[Danny Richie]

Here are the actual measurements of one of the units like yours with the metal grills on them.

The wiggles in the tweeters response 3kHz and 4kHz region are the result of the grills.

[Tyson]

I'm stumped for the moment. The sealed box can't really create a peak like that. The roll off starts at about 150Hz and creates a gradual roll off that ends with a -3db at 62Hz.

What if the box isn't airtight?  What if it lost it's seal?

[Danny Richie]

What if the box isn't airtight?  What if it lost it's seal?

Even if it did then it shouldn't cause a change in output, and certainly not like that.

At the point of his measured peak they should already be 2 to 2.5db down in output.

They are the same woofers as the ones used above as mids. And there are no resistors inline with the mids. So there is nothing to bring the mid's down in output relative to the woofers. It's the woofers that should be down a little compared to the mids due to the higher dcr of the larger inductor.

[Danny Richie]

Let me think on this for a little bit.

I just built out a set of these for my Dad's theater room back in November and the bottom end of his pair was smooth as glass.

[Tyson]

Is the bass in both speakers elevated, or only in one speaker?

[brj]

bconline, that's an impressive attack on the problem - dedicated, thorough, and no small amount of work.  Well done!

(Tyson beat me to a version of the one question I was going to ask!)

[bconline]

Thanks to all of you for the help and suggestions!

To isolate the woofers, I unsoldered them from the crossover but left them connected to each other in parallel. I wired the woofers directly to the amp with a make-do wire through the hole where the terminal cup was. I stuffed a rag in the hole to approximate the sealed cabinet. Like this:



I then measured the drivers in the near-field (1" from the face, no grills). Tyson asked a good question about whether both drivers behaved the same. So I measured each woofer separately and I measured each midrange separately. Next, I reconnected the woofers to the crossover and measured the 2 woofers again. All of this was done in one sitting, with no changes in test conditions.  The results are below. In the diagram:

• Blue is each woofer driven directly, with no crossover
• Red is each woofer driven through the crossover
• Purple is each midrange
• Green is the tweeter

A few comments:

• It is nice to see the consistency between drivers
• Without the crossover, the woofers drive nicely well through the midrange
• Without the crossover, the woofers roll off reasonably on the low end like a sealed cabinet should
• The midrange responses are probably artifically low and wonky because of the open back nature

The response of the woofers with the crossover is very confusing to me.  I thought that it was impossible with a passive crossover to have a response higher than the woofer being driven directly at any frequency. But here I see about a 4 db increase at the peak. How is that possible? Am I missing something?

The other thing I observe is that the rolloff of the woofers on the high side starts at a much lower frequency than I thought it would. And that seems to create this artificial peak. I would guess that if the crossover frequency was at 200 hz instead, and the peak was not there - then the response would be very smooth like Danny designed and like Danny is getting in practice with other X-Statiks.

Could any manufacturing defect in the crossover create these symptoms - like bad solder joints, wrong caps inserted, shorted caps or ???

BTW, both the left and right speakers are behaving approximately the same. But I did not do these detailed measurements on both sides.

[Danny Richie]

I have an idea. It's speculation but it would explain what's going on. I'll shoot it to you in a PM.

[HAL]

Did you measure the DC Resistance of the paralleled woofers after disconnecting them from the crossover?

[bconline]

HAL - Yup. I measured the DCR of the woofers out of the circuit at ~11.4 ohms each, which is exactly what Peerless specs. In parallel they measure just under 6 ohms.

[bconline]

A little progress...

I followed a hunch from Danny that the electrolytic bypass cap may be out of spec. On his suggestion, I unsoldered one leg of the cap from the PCB and re-ran the near-field FR tests.



The new green lines are the woofers with the crossover, and without the electrolytic. The peak is down a little and the low-pass crossover point is out a bit. Still not quite right yet, though.

[bconline]

BTW, here's a picture of the whole crossover PCB.

[Danny Richie]

Going back and looking at all of your measurements has me noticing a pretty elevated area in that same range on most of your measurements. Even the measurements of the mids that should fall like a rock below 200Hz pick back up in that range. It just looks worse on the X-Statik woofers as they only cover a narrow range and drop off on both sides.

So some of that might be your microphone calibration just showing more output in that area and making it look a little worse than it really is.

You could have some room gain right in that area too.

To get the power curve you're looking for might require an inline cap before your amp to roll off the lows a bit.

[bconline]

So some of that might be your microphone calibration just showing more output in that area and making it look a little worse than it really is.

That could well be the case. The weak link my setup is an old Mackie mixer which I use as a microphone preamp. I really don't know what it's response profile is. Because of this unknown, I am trying to focus on the relative response of various speakers measured with the same equipment and conditions.

Is a circuit diagram of the X-Statik crossover available? I could try experimenting with it to see what effects I can get.

[Danny Richie]

That could well be the case. The weak link my setup is an old Mackie mixer which I use as a microphone preamp. I really don't know what it's response profile is. Because of this unknown, I am trying to focus on the relative response of various speakers measured with the same equipment and conditions.

Is a circuit diagram of the X-Statik crossover available? I could try experimenting with it to see what effects I can get.

Shoot me an e-mail and I'll respond with a schematic and a couple of things that can be done to take it in the direction you are wanting to go with them.

[bconline]

Thanks Danny! I sent you a PM.

[bconline]

I got some advice from Danny and did a some research on the internet. This what I ended up doing:

I removed both capacitors from the woofer's low-pass circuit - making this part of the circuit into a first order crossover. That reduced the 80hz peak by 2 - 2.5 db and raised the output above 125hz. The overall repsonse was smoother. Here's the near field comparison:


However, the sound still a little boomy. And the measurement in the driveway showed the remaining peak.


My theory is that the reason for the peak in the first place is that are divers are out of spec. I used an online crossover simulator to simulate the schematic: http://www.micka.de/en/2weg_en.php.  It confirmed that Danny's crossover is an excellent design with very flat response.

So, to further reduce the peak, I tried adding a small amount of mass to the woofers. This lowers the resonant frequency. The tradeoff is lower sensitivity. I tried different masses added on the front.


After settling on 2 grams per woofer, I added that mass to the back, wrapping it around the voicecoil. Like this:


This had the expected effect. Here's nearfield woofer response (added mass is the green line):


And here's the resulting driveway response, measured at the tweeter level, 3 feet away:


And here's a comparison with another pair of speakes that I like the sound of:


And the sound??? Very very nice now! Not boomy any more. The bass is solid and transistions well to the MFW-15 subwoofer when really low bass is needed. Thank you very much to Danny and others who gave me advice!!

A few other comments:
(1) Becuase of the slightly lowered levels on the low end, the speakers are a little bright. The response above 2K is a little high. So, I may try padding the tweeter down a little bit.
(2) This tweeking is only for my particular pair of X-Statiks. I'm not implying that any other X-Statik need this kind of treatment.
(3) In the FR measurements, the dip at 200 hz is a floor bounce effect. I don't hear that dip in listening situations.

[poseidonsvoice]

bconline,

Again, great work.

It's important that you realized that the problem wasn't necessarily the room (indeed it's always a contributor, notice the SBIR issue you have between 100-300 Hz), but actually something that was intrinsic to the design of your particular pair of Xstatiks. This is important for those reading this thread (doing the right measurements ARE important). Not only did you hear the problem with your ears, you measured it. Not only did you measure it in room but you measured them outside to minimize the contribution of the room. And you made didactic comparisons to other loudspeakers that are monopoles (sealed, ported, passive radiator, TL, etc...), to verify that the problem again was intrinsic to your pair of Xstatiks. This probably happens more often than people are willing to admit. Others would have told you to let the speakers 'break in', etc...

Unfortunately you had to increase the moving mass of your woofers and lower the Fs, since they appear to be out of spec. The only way to conclusively prove that is to measure the T/S parameters of each woofer individually, and then compare to the original design. Honestly, I wouldn't bother at this point.

And yes, it appears that when these particular speakers left AV123 in Columbia, the final check wasn't done right. I am glad you are patient as well as diligent. You did somebody else's work! Kudos!

Best,
Anand.