Thiel Owners

3.5' mids replaced with Madisound's '13M replacement' scan-speak 10f with adapter ring.

My experience to date (and a little advice):

Loosening the tweeters needs to be done with just the right screwdriver and much care.  I had a ratcheting screwdiver with a perfect bit.  The screws area deep and tight.  Leave the top screw in and remove bottom two then loosen to be able to remove the mids.

The mids themselves come out easily.  White wire is positive. There is a red paper washer under the terminal on the speaker which I did not notice at first.  Black one on neg. side (duh)

The waveform rings will come off with with a strong sharp folding knife to pry it apart from the driver.  It is perhaps hot glue that holds them together. It is not epoxy as I had feared.


The waveform rings have a ridge that fit the original M13 but do not work on the replacement. You will have to hacksaw it off if you want to use it on the 10f or put the new ones in without them.

Measurements and impressions:  I took some less then laboratory spec measurements and recordings with a condenser mic into Logic Pro.

email me noams@rum comcast.net if you want pics of spectrum.
At 10"  lined up on the mid centers*:
From 200 to 4khz the graphs looked fairly similar. Above 4khz.  The 13m's drop off on an even slope The 10f's fall off less evenly,  drop off a bit at around 7K then bump up around 11k
* (yes I'm picking up some tweeter too)
Measurements done just after driver install - no burn-in 

Listening:
I did not do any burn in but have about 10 hours on them now. Mostly a variety of well recorded Jazz and alt-country.

They are very transparent and detailed but to my ear (at this point) the upper mids through lower hi's sound more pronounced and tad exaggerated. I hear it on vocals through snares and hi-hats. Come to think of it even upper bass is affected.  I did not carefully compare sensitivity and was not super careful running pink noice at the exact same level each go round.   I am suspecting they are bit more sensitive and are boosting everything in their range (which is a lot with these crossovers) in their own very detailed way. **

 I am hoping further break-in will reduce this.  Imaging is as good as ever.  On some material they sound great, and overall the sound is still very impressive - but to my ear the originals seemed more balanced.   They sounded 'just right' on almost everything I threw at them.   Some of this impression may just be what I'm used to. 

One of my original 13m's is still good and the other has a bad surround but is otherwise good.   I found a rubber surround that seems to fit perfectly and will at some point replace the foam on one or both of them.

For now I will keep these in and play the 'SH' out of them for a while. I'll report further impressions after they've had more hours on them.

Thanks to those here who helped with this endeavor.


** I do a bit of my own DIY on tube guitar amps and guitar controls but haven't tweaked crossovers at all.  If further break-in doesn't land me where I want to go I'm wondering if a simple voltage divider or dropping resistor right at the driver terminal could in effect reduce their relative sensitivity to the other drivers without mucking up the overall sound or screwing anything else up.


   

A guy in Rochester NY is selling a CS3.7 for $4K but it is local pickup.

Status update on my CS3.7 refurb.

1) I was able to contact Rob Gillum today and we discussed the resistance measurements on my crossover. One speaker, the good one, had 0 resistance on the tweeter, What the heck happened there. I was thinking I would need to disconnect the crossover and send to Rob. Not something I wanted to do.

However, after opening up the crossover module I see the white positive tweeter wire disconnected from the crossover. Maybe it got disconnected when I moved the speaker to the garage for painting. I did use kid gloves with the transport but maybe not careful enough. This speaker was the one that was measuring stronger on the high frequencies on the SPL measurement. So the wire must have got disconnected after the measurements. The re-soldering of that tweeter wire was tricky but got done. I measured the resistance again and everything was good on both speakers.

I added the 2 new COAXs that I got from Rob. They are the black ones from 2014. I packed up the old COAXs to send to Rob for analysis. If they are good they will be my backup drivers.

2) I finished the paint job. It was a 9/10 until I became more ambitious and decided to touchup a smudge on the aluminum shell of one speaker. I added some of the great 3M painting plastic to cover the wood in a small section. Painted it and came back the next day to remove plastic. Unfortunately, the tape on the plastic stuck too tightly on the new paint on the wood and took some paint off. Awww!. I fixed that with some sanding and touchup but now that side is a 7/10. I can fix it perfectly with some sanding and a full coat on that side but you can only see the touchup if you look closely from an angle. Maybe I will fix that but for now I have the speaker placed so that I never see it.

3) I added the new outriggers I got from Rob,. I like the way that looks and functionally useful so that my son does not topple the speaker. I managed to do all of this work without my son seeing what the drivers look like. Which is a good thing.

As I listen to some John Lennon, the sound now seems correct. I was told that the new COAX needs about 200 hours to settle. However, given the massively messed up nature of the speaker when I brought it home, it sounds the best I have ever heard right now. The COAX is not as smooth as it will get but it sounds so much more coherent across all the frequencies. A bit like the Yamaha NS5000.

All of this cost me about $2000 extra. With paint, 2 new COAX, outriggers, solder kit (and self training), multimeter, and the Recoil kit to fix the COAX thread. What a nightmare but a great learning experience on speaker builds.

I will measure the speakers next week and hope the results are perfect. If so, I can continue with my DRC implementation.

@tomthiel, In reply to your 11/25/2020 post:

Thiel Owners | Audiogon Discussion Forum

Let me say that your measurements are better than mine, as I have none. Though I think Stereophile's component measurements have real value, I am hesitant to use Stereophile's speaker measurements, especially with speakers like Thiel's. I have heard the 3.5's next to the 3.6's very often at my preferred dealer at the time, Innovative Audio in Brooklyn, with various gear such as Adcom , B&K, conrad-johnson, Exposure, Levinson, Krell, PS Audio, Rowland, Spectral, etc., And as they had a policy of only (except for the small budget lines) having one set of speakers in a room at a time, they were regularly moving speakers in and out of their various sized rooms. So there was that too. It's hard for me to imagine a speaker with a port and less bass output as being "more accurate" by "adhering to a more flat frequency and phase response curve" than a sealed box with extended bass response. 
Thiel CS3.5 loudspeaker Specifications | Stereophile.com
Thiel CS3.6 loudspeaker Specifications | Stereophile.com
It might appear as though the 3.6's were .5 dB more linear above sub bass frequency but, when including sub bass frequency they were 1dB less linear, and only to 27Hz as opposed to the 3.5's 20 Hz. I can only wonder what the 3.5's frequency response above 27 Hz might have measured as. It seems incongruous that, all else being equal, a port would be more phase linear than a sealed box.

 I was aware of the significance of numerology in eastern culture. But, it seems as though Thiel's nomenclature only had value within Thiel. The difference between the CS3's and the CS 3.5's to the consumer seemed less than than that between the CS 3.5's and CS 3.6's. The CS 2's were 3 ways. The CS 5's had 6 drivers. The CS 6's had 3 drivers. The CS 7's had 4 drivers. Thiel could have used any number of alternative names to avoid using the #4, such as 3+1, or 2+2, etc.. The baffle was covered by a grill. With apologies to Bill; : "A speaker by any other name would sound just as sweet". 

 Perhaps "rushing to market" was overstated, But meeting market demands still seemed to be the objective. The refinement of the 3.6's was welcome, but expected. The loss of bass response and the let down in time and phase coherence (a Thiel hallmark!) was not, then the extra demands on amplification and concurrent costs was also not. The jaded might suspect it was an attempt to give dealers an excuse to sell their more of their new home theatre sub woofers.

Oh well, time moves on, and so must we.:-)

@thieliste, re: your 11/24/2020 post:
Thiel Owners | Audiogon Discussion Forum
IMHO, what improvements the Thiel CS 3.6’s had over the Thiel CS 3.5’s amounted to 2 steps forward / 3 steps back. Where as the flat co-axial drivers in the CS 3.7’s were a major(!) improvement forward. Still when all (!) is considered; I think Thiel’s CS 3.5’s were Thiel’s best product. When I say all, I’m including ROI or more plainly value. Lets compare:

Thiel CS 3.5’s last sold for $2850 up until 1992
Thiel CS3.5 loudspeaker Specifications | Stereophile.com

Thiel CS 3.7’s first sold for $11,000 in 2006
Thiel CS3.7 loudspeaker Specifications | Stereophile.com

Considering inflation between 1992 and 2006 would make the
Thiel CS 3.5’s cost in 2006 = $ 4,067.03

USD Inflation Calculator - US Dollar (1956-2020) (inflationtool.com)

Now, if one wanted to retrieve all the musical information available on one’s recordings, unlike the Thiel CS 3.5’s the Thiel 3.7’s would need to be supplemented with subwoofers. Let’s use a pair of Thiel’s own bottom of the line Thiel SS1’s, original 2003 cost of $2800 each adjusted for inflation in 2006 = $3046.10

USD Inflation Calculator - US Dollar (1956-2020) (inflationtool.com)

Thiel SS1’s X 2 = $6092.20

That alone costs more than the 2006 inflationary cost of the Thiel CS 3.5’s

When added to the original cost of the Thiel 3.7’s:

Thiel CS 3.7’s: $11,000
+
Thiel SS! X 2 : $6,092.20
____________________________

Total: $ 17,092.20

Or more than $13,025.17 than the 2006 inflationary cost of the Thiel CS 3.5’s

And that doesn’t include the cost of amplification:

Jim Thiel mostly used the Threshold S 500 amplifier for the Thiel CS 3.5’s. Lets use the last version sold; the Threshold S550e which sold in 1989 for $6300 adjusted for 2006 inflation for a cost of $10,289.13

USD Inflation Calculator - US Dollar (1956-2020) (inflationtool.com)

Jim Thiel mostly used the Krell FPB - 600 amplifier for the Thiel CS 3.7’s. Lets
use the Krell FPB - 600 C which sold in 2000 for $13,500
adjusted for 2006 inflation for a cost of $15,867.10

USD Inflation Calculator - US Dollar (1956-2020) (inflationtool.com)

Which would be an additional $5,555.97 in amplification cost,
which when added to the extra costs of the CS 3.7’s with 2 Thiel SS1’s over the 2006 inflation adjusted Thiel CS 3.5’s = $18,603.14

Thiel CS 3.7’s = $11,000
Thiel CS SS1 X 2 =$ 6,092.20
Krell FPB-600C =$15,867.10
________________________________

Total $32,959.30

Thiel CS 3.5’s =$ 4,067.03 -
Threshold S550e =$10,289.13
________________________________
$14,356.16
___________

$18,603.14

Now the Thiel CS 3.7’s combination will outperform the
Thiel CS 3.5 combination in most every way; ease of placement, freedom
from lobing, ultimate loudness, (and with the subs especially
in the bass region), and much, much more,... except time and
phase coherence, and foot print where the Theil CS 3.5’s despite their 20 year age disadvantage still have the edge. Still, the CS 3.5’s wouldn’t be too embarrassed by the comparison. If one wanted to do something similar with the current state of used prices, I think the advantage would be probably proportionately even greater for the Thiel CS 3.5’s.
Considering everything; I think the CS 3.5’s were Thiel’s best product.

Interesting I’m curious what about the cs6 or cs7? I’ll probably keep the cs 3.5 and get the 3.7 for the other system. What were your impressions on the cs3.5 on adcom amps vs the others?