience with CJ but ARC is great stuff, IMO. If I wasn't an Ayre-head, I would probably have ARC (or Aesthetix or a John Curl design). There are one or two Thiel owners here on A'gon who use ARC to drive CS2.4 and, even, CS3.7 using ARC. |
What is "radical divergence"Some people like to listen at very high SPLs and have large rooms. They will need a LOT of power. Some people have smaller rooms and listen at moderate levels. They will need hardly any power. I did some calculations for my CS2.4s. Thiel claimed a minimum impedance of 3 ohms and three reviewers measured the minimum ranging ~2.7-3.1 ohms. Assuming a *constant* 2.9 ohm load (a worst case scenario as the impedance dips this low only for part of the frequency spectrum) and ignoring room interactions, I need about 5.5 watts for 85 dB at the listening position, ~20 W for 90 dB, ~50 for 95 dB, and ~180 W for 100 dB. I estimate my amp clips at ~97 dB. But I very rarely ask for more than 90 dB peaks and most of my listening probably only requires 2-5 W. My amp has plenty of headroom for my listening preferences. Others may need more power. If you demand 105 dB peaks (sorry for your ears), you may need 500-750 W with something like the CS2.4, depending on room interactions! So, someone like me needs <50 W but some others may need an order of magnitude more power. Radical divergence. |
@oblgny If those 2.4s have undamaged drivers and 7+ cabinet they are probably worth the asking price. Doesn’t matter about the SE or not. Most of the SE features were cosmetic (the finish *is* stunning). The only change affecting the sonics was use of Clarity Cap SAs on the coax feeds. In fact, some older 2.4s might have otherwise higher parts quality than the SE. Regardless, I promise you Tom Thiel’s upgrade will be head and shoulders beyond that. I recently had the opportunity to hear a couple of uber $peaker$ and the experience confirmed my impressions: in terms of resolution, transparency, soundstaging, and neutrality, my modded 2.4s sound as good as the best I’ve heard. Not surprisingly, the 2.4s can’t match the bass performance of $$$ designs. And not simply bass extension (the 2.4 has energy down to about 30 cycles which covers nearly all musical content). In particular, a sealed box design I heard has a clear step up on the Thiel’s passive radiator. Bass definition/ articulation was notably better via sealed box. That said, I consider the passive radiator to have a good mix of tradeoffs considering their price point (when I got the 2.4s, the first thing I noticed was tighter bass compared to my Vandersteen 2Ce Sig IIs). The uber designs also have a more tangible image density. One speaker, retailing for about $80k, had an unbounded, open sound that was exquisite. And I’ve previously heard a $peaker that recreated the sense of hall space on live recordings in a way that no other has. Would I trade my 2.4s for the $80k model? Yes, I would. Can I afford the $80k model? Not even close! It’s quite amazing how much SQ Jim Thiel brought to regular working people. The 2.4 is a speaker that working class people can realistically afford. And Tom Thiel’s XO upgrade brings the performance on par with the best I’ve heard in the parameters that are most important to me: resolution and transparency. The weaknesses I outlined above are minor. The $$$ designs showed me a few places where they outperform my 2.4s but they never left me feeling my speakers are broken or deficient in any way. I just bliss out to music on a system that is paid for! Highly recommended :) |
I’m settling in, down, planting myself!This is almost where I am. My modded 2.4s sound superb. I would probably have to drop at least $30k to better them in any meaningful way. My amp is killer, would have to $$$pend to get better performance there, too. The reason I wrote “almost” is because I am in line to get my Ayre DAC upgraded to latest version. After that I have zero plans to change my system in any way. |
How were the improvements measured? Jim Thiel again: "The improved resolution is not the kind of thing that shows up well in measurements; the magnitude of the difference between the CS2.4 and the CS2.4SE is more easily heard than discerned from graphs. The new capacitors allow more nuance, air, detail, and decay to be reproduced by the coaxial drive unit. This was especially evident to us when listening to recordings that contained realistic reverberation, as well as recordings where the instruments were not processed heavily." https://www.soundstageultra.com/equipment/thiel_cs24se.htm |
I bought my CS2.4SEs from TMR. The cabinets were in rougher condition than they showed in the pics and description but otherwise they were good to deal with. Those 2.2 look good in the pics but hard to believe a speaker that age is really a 9/10. For starters, there appears to be a scratch on the binding post plate. |
Try not to despair too much.I, for one, couldn't care less if Obscene Audio charges $50K per foot for a USB cable. No skin of my nose! Even if I could afford it I wouldn't bother (law of diminishing returns). And if the Obscene guys bilk the 1%ers so that they can feed their families, more power to 'em! Meanwhile, my experience is that cables and wire *do* make an audible difference. I heard really nice improvements with better speaker cables and hook-up wire. Less so but still worthwhile with interconnects but only tiny differences with power cables. To each their own. |
under false pretensesProf, right there is where we see it differently. You seem to view high end cables as "snake oil". My position is more nuanced: yes, there are obscenely priced cables whose performance is not commensurate with price but that doesn't make it snake oil (similar examples with speakers, amps, and sources; heck, wine and watches for the matter). Again, my experience is that cables and wire *do* make audible improvements. And I have no issue whatsoever if your experience is otherwise. |
Wow, Tom, thank you for that candid post. Regarding my upgrade, I would not hesitate to do it again. Some may consider it expensive, especially relative to the price of used Thiels, But the result is a speaker that compares favorably against other models up to $30k and is not embarrassed by the very best I’ve heard. I am very curious about your baffle treatment. I need to know more but am a likely candidate! |
I’m still confused as to what you did when you experienced the resonance. Ayre recommends playing the IBE at “moderate” levels which I interpret as less than typical SPL for music. When I did my XO rebuild, I searched for cabinet modes with steady tones (ie, not IBE). At higher SPLs, there were many frequencies that caused distressing distortions. I never determined the cause - amp or speaker - but immediately stopped that exploration! Thank G*d, actual music sounds sublime! |
I do have a question for you beetemainia and holco ,From my 5 March 2019 post in this thread: I replaced every passive part
from the binding posts
to the driver hookup wire. |
