I really can’t say enough about fuzzmeasure; easy to use, great looking graphs, distortion from a chirp measurement… the list goes on and on.  I even noticed that when you view a fuzzmeasure file in the finder it has a great preview and even works with quicklook!  How bout that for mac awesomeness!

As for the sub, it took a really long time to get it to sound as good as it does setting it up by ear, so I am hesitant to change anything, but sometimes I cant help myself.

Click on any image above and use keyboard arrow keys to quickly compare images.

As you can see, the 3rd harmonic doesn’t change apreciably. The 2nd on the other hand increases dramatically with each doubling of the SPL.

This one has a little more informed crossover point.  I ended up with 1800Hz LR4.  Hopefully this will be low enough to attenuate the peak in 3rd harmonic on the woofer (centered @ 1600) as well as get rid of most of the weird peak in high order harmonics in the tweeter around 2k.   The impedance looks ok, only dropping to 4 ohm min in the 3k-4k region.  Listening to one speaker with the crossover emulator built into soundeasy this crossover sounds pretty good.

Alright, crossovers are fun, but I guess I am getting a little ahead of myself.  I am now actually doing the research to find the propper crossover point (instead of guessing).  With the help of some of the guys over at the HTGuide forum I think i will be able to get this sorted.  Time for the measurements:

Minimum Phase
HDS Nomex	9930
Click the links above to compare plots.

Percent Distortion
HDS Nomex – Near-field	HDS Nomex – 10″ mic. dist.
9930 – Near-field	9930 – 10″ mic. dist.
Click the links above to compare plots.

The difference between the near-field and 10″ measurements makes me a little nervous.  The level of the driver did not change, only the mic position.  The huge difference in distortion levels must be a result of the mic distortion as the driver should have been producing the same distortion regardless of mic position.  The mic / preamp did not clip in these tests, but I think it is obvious that they produce a large percentage of the measured distortion themselves.  The question becomes; is the mic distortion consistent?  Or does it vary greatly with frequency?  Can these measurements be trusted at all?  The good news is that they do show aprox. the same patterns, even if the details change with mic position.

So, I took the test box outside for some measurements.  This was my first time trying to measure outside.  My inside measurements were always repeatable, but I never quite trust them.  Like I said, this was my first time measuring outside, so there was a learning curve.  There was some ambient noise, I can hear the freeway from my house, some construction in the distance ect.  Its amazing how much noise there is outside if you actually stop and listen!  I figured that with 6 cycles averaged together I could get rid of most of the ambient.  The thing I didn’t realize was going to be such a problem was the wind.  It wasn’t heavy wind, but with the puffs of air I could go from good measurements to clipping the mic-pre in a heartbeat.  I need a windsock.  Anyways, eventually I got consistent and repeatable results that I am pretty confident in.  Measurements taken at 1.5m using ARTA’s sweep excitation for impulse measurements.

The next step was taking everything into Soundeasy and whipping up a crossover.  After some fiddling I ended up with a LR4 @ 2200.  This was more a choice of convenience of hitting the targets than any precisely contemplated crossover point.  I don’t think the woofer will be having trouble by then, and I think moving the crossover up when using a woofer that can handle it gives the tweeter a little extra headroom.  Here is the basic info:

Like I said, this isn’t a finished crossover.  It is promising though.

The Drivers:

After looking at a lot of different drivers I decided that I should choose a well behaved woofer. After all, my last project never got anywhere because I just never figured out how to tame the drivers complex breakup. So this time I want paper or poly. But which one? I decided to try both. So I have a Peerless 830875 6.5″ Nomex woofer and a Peerless 830874 6.5″ Poly woofer. These will be mated to the Usher 9930 tweeter. I think these will go well together. But we will see.

The Box:

After doing some modeling in Soundeasy I decided to build a relatively exact replica of the parts express .5′ box. I am now wondering if this is a little small for this woofer and since I built the damn thing myself I am wondering why I didn’t just make it a little bigger. One thing I have noticed already… .5 cubic feet is pretty big. I was thinking this would a relatively small speaker. Not the case.

Here is the basic tuning info with a 2

This is the modeled impedence vs. the measured impedence. Modeled on top, measured beneith.

There are already a few things that I think I am going to change before going on to the next step. The baffle is currently removable. I don’t like it. It doesn’t seem as rigid as a completed box and the supports that I am bolting it too keep breaking, even though they are made of oak. I think I will also move the woofer closer to the tweeter. It didn’t look that far away when I drew the baffle on the computer, but I don’t see any reason why it couldn’t be 1″ closser.

Here is a different box and tuning I am thinking of:

And an overlay of the old with the new:

What I hope to find in all of this is a resonant behavior that can be correlated to the subjective sound quality of each plinth.  I have a feeling that the actual behavior that improves the perceived sound quality will be a systemic one.  In other words it will be more complicated than just the plinth, it will involve the plinth and its connection to the top plate, springs and sub-chassis.  Unfortunately, finding a systemic behavior like that is very difficult.  So, instead of looking for the cause of the SQ difference I am going to be looking for a direct correlation between the plinths measured behavior and the SQ, even if it is not the actual causation.

The cover is not hard to remove, or replace for that matter. All you need to do is grab it with some small pliers and pull… that’s it. It is held on by some tacky glue and can be replaced with just the power of the magnet and the tackiness of the glue. The only thing to watch out for is the possibility of the cover being pulled into the delicate dome by the magnet, so just be careful and make sure you have a good grip. These measurements were taken on a poorly planned baffle and therefor include some serious diffraction effects.

Frequency Response with Grid

Frequency Response without Grid

As you can see, the phase shield effects the top octave dispersion pattern and not much else.  Personally I don’t see the reason for the shield from these plots, the response without the grid looks better to me, but I have heard it argued the other way too and obviously SEAS thinks it is better with the shield, or they wouldn’t have put it in there.

So does the phase shield effect any other aspects of the driver?  Here are some distortion plots with and without the shield:

Inter-modulation Distortion with Cover (30 degrees off axis)

Inter-modulation Distortion without Cover (30 degrees off axis)

Swept Distortion with Grid (on axis)

Swept Distortion without Grid (on axis)

Swept Distortion with Grid (30 degrees off axis)

Swept Distortion without Grid (30 degrees off axis)

So, what does it all mean?  It means I wasted a lot of time!  The only difference I can measure is the FR difference above 10k.  So leave it on?  Take it off?  I would take it off.  Testing with my ears is always part of the process too, and there is definitely an audible difference between the two.  I prefer the sound without the phase shield.  The good news is that it is the little sticker inside of the grid that acts as the shield, so if you want you can remove the sticker and keep the shield on.  The even better news is that since it really doesn’t change the frequency response below 10k you can do this with any of the available designs that use this popular tweeter and not have to worry about ruining the crossover.