PDA

View Full Version : Beneficial Modifications to the "Keeper" Crossovers Based on Brnger Data



toddalin
11-13-2007, 04:43 PM
As those who have followed my threads may know, I've been using L200 cabinets with 2205s reconed as 2235s for the woofers, LE175 on HL-91 Horns/Lenses for the mids, and 2402s (075s) for the highs. In the original configuration, I used 130As and N1200/N7000 crossovers. The 130As were replaced in the quest for better bass response. At that point, the woofers no longer met the mids leaving what appeared to be a "hole" in the frequency response. It was thought that the crossovers needed to be better taylored to the new woofers.

The "Keeper" crossovers were then constructed. These were based on JBL specs for the N200B crossover (used in the L200 cabinets), with N7000 circuit added on the board. Modifications from original JBL specs were that the resistor in the Zobal network was reduced from 7.5 to 7 ohms (based on lack of availability), and the replacement of the 16.5 mF cap with two 8.2 mF caps in the midrange. Also, the choke for the mid and high pass was selected at 0.5 mH. (The N7000 uses 0.6 mH whereas the N8000 uses 0.3 mH.) All components are of high quality. All caps are Solen and have Theta Audio-Cap 0.01 mF by-pass caps. All chokes are Erse with large gauge winds. All resistors are Mills 12 watt.


http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Crossover1.jpg

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Keeper_Crossover.jpg

In theory, this should be a near perfect match. Even the JBL crossover specs note the replacement of the L200 woofer is the 2235 (though they note that it is not identical).

With the new crossovers installed, all should be well... or so I thought... But NOOOOOOO :biting:. There was still a substantial hole in the frequency response.

Maybe the Alnico woofers had a problem. But they were reconed with original JBL parts in Dec 2005 and remagnitized at that time.

Grumpy came by and we tested them. All parts were original JBL replacement and OCS did the work. But using Woofer Tester-2 they appeared to have a low BL rating (flux density or such) coming in between 17 and 18 (JBL spec for Alnico woofers is ~21). (A low BL factor can account for a reduced top-end response for a woofer or driver.) So yesterday OCS remagnetized again. (Thanks again Eric! :thmbsup:)

Maybe the crossovers also needed further tweeking. Grumpy ran some simulations and it was determined that a reduction in the size of the Zobal cap would push the woofer up a hundred Hz or so without drastically redesigning the crossover. I also replaced the 7 ohm Zobal resistor with a 10 ohm piece as is used with the 2234 in the 3133 crossover.

Based on hours of testing with the Behringer 61 band analyzer, I determined that there was room for further improvement. The horn AND the tweeter both peaked at about 10K adding too much "spittyness" (sibilance). Reducing the volume of either to reduce this peak cut its frequency across the band including areas that didn't need reduction. The 1.5 mF cap in the midrange circuit was changed to a 4.3 mF unit and now the mid is "all done" by 9 kHz, so provides virtually no sibilance. This is now handled strictly by the tweeter.

On the low end of the midrange, the idea was to better fill the aforementioned "hole" and meet the woofer. The 1.5 mF cap left over from the midrange low pass circuit was added to the 16.4 mF high-pass circuit bring this up to 17.9 mF total. (The 3133 crossover uses an 18 mF cap here, so I'm not pusing it any lower than JBL did.) This does add a a couple tenths of a dB to the bottom of the horn's range (and there was no sense in letting these caps go to waste).

The tweeter circuit was also addressed. A 1.0 mF was placed in series 'ala 3133 crossover. Also, not used in the 3133, I added the 7 ohm resistor to ground. This shunts most of the tweeter's enery such that I can actually use its volume control. With the "Keepers" if I more than "cracked open" the knob, the 2402s, were too loud for the rest of the components. While adding the 1.0 mF cap did reduce the volume somewhat, it was still way too loud to effectively use the L-pad. With the 7 ohm resistor, I can still crank the tweeter louder than it needs to be, but it balances really nicely when the knob is now at "4."

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Keeper_Revised_Crossover.jpg

So what did all of this gain???

The biggest improvement was in the transition between the horn and tweeter. There was too much spittyness making prolonged listening fatiguing. Cymbols now have a much more real and articulated sound without all of the extra "sizzle" that actually covered their true sound. Bands like Yes and Asia sound much more pleasing without all of the false "air."

But, I've still got that frequency hole at the top of the woofer's range that drives me crazy. :screwy:

Maybe a floor bounce, or an interaction in the placement of the drivers, or the whole concept of using a 2235 cone on a 2205 basket, or something in the room, or :banghead:

speakerdave
11-13-2007, 06:49 PM
If, when you say "3133" you are referring to the crossover for the 4333/L300, remember, that speaker used the 2312/H92 (longer) horn and crossed over at 800 Hz. The 2307H91 was never used below 1200Hz as far as I know, so if your effective high-pass for the horn is below that, you may not be getting the SPL you expect in the lower part of the treble range.

David

Zilch
11-13-2007, 07:33 PM
The "Keeper" crossovers were then constructed. These were based on JBL specs for the N200 crossover (used in the L200 cabinets), with N7000 circuit added on the board.N200B, actually....


The tweeter circuit was also addressed. A 1.0 mF was placed in series 'ala 3133 crossover. Also, not used in the 3133, I added the 7 ohm resistor to ground. This shunts most of the tweeter's enery such that I can actually use its volume control. With the "Keepers" if I more than "cracked open" the knob, the 2402s, were too loud for the rest of the components. While adding the 1.0 mF cap did reduce the volume somewhat, it was still way too loud to effectively use the L-pad. With the 7 ohm resistor, I can still crank the tweeter louder than it needs to be, but it balances really nicely when the knob is now at "4."Yup, and why I advocated cascading the filters.

7-Ohm resistor to ground (value not shown on your revised schematic) is messing with the impedance. You need a fixed L-Pad there to bring the variable one into a useable operating range. Tweeter highpass is now third order. Put it in Spice to see what's going on there, now....

toddalin
11-13-2007, 09:56 PM
If, when you say "3133" you are referring to the crossover for the 4333/L300, remember, that speaker used the 2312/H92 (longer) horn and crossed over at 800 Hz. The 2307H91 was never used below 1200Hz as far as I know, so if your effective high-pass for the horn is below that, you may not be getting the SPL you expect in the lower part of the treble range.

David

HL-91 was used at 800 Hz with the N200B in the L200 cabinets. HL-91 was also used down to 500 Hz in the old S7 system in the C50 monitors.

toddalin
11-13-2007, 10:06 PM
N200B, actually....

Yup, and why I advocated cascading the filters.

7-Ohm resistor to ground (value not shown on your revised schematic) is messing with the impedance. You need a fixed L-Pad there to bring the variable one into a useable operating range. Tweeter highpass is now third order. Put it in Spice to see what's going on there, now....


Yes, N200B. Thanks for the correction.

Changes to a 3rd order, but doesn't change the phase shift. Ultimately, it measures close to the same except that I gained control of the volume knob. Maybe it rolls off the highs a little more, but that may be a good thing here.

For the center channel I changed the 5 ohm resistor in the horn circuit to a 7 ohm unit to give the horn more volume. I had the horn volume cranked up all the way with the 5 ohm resistor just to keep up with the woofer..., and it still wasn't enough. That LE175 is from a different lot and the horn is a 1217(?) potato masher. Maybe the horn and driver give up some efficiency to the HL-91s.

Wait until you see what I've come up with for the W10GTI/2425-2370s. Really smooth! (Much smoother than this rig.) Even has HF compensation. ;)

Robh3606
11-13-2007, 10:06 PM
Hello Toddalin

You could try tweeking the Zobel a bit. I ran the network with a Zobel with 33uf and 10 ohms?? It's rolling off the top end response. Try something like the second one on the schematic and see what happens.

Rob:)

speakerdave
11-13-2007, 10:15 PM
HL-91 was used at 800 Hz with the N200B in the L200 cabinets. HL-91 was also used down to 500 Hz in the old S7 system in the C50 monitors.

oh

Well, I think they know better now.