SDX's?

Box is too small resulting in lots of power being used and thus increased power compression. That's just one possibility.

I believe one point made was its single 8" spider being somewhat of a limitation down low (when the excursion becomes higher).

In a sealed box, down low, the only thing that really matters for output (SPL) is excursion. In the smaller box (volume/driver=70L) it will require more power to get the cone to move the same distance as in the larger box (100L). The spider should be inflicting the same evils on both since the excursion is the same, but the power required to get that excursion is different. Thus my vote for power compression.

From my simulations I would say that the SDX could definately use more than the 100 litres... for one. I would do a 7 - 7.5 cu ft for each.

Bring the BL up and the optimal box shrinks...at the cost of $$$...for the more capable motor.

4.5cu ft is all these will get from me,per driver.

With a xmech of 40 mm, the SDX 15 should experience no problems.... and the spider certainly looks bigger than a standard 8".

interesting thing i learned about sdx while talking to the gentleman at CSS today. the recommended watts for the SDX is only 1000w. i know some people here run in 3000w and i do it 1600w myself. i called him about a rattle i was hearing from my sonosub and he said that i may be overpowering it and causing distortion. i took out the driver to confirm that it wasnt the case.

"With a xmech of 40 mm, the SDX 15 should experience no problems"

Just because it can move that far w/o damage doesn't mean it doesn't take a lot of power to do so.

The spider is 8", how you design the spider is very important too, Kevin has the Maelstrom spiders able to move 43mm one way on an 8" landing. Granted a single or pair of 10" spiders would result in lower distortion, but at far greater cost.

If there is a response difference it is T/S related. If you see differences as you scale it with power you can get into cms or BL related compression but it sounds to me like the differences you see are due to either your modeling, or a difference in the published T/S and the enclosure used.

I'm not defending the SDX either. It isn't even my driver, but if there is a difference between what you model, and the measured FR at low output levels then it's an issue with the T/S parameters, the build or the measurement.

The only exception to that is if you are measuring them at the limits of their output. Then you will see compression issues related to Cms, BL or power compression limits.

Thats what makes it so weird, it stops 6dB shy of its projected output down low, up high, it performs admirably. Why, when excursion is the limiting factor, does the driver fail in output so prematurely? Does this mean it isn't moving as much as it ought?

"Why, when excursion is the limiting factor, does the driver fail in output so prematurely? Does this mean it isn't moving as much as it ought?"

Yes, if it's producing less output, excursion is lower, possibly because of higher Re/lower BL from heat and nonlinear suspension (stiffness increases w/excursion),and port compression if it has one.

It was sealed.

The enclosures are too small IMO, and require too much power, which is likely why we're getting the compression that we're seeing. That's also one of the reasons why I like larger (.6 or lower Qtc) sealed enclosures. Less power required = less compression. (And I'm not saying you do, I am only speaking for myself and not disputing anything anyone else is saying)

The issue here though is that this is the only driver that we see this issue from, others have the same higher Q as the SDX, yet they don't fail like the SDX does.

TC2k vs SDX. The TC has a more heat friendly motor due to more copper in the VC, and what is it's former made of? What about other heat wicking measures?

What I'm saying is XBL is not a power friendly design due to less wire used in the VC. There's obviously more to it then that, but all things being equal, a larger VC with more metal (and more surface area to disipate heat) will handle more power with less compression...until saturation, but everything has to deal with that eventually.

So the SDX needs more box and less power. No biggie.

TC 2k former is aluminum, and is attatched to the aluminum cone with thermally conductive epoxy. The VC wire is aluminum, not copper. Regardless, there is more surface area on the TC driver to dissipate heat.

The cone does a surprising amount of thermal dissipation by itself.

That Al former is really benificial! The SDX has a fiberglass former, so that's not disipating much heat. The TC2k with a fiberglass former wouldn't conduct nearly as much heat to the cone.

Just thinking out loud, but wouldn't the power compression due to a heating VC be visible higher up in frequency too? Was this a sweep that started from low frequencies to high? If so, once it got hot it should show compression until the end of the sweep, with it maybe getting gradually better if it was able to start cooling at the higher frequencies.

I dunno, just seems to me that if it did ok until a low frequency that it may be more excursion related than heat?

Yeah if he went from low to high, there goes my power compression theory, but I thought he went from high to low and thus why I considered power compression.

Ilk uses a 30 second reverse sweep. That means from high to low. He chose the reverse sweep to save the hardest part of the sweep for the driver until the end, favoring the results above 20Hz.

The size of the box is, in this case, irrelevant. Loss of 1, maybe 2dB below 20Hz, no more, and should be relevant to a good model with slight variations due to stuffing and true input power vs rated input power.

Thermal compression at the end of a maximum power 30 second reverse sine wave sweep is the same for both drivers. The difference in the amount of windings in the VCs is moot when multi-KW power is applied for that long to both of those drivers. Since neither driver suffered a shorted coil, all that's left is the temperature, which is high enough in both cases.

The TC 2K (with a much better amplifier, at a better load) was bested by the Tumult across the board, except at 12.5Hz, where the 2K gave 4dB more, so I would tend to eliminate the differences in motor topology as to thermal capacity.

The CEA 2010 lists use 1/3 octave tone bursts with a cap tied to harmonic distortion, so that is where I think the difference between the drivers and the SDX (XBl^2) vs its model lie.

I don't think 12.5Hz and 16Hz are included in the CEA 2010 specs and I don't know what that may have to do with Ilkka's results. I also don't know if Ilk does any sort of extrapolation from his sine sweep max output test or he actually performs the CEA spec'd tone burst tests.

In any case, I believe that the SDX is listed on Ilk's CEA lists as having performed below the model predictions down low because of its excessive non-linear distortions vs the 2K, not its max output abilities. A model will not reflect non-linear distortions, but a THD+N limit to the results of a real world test, for example, will.

I'm not familiar with the specs of the SDXs cone, spider, surround or production offset tolerances , but I would suspect the answer lies there.

Bosso

I'd like for one of his fully working 2Ks (I forget entirely what was up with some of his 2Ks) with the new top assembly and an MK2 Tumult to be compared head to head, each with a CE4000. Won't happen though.

Hi Ricci,

You can't really compare the CEA-2010 max output figures with the WinISD "max output figures" because the CEA-2010 numbers are THD limited (please see this PDF for explanation on the CEA-2010 standard) and are using a burst signal instead of simple sine wave approximation of the WinISD. Also the WinISD approximation is only a function of the maximum power handling of the driver (Pe) or the maximum "linear" excursion (Xmax), whichever is reached first at each frequency. In other words it's EXTREMELY simplified approximation which basically doesn't always (ever) add up with the real word tests, even when not using any THD limiting or burst signals. Simulations are only simulations based on small signal T/S values which change a lot with larger input signals. Also basic simulations don't simulate the effects of any non-linearities (power compression etc.)

But when it comes to the SDX15, the CEA-2010 output down low is limited by the high 2nd order distortion component. This means that the compliance (Cms) non-linearity is quite high i.e. the suspension is causing a lot of distortion which limits the CEA-2010 numbers at the low end. The 3rd order harmonic is quite low which tells that the Bl related distortion is low, thanks to the XBL^2 motor.

Will,

The surround got detached from the alu cone. Bad glueing I suppose. I fixed it with some special Loctite glue and now it's better than new.

Tumult MK2 would be a nice driver to test but unfortunately very difficult to find.

Hey Ilkka!

Nice to see you around

While that does relieve me of my power compression woe's, it reminds me of the suspension uglies.