crossover mod project!
12-01-05, 05:26 PM
#1
Moderator - Electronics Forum
Thread Starter
crossover mod project!
Hello all!
Working on a crossover modification - think you'll like this. On the left is a pic of the modded crossover. On the right, a stock crossover. The stock xover is used in the Dynaudio Audience 50 home speaker. The modded one hasn't been powered up yet and is in the works.
If you look at the pic of the stock xover, you'll notice small coils and smallish capacitors. The resistors are the white blocks, which are wirewound and sandcasted to help tolerate the heat. Dyn built the speaker as an "affordable" alternative, and it shows. The one BIG no-no that they did was to use an IRON CORE INDUCTOR. With iron core inductors, you can get the same mH (millihenry) ratings as a larger air core inductor...but with one problem. Hysteresis.
Definition of hysteresis from Northcreek audio...
"What about ferrite core inductors, which have very low DCR?
The problem with ferrite core inductors - all of them - is hysteresis distortion. Hysteresis is related to the energy loss incurred by the core material during the magnetization polarity shift that occurs in every cycle of the signal flowing though the inductor wire - for a 500 Hz signal, this is 500 times the core polarity has to be reoriented, per second! Hysteresis distortion sounds like compression - and also measures as wave-form compression. Furthermore, when the current though the inductor is large enough, the core will saturate and the value of the inductor dynamically shifts to a much lower value. This is also clearly audible.
The only practical application for ferrite-core inductors in high end audio is as low DCR "trap" coils in tweeter resonance trap circuits, where the circuit is functioning in a very narrow frequency window and the current flow is extremely small."
Unfortunately, the Dyn crossover for this application is an iron/ferrite core. If you see ANY iron cores in a crossover, it means that they're cutting corners and not making a closer to ideal crossover. IMO, you should not use an iron/ferrite core inductor when it comes to the audio signal.
The xovers in the car audio Dyn setups are all air core, which is a step in the right direction. Solen caps are used, which are ok. In the Dyn/Audience setup, they didn't even use a Solen cap, but rather some off the shelf no name cheapie. The BIGGEST OFFENDER is the Rainbow Reference xover. LOTS of iron/ferrite/laminate core inductors all over the place...with discrete no-name caps. Is this REALLY worth the 7 to 8 grand paid for their "premium" speakers? Will have to scan in a picture for you guys/gals to show you the HORROR of their xovers.
Switch over to the left side of the picture. Notice that the coils are HUGE in comparison to the little guys on the right. They're the SAME values, but the larger ones are made out of copper foil. Much lower dcr (dc resistance) and the build quality is fantastic. They're made from Alpha Core/Goertz. Many speaker manufacturers (including Dynaudio) almost never reach this level of inductor. With Dyn, you have to pay nearly 50 grand to get these in the speaker xover (home use/Dynaudio Temptation).
Very large capacitor is made by Hovland (Musicap). Made a TREMENDOUS difference in a test circuit - very smooth, detailed, revealing. The cap is nearly 3 inches long and 1 inch in diameter. About 45 dollars each. The cheapest cap is made by Solen and that's used in a trap circuit - 6 dollars.
Best way to wire a xover is doing it by "point to point" method with some really good wire. Use silver solder if possible. With point to point, and if done correctly, more current can flow. A typical home speaker setup would probably have huge speaker wire runs only to narrow down to 18 guage in the speaker internals. Some may argue that the 18 guage doesn't make a difference in a short run, but it does. Check out the xover (fully modded) on this page...
http://northcreekmusic.com/801web.html
I would have used the Northcreek inductors (8 guage), but it wouldn't fit in the speaker box. Maybe for my next evil project though...
Percy
Working on a crossover modification - think you'll like this. On the left is a pic of the modded crossover. On the right, a stock crossover. The stock xover is used in the Dynaudio Audience 50 home speaker. The modded one hasn't been powered up yet and is in the works.
If you look at the pic of the stock xover, you'll notice small coils and smallish capacitors. The resistors are the white blocks, which are wirewound and sandcasted to help tolerate the heat. Dyn built the speaker as an "affordable" alternative, and it shows. The one BIG no-no that they did was to use an IRON CORE INDUCTOR. With iron core inductors, you can get the same mH (millihenry) ratings as a larger air core inductor...but with one problem. Hysteresis.
Definition of hysteresis from Northcreek audio...
"What about ferrite core inductors, which have very low DCR?
The problem with ferrite core inductors - all of them - is hysteresis distortion. Hysteresis is related to the energy loss incurred by the core material during the magnetization polarity shift that occurs in every cycle of the signal flowing though the inductor wire - for a 500 Hz signal, this is 500 times the core polarity has to be reoriented, per second! Hysteresis distortion sounds like compression - and also measures as wave-form compression. Furthermore, when the current though the inductor is large enough, the core will saturate and the value of the inductor dynamically shifts to a much lower value. This is also clearly audible.
The only practical application for ferrite-core inductors in high end audio is as low DCR "trap" coils in tweeter resonance trap circuits, where the circuit is functioning in a very narrow frequency window and the current flow is extremely small."
Unfortunately, the Dyn crossover for this application is an iron/ferrite core. If you see ANY iron cores in a crossover, it means that they're cutting corners and not making a closer to ideal crossover. IMO, you should not use an iron/ferrite core inductor when it comes to the audio signal.
The xovers in the car audio Dyn setups are all air core, which is a step in the right direction. Solen caps are used, which are ok. In the Dyn/Audience setup, they didn't even use a Solen cap, but rather some off the shelf no name cheapie. The BIGGEST OFFENDER is the Rainbow Reference xover. LOTS of iron/ferrite/laminate core inductors all over the place...with discrete no-name caps. Is this REALLY worth the 7 to 8 grand paid for their "premium" speakers? Will have to scan in a picture for you guys/gals to show you the HORROR of their xovers.
Switch over to the left side of the picture. Notice that the coils are HUGE in comparison to the little guys on the right. They're the SAME values, but the larger ones are made out of copper foil. Much lower dcr (dc resistance) and the build quality is fantastic. They're made from Alpha Core/Goertz. Many speaker manufacturers (including Dynaudio) almost never reach this level of inductor. With Dyn, you have to pay nearly 50 grand to get these in the speaker xover (home use/Dynaudio Temptation).
Very large capacitor is made by Hovland (Musicap). Made a TREMENDOUS difference in a test circuit - very smooth, detailed, revealing. The cap is nearly 3 inches long and 1 inch in diameter. About 45 dollars each. The cheapest cap is made by Solen and that's used in a trap circuit - 6 dollars.
Best way to wire a xover is doing it by "point to point" method with some really good wire. Use silver solder if possible. With point to point, and if done correctly, more current can flow. A typical home speaker setup would probably have huge speaker wire runs only to narrow down to 18 guage in the speaker internals. Some may argue that the 18 guage doesn't make a difference in a short run, but it does. Check out the xover (fully modded) on this page...
http://northcreekmusic.com/801web.html
I would have used the Northcreek inductors (8 guage), but it wouldn't fit in the speaker box. Maybe for my next evil project though...
Percy
Last edited by Percy; 12-01-05 at 05:33 PM.
12-04-05, 10:45 PM
#3
Moderator - Electronics Forum
Thread Starter
_ear,
Just had a chance to ABX test the stock vs modded xovers. Had my little brother do the listening between the two since he didn't know they were modified in any way. He just thought they were STOCK speakers. One speaker had the modded xover, the other was a completely stock xover. Both speakers were playing LEFT channel information as you really don't want to try to reproduce a stereo image with one stock and one modded speaker. It'll drive you nuts.
Source was a Sony SCD-1 (modded) SACD/CD player running through a Denon AVR2500. First track was Billy Joel "An Innocent Man". He listened to the first 30 seconds and pointed out that one of the speakers sounded BRIGHT, the other one sounded more mellow. Second song of "Leningrad", also by Billy Joel yielded the same results. He would always choose the more mellow of the two. Third track from the Beverly Hills Cop soundtrack, Glenn Frey, "Heat is on". He then said "hey, waitaminute...were these speakers changed in any way?" I didn't tell him at first but rather asked him what he heard. He described the modded (didn't tell him it was modded until later) speaker as more "open" and a better midrange on the modded xover. I asked him how much as a percentage number...he said "about 25%".
This was with 3 quick 30 second clips from the above tracks. I then told him that the speaker he went with was the modded one. He has a good ear as he was a first violin in orchestra for many years.
Here's what I heard during the ABX.
Stock speaks are normally on the bright side, call them "peaky" if you will. "S" sounds are quite pronounced, which is near the 8khz region. Tweets are seemingly directional, not too forgiving on position of the listener. Mids are clean, but a bit muddy/veiled, sorta like taking a semi thick down comforter and covering the speaker. Small details in the soundstage get lost. You hear the main note, but not too much of the decay after each note. (Piano) Voices are realistic, but are more "flat" than anything else. It's like an oil painting with it's textures reproduced on a flat sheet of cardboard. Very little dimensionality.
Modded xovers - midrange is more open and subtle textures become quite apparent. What was easy to miss in the stockers was more "visable" with the modded. Treble region was also MUCH smoother, much less hard. "S" notes weren't spitty, though they were still a bit on the bright side. I'm not surprised since they're Dynaudio D28/2 tweeters with new domes. As the domes play more through them, the suspension breaks in more and becomes more mellow. The soundstage also revealed much more depth and dimensionality. With the stockers, the depth of soundstage seemed like 5 feet. The modded...more like 25 feet.
Well worth the effort in modification. BTW... the modded xover BARELY fits into the speaker cabinet now. If I used the 8 guage Northcreek inductors, I would have had to place the xovers on the outside of the cabinet!!!
Better inductor = more texture in midrange, stick with AIR CORE ONLY!!! Better capacitor = more spacious treble. These are 6db/oct impedance corrected networks.
Also, keep the inductors AWAY from each other as they will have interaction with each other if placed too close. Their inductive value actually changes.
Scott Buwalda (IASCA guru...heavy competition) did try to make custom crossovers for his vehicle using Alpha core inductors. He mentioned to me that it was a headache. Don't think he used Hovland Musicaps due to their size...they're almost 3 inches in length and 1 inch wide. A little stick of dynamite?
Have to say the "cost nearly no object" crossover was a success!
Now to mod out some Confidence C1 speaker xovers...evil project for 2006.
Or maybe the Alpine F#1 Status?
Percy
heading back to the soldering (4% silver bearing) iron!
Just had a chance to ABX test the stock vs modded xovers. Had my little brother do the listening between the two since he didn't know they were modified in any way. He just thought they were STOCK speakers. One speaker had the modded xover, the other was a completely stock xover. Both speakers were playing LEFT channel information as you really don't want to try to reproduce a stereo image with one stock and one modded speaker. It'll drive you nuts.
Source was a Sony SCD-1 (modded) SACD/CD player running through a Denon AVR2500. First track was Billy Joel "An Innocent Man". He listened to the first 30 seconds and pointed out that one of the speakers sounded BRIGHT, the other one sounded more mellow. Second song of "Leningrad", also by Billy Joel yielded the same results. He would always choose the more mellow of the two. Third track from the Beverly Hills Cop soundtrack, Glenn Frey, "Heat is on". He then said "hey, waitaminute...were these speakers changed in any way?" I didn't tell him at first but rather asked him what he heard. He described the modded (didn't tell him it was modded until later) speaker as more "open" and a better midrange on the modded xover. I asked him how much as a percentage number...he said "about 25%".
This was with 3 quick 30 second clips from the above tracks. I then told him that the speaker he went with was the modded one. He has a good ear as he was a first violin in orchestra for many years.
Here's what I heard during the ABX.
Stock speaks are normally on the bright side, call them "peaky" if you will. "S" sounds are quite pronounced, which is near the 8khz region. Tweets are seemingly directional, not too forgiving on position of the listener. Mids are clean, but a bit muddy/veiled, sorta like taking a semi thick down comforter and covering the speaker. Small details in the soundstage get lost. You hear the main note, but not too much of the decay after each note. (Piano) Voices are realistic, but are more "flat" than anything else. It's like an oil painting with it's textures reproduced on a flat sheet of cardboard. Very little dimensionality.
Modded xovers - midrange is more open and subtle textures become quite apparent. What was easy to miss in the stockers was more "visable" with the modded. Treble region was also MUCH smoother, much less hard. "S" notes weren't spitty, though they were still a bit on the bright side. I'm not surprised since they're Dynaudio D28/2 tweeters with new domes. As the domes play more through them, the suspension breaks in more and becomes more mellow. The soundstage also revealed much more depth and dimensionality. With the stockers, the depth of soundstage seemed like 5 feet. The modded...more like 25 feet.
Well worth the effort in modification. BTW... the modded xover BARELY fits into the speaker cabinet now. If I used the 8 guage Northcreek inductors, I would have had to place the xovers on the outside of the cabinet!!!
Better inductor = more texture in midrange, stick with AIR CORE ONLY!!! Better capacitor = more spacious treble. These are 6db/oct impedance corrected networks.
Also, keep the inductors AWAY from each other as they will have interaction with each other if placed too close. Their inductive value actually changes.
Scott Buwalda (IASCA guru...heavy competition) did try to make custom crossovers for his vehicle using Alpha core inductors. He mentioned to me that it was a headache. Don't think he used Hovland Musicaps due to their size...they're almost 3 inches in length and 1 inch wide. A little stick of dynamite?
Have to say the "cost nearly no object" crossover was a success!
Now to mod out some Confidence C1 speaker xovers...evil project for 2006.
Or maybe the Alpine F#1 Status?Percy
heading back to the soldering (4% silver bearing) iron!
12-05-05, 04:07 PM
#4
Originally Posted by Percy
Well worth the effort in modification. BTW... the modded xover BARELY fits into the speaker cabinet now. If I used the 8 guage Northcreek inductors, I would have had to place the xovers on the outside of the cabinet!!!
What, you think it's a problem not being able to fit the crossover in the speaker cabinet??? It's a feature!!! You could make a nice plexiglas box for the top and put the crossover on prominent display and charge an extra $3K for the speakers. Eh? Make sure you solder neatly.... ;-)
Jerry
Thread
Thread Starter
Forum
Replies
Last Post
olddog
Lexus Audio, Video, Security & Electronics
8
12-01-06 08:38 PM
LEXISM
Lexus Audio, Video, Security & Electronics
1
02-12-04 11:13 PM
