advertisement


SBL cross-over

If turned on a lathe then aluminium mate - acrylic isn't much fun to turn. If it just an adaptor ring though I can cut it from off cuts of 3mm black acrylic very easily (no charge) - just say (and send me the measurements) - or if local to Sheffield I could come and measure myself (also hear your SBL's :)).
 
I'm not sure they go loud enough ;-) I'm South Derbyshire if you're passing? Doors alway s open, just don't look at the mess.

I'll give them some time, not sure I'm mixing up subtlety with lower output, but measurements will give some answers.
 
Pardon........ Has somebody told you I listen very loud?

There is some software I use REW (obviously you need a microphone) the AV boys use it a lot and it can be downloaded for free - I use Windows on the PC but there is probably a MAC version.
 
I've downloaded REW, used a non calibrated mic off an AV amp.

I measured;
standard SBL + PXO
SBL + Audio42
SBL + D2905 + A42.

The surprise to me was the fairly steep roll-off @ 10k with all options.
Original Tweeter has a higher output +4db @ 20k.
The A42 XO filled a hollow 1500-5k.

sbl responce by levs the diver, on Flickr

Measurements taken in room, I have a solid floor but plasterboard stud wall. Far from lab methods, but interesting none-the-less.

Mark, if you have any REW recommendations I welcome comments.

I wanna try another speaker now ;-)
 
The SBL tweeter has a rising impedance at 10k which causes the roll-off that you measured. A well designed conjugate network in the crossover filter corrects this and allows a flat response to 18k. From what I remember this can be done with 2-3 additional components.

Here is the impedance sweep of the 2905 (in red)
https://www.dropbox.com/s/mjsy430iq37ee9l/impedance.jpg?dl=0

This I think exactly mirrors your measurements.
 
Thanks Ron, its nice to get additional confirmation of results.

TBH my ears don't work much beyond 12k, probably explains my puzzled look when my AT cart was accused of raised top end ;-) My ears tell me the 2905 is better.

Its fascinating comparing graphs vs hearing. I've just done a sweep of another speaker, my untrained reading would suggest the graph is better, its flatter/smoother and more extended hi and low, yet its a small budget speaker. (and it sounds it TBH)

504 by levs the diver, on Flickr
 
I am currently starting to modify the SBL xovers and have the following questions:

1. What sound improvement can i expect from changing the resistors to bétter ones? What brand and type is recommended?
2. For the coils what type is considered best? Copper foil/wax paper? And what improvement can i expect from changing the coils?
3. For capacitors I am using Wondercap 4.7 uf (HF) and SCR cap for 22uf (HF) and SCR 12 uf (LF). Are both Wondercap and SCR considered to be just average quality or do they belong to the better category?
4. Currently still using the original 1.5 uf capacitor (HF)What brand is recommended for this application?
5. There is one 12uf Alcap still in the LF network. Should i change this one as well. I was told that this one is not really crucial.

Currently i am still burning in the crossovers but it has been quite an interesting journey.......sometimes there is much improvement, sometimes very muddy sounding and sometimes no PRAT ! i hope this is normal.....

Looking forward for your comments and opinions.

Thanks and Regards,

David
 
1. I would go with the Mills ones....very good build and sound and probably the best VFM that you could get. It is true that there are better, but you would have to spend 15x the cost of the Mills and without an x/o redesign, this is probably redundant. The Mills should sound considerably cleaner, smoother and more detailed than the cheap sand cast ones that come standard.
2. Leave the coils alone for now unless you have access to modeling and/or a best bench with a calibrated microphone.
3. Wondercap was better than average-twenty years ago. I tried them in the original Spica TC-50 speakers and they really were nothing special.
4. The HF cap is probably the most critical. Try a Mundorf Silver Supreme Paper in Oil (PIO). There are better to be had, but it will cost 10x plus. Avoid Solens, especially for HF sections.
5. LF caps are less important. For a flavor of what gains could be had, try bypassing the 12uF with a 0.22uF high quality film cap. If you like what you here consider replacing them.
 
divedeepdog said:
I've just tried an added Zobel network, but measurements suggest an even earlier slope @ 10K. Have I got the wrong components?

Online calculator suggest 2.02 uF and 5.8ohm resistor, I've bought 2.2uf + 5.6 ohm yet no raising of the response curve.

I've used this;
http://diyaudioprojects.com/Technical/Speaker-Zobel/

Re = 4.7
Le 0.07

I've sited it at the speaker terminals
Click to expand...
In my experience, a zobel flattens the rising impedance but does nothing to change its frequency response. If you want more extended HF, you should go with the 9500 or 9700 variants of the 2905.
 
I'll bow to Ron's experience, but Solen's worked well in my first DIY build. No other reason to using them apart from me using them before. I hadn't done any cap rolling :)

It was obvious the Audio42 with Audyne caps and Jentzen foils are better, but @ £400 vs £100 (component estimate) on £400 worth of speakers has to be considered?

Had I not fallen on the A42, my next step was to split the cross-over to allow bi-amping, and some nice Clarity caps.



The D2905-3000 work well, but lacked a bit of sparkle/atmosphere. Following a visit form YWAON I've got some Dayton mini ribbons, same as he uses on his NS1000 project. Its brought some sparkle without being harsh.
 
Dear Ron,

Many thanks for your input. I will order the mills resistors to try on the xovers.

rontoolsie said:
1. I would go with the Mills ones....very good build and sound and probably the best VFM that you could get. It is true that there are better, but you would have to spend 15x the cost of the Mills and without an x/o redesign, this is probably redundant. The Mills should sound considerably cleaner, smoother and more detailed than the cheap sand cast ones that come standard.
2. Leave the coils alone for now unless you have access to modeling and/or a best bench with a calibrated microphone.
Noted on this

3. Wondercap was better than average-twenty years ago. I tried them in the original Spica TC-50 speakers and they really were nothing special.
4. The HF cap is probably the most critical. Try a Mundorf Silver Supreme Paper in Oil (PIO). There are better to be had, but it will cost 10x plus. Avoid Solens, especially for HF sections.

Noted on this one. Is the Mundorf cap you recommend is the Mundorf MCap Supreme Silver Oil or the Mundorf MCap Supreme Cap?


5. LF caps are less important. For a flavor of what gains could be had, try bypassing the 12uF with a 0.22uF high quality film cap. If you like what you here consider replacing them.
Click to expand...

I will try this one as well.

Thanks
 
Naim2 said:
I am currently starting to modify the SBL xovers and have the following questions:

1. What sound improvement can i expect from changing the resistors to bétter ones? What brand and type is recommended?
2. For the coils what type is considered best? Copper foil/wax paper? And what improvement can i expect from changing the coils?
3. For capacitors I am using Wondercap 4.7 uf (HF) and SCR cap for 22uf (HF) and SCR 12 uf (LF). Are both Wondercap and SCR considered to be just average quality or do they belong to the better category?
4. Currently still using the original 1.5 uf capacitor (HF)What brand is recommended for this application?
5. There is one 12uf Alcap still in the LF network. Should i change this one as well. I was told that this one is not really crucial.

Currently i am still burning in the crossovers but it has been quite an interesting journey.......sometimes there is much improvement, sometimes very muddy sounding and sometimes no PRAT ! i hope this is normal.....

Looking forward for your comments and opinions.

Thanks and Regards,

David
Click to expand...

Hi David,

a) When you replaced the 20uf with a 22uf, did you remove the1.5uf that is in parallel with the 20uf, so that 20uf + 1.5uf is replaced by 22uf?


b) Have you changed the 1 and 27 ohm resistors to compensate for the lower ESR of the 4.7uf capacitor? I use 2.2 and 10 ohm. If not the HF will be too loud. See:

http://i575.photobucket.com/albums/ss196/isnchys/sblcrossovercircuit.jpg


c) For the final version of my modified crossover see:

http://i575.photobucket.com/albums/ss196/isnchys/sblcrossover09mar2014.jpg

There was some benefit in changing all the capacitors to polypropylene types. I prefer to use ClarityCap ESA 250v which are reasonably compact, though I also used two Mundorf Mcap because they fitted the available space. The 16uf on the LF network was replaced with a 15uf since 16uf is not a standard value for poly caps. I figure that a 15uf 5% tolerance cap is close enough to a 16uf 10% tolerance cap, and it works fine.

Wire, air core coils were used because they are smaller and lighter that foil coils. The only problem is that the 135uh is not a standard value and was contructed by unwinding a larger value.

Replacing the coils gave an larger improvement than changing the HF capacitors!


e) I now use a new build of the crossover that uses foil coils, see:

http://i575.photobucket.com/albums/ss196/isnchys/SBLackBoxPxo2.jpg
http://i575.photobucket.com/albums/ss196/isnchys/SBLackBoxPxo3.jpg
http://i575.photobucket.com/albums/ss196/isnchys/SBLackBoxPxo1.jpg

This crossover is a significant improvement over the earlier version. While some improvement will be gained by hardwiring the components (shorter signal paths and fewer solder joints), star wiring, and better resistors, I suspect that the foil coils contributed most of the improvement.

All the required values of foils coils including the 135uh (listed under cross coils) were obtained from: http://www.audio-components.co.uk/


f) The inconsistant perfomance that you reported is not normal. Might be a dry joint in the crossover?

Or the input sockets of the crossover can be a problem They are only held in position by solder which can crack or come loose. Insert a Naim speaker plug if resoldering the socket to hold them parallel.

Or its something else:confused:

Regards,
John
 
Happy new year 2015 to you all!!

I hope you have an enjoyable Christmas and New Year Holidays.

It was quite an experience in listening to Christmas music while burning in the modified SBL xovers……..it was fun and frustrating at times but at the end it was worthwhile and enlightening one.

So… below is my burn in experience..:

The first 10-15 minutes: The sound was very nice, clear,not grainy, lots of PRaT, a definite improvement over the old crossover. Then the next half hour or so the sound suddenly became muddy, slow, no pace, unbelievable “ugly”, lack of top end, bottom end was grainy.
Then suddenly the Top End came back. It was like someone lifted a veil going from the left speakers to the right speakers. A rollercoaster ride for the 1st hour.

From the next hour to the next 30 hours it was all the “fun” : grainy not grainy, PRaT no PRaT, Muddy – Clear, Coarse mids – Lovely mids. But there was incremental improvement to the sound for every block of 10 hours or so of burn in time. The first 10 hours was frustrating but at the 20th hour there is a tiny light at the end of the tunnel.

At more or less 30 hours the highs becomes clear and mids starts to flow nicely.
At 40 hours the bass becomes clear and tight.

Then comes the delightful 45th hour of burn in……. PRAT came back!

Then….. the Magical 48th hour…….its ALIVE…:D now SBL sounded very very good.
Details are there (tons of details were not presented by the old xover), Bass is tight, human voice sounds more natural. Tweeter brightness/forwardness is much reduced (Naim CDX is now much more fun to listen to). Overall it was an excellent improvement over the original (Old – 17 years!) crossover. At least to my ears it was at least as good as going from bare nac72 to nac72+Hicap. It gives me a direction on how much can be improved from “Stock” components and how it impact the sound quality.

I am only using so-so quality films caps for now (SCRs and Wondercap). I can only imagine how much improvement can be achieved once good quality components is in the crossover such as ESA Claritycap, Mundorf Silver Oil, good resistors, and coils.

I wonder do resistors and coils need to be burn-in as well?

Many thanks,

David
 
I am beyond confident that the addition of upper echelon components will yield even bigger improvements over those you have already gotten. If you do end up with Mundorf Silver/Oil the burn in will be substantially longer-at least 200+ hours.

I am absolutely certain that a well designed passive crossover using premium components will better any Snaxo/*cap, although it may up costing as much (but requires less shelving space and no need for service/recaps etc).

I would also be VERY tempted to use one of these resistors in the tweeter network. There is anectodal evidence that these may be better than even the Duelund CAST resistors at half the price...

http://partsconnexion.com/resistors_path.html
 
I'm with you Ron, £100 worth of components put my SNAXO in the classified :D

David, be great to see some pictures ;-)
 
Sorry for slow response, work overload........:)

Ron,
Very tempting idea. Do you think Duelund caps are really worth it for SBL? They are ultra expensive. Or Mundorf Silver oil is already 95% of the Duelund caps?

Divedeepdog,

I will take some pics later on.

FYI I am using Hiquphon tweeters OWII, which my ears tell me they are much better compared to the old original ones.
 
Hi John,

Thanks for the information. Below is my answer to your questions in italics.


a) When you replaced the 20uf with a 22uf, did you remove the1.5uf that is in parallel with the 20uf, so that 20uf + 1.5uf is replaced by 22uf?

I did not replace the 1.5 uf so it is 22 uf + 1.5 uf and I added a capacitor bypass.I will post a picture of the xover soon.

b) Have you changed the 1 and 27 ohm resistors to compensate for the lower ESR of the 4.7uf capacitor? I use 2.2 and 10 ohm. If not the HF will be too loud. See:

http://i575.photobucket.com/albums/s...vercircuit.jpg

Not yet but I plan to change them with the new resistors. I am using Hiquphon OWII which resistors value do you suggest?

c) For the final version of my modified crossover see:

http://i575.photobucket.com/albums/s...r09mar2014.jpg

There was some benefit in changing all the capacitors to polypropylene types. I prefer to use ClarityCap ESA 250v which are reasonably compact, though I also used two Mundorf Mcap because they fitted the available space. The 16uf on the LF network was replaced with a 15uf since 16uf is not a standard value for poly caps. I figure that a 15uf 5% tolerance cap is close enough to a 16uf 10% tolerance cap, and it works fine.

Wire, air core coils were used because they are smaller and lighter that foil coils. The only problem is that the 135uh is not a standard value and was contructed by unwinding a larger value.

Replacing the coils gave an larger improvement than changing the HF capacitors!

Very important point , thanks John
Does coils need burn in as well? how about resistors?

e) I now use a new build of the crossover that uses foil coils, see:

http://i575.photobucket.com/albums/s...ackBoxPxo2.jpg
http://i575.photobucket.com/albums/s...ackBoxPxo3.jpg
http://i575.photobucket.com/albums/s...ackBoxPxo1.jpg

….

John, how does the sound compared between Wire air coils vs. Foil Coil? I hope you have a chance to do a comparison and took some notes.:)

f) The inconsistant perfomance that you reported is not normal. Might be a dry joint in the crossover?

Or the input sockets of the crossover can be a problem They are only held in position by solder which can crack or come loose. Insert a Naim speaker plug if resoldering the socket to hold them parallel. Or its something else .

Well , Apparently it was the “burn in” process. After the 50th hours onwards the sound stabilized with much much improvement.


More questions to you John,

I noticed from you photo bucket album that you had a few iterations of modification of the X-Over

Could you describe the changes in sound quality going from :
1. original to the SBL xover 1st update (changes only of the HF Caps)
2. from SBLxover 1st update1 to sblcrossover25jan2014.jpg.html (changes to wirewound resistors (mills?) and the LF caps + wire coils but still use Bennic resistors in LF and HF)
3. from sblcrossover25jan2014.jpg.html to sblcrossover09mar2014.jpg.html (changes in LF caps but still use Bennic resistors in LF and HF)
4. sblcrossover09mar2014.jpg.html to the SBlackboxPxo (point to point soldering, Foil coils, All ESA Claritycap, Mundorf ? resistors)

On each improvement what are the most noticeable and prominent improvement that you hear? Do you use NACA5 for the point to point connection in the SBlackPxo?



Many Thanks David
 
You must log in or register to reply here.


advertisement


Back
Top