DIYers - Need your Crossover Help... XO and tamming the tweeter...

Sonny · « **on:** 16 Oct 2007, 10:39 pm »

Hi there, I am looking for some advice from those of you who are familiar with the Scan Speak D2905/9700 tweeter used with the 18W8545-K Scan Speak woofer. Currently, my speakers "tweeter" is sound a bit harsh, bright and hot in my very reflective room. This only occurred as I switched from the 4 ohm tap on my tube amps to the 8 ohm tap.

On the 4 ohm tap, I seem to get more body and less brightness from the speaker, but more destortion...on the 8 ohm tap, I get better response in terms of playing louder without destortion and more dynamics, but with the 8 ohm tap, I am getting a little leaner sound, therefore rendering my tweeter from detailed to "hot", "bright" and "harsh"...how do I tame the tweeter????

I would like to keep things on the 8 ohm tap, so I would like to tame the tweeter. Here is the XO design below on my current speakers...

Any help is appreciative....
T

Kevin Haskins · « **Reply #1 on:** 16 Oct 2007, 11:59 pm »

Do you have a FR & Z plot?

Sonny · « **Reply #2 on:** 17 Oct 2007, 12:11 am »

Kevin, unfortunately, NO...
I don't have a software to plug in the data to get it either...
The original XO design was this:

but I never tried it this way....was told by the designer that the former schematic, in original post was more transparent...etc...
Thanks
T

*Scotty* · « **Reply #3 on:** 17 Oct 2007, 01:31 am »

Sonny, I would try using the original circuit your amp may like it better and have fewer response anomalies due to the impedance variation at the crossover frequency.
Scotty

aerius · « **Reply #4 on:** 17 Oct 2007, 01:39 am »

Unless I'm reading the schematic wrong, the first x-over has the two 6.8uF capacitors for the tweeter in parallel, effectively making it a 13.6uF cap. In the 2nd x-over, the 6.8uF caps are in series, making them a 3.4uF capacitor. Which means the x-over frequency you have now is a lot lower than in the original, which means your tweeter's getting flopped around too much and distorting. Try yanking out one of the 6.8uF caps in your current x-over and see if things get better.

*Scotty* · « **Reply #5 on:** 17 Oct 2007, 02:10 am »

Sonny,Have you contacted the designer about your problems and did his original design have FR and Z curves associated with it?
Scotty

Occam · « **Reply #6 on:** 17 Oct 2007, 02:17 am »

Sonny,

The first crossover in both versions between tweeter and midrange are series crossovers -
http://www.humblehomemadehifi.com/Crossover.html
The crossover for bass driver in both is a conventional 12db/oct parallel crossover.
Both versions of the series crossovers are discussed in the above reference.
The easiest (not necessarily optimal) way to tame the tweet would be to increase R1, the 1.5 Ohm resistor.

Perhaps someone with more experience/expertise with series crossovers will chime in.....

Sonny · « **Reply #7 on:** 17 Oct 2007, 06:00 am »

Hi all,
Thanks for the advice and links...
I was out at the Hillary Hahn concert at Zellerbach in Berkeley, CA...
I will work on some of the sugestions when I get a moment, hopefully this weekend...but Keep them Coming!!!

She was amazing!!!

What Grace, Beauty and above all, masterful on the Violin...

BobM · « **Reply #8 on:** 17 Oct 2007, 11:43 am »

Quote from: Occam on 17 Oct 2007, 02:17 am

Sonny,
The easiest (not necessarily optimal) way to tame the tweet would be to increase R1, the 1.5 Ohm resistor.

That's what I was going to suggest

But why did you move that 6.8uF cap up to the tweeter circuit? You very likely messed with the balance that the original designer had in mind. Was it too bright before? Was that your first attempt to adjust things? If "yes" then I would suggest putting it back the way it was and play with R1.

Enjoy,
Bob

Dan_ed · « **Reply #9 on:** 17 Oct 2007, 01:38 pm »

Yes, by moving that cap into the tweeter circuit you've added an additional phase change that wasn't there in the original circuit. That could easily explain the difference in what you're hearing.

I'm not at home now or I could plot the freq/phase responses for you and try to post them so you could see what changed.

BTW, it may help to simplify things by forgetting the components that are parallel to the mid and bass drivers. I don't mean physically remove them, just don't worry about them for now. They're only there for impedance correction.

Sonny · « **Reply #10 on:** 17 Oct 2007, 03:17 pm »

Quote from: BobM on 17 Oct 2007, 11:43 am

Quote from: Occam on 17 Oct 2007, 02:17 am
Sonny,
The easiest (not necessarily optimal) way to tame the tweet would be to increase R1, the 1.5 Ohm resistor.

That's what I was going to suggest

But why did you move that 6.8uF cap up to the tweeter circuit? You very likely messed with the balance that the original designer had in mind. Was it too bright before? Was that your first attempt to adjust things? If "yes" then I would suggest putting it back the way it was and play with R1.

Enjoy,
Bob

Bob et All,
Thanks for the advice...
Well, the reason I moved the 6.8 from the tweeter down to the mid was because that was what was recommended by the designer himself...to him, this yield a better sense of detail and high freq. extension...Perhpas this weekend I should try an put things back to the original...if I do that, should I reinstall the 22ohm resistor across the Tweeter?
Thanks
Tuan

Sonny · « **Reply #11 on:** 17 Oct 2007, 03:18 pm »

Quote from: Dan_ed on 17 Oct 2007, 01:38 pm

Yes, by moving that cap into the tweeter circuit you've added an additional phase change that wasn't there in the original circuit. That could easily explain the difference in what you're hearing.

I'm not at home now or I could plot the freq/phase responses for you and try to post them so you could see what changed.

BTW, it may help to simplify things by forgetting the components that are parallel to the mid and bass drivers. I don't mean physically remove them, just don't worry about them for now. They're only there for impedance correction.

Dan, thanks...if you plot it, that would be much appreciative...
Thanks!!!

Dan_ed · « **Reply #12 on:** 17 Oct 2007, 03:30 pm »

Sonny,

did the designer recommend removing that 22 ohm resistor, R2? Seems to me that R1 and R2 are functioning as an L-pad. Without the Z value of the tweeter it is not possible to know what the attenuation is. Removing that resistor, R2, could make the tweeter sound pretty hot.

Sonny · « **Reply #13 on:** 17 Oct 2007, 03:38 pm »

Quote from: Dan_ed on 17 Oct 2007, 03:30 pm

Sonny,

did the designer recommend removing that 22 ohm resistor, R2? Seems to me that R1 and R2 are functioning as an L-pad. Without the Z value of the tweeter it is not possible to know what the attenuation is. Removing that resistor, R2, could make the tweeter sound pretty hot.

Well, the designer did condone the new XO design...that I know...He, Tony of Humble Homemade Hifi, suggested that I 1) increase the value of the 1.5 ohm (R1) to something like a 1.8 or 2.2 ohm and/or 2) add a LCR network across the tweeter with a L of 2.2mH, a C of 47uF and a R of 4.7 ohm...

The funny thing is that on my 4 ohm tap, the tweeter isn't so hot, it's actually great there...

Sonny · « **Reply #14 on:** 17 Oct 2007, 03:41 pm »

Quote from: aerius on 17 Oct 2007, 01:39 am

Unless I'm reading the schematic wrong, the first x-over has the two 6.8uF capacitors for the tweeter in parallel, effectively making it a 13.6uF cap. In the 2nd x-over, the 6.8uF caps are in series, making them a 3.4uF capacitor. Which means the x-over frequency you have now is a lot lower than in the original, which means your tweeter's getting flopped around too much and distorting. Try yanking out one of the 6.8uF caps in your current x-over and see if things get better.

Thanks...I will try a quick bypass of one of the 6.8 caps this weekend and see...
Thanks!

Sonny · « **Reply #15 on:** 17 Oct 2007, 10:45 pm »

I think the issue here is "impedance" of the tweeter...which based on the switching from the 4 ohm to 8 ohm taps on the Tube Amps, makes a huge difference...Perhaps a LCR network across the tweeter will help flaten out the Impedance of the tweeter....
Thoughts???

*Scotty* · « **Reply #16 on:** 18 Oct 2007, 05:25 am »

Sonny,I think it is a question of your amp not liking the impedance it sees at the crossover frequency between the woofer and the tweeter.
Adding an impedance flattening network may do a lot of good good in this case. You can also fudge the impedance of the tweeter up to 8ohms with the L pad. If you do this the crossover parts values will have to reflect this change. I was wondering what the designed crossover frequency of your speaker is? If I was playing with these drivers I would follow the tweeters recommended crossover of 2.8Khz. at 12dB/oct. to avoid distortion problems and maybe cross the woofer over at 1.3kHz at 12dB/oct. to try to suppress the approximately 7dB spike in it's output at 3.2kHz. The idea with the 1.3kHz. woofer xover point is to hopefully wind up with an acoustical roll off of 12 dB/oct at 2.8kHz. thus matching up with the tweeter. The tweeter looks like it will need about 2dB of attenuation to equal the efficiency of the woofer. The original XO design may
accomplish these goals with a better behaved impedance curve at the xover frequency.
Bear in mind that most of my post regarding the drivers and how I would design the crossover is based on the mfg published data does not reflect the drivers real world measured behavior and their interaction with each other and the crossover parts within the context of a series network.
Scotty

Sonny · « **Reply #17 on:** 18 Oct 2007, 06:21 am »

Scotty and All...
Thanks for your responses...
Tonight, I got home and put back the 22 ohm in an L-Pad with the 1.5 ohm....
So far, this seems to do the trick and right now, I am listening to Chet Atkins's "Stay Tuned" and it sounds great!!! I am happy at the moment...

Earlier, I listened to the Patricia Barbers's "Companion" and on cut 2, 5, and "Light My Fire", what was annoying in the highs is now listenable and enjoyable...I don't know why that first time I put in the 22 ohm, I didn't feel the difference...but with it in now, it seems to take the tweeter down about 2 db and that, so far, is doing the trick for me...

FYI, the XO design was done by Tony Gees and some others....

DIY is fun, but can also be a heartache...definitely not your "buy good speakers, hook up, placement and toe-in" and listen procedures...but DIY is absolutely rewarding!!!

Perhaps this weekend, I will move one of the 6.8uF Cap in up to the tweeter like the Original XO design (XO #2 in this thread)...

Thanks All...and stay tuned!!!

Sonny · « **Reply #18 on:** 20 Oct 2007, 04:24 pm »

OK, what's the latest????

Well, spending my Friday night playing with the XOs is not what my GF was liking from me...if it's not X and Os in terms of sports, it's XOs in terms of audio...she puts up with it, but only for so long....

Right...so...here we go....

Last Chatper, I left off with doing nothing but adding the 22 ohm back on the tweeter in the L-Pad to attenuate the tweeter, and I said I was happy, and I was! But then, I got to thinking, what if I reverted back to the the Original XO with the 6.8uF cap back in the tweeter Series and the 1.5 ohm and 22 ohm???
WHAT IF??? So, I de-wired and re-wired the XOs, and was all excited to plug it in...and what happened????

What did the Original XO reveal??? Well,

no really anything good...the bass was not muddled (I guess pulling that 6.8 uf from the Midrange really messed with the balance of the whole speaker...

Then, well, the tweeter was still hot and just sounded compressed, no air, nuttin'

No wonder why they made the change.... So, back to the de-wiring and re-wiring, back to the current with the added 22ohm...now, things are better, clearer, more transparent, balanced, bass / mid is nice, not muddled, tweeter is revealing...and well, everything is back to normal, less a tweeter at approximately 2 db attenuated down...

Ok, for now, that's it! Next Job, get some listening in!!!!

Keep the ideas coming if you have any...I'd be open in trying to design a parrallel 1st order XO for this speaker...here are the details...of the drivers (all Scan Speak):

Tweeter: D2905/9700 6ohm nominal 4.7ohm dc resistance Le of .01
http://www.tymphany.com/datasheet/pdf/ss/D2905_970000.pdf

Mid: 18W8545K 8ohm 5.5 .4
http://www.tymphany.com/datasheet/pdf/ss/18W_8545K00.pdf

Woofer: 21-8555-00 8ohm 5.5 .4
http://www.tymphany.com/datasheet/pdf/ss/21W_8555-00.pdf

Thanks!!!
Tuan

*Scotty* · « **Reply #19 on:** 21 Oct 2007, 03:17 am »

Sonny, neither SCANSPEAK woofer is well suited to a 6dB/oct network. The 21-8555-00 goes nuts at the upper end of it's passband with a peak that goes clear off the chart exceeding 13dB at approximately 3.5kHz. The 18W8545 has that nasty + 6dB peak at 3.2kHz, right about where you might want to start turning it off when crossing over to a tweeter. Both of these drivers really require a steeper slope than 6dB/oct. to adequately suppress their deviations from flat response. Or in the case of your 8in. woofer cross them over far enough away from their bad spot so that you don't have a problem.
I think you should resist the urge to fix something that is not broken. If you are able to just listen to music without focusing on the mechanisms involved you are done.
Scotty