Aftermarket crossover for Naim speakers (DBLs)

"In what way(s) do you believe the X/o is compromised? Transfer function, parts quality, both? And are you referring to the 4ohm, 8 ohm or both?"

What Richard said. Also, If I were you I'd take James up on his offer. With the help of CAD, I'm sure a new Sara PXO would not only sound better, but present an easier load as well.

"And is this common to other Linn speakers of this vintage?"

Definitely.

"Upgrading the Sara x/o's is a tempting proposition; pulling the things apart is not!"

LOL... The pair I had at home for a week belonged to a friend of a friend who wanted me upgrade the XO's with film cap's. I seem to remember they were a bitch to get apart and I phoned him to warn of possible damage - he wussed out and I wasn't sorry!

Mr Tibbs

Inspired by this I opened the box at the back of my IBLs; judging from the parts used it was quite obvious that this XO would gain quite a bit from thoughtful part swapping. My goal was to improve on the sound without altering the speakers' positive attributes (neutrality, coherence).
The treble has to pass through 4 parts directly (C1-3,R1). C3 is a reasonable MKT type, so I decided not to touch C2/C3.

I went for C1 and R1 instead, changing C1 for a Solen of the same value made a big difference; the problem is that the tweeter is used quite deeply into the midrange, and I've always found Solens somewhat phase-manipulating (forward mids, slightly nasal voices, etc.).

My current configuration involves a Mundorf MCap (4.7uF) and R1 was changed to a MOX of higher spec (3.7Ohm) because of the much lower ESR of the MCaps. The excess 0.3 Ohms don't bother me, as the speaker was sometimes a little bright anyway.

Current findings:
*much reduced distortion
*better (quicker) dynamics, quite detectable with horns
*much more detail
*better defined bass

In general, they sound somewhat 'lighter on the foot'. Honestly, I think this is one serious upgrade for IBL owners. The Mundorf caps match the revealing nature of the IBLs perfectly.


IBL Crossover (Tweeter)
+
O---||----o----o--||--o----^v^v----o----o----o
....C1....|....|..C2..|.....R1.....|....|....
..........|....---||---............>....>....
..........O.......C3...............<.R2.<.R3.
..........O.L1.....................>....>....
..........O........................|....|....
..........|........................|....=.C4.
..........|........................|....|....
O---------o------------------------o----o----o
-

Parts:
C1: Electrolyte, 4.7uF
C2: Electrolyte, 20uF
C3: MKT, 1.5uF
R1: Cement Resistor, 1.5R
R2: Cement Resistor, 18R
R3: Cement Resistor, 6.5R
C4: MKT, 1.5uF


Have fun,

Oliver

08.04.2006: I have been listening for a long time with the following configuration now, which seems a natural upgrade to me:



* the big blue MOX resistor measures 3.3R (this hunk can withstand 8W; 4W would be enough here)
* the black MKT cap next to it is made by Intertechnik and measures 22uF
* the small white MCap on the far right measures 0.47 uF
* the bigger white MCap above the small one measures 4.7uF

Another mod I've done: put some additional sheep wool into the cavity behind the tweeter; this will damp some of the 'nasality' the IBLs sometimes show. Cut a hole into the wool, so that the tweeter fits in, leaving no unfilled space behind it.

If the velcro becomes loose and needs to be renewed: I used this stuff from Conrad and cut it lengthwise. It has a strong adhesive, so be careful when applying. ;-)

Best again,
Oliver

Edit: spelling
Edit: fixed schematics
Edit: added image and description of XO
Edit: added source for velcro strips

Interesting project Ron! While you're taking things up a notch, I wonder what a suspension system (like the mainboard in a CDS) would do for the new x-over boards? Perhaps a test for audible microphonics would be a pillow underneath the x-over board while they're lying on your hardwood floor.

Keep us posted regardless....

dave

<<Rusty, if you are interested I can measure the native response of your Saras (when I'm measuring the E-V and PFM-Special) and redesign a PXO using LspCAD for them if you like.>>


Thanks for the offer James, but that would of course mean a full strip-down to access the driver inputs.

Ideally I'd like to get my hands on a second pair for experimentation.

Long shot, but does anyone have the X/O schematic for the original Sara?

Rusty,

"that would of course mean a full strip-down to access the driver inputs."

If you can get access to the XO circuit, then that should be OK. I would measure the paired B-200s as a single driver, and not separately. But I hear even the XO board is a bitch to get to.

James

Inspired by this I opened the box at the back of my IBLs

Click to expand...

Euugh! - is that what's in my astounding little IBL's crossover?

They sound bloody marvellous stock, so getting rid of those 'orrible 'lytics is seriously tempting, they should be transformed!

How easy was it to peel the silicone off that secures the boards?

Andy.

Andy,

Andrew L Weekes wrote:
[...]They sound bloody marvellous stock, so getting rid of those 'orrible 'lytics is seriously tempting, they should be transformed!
How easy was it to peel the silicone off that secures the boards?

Click to expand...

in fact quite easy; all you need is a sharp pointy knife and a little patience. Once you hold the plastic box in your hands try to press the walls out a bit with your thumbs. You will probably hear a slight cracking noise, that is the glue loosening from the side walls. Repeat with all sides.
Then try to lift a piece of glue somewhere in the middle and once you get hold of a chunk, just pull. Mine was partly flexible, so I could rip out huge chunks in one go. The corners are a bit trickier, this is where I had to lift small bits with the knife. The glue does not really stick to the PCBs, so I managed to get them out unscratched.

I really think it's worth it; 10,-EUR for the caps + 4,- EUR for the MOXes could save a lot of dough upstream here.

Best of luck,

Oliver

A wunderbare thread.
Also been looking at upgrading the x-overs in my Credos (again the contents are just bipolar electrolytics and other cheap nasties inside) after looking at the avondale, audio42 and various other sites and threads. And one thing I do know is film caps beat the pants off electrolytics so it seems an obvious mod to do.

However have no idea about xover design so a quick question for teccy people. The credo xover is the same as for sbls etc (pic swiped from avondale site for speed; hope you don't mind Mr W or I'll edit it off the thread again) but the components are not all common values. So does anyone know if components with values roughly +/- 5% would be ok to use?
particularly I want to use:
a)680uH inductor instead of 700uH
b)a single 22 uf cap instead of 20 + 1.5uf caps in parallel
c)all inductors changed to air-cored copper foil, 350W 16awg

Would these changes be fine or would the odd Ohm or uf variation mess up crossover points and frequency responses noticeably?
I'm assuming not, as most of the parts are at least +/-10% tolerance anyway though they could be matched. Are there any parts that should be closely watched? Any advice would be gratefully received.

This link might also be of interest to xover makers http://home.zonnet.nl/geenius/Cap.html

cheers
Ced

Ced,

ced wrote:
[...]
b)a single 22 uf cap instead of 20 + 1.5uf caps in parallel

Click to expand...

I already have these babies lying on my desk (22uF); reasonably cheap and better than both standard caps. Could be really interesting to use them in parallel with some really nice, but small cap. I'll report.

Best regards,

Oliver

P.S.: the similarity between the IBL's XO and your Credo's is amazing; what tweeter does the Credo use?

Edit: typo

It seems somewhat suboptimal to have two entirely different speakers using the identical crossover. Even if the drive units are identical (which they are not), the fact that they are operating into entirely different cabinets, volumes, Q's imply that an optimized x/o would have a specific transfer function for each. Simply assigning somewhat artibrary crossover points and orders is (from what I am beginning to learn) trying to hit a target in the dark.

As it has already been pointed out, simply swapping inductors may change the damping as the air-coiled ones typically have higher DC resistances over the ferrite cored ones (my massive tape wound ones were specified from 12 ga rather than standard 14 ga to lower the DCR. Upgrading the electrolytic caps for film caps cannot help but make things better though and is probably the first place to start.

The task of getting it right is quite thorny, especially if you are the end user. But it is well worth it- in the week or so since I have been running my custom DBL x/o I am listening to more music in a day than I used to in an entire week before that. The pleasure that having a sound that is so holistically right in every way cannot adequately be qualified or quantified. But I have never had a system that allowed me access so deeply into the musical threads and intents.

I had an audiobuddy come and visit over the weekend. We swapped back and forth between the stock and the custom x/o a few times. He could not believe how much coloration, hardness and smear the stock units indelibly branded onto the sound. When he left he immediately paid the speaker builder a visit with the intentions of having him fashion a custom x/o for his Proac Response 3.8s.

The credo xover is the same as for sbls etc (pic swiped from avondale site for speed; hope you don't mind Mr W or I'll edit it off the thread again) but the components are not all common values. So does anyone know if components with values roughly +/- 5% would be ok to use?

Click to expand...

OK, sounds like caps first is a good idea and then hope with the inductors or make sure the dc resistance is similar at least. Ta for the tip.

I get the point that the optimum is to have a crossover custom designed to the speaker measured in situ. Ahhh if only I....

But if I understand correctly not only the credo and sbl but also the SL2 have the same crossover circuit. And if the Credo then probably the old intros also. Makes clear manufacturing sense and hey, the naim poweramps and preamps are all basically the same circuit. I have no real idea about speakers but I get the impression the above are all basically the same speaker just the decoupling between the tweeters is different. Possibly internal volumes and the 'gap' are similar or can be tweaked to make this possible or account for different drivers?

Ironically I remember a few years back seeing some German Naim fan's site where he stuck his sbl tweeters to the wall behind the box via a metal rod or similar. Voila a few years later the SL2 with more than a leet-le similariy? In fact just found it: http://hu.fzk.de/hu/hifi/naim/sbl/sbl_english01.shtml

re credo tweeter, no idea what it is Oliver but yeah I noticed the similarity too.
laters
Ced

It is no coincidence that Naim choose to use essentially the same drivers for all their loudspeakers (the new Arriva excepted). The transfer function of their crossovers could reliably be adapted from one loudspeaker to another. They even inverted the midbass driver of their Credo/Allae/Intro so that the plane of the drivers' acoustic centres approximates that of the SBL and SL2. This is essential if one is to adapt and existing set of transfer functions from one loudspeaker system to another.

However, I don't think there is anything sinister in Naim's intent. After all, they are a manufacturer who will endeavour to streamline production and minimise unnecessary overheads. So a cookie-cutter works for them, to an extent. No doubt helped by a firm belief that loudspeakers do as they are told by their amplifiers.

But as Ron Toolsie has clearly articulated and from my own experience in designing and building loudspeakers from scratch, there is a lot to be gained in any system (mullet or otherwise) in effort to optimise a passive crossover network for a given complement of drivers. I would go so far as to say that the best drivers used in conjunction with a mediocre, designed-by-textbook crossover network will undoubtedly sound worse than a complementary set of competent drivers connected to a well designed and thoroughly optimised crossover network.

James

I had the chance to listen to a pair of 30-year-old RadioShack speakers, made with the cheapest all paper (even the voice coil) drive units, which sell these days for about $2 each both before and after some extensive x/o R&D. The speakers in their stock state sounded miserable, harsh, 2 dimensional and totally lacking any bass at anything less than full tilt SPLs.

Two days later I heard them again...one had been reworked keeping the identical crap drivers, but hand tuned with a premium quality x/o (housing some $600 in parts, Solen caps, air cored inductors etc), where the order, crossover points and transfer function was optimized for those drivers in their cabinet. Moving the balance from side to side easily showed how vastly better the reworked on sounded. Touching the tweeter surround of the stock unit showed it to be buzzing away in the background, while the reworked side was almost still to the touch- the stock x/o was cutting in way too early and allowing frequencies as low as 300 Hz.

and two days ago I had the chance to hear them as a matched pair with both x/o having been redone. Driven with a restored vintage Radioshack receiver of equal annuity they sounded not only decent, but very good. Deep tuneful bass, a wide open uncolored mid and hash free but extended top end, not to mention a deep soundstage. In other words they sounded like a good pair of $1500+ speakers. And in a way they were, for they had the abovementioned $600 in x/o parts and about the same again in labor spent designing and fabricating their crossovers. In spite of drivers costing less than $2.00 a piece ($1.87 I think they sell for in the catalog).

It appears that in any speaker, the 'source first' doctrine applies to its crossover-because the signal hits there before it gets to the drivers or has a chance to resonate within the cabinet.

I would go so far as to say that the best drivers used in conjunction with a mediocre, designed-by-textbook crossover network will undoubtedly sound worse than a complementary set of competent drivers connected to a well designed and thoroughly optimised crossover network.

Click to expand...

"It appears that in any speaker, the 'source first' doctrine applies to its crossover-because the signal hits there before it gets to the drivers or has a chance to resonate within the cabinet."

... and in that regard, the PXO has the same duty of care as a preamplifier. To ensure that the amplified signal is correctly handled (without adverse effect on the driving amplifier), split, and delivered intact to the complement of drivers that will reconstitute the full bandwidth signal and convert it into music, speech or noise.

James

I suspect the real reason that Naim use an identical/similar passive X/O is so that their active X/O can be easily applied to all these speakers, with only the gain trim-pots to fiddle.

Don't think the SL2 uses the same X/O though; given the vastly different sound of these vis-a-vis SBL's, that would be incredible.

Naim initially had designed a very simple crossover for the Intro, IIRC. The speaker was unsaleable.

Someone put an SBL crossover on them and they got much better. So Naim hastily modified the speaker to include the same crossover as the SBL. It cost them a few quid more per pair, but it put the Intro on the market.

I believe at the time Naim had a designer who was very good at the mechanical aspects of speaker engineering, but pretty useless at the electrical side. Since then, they have grown quite a bit and now are able to employ people with a number of different skills.

One of the reasons that stock Naim speakers sound better active than passive is that Naim do indeed (or at least used to) design their sppeakers for active operation. They then replicated the transfer function of the active crosover for the passive one, apparently failing to take into account the loss of electrical damping and rise in distortion that goes with a ("commercial grade") passive crossover.

As a result, passive Naim speakers have a repuation for brightness. It is my educated guess that what one hears as brightness in Naim speakers is more often than not not frequency response-related, but distortion. Improve the crossover elements, reduce the distortion and hey presto, less perceived brightness.

Markus,

Markus Sauer wrote:
[...]As a result, passive Naim speakers have a repuation for brightness. It is my educated guess that what one hears as brightness in Naim speakers is more often than not not frequency response-related, but distortion. Improve the crossover elements, reduce the distortion and hey presto, less perceived brightness.

Click to expand...

drawing conclusions from my humble experiments I'd support your theory.

The pervasive brightness of the passive DBLs with the stock crossover is most certainly not due to a tipped frequency reponse- for they start to roll off at 12k and are 5dB down by 20k. The redesigned x/o is now flat to 18k, yet sounds far smoother but vastly more detailed and coherent. I imagine that at least some of the bright sound is due to tweeter and crossover resonance. Pushing the x/o point of the tweeter to 4k gets rid of the resonances lower down in the spectrum, and replacing the ferrite cored inductors with air cored ones gets rid of the crossover saturation and resonances.

I have taken the plunge. Should be here on Thursday....


Regards,
Marc

Received them today. JUST BRILLIANT!!!
If you own a pair of DBLs get these made. Its what DBLs should sound like. I am not sure if I need to go active? It might be very close.....Maybe better...

If your in the area? Come on by for a listen...

Marc