The original CS2.4, as best as I can tell, had 13uF + 1uF and 27uF + 1uF in the tweeter feed. In the SE version, the by-passes are replaced by single 14uF and 28uF caps as seen here: http://www.hifishock.org/gallery/speakers/thiel/cs2-4se-2-thiel/I am unable to find these values in single caps from Clarity or Mundorf, altho' you can achieve the same capacitance with, for example, a 12 and 2uF cap. Rob Gillum at Coherent Source Service has SE upgrade kits so that you can transform you CS2.4's performance. Presumably these include the full-value caps seen in the linked pic. He also has the SE outriggers with SS spikes. Please support Coherent Source Service! I have a pair of SE version, so already with the Clarity SA caps. I might upgrade those to a newer version (eg, Clarity CSA) but will probably upgrade the caps in the woofer feed first. From the schematic supplied by Rob Gillum, it appears there is a 33uF cap in the direct feed. Again, I can't find that exact value in a single cap so will have to get there using a combination. Rob Gillum wrote me that the CS2.4's bottom panel is glued on. Access to the crossovers is via the passive radiator. I am leery of removing these for fear of difficulty in reattaching them but Rob wrote that you can grasp the radiator by its surround while re-screwing. I will report back as I upgrade my Thiels (won't be anytime soon). |
After much research and consideration, Tom Thiel has settled on Clarity CSAs for the upgrade. He even has a custom low voltage version to replace the electrolytics. I am going to also try Multicap RTX bypasses (1%) on the coax board. Progress on my 2.4s has been slower than envisioned last spring but I hope to have one channel done by the end of the month.
The passive radiator is hardier than it looks, but not easy to get the proper angle for removal. Pretty snug in there. |
Yes, please share a description of sonic improvement.
Nice work, though. And thanks for the pic. I suggest replacing the resistors, too. I upgraded my sandcast OEMs to Mills MRA 12s. A nice sonic improvement at low cost ($100). My resistors showed signs of overheating, so it’s possible my old ones were damaged more so than the Mills being better? Yours appear clean. Still, a cost effective upgrade.
With Tom Thiel’s tutelage, I will soon be completely rebuilding my boards, even replacing the coils. Everything will be new from the binding posts to the XO output wire (other than I will recycle the Mills resistors). I will post my progress and impressions on the Thiel owners thread. |
@andy2 holco has uploaded the schematic. The 14 and 28 uf caps are in the feed. Note the SE version gets full values in single caps whereas the original added 1uf bypasses (ie, 13+1=14). The 43 is a subfeed. Those big ELs are in the shunt position, so they probably make less audible difference but might still offer a sonic improvement with an upgrade. Even if not, it’s worth replacing older ELs.
The decision to go with CSAs is partly a choice of price. I mean, does it make sense to put $8000 worth of Dueland caps in a speaker that retailed for $4000? Mundorfs and many others were considered. Tom and I spent many weeks researching and we each spoke with several people with good knowledge before landing on the CSAs. I was pushing for the top of the Clarity line CMRs but Tom got some good info that the CSAs, possibly with small Multicap bypasses, would hit the sweet spot and for much less money. Toms decisions are very well considered and one of the considerations was price. At the end of the day, Tom wants to not stray too far from the Thiel tenet of value while still providing a very worthwhile upgrade. Might you eke out even better performance with Dueland and such? Maybe! Let us know whatever you try.
|
I gave heavy consideration to the Mcap supreme. It’s said to be quite good. One reason I’m going with Clarity is that Jim Thiel et al. did many comparisons when picking caps for the SE version. I don’t know if they tried Mundorfs but probably. They chose the Clarity SA as the best and that cap has been upgraded by two generations now. |
Regarding access to XOs for the 2.4, here are Rob Gillum’s instructions to me:
“The crossover can be accessed through the passive radiator opening. The base is not removable, as it is screwed, and glued. To access the crossover, you must remove the passive radiator screws and let the passive radiator drop into the cabinet. It can be rotated 90 degrees, and removed while servicing. To re-install the passive radiator, you can place your fingers at the surround, pressing outward to hold in place while re-installing the screws that hold bit in place.”
I would add two things, 1) be patient rotating the radiator to get the right angle for removal; 2) you don’t need to be overly paranoid grasping the radiator surrounds as I was initially. Not nearly as delicate as I first imagined, albeit still need to exercise care.
|
Replace with the same? Why? The only ones that will wear out are the electrolytics in the coax shunt position. In my SEs, these are CYC and those might be the last I would use in an upgrade. Erse has an affordable 100 uF metal polypropylene. And if you’re gonna to replace anything else, might as well put in something of higher quality than the OEM caps.
Jim Thiel used Clarity SAs in the coax feeds and Tom Thiel is planning to use Clarity CSA (2 generations newer than SA) for most caps in his 2.4 beta upgrade. More info on the Thiel Owners thread. |
Electrolytic caps (those 100 uF in a Thiel) will drift after, say, 15-20 years. But everything else should be good for even longer. As far as changing the character, yes, you can make the speaker sound even better by upgrading the quality of the passive parts. The OEM caps and resistors are far from top quality available, especially today. Thiels were built to a price point so that regilar people could afford them. They are really good but can be made even better by improving parts quality in the crossovers. Go read the Thiel Owners thread starting about February when Tom Thiel started posting. |
@desmond888
Your Clarity Cap plan sounds good. You can save money and space by going with the 250 V version for the larger capacitance value (ie, 10 and 18 uF) and 630 V version for the smaller value (3.9 and 10). Another path to 28 uF is to get 25 + 3. Madisound has the CSA 250 in 25 uF and I think Partsconnexion has the 630 V version of 3 uF. You might also get add a small bypass to each, say 0.1 uF. I’m going to try Multicap RTX at ~1% but Cornell-Dubilier is a worthy contender, among others.
You might also upgrade the 43 uF cap which is a subfeed. That cap is downstream from the 16 ohm resistor so you can go with a lower voltage. Until I started working with Tom Thiel, I considered a 39 uF ERSE Pulse X (100 V?) + 3.9 uF CSA250V. Finally, those 100 uF electrolytics should optimally be replaced with MPT-type capacitors. Again, ERSE PulseX is a worthy contender but you might have trouble fitting those on your boards. At the least, I suggest replacing the electrolytics with fresh ones (see Tom Thiel’s recent post on the Thiel Owner’s thread).
While you’re in there, I highly recommend replacing all resistors with Mills (or Vishay-Mills) MRA-12s. Sonic Craft has all the CS2.4 values although you have to mix and match the old (Mills) and new (Vishay-Mills). Only about $100 for all and makes a worthwhile improvement in SQ.
|
630 V version has thicker film. Tom Thiel (and his sources) is confident that will sound better. We are putting 630 V on the coax feed (ie, 14 and 28 uF caps) for maximum SQ. Where you are runnng in parallel to get proper capacitance, you might use a 250 V for the heavy lifting and 630 V for the final icing on the cake. Eg, 25 uF 250V plus 3 uF 630 V for 28 uF cap. The 43 uF subfeed cap is downstream from a 16 ohm resistor and will be a 250 V CSA. Elsewhere, I will be using 250 or 160 V caps, except for the Multicap bypasses which are 600 V.
That said, have fun and please relax about getting this “right”. CSA 250s are certainly going to be a notable upgrade over the OEM caps!
MRA-12 will give you more headroom compared to the OEM 10 W resistors. That is good!
|
@desmond888 You might post this over on the Thiel Owners thread where Tom Thiel will see it. Your perceived shift in balance *could* be an artifact of altered total resistance on the coax board compared to woofer board. For example, the CSAs might have less resistance than the original caps. Might need to change the resistors to bring the balance back. But get Tom’s advice before proceeding! It might also be that your experience is shifted more so than actual frequency response . . .
Your boards appear to be second generation for the CS2.4. Printed circuit board but with good quality caps (MPT) and coils. Again, I strongly recommend upgrading the resistors, after you consult with Tom Thiel regarding possible value change. But good job!
The radiator can be a real challenge. It seems that there is one sweet spot where you can get it in or out and finding that spot is frustrating! |
@desmond888
Thiel used unleaded solder (high melting point) but I do not know the brand. I am using Cardas (leaded) for my new boards. In my case, I plan to replace everything from the terminal posts to the driver wiring. Posts and wire will be Cardas, so OEM solder is completely omitted.
I do not know the brand of caps Thiel used on your boards. Solen is a good guess but I suggest you ask Rob Gillum.
Are you still hearing an imbalance with the woofer? Did you change anything on the woofer board? The 2.4 layout I have says the DC resistance is 32 ohms for the coax (measured from hot in to hot out). @holco posted the same layout but someone wrote 31 Ohms DCR. My new board checked in at 30.4 ohms. I will measured the OEM board very soon to see how it compares. DCR on the woofer is <1 ohm. My new and OEM woofer board measured 0.4 ohm. Changing caps and coils can affect DCR even if the resistors are the same. This could have small effect on the sonic balance but shouldn’t change to the degree you reported. Perhaps something else is wrong?
Also, I am surprised you report no problems with your room. I suggest a few simple wall coverings, such as throw rugs to the side and rear wall, to dampen early reflections.
|
@desmond888
interesting report, thanks. Holco’s report is earlier on this thread.
My preliminary take is that the CSA is more transparent and resolved for pretty much all instruments and vocals. More open and clear, with notably better microdynamics. I will have more to say later, on the Thiel Owners thread.
But my subjective take should differ from yours. I have a late SE version with XOs sourced from FST. The coax feeds are Clarity SA but all other caps are CYC MKT, which is a lower quality. Also, I’ve been running Mills resistors since May. And my upgrade is a full rebuild with Multicap RTX bypasses on the coax feed CSAs. There is nothing from the OEM boards remaining, not even the coils. So, my starting and ending places are both substantially different than yours.
|
@holco fair enough. Unless you have an FST-sourced XO like me (and your pics indicate you have a Lexington XO), the OEM coils are already excellent. I agree your boards will most benefit from better caps and resistors. |
@desmond888 I now have CSAs in both channels (although 2nd channel without bypass caps). At first, I thought I heard a subtle “hallway echo” on male vocals and “silvery sheen” on female vocals. But now it all sounds clean, so I assume it was a burn-in thing. And all I can say is “WOW”! The transparency and resolution are top shelf. Super clean sounding with textures previously unnoticed. Intelligibility of lyrics previously mumbled.
Now, my upgrade is much different than yours but I give CSA a big thumbs up! |
@holco curious why you’re putting a foil coil on the woofer board and not replacing the one in the coax feed (0.15 uF).
One caution for the DIY crowd from Tom Thiel is to pay attention to the DC resistance on the boards as new parts can change these values and upset the balance. In my case, the new parts matched original values within 0.1 ohm. As Tom’s beta tester, I have full value capacitance in single caps. Only my bypass caps (~1% Multicap RTX) are in parallel, but those did not add measurable resistance. |
On my boards, that coax feed coil is an ERSE FoilQ. It is availble in 16 gauge at 0.15 mH. The OEM aircore one is 18 ga but I cannot tell it hurt anything to increase the gauge and was OKed by Tom Thiel. Then again, I’ve changed so many parts at once that it is near impossible to say where the biggest improvements came from. Well, I did change the resistors to Mills as a first step, so I know what those add to SQ. I will soon compare the Clarity CSAs with and without bypasses on the coax feed. |
The Mills gave a richer, fuller sound. I guess “more saturated “ is one way to describe it. Most importantly, the Mills removed a subtle “glassy” quality in the midrange and resulted in an ease of presentation, ie, reduced fatigue during long listening sessions. An easy recommendation, especially given the relatively low price. |
|
@desmond888
Are you asking about capacitance or resistance? The DC resistance on my FST coax board (with Mills resistors) was 30.5 ohms, 0.4 ohm on the woofer board. My new boards matched those values within 0.1 ohm. On my copy of the 2.4 layout, someone hand wrote 32 ohms for the coax, <1 ohm for the woofer. But that layout is for point to point on masonite. Tom Thiel discovered the values were later adjusted for the PCB environment. Holco’s copy indicates 31 ohms for the coax. Ideally, you would have measured that before you started your upgrade.
This is measured simply by connecting your ohm meter from the + in to the + out. |
Probably a question for Tom Thiel . . . I do know for individual components +/- 3-5% is acceptable, but maybe the tolerance is tighter for the global circuit? As a check, l compared the frequency response of a tone glide with and without the upgrade (well, OEM boards with resistors replaced). All good in my case. If the balance sounds right, you’re probably good. If you’re paranoid, ask Tom on the other thread. He *might* recommend modifying a resistor value.
|
@holco
yes, I literally had a grin on my face after I had the second channel upgraded. I was able to hear the improvements in mono but going to stereo put me in musical bliss. Wow! |
I did not attempt to compare Mills and Vishay Mills. I read one opinion that the older Mills are better but Tom Thiel thinks this is not worth worrying about. And it’s mostly academic because the CS2.4 values force you to buy some of each - impossible to make this comparison even if you wanted to. |
Has you opinion of the CSAs changed? |
think I have taken the maximum out of the CS2.4
I think I'm there, too. If I had ample money and time, it would be interesting to experiment with ultra caps (Jupiter, Dueland), resistors (Path), and try the wax foil coils. But if I had that much money I would probably just buy Vandersteen 7s and call it a day! I think my modded 2.4s sound as good or better than anything new up to at least $20K and for a fraction of the price. |
very curious about your findings!!
I will post my subjective impressions. Hope to find time to compare this weekend. |
@holco
I now have Cardas hookup wire plus double binding posts in one channel. Some of the job was challenging but everything seems good in the final result. Going to let it cook for a few days before I compare to the other channel in mono.
|
very curious about your findings!!
@holco
I was able to compare the Cardas and FST channels in mono
over the weekend. Not an apples to apples comparison because I used the Cardas
in biwire configuration, so changed two variables at once (actually three
variables: hookup wire, binding posts, and biwire cabling). I do have jumpers
and will compare biwire to single speaker cable later (in stereo instead of
mono) but wanted to maximize any differences for this test.
One thing I did *not* hear was increased bass impact as you
reported. There was one song where I *did* hear much more prominent bass (and a
less than pleasing glow on vocals and such) but it turned out that I failed to
disconnect the FST channel, so was hearing the output from both speakers (the
FST speaker was behind and slightly to the side of the Cardas speaker at the
time). Laughing out loud.
First off, the FST channel sounded great and I would have
been happy if that were the final configuration. The Cardas didn’t reveal FST
wire to be “broken”. That said, the Cardas wire/binding posts bring what I
would call an ultimate level of refinement. The differences were subtle in mono
but important for the sound I’ve been seeking. I heard more immediacy/presence,
more flesh. A bit richer tonal saturation. This was most apparent on vocals and
instruments sharing the frequency of human voice (guitar, violin, woodwinds,
etc). Also, acoustic bass, kettle drums and toms. I did *not* hear any
differences on cymbals, chimes, and bells.
Yesterday I installed the Cardas in the second channel. Listening
in stereo revealed that this effort and expense was well worth it. It’s hard to
describe the sonics, but I’ll try. The Cardas wire/posts (and biwire speaker
cables) “snapped everything into focus”, gelled the sonic picture into a
cohesive whole. I would say more openness and clarity, more immediacy, denser
images, more holographic, and a sweeter/richer midrange. I had been considering
an ultra bypass cap to sweeten the midrange – no longer on my to-do list.
There is no end to combinations to investigate. My
configuration came mostly from the mind of Tom Thiel who brought his wealth of
knowledge and experience, and spent many months researching the possibilities
for the CS2.4. With more time and money, I thought I might further investigate
some “ultra” possibilities (Clarity CMR caps, possibly mixed with Jantzen
Alumen Z; Jupiter copper foil or Audyn True Copper Max bypasses; Path Audio
resistors in the coax and woofer feeds). But that was before I heard my
configuration with the Cardas. The sound I am getting is astounding. Other than
low bass, I really can’t ask for better performance. Just, WOW!
|
Btw, the tighter/more bass was with the wax coil's
Ah, please remind me how your new wire changed the sound. |
@andy2
I replaced every passive part from the binding posts to the driver hookup wire.
Binding posts: Cardas "CPBP" (rhodium over silver), dual for biwire
Woofer in: Cardas 2x15 ga twisted pair plus 2x17.5 ga hand twisted (~14 ga equivalent)
Woofer out: Cardas 2x15 ga twisted pair
Coax in: Cardas 2x15 ga twisted pair
Coax out: Cardas 2x17.5 ga hand twisted
(wire gauge per Tom Thiel recommendation; OEM wire was 18 ga throughout)
|
Early on, Tom and I considered outboard and I canvassed the Thiel Owners thread to gauge demand. Plusses include heat management, vibration control, and freedom to use large parts. Minuses include reduced WAF and added cost. I have no regrets with my configuration.
|
