Celestion SL6

[Pete B]

INTRODUCTION
The Celestion SL6 was introduced in 1982 and is a very high quality mini-monitor design made
in England with high quality parts and construction. The woofer and tweeter face plates
are aluminum metal as is the woofer chassis. All mounting hardware is backed by metal thread
inserts. The crossover is on a printed circuit board (PCB) and all wiring is soldered to provide
gas tight connections.

WOOFER ENGINEERING
Celestion engineers employed laser interferometry as a means to view cone and dome motion
across the radiating area vs. frequency in order to minimize breakup and non-pistonic
behavior. Just one optimization is that the woofer cone is molded in one piece including the
dust cap. In later versions, the woofer outer edge is a split design made from two different
rubber materials to provide a further optimization.

TWEETER ENGINEERING
The tweeter is a 1.25" copper dome made of one piece with the voice coil former and also
optimized to minimize and push breakup modes above the audio spectrum. However, the
stiff metal break up while much higher in frequency is much stronger such that it must be
notched in the crossover. Copper being heavy resulted in lower efficiency than most tweeters
and as a result the SL6 has a somewhat depressed top end.

ADVANCED CABINET MATERIAL FOR DERIVATIVE DESIGNS
Seeking a better cabinet material, Celestion engineers chose an aircraft aluminum with an
internal honeycomb structure called Aerolam. The SL600 is an SL6 in an Aerolam cabinet
with a significant price premium due to the difficulties in constructing such a cabinet. The
SL600Si is a minor update to the SL600 with bi-amp/wire inputs and star grounding in the
crossover but is otherwise identical, as far as I know, with the early drivers. SL600Si Review
in Stereophile note the price of $1995/pr not including stands:
https://www.stereophile.com/standloudspeakers/744/index.html

The SL700 employs the later aluminum dome tweeter and split edge woofer in an Aerolam
cabinet: https://www.stereophile.com/standloudspeakers/488/index.html
The SL600, SL600Si, and SL700 were John Atkinson's at home reference speakers for many years.

Back to the SL6:
Reviewers have made favorable comparisons to Quad electrostatics finding them to be very
similar but noting the slightly depressed top end of the SL6. All later versions of the SL6
including the SL6S, SL6Si, and SL700 employed the next revision 1.25" aluminum dome
tweeter with higher sensitivity and were voiced to correct the depressed top end.

Several reviews, linked below, provide much more detail about the Celestion SL6:

HiFi News and Record Review - First published in 1982:
Note that their measurements were made with the microphone half way between the tweeter
and woofer. I've found that the SL6Si and SL700 measure best on the tweeter axis and they
probably would have gotten better results.
https://www.hifinews.com/content/celestion-sl6-loudspeaker

More info on the SL6, not that I agree with all the comments:
https://www.muzines.co.uk/articles/celestion-sl6-speakers/4076

https://audionostalgia.co.uk/celestion-sl6-review/

SL6S (NEWER MODEL) review in Stereophile 1987, $900 a pair - the new aluminum dome tweeter is mentioned:
https://www.stereophile.com/content/celestion-sl6s-loudspeaker

This site provides a tweeter repair service and a lot of information and parts:
https://carpwrangleronline.weebly.com/


SPECIFIC SL6 PAIR COVERED IN THIS THREAD
I bought a pair of SL6 s with drivers looking in good shape but with one tweeter not working.
Update 4-8-2022:
I rejected replacing the copper tweeter due to the depressed top end and the fact that the
crossover must be factory aligned to the high Q metal dome resonance. I then considered the
later aluminum dome tweeter from Celestion but after seeing the measurement of that tweeter
in Stereophile and the high cost these were also rejected. See the poor tweeter response
above 10KHz in Figure 1 - I wonder if this was a single defective sample of that tweeter:
https://www.stereophile.com/content/celestion-sl700-loudspeaker-1991-measurements

I decided to repair these with Dayton RST28F (silk dome) tweeters. I had these on hand with
the truncated face plate and there is also an aluminum version that could be adapted for those
who wish to retain an aluminum dome tweeter. The Dayton tweeter replacement will be
covered in another thread.

TWEETERS
It's interesting that one copper tweeter dome looks shiny, the other looks like it was
subjected to high heat and is badly discolored. The copper tweeters require factory matching
of the crossover notch filter to the particular tweeter. The "factory alignment" involved two
inductors, one to position the notch filter, and another to trim the output at high frequency
probably 15KHz or so. The later Celestion aluminum dome
tweeter is claimed to have a dome resonance above the audio band that does not require a
notch filter. However, as stated above this later tweeter does not measure well. I purchased a
single SL6Si with the aluminum tweeter in order to have an example of the newer woofer,
tweeter, and crossover. the SL6Si crossover does not require any factory alignment and the
crossover was simplified with two inductors and one capacitor eliminated. The SL6Si will be
covered in another thread.

WOOFERS
I pulled one woofer and the spider looks like it has sagged by 1/8" or more, really
looks more like 3/16", is this a common problem? I've noticed it to be fairly common
with rubber edged woofers for them to sag. Has anyone else noticed this problem?
I'll try to get a picture of the sag.

CABINETS
Strange that one cabinet has heavy scratches and dented corners, it's as if one sat on a
concrete floor and was slid around and even turned over and slid more, perhaps the other
was on top and not as badly beat up with only minor scratches.

 

[Pete B]

SL6 WOOFER PROBLEMS
One Woofer edge has separated from the cone, is this common?
Edit: Glued this with Loctite Extreme glue, as suggested on another thread,
and it came out looking perfect.


SPIDER SAG DUE TO EDGE PROBLEMS
Straight edge against spider, note clear view of about 1/8" sag, both woofers were like this:


It was clear that the cone was offset since there was only about 1/8" further inward movement,
and over 1/4" outward movement. It was also easy to push the cone forward where it was then
flush against the straight edge.

PROPPED UP CONE - DOES NOT WORK FOR THESE WOOFERS
Paper towel folded and wedged against the frame to push the cone forward about 1/8" past
the rest position, leaving this for a few days might correct the sagging spider. I don't know if
it will hold and if it will last for say a year or more. Clearly, something is changing, some rubber
edges shrink over time, not sure if this is the problem:

Monday 3/28/2022 a few days later, woofers look much better one is flat, the other is almost
there about 1/32" more to go.
Wednesday 3/30/2022 both woofers have gone back to having close to the same amount of
sag. Trying the treatment again and then watching them, they seem to go back over the
course of an hour or so.

WINTERGREEN OIL
Coated both edges, staying away from the glue joint with straight Wintergreen oil, seems to
have softened them quickly, up to 3 applications but I'm going to try to figure out how to soak
them face down in some sort of jig. Most say to use 50/50 or 1/3 Wintergreen to Isopropanol
which I might try.
This worked shockingly well, after about 8 applications woofer Fs went down from 49 to 27 Hz,
that is a huge shift in compliance and all of the spider sag was gone. BUT after a few days it
came back, so perhaps a long term soak will be better.

 

[Ds2000]

One Woofer edge has separated from the cone, is this common?
View attachment 2522959

Yes, the glue gives up between the poly cone and the surround in these and on the Si they just split. Best bet is to glue back down, those surrounds usually respond well to the repair.

 

[Pete B]

Good tweeter compared to fried tweeter:


Fried tweeter with front plate and dome removed, note that you must lift the terminal
plate along with the dome and front plate otherwise the voice coil wires will rip. Notice
that the coils are completely burnt and falling off of the VC former. It is also interesting
that the VC former and dome are one piece:


Magnet, the center pole is dished with a felt pad employed to absorb the back
wave. The felt pad was not glued in and could easily come out of the dish which would
be an issue if it touches the dome. Ferro-fluid is not used:

 

[Pete B]

SL6 CROSSOVER WITH 2 FACTORY ALIGNED INDUCTORS SP4156
One pair of nuts hold the binding posts shafts to the enclosure, the crossover slides on
making contact and then another set of nuts hold the crossover in place by the binding posts.
Both nuts were slightly loose in both speakers, this is something to always check.

I'm going to guess that the power amp driving one channel went into oscillation causing
the tweeter and associated capacitor to fry. Note the aluminum can showing for one cap,
there should be a blue plastic coating on that cap. It was probably fried due to heat in the cap.

Crossover front and back pics:
There is one inductor and one cap in the woofer network, and 3 inductors and 3 caps in the
tweeter network. Part number SP4156 is under the foam tape, Issue 3 on the back.

 

[Pete B]

Dents in the top corners of one cabinet. Used walnut color Howard's Restore
a Finish to clean this up and at least add color to the damaged corners:

Top Left:


Top Right:


Top Rear:

 

[Pete B]

Place holder for crossover schematic:

 

[Lavane]

You might post any questions you have in the Celestion owners thread in the British Audio forum. There are a few very knowable folks that should be able to help. There are a couple people that restore this series of Celestion speaker.
They are in Australia I think? They were helpful when I was messing with my Celestion Dittons. Existing posts may be of help also.
https://audiokarma.org/forums/index.php?threads/the-celestion-owners-thread.804597/

EDIT, a response may take a few days.

 

[Pete B]

Place holder for woofer measurements:

 

[Pete B]

CROSSOVER SP4156 SCHEMATIC FROM XSIM
The woofer network is a simple second order network with a 2.8 mH series inductor and an
8 uF shunt capacitor. The woofer is large and provides about 6 dB of baffle step as can be
seen in the XO transfer function plot below.
The tweeter network is also a simple 2nd order network with a series 4.7 uF capacitor and a
.25 mH shunt inductor. The network is followed by a notch filter (L4 and C3) to compensate
for the metal dome resonance, and a lowpass filter (L3 and C6) to reduce the response at the
very top end of the spectrum.
L3 and L4 are made from heavy gauge wire and wound on iron cores, the top layer winds are
aligned at the factory by pushing the top layer further apart or closer together. Glue is then
applied to hold the position. They're aligned to match the specific copper tweeter:


WOOFER AND TWEETER INPUT IMPEDANCE PLOTS
Plots of .ZMA files:


CROSSOVER VOLTAGE TRANSFER FUNCTIONS
Note that Xsim has an upper plot frequency of 20 KHz, and the notch for the tweeter dome is
at about 21 or 22 KHz the start of the notch can be seen but only up to 20 KHz.
70dB on the plot below represents the output voltage from the amplifier and the very slight loss
below 60 Hz is due to the DC resistance of the woofer inductor. The slight bump around 100 Hz
is due to the woofer input impedance resonating with the crossover network. The gentle downward
slope from 150 Hz to 1KHz is full baffle step providing solid bass response even
when the speakers are placed well away from boundaries in an audiophile position. The flat
passband from about 1 to 1.5 KHz should be viewed as the passband then about a 6 dB drop
can be seen at the crossover point of 2.8 KHz. This is a textbook crossover response for a
mini monitor speaker with a well behaved woofer.
The tweeter crossover response is tilted down by 6 dB/oct from about 10 KHz to 2K eventually
becoming closer to the expected 12 dB/oct below 1 KHz. This is not the typical response
but must have been required for the unusual high moving mass copper tweeter. The tweeter
notch is just above 20 KHz and the response is well into it by 20 K, the upper limit of Xsim:

 

[djordan]

Bump!

Hi Pete,

Great restoration project - lots of valuable information and restoration tips and tricks.
I'm a HUGE Celestion fan and I've owned many of this series already - but I never knew about the factory aligned notch filter!

Question - you state:
"The network is followed by a notch filter (L4 and C3) to compensate
for the metal dome resonance, and a lowpass filter (L3 and C6) to reduce the response at the
very top end of the spectrum." and in the schematic, you note:
"L3 and L4 are factory aligned to match the tweeter"
So...

How critical is this notch filter?
How critical is the precise factory alignment? (presumably each crossover is matched to each individual tweeter).

Asking because I was planning on bypassing the internal crossover entirely - and using a Marchand, line-level tube crossover intead - to biamp the speaker.
(The chosen crossover specification, based on the SL600 with the same drivers is 2.3kHz, second-order, 12dB/octave slopes).

Wouldy you think this would produce a good result?
If I already have that crossover, should I consider a different tweeter?

Thanks!
DF

 

[djordan]

PS
What I'm doing should make more sense with these photos.
I'm essentially harvesting SL6 drivers from a complete speaker and dropping them into an SL600/600si aerolam cabinet.
AND I'd like to perform the bandwidth managment / crossover function electronically.

(BUT given the notch filter and unit-specific fine tuning, I'm having doubts about whether the e-crossover part of the plan will work)

 

[Pete B]

A few questions:
What state are you in? I'm in CT and perhaps I could help you.

Do you have the crossovers that match the tweeters that you have? Thinking if you measure them,
either component values or frequency response, then you'll know what you have to do electronically.
It seems that if those SL6 s are complete then you'd have the correct crossover.

I talked about how the edge stiffens, do your woofers have cone sag?

Do you only have the SL600 boxes? You could simply make passive SL600 s by transplanting all the parts.

Have you considered something like a MiniDSP that will make the complex response easy to do?

I think that the aluminum tweeter would make more sense since it didn't require a notch filter,
but actually if you look at Stereophile's measurements the peak is quite bad.

The SL700 has the aluminum dome tweeter and the split edge woofer. I seem to recall that the split
edge woofer also has lower voice coil inductance perhaps due to copper in there. I have a single SL6Si
that I bought to measure the woofer, tweeter and trace out the crossover. The SL6S and SL6Si have
the same crossover but the S has NPE caps and the Si has all film. Both have the aluminum tweeter
and split edge woofer.

 

[djordan]

A few questions:
What state are you in? I'm in CT and perhaps I could help you.

Hello neighbor, Pete.
NJ metro area right accross the river from Manhattan!

Do you have the crossovers that match the tweeters that you have? Thinking if you measure them,
either component values or frequency response, then you'll know what you have to do electronically.
It seems that if those SL6 s are complete then you'd have the correct crossover.

Original crossovers so YES - matched to drivers (as long as I didn't accidentally swap tweeters when I recapped - I don't think so!)

I talked about how the edge stiffens, do your woofers have cone sag?

I suspect I have the same sag you've shown - will pull out woofer and send update
NOTE - I have another set of SL6/600 woofers that show the same sag - perhaps more.
I also have a set of SL6S/700 woofers that show a tiny big of sag - but less than the SL6/600 woofers.
Will post some photos later.

Do you only have the SL600 boxes? You could simply make passive SL600 s by transplanting all the parts.

Yes. Agree that this is the sensible plan - at least for now.
Even beyond notch filter and the custom alignment, I'm also seeing warnings against connecting tweeters directly to amps without a capacitor or other way of protecting the tweeter from DC current.
Perhaps this alone is not insurmountable, but yet another complication.
NOTE. Funny you should inquire about my Celestion collection.
I'll be shortly adding a post with my other Celestion stuff.

Have you considered something like a MiniDSP that will make the complex response easy to do?

Yes. I've seen the Behringer and other devices that offer very sophisticated DSP. My resistance, though, is I'm such a "digital purist"
Digital playback for me has to be bit-perfect and NOS (non-oversampling).
(My Metrum Octave NOS DAC was recently sent to the factory in the Netherlands for repair - I'm temporary using a Gustard DAC with a Sabre ES9028pro DAC and I'm really not enjoying the sound anywhere near as much as the Metrum)

I think that the aluminum tweeter would make more sense since it didn't require a notch filter,
but actually if you look at Stereophile's measurements the peak is quite bad.

Agree. I've owned the aluminum versions of the SL series - the SL6s,SL6si and SL700.
They all have a more "conventional" sounding treble. They are all excellent, but not unique.
None have the appeal of the rolled off sounding copper-domed SL6/600.
I could likely find a set of aluminum SL tweeters - but then I'm basically creating more of an SL6S (or SL 700) speakers. And perhaps if I'm going that far off spec, there may be better "modern" tweeter replacements available.
But then its a slipperly slope - and I'm already re-designing a speaker which few of us are qualified to do. i.e., I mean like YOU, but not me!

The SL700 has the aluminum dome tweeter and the split edge woofer. I seem to recall that the split
edge woofer also has lower voice coil inductance perhaps due to copper in there. I have a single SL6Si
that I bought to measure the woofer, tweeter and trace out the crossover. The SL6S and SL6Si have
the same crossover but the S has NPE caps and the Si has all film. Both have the aluminum tweeter
and split edge woofer.

Precisely.
The ALL film cap SL6si is a notable advantage for folks looking for pre-owned vintage bargains - because you don't have to consider replacing components or servicing the crossover.

 

[Pete B]

The first important thing to consider is that if we cannot fix the early woofer cone sag problem
then there's little point in moving forward. As I've said, Wintergreen oil works but it only lasts
a few days to a week. I've bought several rubber reconditioners but have not tried them, and
might just jump to brake fluid as drastic as it is. I had been meaning to get back to these with
the brake fluid for the past few weeks and the timing is good. I don't know how bad the split
edge woofers are, I think I read you correctly that they are betting with regard to cone sag so
this would be another reason to do something like an active SL700 with EQ to sound like the SL600.
If I were ever to try brake fluid on a split edge woofer I'd be very worried about breaking the glue
bond that holds the two halves together. I'd probably apply it with a Q-tip carefully avoiding the
bonded area.

The next thing that I'm reading from you is that you prefer the rolled off top end of
the copper tweeter. Since you have so much gear and are considering such an extensive
conversion to bi-amping why not experiment with EQ of SL6S s or SL700 s to make them
sound like the SL6/600? If you can determine the EQ required then you can build with the
easier to use aluminum dome tweeter and emulate the sound of the copper dome. I'm
certain that it is a frequency response difference.

 

[Pete B]

I completed the mod to install the soft dome Dayton tweeters in my SL6 s. I did not want the
rolled off top end. I have access to a pair of SL700 s and I drove just the tweeter section and
worked to make the Dayton tweeter match as close as possible with a tweaked but very
similar crossover.
The goal was to have something that sounds like an SL6Si/SL700.

When they were completed I A/Bed the SL6 s to the SL700s and found the SL6 to have less
punch, and a bit of veiling in comparison. I've not investigated this much further but came to
the conclusion that it was probably differences in the woofers.

Much time went by, and we had a Connecticut Audio Society (CAS) meeting where a friend brought
his SL6S pair that he upgraded with film caps, and a bit more damping on the walls. Essentially
SL6Si s. I brought the modded SL6 s and heard exactly the same thing as when compared to the SL700.
At this point I don't have access to a pair of unmodified SL6 s to compare to my mod or the SL700 s but
it would be interesting.

 

[Pete B]

CELESTION ALUMINUM TWEETER HF PEAKING

I mention this in the first post but will repeat it here. "I then considered the
later aluminum dome tweeter from Celestion but after seeing the measurement of that tweeter
in Stereophile and the high cost these were also rejected. See the poor tweeter response
above 10KHz in Figure 1 - I wonder if this was a single defective sample of that tweeter:"
https://www.stereophile.com/content/celestion-sl700-loudspeaker-1991-measurements

Copied the Figure to put it inline here, note that while output is strong to 10 KHz it is grossly
attenuated by 12 to 15KHz - very bad: