Zaph|Audio - NHT XDS Passive Crossover

Jack's crossover works well. Don't be too afraid of the breakup node that still pokes up, it's still 20 dB down and for many that will be enough. For those that build it, have no fear, you have a crossover that sufficiently does the job. What follows below is my own version, which is another way of doing it. For this little exercise, I took all of my own measurements to work with. The good news is that they look pretty close to the supplied measurements.

Raw in-box driver measurements

For standard test conditions, please see the various test groups elsewhere on this website. Note that the woofer measures extremely flat on my test baffle which is an important point as you read into the next section.

I had noted elsewhere that I've seen some measurements of Excel drivers where there appears to be a shelf in the response between 600 and 900 hz. I've never come across that issue myself. My first measurement of the XDS woofer removed from the enclosure and mounted to my test baffle was perfectly smooth. I put it back in the enclosure and surprise, there was the 600-900 Hz cliff, though to a slightly smaller extent than what is shown in NHT's measurements. Note that these are 1cm near field measurements, with the mic placed closer to the center of the cone than the edge. As such, baffle step and diffraction are largely removed from the picture. A very large gating window allows a lot of detail to be shown, although this very near field measurement rolls off the top end of the drivers response.

I have done several Excel W15 systems myself and none of my systems have shown a cliff of that magnitude. I just happened to have a couple of W15 systems in the house at the moment so I dropped the XDS woofer into one and measured it. The cliff was gone. It is absolutely, most definitely caused by the NHT enclosure and is not any sort of defect or resonance in the driver.

Comparing my W15 enclosure to the NHT XDS W15 enclosure shows only one real difference. My woofer hole has a back side chamfer while the NHT does not. The NHT XDS enclosure has a 3/4" straight through hole. I knew that kind of mounting did have some effects from previous tests, but it appears that some drivers are more sensitive to a flat non-chamfered woofer hole. It's still not exactly clear if this is a restriction or reflection issue.

Straight woofer holes do not always appear to have that big of an effect. In the SR71 design with the Seas Reed cone woofer, the baffle was a Madisound prefab standard with a straight hole and I did not see much of a response issue as a result, perhaps because of the increased open area of the 17cm woofer. But I think it's pretty obvious that the smaller a driver gets, the more it needs to breath. The NHT XDS enclosure seems to be a good example of the importance of a rear woofer hole chamfer.

It's a little disappointing to know that not every detail was considered in the design of such a beautiful enclosure as this, but it's not that big of an issue anyway. When you read on below, you'll see that I did not feel the need to compensate for this woofer response issue.

After listening to Jack's crossover for some time and becomming familiar with it's sound, my crossover reflects a couple of minor changes that I prefered - A bit more baffle step compensation and more control of the woofer's breakup node. The woofer filter is simply a series inductor with a 2 component LC bottomless notch.

I provide a vertical polar response because it's a far better way of evaluating the phase tracking of even order crossovers than the reverse nulls so many people do. For this system, the design axis is ears level with the tweeter, but the polar response maintains a flat response above this. In other words, you're not going to have a hole in the response when you stand up.

Jack's tweeter crossover worked almost perfectly, rolling off the tweeter just before 3rd order distortion issues. I found no real need to change it too much, except for adding a shunt resistor as a level adjustment tool and tweaking the values to available standards.

I found it hard to evaluate these speakers without a subwoofer in place, mostly because bass extension was lacking so much as to overshadow the speaker's other issues. It would be easy to mistakenly select too much baffle step compensation in an effort to fix the lean tonality. Once a sub was in place and properly set up with a steep active crossover, the only thing I noticed was a midbass that was a bit low in level. Male vocals with accompaniment brought out this issue. Upsizing the series inductor to 2.0 mH seemed just about right, and also helped make the 600-900 Hz cliff less noticeable as glare in female vocals. On the other hand, the speaker just lost another dB or two of sensitivity, so be sure to have lots of power to make up for it. With a sub handling everything under 100Hz and a huge amp, these speakers got way louder than I would ever want and really showed one thing Seas has always been good at: dynamics and power handling.

Ideally, I'd recommend at least 100 watts to each satellite. If you don't have that, you might want to bump the series inductor back up to 1.7 mH, adjust the tweeter level up, and live with the thin midbass as a good compromise. Flea power tube amps should not be used, especially since I didn't work towards a flat impedance anyway. Regardless of the amplifier watts available, I still recommend tweaking to personal taste. These options are provided for that reason.

Other comments

This design was optimized for my standard conditions - mounting on stands a couple feet out from the wall, a few feet from the side wall and a medium size room with carpeting.

Given the easy connection to the drivers with banana plugs and the cramped cabinet interior, I'd probably recommend an outboard crossover. It could be mounted in a box and placed on the floor behind the speaker, or built into the stands.

It's too bad the normal Excel copper phase plug has been blacked out in this design. It would look great with the vanilla and maroon cabinet colors. Excel woofers can have the phase plug removed with a screw in the back of the motor, though I failed to notice if it was accessable with the sheilding cup in place. If you can get at it, maybe you could mount it to a lathe and sand off the black.

The Excel W15 woofer is one of my long time favorites, and this system easily reminds my why. It's very clean, dynamic and smooth. This system's tweeter is also excellent, at least above 2300 or so.

Page done by John "Zaph" Krutke © 2009
Also visit -Zaph|Audio-

Related Speaker Designs

• SR71 - 2-way Bookshelf
• ZBM4 - MTM Design
• ZMV5 - Compact 2-way
• C500 - Tower Speaker
• DLR - Full Project