Renovating Thiel CS3.5 Speakers

[Kotter67]

Hi Vince2. I was wondering if it would be possible to get the .frd files you generated. I was contemplating pursuing a passive replacement strategy with the mids (new driver + crossover modifications) and actually considered this wavecor, but ended up getting the RS125P. I have DATS and a UMIK-1, but the effort stalled out when it came to taking the driver frequency measurements. Just have not had a need, as I have multiple replacement drivers. I could use the frd files to begin modeling, as I have everything else. I have the DATS data on good drivers (your "good" tweeter is somewhat out of spec, fs should be closer to 600Hz and the woofers fs at 45Hz is higher than the 25Hz I have measured on two sets of woofers). The mid data looks spot on with three working units I have. What tweeter did yours come with? The CS3.5s were shipped with both the D28AF and D28/2 (later units). I have collected a great deal of info on the CS3.5 and its drivers over the years. There are allegedly about 3000 sets of these out there and the mid failure issue has been taking a good number of them out of commission. I know Rob Gillum was reconing them for $250 each a few years back, but he may have run out of parts by now.

 

[Vince2]

Do you need the frd files of all 6 drivers? In txt format from rew? These were measurements made from close range, about 1 " or less. My drivers measured all differently so I eq'd each separately. Were your measurements taken while the drivers were installed? I believe that that I have 2 older and one newer tweeter. As Rob is no longer able to service the tweeter or the midrange, I went with the wavecor was I also like their tweeters.

 

[Kotter67]

Actually, the .frd for the midrange would be all I need. The .zma from DATS for the same driver if possible, but I have that on mine and they are probably close. Regardless, I'm going to have to take my own measurements, but would give me something to play with. Here are my DATS measures for the mid ranges I have. Good correlation within mine and with yours also. I also have four bad cores. These seem to either work well or not at all.

 

[Kotter67]

Here are two measurements of good D28AFs and two that still function, but test out of spec. I cannot say if the "bad" ones sound any different, have not heard them. The peak ohms tend to drop and fs shift out when the fluid is drying out. Seen this on alot of older tweeters. I plan on replacing the fluid on these at some point. Already have it, just can't seem to get motivated to finish the job. Also some photos of a disassembled unit I found elsewhere. The D28AF does not have any way to align for the re-install, you will have do it by trial and error. Screws with a different head with more grab may help and then swap them out for the originals once properly set. A test tone while making minor adjustments. Should work. Just be careful with the fluid, it might all get sucked out by the magnet when you pour it out.

 

[Kotter67]

The CS3.5 Tweeter:
The D28AF was initially used and replaced by the D28/2 later in the production run (got this from Thiel way back) with no crossover modifications. The advantage of the D28/2 is that replacement voice coils were once available (Madisound used to have them, but may have run out). The D28/2 has alignment pins that facilitate reinstallation of a removed coil for replacement or fluid refresh. The MDT-32S is reported to be a good substitute for the Dynaudios, but it is not a drop in replacement. The back cup is deeper and will require a spacer behind the face plate to bump it out due to interference with the housing that is integral to the front baffle. Also, all of these tweeters have been around awhile, I have several of each, and most of them exhibit the lower impedance and higher fs symptom. Buying older units is a gamble. So if a fluid refresh of the units you have does not restore them, the best bet is to purchase new MDT-32S units, as they are still in production. Other substitutes for the tweeter include the D2905/9300 (recommended by Solen, see attached, plate is 105mm), which I think is also still being produced (idea here is to purchase new to get a properly spec'ed one).

 

[Vince2]

Thanks, currently otherwise occupied, but when I pick this project back up may give this a closer look.

 

[GettnBetter]

I just opened this can of worms with my CS 3.5's. My right mid range was dead. I found this https://www.madisoundspeakerstore.com/approx-5-midrange/scan-speak-13m-midrange-replacement-speaker/ and bought the last one, they indicated they are getting more. Was a perfect replacement physically but clearly not the same as the original, the magnet is much smaller and is clearly not as strong as the original 13m/8531. Written on the original label is also PW: 420 which I assume was some kind of power rating. The replacement driver magnet is small enough to handle may be 30w if that. I don't have any means of measuring but they ironically sound quite similar but I'm no audiophile. In any event, this replacement midrange is working for the time being.

I also went down the rabbit hole of researching replacement caps for the crossovers. It's been something like 35-40 years and these things get used every day. I assume the non polars have dried out somewhat. Not sure I need to but I'm planning to replace the non-polars and maybe the MKP's if I see any physical damage. Was surprised how big the crossover boards are. Madisound helped me locate a thread over at DYIaudio about this. The only difference is my crossover has an 8.2uf MKP vs. the one there showed 8uf. I assume this was either due to availability or a driver change. I ordered up a bunch of non-polars and the MKP's. Might start with the non polars and stop if I don't hear any real differences or if anyone recommends otherwise, I'm no expert in this regard.

The reason I'm posting however is something very interesting in my diagnosis of the problem with the original midrange driver. The voice coils in the driver appear fine, I can measure resistance in the tinsel leads as they exit the spider. Now here is the interesting part, they measure 0 outside the terminals. The leads were still secured just fine, not broken. Whatever happened, it appears as if the leads died inside the glue as they pass through the terminals. I just drilled out the center of the terminals and soldered my own tinsel lead extensions and used a bit of silicone to secure the leads and the driver works great again!

Is it possible blue loctite or something similar was being used to glue the tinsel leads in the terminals and over time, the galvanic process essentially cut off all electrical transmission as the leads passed through the hole in the terminals?

 

[Vince2]

I never tried to make any changes to the xo as I went the active way. For info on this and your experience with the lose connection for the driver, contact Rob at https://www.coherentsourceservice.com/contact/

 

[mitchalee]

I recently bought Thiel CS3.5 loudspeakers as part of a lot. The drivers are all original. Both wooferts are OK. Both mids are dead open. One tweeter is dead open. The other seems to work. Both woofers and the possibly OK tweeter measure as 4 ohms. I know the woofers are OK. And the tweeter seems OK. I know the original CS3.5 were 4 ohm speakers. Here's my puzzle: why does everyone recommend 8 ohm replacement mids and tweeters for these speakers? Are you sure the original mids and tweets were 8 ohms?

Tomorrow I will re-test the tweeter and re-measure it with a digital meter at my shop. But the junk meter I have at home says 4 ohms.

I know that one can use a 4 ohm woofer with any mid or tweet one wants and then just match with crossover design. The woofer will set the measured resistence of the whole system. However my experience is that it is generally cleaner and simpler to stay with matching drivers. So I am surprised that Thiel chose to mix and match. Am I missing something?

 

[Vince2]

If you look at the dats measurements in the earlier posts, you will see the resistance/frequency measurements of the original drivers. Don't know where to begin in clearing up your confusion. Are you familiar with the design goals of the Thiel speakers? Not that I'm an expert, but tweaking the crossover is not a trivial pursuit.

 

[mitchalee]

If you look at the dats measurements in the earlier posts, you will see the resistance/frequency measurements of the original drivers. Don't know where to begin in clearing up your confusion. Are you familiar with the design goals of the Thiel speakers? Not that I'm an expert, but tweaking the crossover is not a trivial pursuit.

I was confused as to the impedance/resistance of the individual drivers in the CS3.5. Not so much the details of how the impedance varies with frequency but whether to consider drivers marketed as 8 ohm or 4 ohm. The concensus seems to be that the mid and tweet are 8 ohm drivers.

What through me off initially was the fact that the CS3.5 is sold as a 4 ohm speaker. So I was not sure if it used a 4 ohm woofer or not. My mulitmeter only has DC restsance. So 4.9 on the tweeter had me confused until I realized that it was close enough to 5 for the tweeter to be an 8 ohm model.

I decided to go with the
Madisound:
Morel MDT32S $112.00 each = $224.00
Scan-Speak 13M Midrange Replacement $210.00 each = $420.00
Subtotal: $644.00

Scan-Speak 13M Midrange Replacement is a Scan-Speak 10F/8424G00 midrange mounted on a custom machined aluminum faceplate. It should drop in.
I have been told the Morel may have minor depth mounting issues. This is a bit dissapointing just because I am lazy carpenter! And want the drivers to a stay at exactly the right distance without having to change the tilt of the cabinet. Oh, well . . . We shall see.

Thanks for responding!

 

[Vince2]

Yes those are currently the recommended replacement drivers. The cs3. 5 are rated as 4 ohm, but they stay above 5 ohm over the entire range and are easy on the amp to drive. You need the bass eq for best performance if you don't use other eq methods. These speakers produce enough dB for regular spaces but we're not known for extreme high loudness.

 

[Pineapple Dave]

I was confused as to the impedance/resistance of the individual drivers in the CS3.5. Not so much the details of how the impedance varies with frequency but whether to consider drivers marketed as 8 ohm or 4 ohm. The concensus seems to be that the mid and tweet are 8 ohm drivers.

What through me off initially was the fact that the CS3.5 is sold as a 4 ohm speaker. So I was not sure if it used a 4 ohm woofer or not. My mulitmeter only has DC restsance. So 4.9 on the tweeter had me confused until I realized that it was close enough to 5 for the tweeter to be an 8 ohm model.

I decided to go with the
Madisound:
Morel MDT32S $112.00 each = $224.00
Scan-Speak 13M Midrange Replacement $210.00 each = $420.00
Subtotal: $644.00

Scan-Speak 13M Midrange Replacement is a Scan-Speak 10F/8424G00 midrange mounted on a custom machined aluminum faceplate. It should drop in.
I have been told the Morel may have minor depth mounting issues. This is a bit dissapointing just because I am lazy carpenter! And want the drivers to a stay at exactly the right distance without having to change the tilt of the cabinet. Oh, well . . . We shall see.

Thanks for responding!

Sent you a PM about the dead drivers.