Klipsch 6.5" - Peerless 4" Nomex - Vifa NE19VTS Xover attempt

[J_Shea]

The frd and zma files for the Klipsch 6.5" and the Vifa NE19VTS I got courtesy of Zaph. The frd and zma files for the Peerless 4" HDS 830872 are courtesy of Augerpro's photobucket here: (then traced with SPL Copy)
http://s139.photobucket.com/albums/q288/augerpro/Peerless%20830872%20HDS%20Nomex/

I'm struggling to learn how to use PCD 7 and hoping that somebody will point out any egregious errors I've made so far. This crossover attempt is just to get a rough idea how to use the software, and it does not account for BSC or baffle diffraction yet.

http://thumbnails26.imagebam.com/11675/18e1ee116742234.jpg (http://www.imagebam.com/image/18e1ee116742234) http://thumbnails35.imagebam.com/11675/2311ad116742254.jpg (http://www.imagebam.com/image/2311ad116742254) http://thumbnails14.imagebam.com/11675/98f1f6116742259.jpg (http://www.imagebam.com/image/98f1f6116742259) http://thumbnails37.imagebam.com/11675/f5d2c4116742264.jpg (http://www.imagebam.com/image/f5d2c4116742264) http://thumbnails30.imagebam.com/11675/0fc9b8116742269.jpg (http://www.imagebam.com/image/0fc9b8116742269) http://thumbnails26.imagebam.com/11675/4531b7116742273.jpg (http://www.imagebam.com/image/4531b7116742273) http://thumbnails9.imagebam.com/11675/8cb0f2116742281.jpg (http://www.imagebam.com/image/8cb0f2116742281)

[arlis_1957@yahoo.com]

Your off to a good start, but you should put them through rm first.

[tom_s]

You also need to look at the phase of each driver and how they align. As Arlis is alluding to, once you run the drivers through RM positioned on a proposed baffle, everything will change. I would use a little padding on the mid & tweeter. It might sound a little lean without.

[LouC]

You will like the mid & tweeter. I used that mid on the Braptor (http://www.lonesaguaro.com/speakers/Buyout/BravoxTower.html) and the tweeter on the Piccolata (http://www.lonesaguaro.com/speakers/TB4V/Piccolata.html). I haven't used the Klipsch yet, but I have 4 out there.

Let me know if you want any of my measurement files (in the box)

[J_Shea]

You will like the mid & tweeter. I used that mid on the Braptor (http://www.lonesaguaro.com/speakers/Buyout/BravoxTower.html) and the tweeter on the Piccolata (http://www.lonesaguaro.com/speakers/TB4V/Piccolata.html). I haven't used the Klipsch yet, but I have 4 out there.

Let me know if you want any of my measurement files (in the box)Cool, thanks Lou.
You're right I do like that mid and tweeter. I've been listening to them in a test box crossed over with a rudimentary 3kHz lr2 textbook, and fiddling with equalizer settings for the last month. The tweeter sounds very clean, and seems to need about 2-3db of padding from the mid when it's used in a 2-way. That's by ear, I don't have measuring equipment yet. The mid has a smooth sound that I like a lot, but there is a 'flatness' to the bass. I don't know how to describe it better. Also, that peak at around 6kHz is clearly audible.

I like the Klipsch woofer a lot too, so long as it's kept below ~1kHz. It has a meaty sound on bass instruments that is great imo. :D

So far, using Response Modeler is even trickier than using PCD. Is there a tutorial somewhere that specifically addresses Response Modeler? I think I'm doing it right, but it's hard to be sure.

[ligs]

I like the bass of Klipsch 6.5 too. "meaty" is a good description and I like that.

[J_Shea]

Here's the second attempt (not counting the hundreds in between) at a crossover for these drivers. This time the full 6dB of BSC is accounted for along with the diffraction hump for the intended baffle for all drivers in their respective positions. The ported box response for the woofer, and the sealed internal box response for the midrange were spliced before adding the BDS to the model.
The inductors are all 20g except the low pass for the woofer which is 18g.

How does this look? Any help would be greatly appreciated.

http://thumbnails34.imagebam.com/11782/12e865117814274.jpg (http://www.imagebam.com/image/12e865117814274) http://thumbnails15.imagebam.com/11782/673a91117814278.jpg (http://www.imagebam.com/image/673a91117814278) http://thumbnails36.imagebam.com/11782/f1519f117814286.jpg (http://www.imagebam.com/image/f1519f117814286) http://thumbnails35.imagebam.com/11782/670a6b117814293.jpg (http://www.imagebam.com/image/670a6b117814293) http://thumbnails39.imagebam.com/11782/7425bf117814298.jpg (http://www.imagebam.com/image/7425bf117814298) http://thumbnails35.imagebam.com/11782/4da3db117814303.jpg (http://www.imagebam.com/image/4da3db117814303) http://thumbnails16.imagebam.com/11782/405038117814314.jpg (http://www.imagebam.com/image/405038117814314)

[Wolf]

That's much better, but I think you need to bump up your tweeter one order.
Due to differing Z-offsets, the asymmetrical approach on xovers is a good idea. The one order shift on the tweeter to make it 3rd on tweeter and 2nd on mid will help align your phase much more efficiently.

If you want more tips, I have a few sections in my blog here at PE.
Later,
Wolf

[J_Shea]

So, I changed the tweeter high pass to 3rd order which made it a little easier to deal with a dip in the FR from 7 to 9 kHz. But, everything changed. Or nearly everything did. I moved the midrange crossover point down, and switched the inductor on the woofer to a low Re Jantzen P Core. For some reason that made it easier to get the phase to behave while avoiding creating too big a hump at around 400 to 600 Hz.

This is very difficult. Each little change has an effect on multiple other things, and sometimes a small beneficial change one place makes a big change for the worse somewhere else. The design tools are scientific, but using them seems to be an art form. It is fun, but my brain hurts.

EDIT: Oh, and I eliminated some resistors in the tweeter high pass, and ended up with almost 2dB higher sensitivity everywhere.

How does this iteration look?


http://thumbnails15.imagebam.com/11785/f33b88117844647.jpg (http://www.imagebam.com/image/f33b88117844647) http://thumbnails29.imagebam.com/11785/dd7df2117844652.jpg (http://www.imagebam.com/image/dd7df2117844652) http://thumbnails30.imagebam.com/11785/89808d117844662.jpg (http://www.imagebam.com/image/89808d117844662) http://thumbnails37.imagebam.com/11785/c86116117844665.jpg (http://www.imagebam.com/image/c86116117844665) http://thumbnails40.imagebam.com/11785/01b931117844673.jpg (http://www.imagebam.com/image/01b931117844673) http://thumbnails39.imagebam.com/11785/fb9011117844678.jpg (http://www.imagebam.com/image/fb9011117844678)

[MSaturn]

Taper that high end with a zobel circuit on the tweeter.

[J_Shea]

I tried to do that, but it doesn't seem to work very well for this tweeter. There's a big 2dB trough in the FR from 7 to 9 kHz, and the impedance rises very little from 2 kHz to 20 kHz. It just hits 4 ohms at 20 kHz. The zobel just seems to drop response everywhere by about 3 db, and has no effect on the tilt. Tweaking the zobel values for a flatter response sends the phase all over the place.

[Pete Schumacher ®]

I tried to do that, but it doesn't seem to work very well for this tweeter. There's a big 2dB trough in the FR from 7 to 9 kHz, and the impedance rises very little from 2 kHz to 20 kHz. It just hits 4 ohms at 20 kHz. The zobel just seems to drop response everywhere by about 3 db, and has no effect on the tilt. Tweaking the zobel values for a flatter response sends the phase all over the place.

Don't even try to taper the response. Just listen slightly off axis, 5-10 degrees, and you'll be golden.

If you MUST taper, you'll need a very small series inductor.

[J_Shea]

That's a good idea, thanks Pete.
These are for my brother, and it looks like his seating position will be ~10 degrees off axis if the speakers are toed in about 5 degrees. Now I need to come up with a center channel speaker design using the same drivers for him. Unfortunately those 4" HDS nomex speakers are NLA, so I can't use them as a basis for the rear surrounds. I only have 4 of them on hand, so I would need one more to finish a 5.1 setup. The similar 5 1/4" nomex drivers that I had in mind for rear surrounds are also NLA.

[bmaupin]

Here's what I did with the NE19 to deal with the rising top end. L11 is 0.11mH and R13 is 8ohms.

https://home.comcast.net/~bradmop/response_files/NE19_4th@2200.jpg

Here's the tweeter's measured in-box response:

https://home.comcast.net/~bradmop/response_files/NE19.jpg

[bmaupin]

Unfortunately those 4" HDS nomex speakers are NLA, so I can't use them as a basis for the rear surrounds. I only have 4 of them on hand, so I would need one more to finish a 5.1 setup.

There is one for sale in the for sale section: http://techtalk.parts-express.com/showthread.php?t=222722

[J_Shea]

Here's another try. I made lots of changes. First I added a notch to the midrange at the first breakup point, and changed the high pass to 3rd order. Those changes made it easier to blend the mid with the woofer and tweeter. Second I added a zobel to the woofer to smooth out a hump in the midbass. Third I added a sort-of filter to the tweeter to tame the rising response. Fourth I changed the geometry to make matching the phase easier. The layout stays the same on the baffle, but the baffle is now angled backwards 4 degrees to bring the woofer's acoustic center in line with the tweeter's.

Here's what it looks like. As you can see the FR is pretty flat, but the tweeter's phase doesn't track too well with the mid:

http://thumbnails30.imagebam.com/11801/f41a94118001607.jpg (http://www.imagebam.com/image/f41a94118001607) http://thumbnails34.imagebam.com/11801/8781da118001614.jpg (http://www.imagebam.com/image/8781da118001614) http://thumbnails29.imagebam.com/11801/16592f118001618.jpg (http://www.imagebam.com/image/16592f118001618) http://thumbnails27.imagebam.com/11801/fae7f6118001621.jpg (http://www.imagebam.com/image/fae7f6118001621) http://thumbnails36.imagebam.com/11801/1a748e118001626.jpg (http://www.imagebam.com/image/1a748e118001626) http://thumbnails17.imagebam.com/11801/e873f8118001632.jpg (http://www.imagebam.com/image/e873f8118001632) http://thumbnails27.imagebam.com/11801/13f076118001636.jpg (http://www.imagebam.com/image/13f076118001636)

[2leftthumbs]

I see with interest that the K6.5 diver is now sold out.http://www.parts-express.com/pe/showdetl.cfm?Partnumber=299-101&FTR=299-101

[J_Shea]

Yeah, it's a shame those drivers are all gone. They are really neat and at the original price of $15 each, a heck of bargain.

All of these crossover attempts I posted are complete bollocks by the way. I failed to extract minimum phase for the .frd and .zma files in Response Modeler before importing them into PCD. Trying to get these three drivers to behave with each other is much, MUCH more difficult than I thought.
I'm still plugging away at it with some progress, but the very low impedance of the tweeter is problematic.

[ocdSCHACK]

If you have too much trouble... just trash the Klipsch drivers. They have a big magnet and a crappy sound. I thought I liked them, until I tried something else. They can give a designer headaches that aren't worth the final product. Not that they can't sound good, but after the costs of crossover components and frustration, they can't go near anything decent.

[johnnyrichards]

Trying to get these three drivers to behave with each other is much, MUCH more difficult than I thought.
I'm still plugging away at it with some progress, but the very low impedance of the tweeter is problematic.

Well, all I can say is that shooting for a 3-way design can be frustrating. WRT the tweeter impedance, you can try adding an L-Pad plus an additional series resistor. If you are crossing high enough (~2.5x the Fs or higher) you might be able to skip the L-Pad. I rarely do much more than a single series resistor in tweeter circuits myself. Then again, my listening levels are such that any uglies from not suppressing resonance peaks in the tweeter rarely show up.

I am as big a proponent of measurements and simulations as the next guy, but sometimes I think we place too much emphasis on how they look rather than giving it a listen and correlating what we like with the measurements rather than the other way around. I truly beleive that we can measure what we like, it is just making that correlation that can prove to be difficult.

[ligs]

Courtesy of Paul Carmody here is a 2-way design using Klipsch 6.5 with BG Neo-3PDR. I would say the XO is fairly simple and it sounds good.

You can simplify the XO further if you use BG Neo8 PDR as a tweeter.

In my previous career I was a scientist, simulations and measurements were needed to justify what I was doing. Now retired, I try to be more right brained and holistic and just use my senses:)

As with anything memorable, it is the process not the product that matters. So keep tweaking away!

[ligs]

the attatchements for Paul Carmody's design for K 6.5 and Neo3 PDR.

[J_Shea]

If you have too much trouble... just trash the Klipsch drivers. They have a big magnet and a crappy sound. I thought I liked them, until I tried something else. They can give a designer headaches that aren't worth the final product. Not that they can't sound good, but after the costs of crossover components and frustration, they can't go near anything decent.

Did you try them as the woofer of a compact 3-way? They sound really good to me in that role, even with a crossover that's a long, long way from 'ideal'. (phase 100+ degrees out throughout the crossover region so the voicing of the speaker changes dramatically as you move from +30 to -30 degrees off axis left and right... amongst other problems)