Xmax Investigation

[Contrasseur]

Prompted by all the Klippel arguments around these forums, I used the real world CEA-2010 max burst results to calculate piston excursion for different enclosures. Between all of us here, we've modeled thousands of theoretical enclosures based on manufacturer's Xmax, mathematical Xmax, Klippel Xmax, etc. It is certainly relevant to our interests if we can figure out how to model the large-scale performance of various systems.

 

I used the piston excursion formula presented here and solved for Xmax.

http://www.baudline.com/erik/bass/xmaxer.html

 

I used Ricci's Sd measurements, c=343m/s, and air density =1.275kg/m^3. I added 6db to each SPL measurement to get the 1m SPL.

 

My results are strange and I need some of your thoughts interpreting these data.

 

I calculated based on full-space radiation. Half-space radiation would result in excursion values half as high as calculated. It seems that as a general trend, frequencies 20Hz and below radiated into the ground, behaving as full-space radiation. Calculated excursions are unrealistically high as frequency approached 63Hz. The sound seems to have reflected more and more off the grass, behaving as half-space radiation. All results 80Hz and up were amp-limited, but track closer to the half-space modeled behavior of these systems.

 

***EDIT*** THESE NUMBERS ARE FLAWED. SEE POST #10 FOR CORRECT DATA POINTS!!

 

 

The LMS Ultra 5400 and the XXX 18D2 Split Coil make for an interesting comparison. XXX is clearly the higher excursion driver at nearly all frequencies, but usable excursion steadily declines the further it's used below the impedance peak. I originally thought that intermodulation distortion from the great current passing through the driver is causing enough distortion to push it over the threshold at lower volumes than the LMS Ultra. The LMS Ultra has much better managed inductance, which might explain why the XXX is only getting full excursions where the impedance peak restricts the current. To support this, the even higher inductance XXX 18D2 Overhung is capable of greater excursions at 25Hz and 31Hz because it's higher impedance peak reduces current and intermodulation distortion more than enough to offset it. For further support, it's performance below 20Hz falls apart. Some of this could be due to greater BL variation, but down to 10Hz this "54mm Xmax" driver is only moving 20mm one way before it hits the distortion thresholds. The lowly Ultimax matches it's performance at 10 and 12.5Hz.

 

Then I looked at the Rockford Fosgate T3S2-19 and the Dayton Ultimax, both of which have some of the lowest Le/Res of high-excursion drivers tested. Both of these had the same excursion tracking as the exceptionally high inductance XXX drivers. When adjusting for the hypothesized full-space to half-space radiation transition from 20Hz-40Hz, these two drivers are distorting at nearly the same excursions at 12.5Hz as they are at 31.5Hz! This follows with just what we would expect from such low-distortion, linear drivers. Plus the Ultimax nearly ties the performance of the XXX 18D2 Split Coil at 31.5 Hz, right at the XXX's impedance peak where current is the lowest, so it's hard to support the intermodulation claim.

 

Really the LMS Ultra 5400 is more of an outlier in that calculated excursion (after adjusting for full-space to half-space transition), actually INCREASES lower frequency! The LMS Ultra 5400 has one of the most linear BL curves from it's advanced motor, and what I (like many others) would assume to have closer to ideal pistonic motion. Combined with well-managed inductance, one would expect usable excursion to be the same all the way down to very low frequencies. The calculated results are in fact, consistent with what one might expect if this driver was only radiating into full-space through the whole bandwith of testing.

 

I'm sure there HAS to be a full-space to half-space transition, because I can't believe that the Ultimax can do 48mm one way at 31Hz, or the 21SW152-4 can do 37mm at 50Hz, or the BMS 18N862 can swing 42.4mm at 50Hz. The Aurasound NS18-994-A would be well beyond the gap (or more likely destroyed) if it were producing excursions of 47mm at 31.5Hz!

 

In regards to Klippel, the HS24-D2 was (I thought I read somewhere) 25mm. However, you can see that an ideal piston of 2130cm^2 could only generate these SPLs by moving some 50mm one way!

 

One thing we know for sure is that Xmax is a moving target. The same driver can cross the distortion threshold at very different excursions depending on frequency. One driver can lead in the infrabass and midbass, but lose handily to the same comparison driver somewhere in between. Other than that, I'm a bit at a loss. Thoughts?

[Funk Audio]

At and near resonance some very interesting and complicated things happen in the prevalent moving coil, fixed magnet motor design(potential here for some DEEP discussions). And these affects often allow the driver to produce higher than the theoretical excursions at distortion limits, that is why a good safety margin past Xmax, to Xmech is very important. As in the case of the LMS5400 the motor is strong and linear to over 30mm but it is mechanically limited at ~39mm so even though the motor could likely go past that at some frequencies, based on the makeup of the distortion at those frequencies, the mechanical parts will not allow it. In the case of getting lower than theoretical excursion at passing distortion limits this is usually a case of the driver not having enough motor force to over come the air spring of the box, although there are many factors in this regard as well. With most drivers the trend is as you have found it, around resonance many drivers will be able to get to or near mechanical limits with passing distortion, sometimes the motor could go further as is the case when you see "Mechanical noise" or very high order distortion as the limiting factor this usually means mechanic limit has been hit.

[Electrodynamic]

In regards to Klippel, the HS24-D2 was (I thought I read somewhere) 25mm. However, you can see that an ideal piston of 2130cm^2 could only generate these SPLs by moving some 50mm one way!

 

A member of a different forum put it best: "magic-sauce". I wish I knew what "magic-sauce" was and how to use it but I apparently use it on my 24" woofers, haha. 

[Contrasseur]

That makes sense and all, but that's not what we're seeing. The data doesn't show other drivers OVERPERFORMING at fb, it shows the LMS Ultra 5400 UNDERPERFORMING at 25-40Hz.

 

You have to subtract 6db (divide excursion by 2) from the 40Hz measurements because the drivers are radiating in half-space at that point. Look at the RF T3S2-19 especially. It's putting out as much bass as 71mm of excursion in full-space. We know it can't because once it hits ~45mm it hit's the other gap and starts braking HARD. Even if it could go further, the mechanical limit at the spider landing is ~54mm. However, Josh Ricci's measurements show a clean (still below CEA-2010 thresholds) output of 125.4db. If we assume that it's radiating in half-space, then it's only moving 35.5mm one-way, consistent with both the physical construction of the driver and Ricci's clean measurement.

 

Given that the LMS Ultra 5400 was tested on the same grass plot as the rest of these systems, it must also be radiating in half-space at 40Hz. The SPL generated means the cone was moving at 18mm one-way. This is quite poor for a a driver with a manufacturer's Xmax listed at 34mm. The Aurasound NS18-944-A is a much lower excursion driver with a geometric Xmax of a mere 12.7 and manufacturer's Xmax of 19mm. Yet at 31.5 and 40Hz it's capable of MORE clean output than the LMS Ultra 5400.

 

I'm noticing that while the LMS Ultra is underperforming at 25-40Hz, the trend among most of the other drivers is to be significantly underperforming at 10-12.5Hz. The XXX Split coil, for all it's 54mm of claimed Xmax, breaks the CEA-2010 thresholds at a mere 27.6mm at 10Hz in full-space radiation.

 

You do bring up a good point about mechanical limits though. Looking through Ricci's Max SPL- No THD Limit measurements will show us how much further the cone has left to travel. I will have to add these numbers to the data set.

[Contrasseur]

A member of a different forum put it best: "magic-sauce". I wish I knew what "magic-sauce" was and how to use it but I apparently use it on my 24" woofers, haha. 

haha,

but seriously!!!

 

We're seeing empirically demonstrated magic sauce as measured in real life! Ricci has gone through great lengths with very high-end test equipment, but we're still seeing these unbelievable results! Seriously, as in, I can't believe this is possible. Ricci's published all his gear and techniques and I'm sure his data must be accurate. It's just hard to accept because it defies common sense!

 

If we can somehow figure out what magic sauce you dropped in the 24" cauldron, then this planet is about to get a whole lot bassier!

[dgage]

If we can somehow figure out what magic sauce you dropped in the 24" cauldron, then this planet is about to get a whole lot bassier!

 

Please don't.  I'd prefer if people don't understand the magic sauce that is the BHS/HS-24, it is much easier for me to sell a special subwoofer in a magic box.   Actually, it must not be the sub, it is the magical enclosure I built.  Yep, that's it.   Of course, let's not talk about how my magical enclosures aren't quite as nice as the Funk Audio cabinets, but they don't cost as much either.  Actually, I've done a few custom enclosures with solid wood corners and I understand why Funk Audio cabinets cost as much as they do, definitely not easy.

[Contrasseur]

At and near resonance some very interesting and complicated things happen in the prevalent moving coil, fixed magnet motor design(potential here for some DEEP discussions). And these affects often allow the driver to produce higher than the theoretical excursions at distortion limits, that is why a good safety margin past Xmax, to Xmech is very important. As in the case of the LMS5400 the motor is strong and linear to over 30mm but it is mechanically limited at ~39mm so even though the motor could likely go past that at some frequencies, based on the makeup of the distortion at those frequencies, the mechanical parts will not allow it. In the case of getting lower than theoretical excursion at passing distortion limits this is usually a case of the driver not having enough motor force to over come the air spring of the box, although there are many factors in this regard as well. With most drivers the trend is as you have found it, around resonance many drivers will be able to get to or near mechanical limits with passing distortion, sometimes the motor could go further as is the case when you see "Mechanical noise" or very high order distortion as the limiting factor this usually means mechanic limit has been hit.

NEW DATA POINTS!!

 

Josh Ricci's No THD Limit measurements provide the perfect reference! He seems to be pushing them to consistent excursions at the mechanical limits

 

Funk Audio right on about the LMS Ultra being linear right up to excursion maximum. Also right on about the interesting and complicated things at resonance. LMAO that "DEEP discussion" pun just hit me! But seriously, I would like some more discussion on it! This behavior is most interesting and complicated indeed. Near resonance, many of these systems are demonstrating linearity almost all the way to the mechanical limits. We're still seeing significant dropoff at infrabass for many of these systems.

[Contrasseur]

GOD DAMMIT

 

 

My spreadsheet is flawed. All of these data points are for the SD of the RE Audio XXX 18D2 Split Coil. I locked that parameter so autocomplete wouldn't change it, but I never changed it when copying the formula to the other drivers. All the calculated excursions for 19", 21", and 24" drivers are wrong. 

 

Electrodynamic this means you're not God unfortunately. But still pretty awesome. Cthulu maybe?

[Funk Audio]

To he honest I did not closely examine your chart. I was just going off your comments and I have made up my own charts for this exact thing in the past so I was just going off the info I am familiar with. looking at your chart now there are some things that don't look quite right there.

[Contrasseur]

Here's the new data with correct Sd values. 

 

 

It makes a lot more sense, but there's still some very odd results. Regardless of my calculations, we're seeing the same behavior in Ricci's measurements. Certain drivers can move much more air than others, but only at certain frequency ranges.

[dgage]

Looking at your updated values, the ones that I have experience with make sense to me for the LMS-5400 and HS-24.  I found the LMS-5400 to be a very linear driver across the entire range.  My experience with the LMS-5400 was in the room, no ground plane testing, but with multiple LMS-5400s around the room, I didn't need to DSP much in the sub region. 

 

For the HS-24, your comment about following the inductance curve as well as output rolling off below the inductance curve matches my experience with it in a few different size enclosures.  Of course, my Mariana 24 enclosure exacerbates the inductance hump since it is a little too small but the BHS-24 does really well.  I need to play with the BHS-24 subwoofer more but I really, really like the sound and performance of it.

[Ricci]

I have scratched my head over some of this for years. Wondered how long it would take for someone else to ask. There are some outliers that seem to result in impossible amounts of displacement and I'm not sure why. Even if we throw in perhaps a 1dB error in calibration or atmospheric conditions it's still can't account for all of it. Not sure I agree with a fullspace to halfspace transition though.

 

The results at the impedance maximum are nearly always the highest. Efficiency is highest and current through the coil is lowest. Distortion is low so the driver can be pushed nearly to mechanical limits before distorting badly. Down near 10Hz current is very high and the signal is longer duration as well. Current induced distortion in the motor is a factor. Drivers with low motor force and thus lower efficiency do worse away from the system fB as do those with high amounts of inductance variation. Drivers like the XXX distort more while those like the RF and IPAL do well. Of course those are also drivers with a very low system qtc related to the extreme motor force.

 

The LMS is limited by mechanical excursion rather than THD at virtually all bandwidths. I may have prematurely stopped it at a few bands but past experience taught me that you stop when it starts making non harmonic noises. Otherwise you end up with a giant paperweight.

 

Really high THD can account for some extra SPL with less excursion than calculated for a pure sine tone as well but I don't think that can account for all of the discrepancy. Perhaps the SD values are somehow under represented too but that is a fairly concrete and easy to measure parameter.

[Ricci]

Another thing that I've noted in some drivers behavior is what appears to be mechanical distress due to acceleration forces. These would occur higher up in frequency in the 40-63Hz bandwidth. If you calculate out the acceleration for some of the drivers in that range it gets very high. I've heard some make bad mechanical noises at a point where excursion should be nowhere near the limits. My guess would be that the coil is rocking or the cone or other suspension pieces may be flexing or sucking in due to the acceleration forces or the vacuum created in the cabinet. One driver blew a dust cap about 5ft off of the driver.

[Funk Audio]

I have scratched my head over some of this for years. Wondered how long it would take for someone else to ask. There are some outliers that seem to result in impossible amounts of displacement and I'm not sure why. Even if we throw in perhaps a 1dB error in calibration or atmospheric conditions it's still can't account for all of it. Not sure I agree with a fullspace to halfspace transition though.

 

The results at the impedance maximum are nearly always the highest. Efficiency is highest and current through the coil is lowest. Distortion is low so the driver can be pushed nearly to mechanical limits before distorting badly. Down near 10Hz current is very high and the signal is longer duration as well. Current induced distortion in the motor is a factor. Drivers with low motor force and thus lower efficiency do worse away from the system fB as do those with high amounts of inductance variation. Drivers like the XXX distort more while those like the RF and IPAL do well. Of course those are also drivers with a very low system qtc related to the extreme motor force.

 

The LMS is limited by mechanical excursion rather than THD at virtually all bandwidths. I may have prematurely stopped it at a few bands but past experience taught me that you stop when it starts making non harmonic noises. Otherwise you end up with a giant paperweight.

 

Really high THD can account for some extra SPL with less excursion than calculated for a pure sine tone as well but I don't think that can account for all of the discrepancy. Perhaps the SD values are somehow under represented too but that is a fairly concrete and easy to measure parameter.

 

The distortion components not only add to the measured total SPL of the fundamental +all distortion orders, as is commonly known, but what also needs to be considered is when you take a waveform and add the multiple distortion components the actual peak amplitude of the wave is now decreased, even moderate amounts of distortion combined with the fundamental will reduce the actual peak travel required to make a given total SPL, this combined with the additive output of the distortion components will make up the differences, I have calculated this affect out for some data points on a couple drivers and what I have seen is the corrected waveform correlates to the measured output and indicates actual excursion at or very near mechanical limits for the frequencies and outputs that seem to defy the measured mechanical driver limits, this method could be used to calculate actual real cone travel for all tested units at all tested frequencies, but it is a significant amount of work,

 

Note; although I have not simulated enough points too see it, I do believe in some cases this affect could be reversed in some cases depending on the total spectrum of distortion and the actual cone travel could be higher than the measured SPL would indicate.

[Contrasseur]

I have scratched my head over some of this for years. Wondered how long it would take for someone else to ask. There are some outliers that seem to result in impossible amounts of displacement and I'm not sure why. Even if we throw in perhaps a 1dB error in calibration or atmospheric conditions it's still can't account for all of it. Not sure I agree with a fullspace to halfspace transition though.

Do you suspect these are radiating in half-space and significantly underperform in the infrabass? That means the XXX 18D2 Overhung Coil broke the CEA 2010 threshold at only 10mm at 10Hz.

Or do you think these are radiating in whole-space and significantly overperforming in the midbass? This means that the Aurasound NS18-994-A was producing a clean signal at 51mm one-way at 31.5Hz.

 

What did the cone movement look like when you were running the tests? Could you approximate the observed excursions?

 

The LMS Ultra 5400 consistently distorted at the same calculated excursion, which is consistent with yours and Funk Audio's observations that it's a very linear motor followed by a very hard limit. This means that the drivers radiate in full space and the LMS Ultra has 38mm excursion before that hard limit.

[Contrasseur]

The distortion components not only add to the measured total SPL of the fundamental +all distortion orders, as is commonly known, but what also needs to be considered is when you take a waveform and add the multiple distortion components the actual peak amplitude of the wave is now decreased, even moderate amounts of distortion combined with the fundamental will reduce the actual peak travel required to make a given total SPL, this combined with the additive output of the distortion components will make up the differences, I have calculated this affect out for some data points on a couple drivers and what I have seen is the corrected waveform correlates to the measured output and indicates actual excursion at or very near mechanical limits for the frequencies and outputs that seem to defy the measured mechanical driver limits, this method could be used to calculate actual real cone travel for all tested units at all tested frequencies, but it is a significant amount of work,

 

Note; although I have not simulated enough points too see it, I do believe in some cases this affect could be reversed in some cases depending on the total spectrum of distortion and the actual cone travel could be higher than the measured SPL would indicate.

This could make some of the difference. Depending on the phase of the harmonics, the actual amplitude peak will be either higher or lower. I hadn't thought of this and I'm interested to see what difference it would make.

 

These are CEA 2010 passing numbers, so let's assume a worst case scenario where it simultaneously breaks all distortion thresholds. I arbitrarily choose an 18" driver hitting 110dB at 20Hz for this scenario. The CEA 2010 testing scheme band-passes the result, centered on the fundamental, and then calculates the amplitude from the amplitude of only the fundamental.

Therefore if it says we broke all thresholds at 110dB we have the following excursions.

Fundamental is 110dB at 20Hz - 27.0mm

Second harmonic is 100db at 40Hz - 2.1mm

Third harmonic is 95dB at 60Hz - .53mm

Fourth harmonic is 90dB at 80Hz - .17mm

Fifth harmonic is 90dB at 100Hz - .11mm

Sixth harmonic is 80dB at 120Hz - .02mm

Seventh harmonic is 80dB at 140Hz - .017mm

Eighth harmonic is 80dB at 160Hz - .013mm

Ninth harmonic is  70dB at 180Hz - .003mm

Higher harmonics will produce negligible excursion at -40db.

 

Assuming all these amplitudes have identical phase, the amplitude peak is 2.963mm. If they're all in phase with the fundamental, then our actual excursion is 11% higher than what is calculated from the fundamental. If they're all out of phase with the fundamental the peak amplitude isn't reduced the full 2.963mm because those higher harmonics will rebound before the end of the fundamental's half-cycle. I'm not sure how to calculate how much exactly.

 

We can say that due to distortions, the calculated Xmax values could be off from the actual excursion by up to 10%. However, these Xmax values are valid for simulations. Sims only model the fundamental, and these calculated Xmax values represent only the portion of the excursion generated by the fundamental.

[Funk Audio]

This could make some of the difference. Depending on the phase of the harmonics, the actual amplitude peak will be either higher or lower. I hadn't thought of this and I'm interested to see what difference it would make.

 

These are CEA 2010 passing numbers, so let's assume a worst case scenario where it simultaneously breaks all distortion thresholds. I arbitrarily choose an 18" driver hitting 110dB at 20Hz for this scenario. The CEA 2010 testing scheme band-passes the result, centered on the fundamental, and then calculates the amplitude from the amplitude of only the fundamental.

Therefore if it says we broke all thresholds at 110dB we have the following excursions.

Fundamental is 110dB at 20Hz - 27.0mm

Second harmonic is 100db at 40Hz - 2.1mm

Third harmonic is 95dB at 60Hz - .53mm

Fourth harmonic is 90dB at 80Hz - .17mm

Fifth harmonic is 90dB at 100Hz - .11mm

Sixth harmonic is 80dB at 120Hz - .02mm

Seventh harmonic is 80dB at 140Hz - .017mm

Eighth harmonic is 80dB at 160Hz - .013mm

Ninth harmonic is  70dB at 180Hz - .003mm

Higher harmonics will produce negligible excursion at -40db.

 

Assuming all these amplitudes have identical phase, the amplitude peak is 2.963mm. If they're all in phase with the fundamental, then our actual excursion is 11% higher than what is calculated from the fundamental. If they're all out of phase with the fundamental the peak amplitude isn't reduced the full 2.963mm because those higher harmonics will rebound before the end of the fundamental's half-cycle. I'm not sure how to calculate how much exactly.

 

We can say that due to distortions, the calculated Xmax values could be off from the actual excursion by up to 10%. However, these Xmax values are valid for simulations. Sims only model the fundamental, and these calculated Xmax values represent only the portion of the excursion generated by the fundamental.

It's much more complicated than that, not just additive/subtractive. If I simulate the waveform as you describe hitting all CEA2010 limits at the same time the peak amplitude is actually about a db higher than the 20hz fundamental would be alone, and shifted in time, but by leaving everything else and actually dropping the second order distortion an extra 10db the peak of the actual waveform is reduced about 1db. Remember we are looking at it a little backwards, the distortion is not whats reducing the peak travel but rather the peak travel being limited by the driver limits is causing the harmonics to align into the peak reducing combination, and if it happens in just the right combination you can get more measurable output at the fundamental(consider area under the curve is what produces the total air pressure we see/measure) with lower peak cone travel, before the distortion components hit their respective limits.

[Funk Audio]

Here I have shown a clean 10hz wave on the left, that same wave with distortions added as per the RFT19 at 10hz CEA2010 passing output on the right, as you can see the peak of the wave is over a db below what it should be, although the fundamental is still "there" and will read SPL as per the pure wave for 10hz content. When I measure the area under the curve for both the right side is actually about 7% higher, this correlates to the fundamentals output +harmonics, total combined sound pressure being higher, while each frequencies output can be measured independently the combined waveform indicates the actual tracking of the cone position over time required to produce all these frequencies at the same time, at the measured amplitudes.
 

[dgage]

It's much more complicated than that, not just additive/subtractive. If I simulate the waveform as you describe hitting all CEA2010 limits at the same time the peak amplitude is actually about a db higher than the 20hz fundamental would be alone, and shifted in time, but by leaving everything else and actually dropping the second order distortion an extra 10db the peak of the actual waveform is reduced about 1db. Remember we are looking at it a little backwards, the distortion is not whats reducing the peak travel but rather the peak travel being limited by the driver limits is causing the harmonics to align into the peak reducing combination, and if it happens in just the right combination you can get more measurable output at the fundamental(consider area under the curve is what produces the total air pressure we see/measure) with lower peak cone travel, before the distortion components hit their respective limits.

Nice! I was thinking some of this but couldn't put it together as cohesively and succinctly as you did.

[Funk Audio]

@Ricci maybe you can explain the XXX overhung 10hz and 12.5hz signal spectrum charts, as the burst chart shows a passing output but the spectrum shows all the harmonics way over the tresholds? If I simulate the waveform for the spectrum as shown in those charts I get exactly 6db higher peak on the wave, indicating the cone was actually traveling twice as far as the 10hz fundamental output would indicate, to produce the spectrum as shown.

[MemX]

Another thread that makes my head hurt

 

 

Keep up the good work, chaps

[Contrasseur]

Here I have shown a clean 10hz wave on the left, that same wave with distortions added as per the RFT19 at 10hz CEA2010 passing output on the right, as you can see the peak of the wave is over a db below what it should be, although the fundamental is still "there" and will read SPL as per the pure wave for 10hz content. When I measure the area under the curve for both the right side is actually about 7% higher, this correlates to the fundamentals output +harmonics, total combined sound pressure being higher, while each frequencies output can be measured independently the combined waveform indicates the actual tracking of the cone position over time required to produce all these frequencies at the same time, at the measured amplitudes.

 

Right, that's why the 11% calculated are the absolute maximum variation. It is certainly less in real life. 

 

In a realistic excursion limiting scenario, the sine wave basically becomes more like a square wave. In a square wave, all harmonics have the same phase, and each subsequent harmonic has a valley where the composite waveform had a peak. Adding each harmonic therefore lowers the amplitude peak (other than Gibbs phenomenon). To turn it completely into a square wave, the fundamental will have an amplitude 4/pi times as much as the resultant square wave. So to put numbers on this, the actual excursion can be as little as 78.5% of the fundamental alone. This is for a perfect square wave, which would be well over the distortion limits.

 

In order to take a sine wave and make it as square as possible without breaking the CEA 2010 thresholds, we're left with just the odd harmonics calculated before, and assume they're all in phase at zero. In our 18" 110dB 20Hz scenario, the fundamental's amplitude peak (90 degrees phase) is 27mm- .53mm + .11mm - .017mm + .003mm =26.566mm, or only 1.6% lower than CEA 2010 would calculate the fundamental to be.

 

In reality we can add even harmonics as well in order to reduce either the positive or negative amplitude peaks, but it will increase excursion in the other direction. This can give us a little more output when a driver is limited on the inward stroke but not on the outward stroke.

 

I agree that this effect is happening and I agree that it's more complicated than my novice back-of-the-envelope calculations. What I'm saying though is that even in the theoretical worst-case scenarios of my back-of-the-envelope calculations, the numbers are too small to account for the measured SPLs! Ricci suggests that due to weather/calibration issues the numbers could be 1dB off, which is only 12%. You suggested the distortions could alter the actual cone movement, which only accounts for a maximum of 11% (although certainly lower in the actual results). I'm suggesting a half-space to full-space transition which can double or halve the observed numbers. This is of great enough effect to rein the numbers back into the realm of reasonability. Ricci doesn't think so, for good reason. How can this be the case when some systems don't exhibit that effect when tested on the EXACT SAME LOCATION??

 

So we can explain our numbers a little better now, but they're still unreasonable. We're getting closer but still haven't fully explained the behavior.

[dgage]

Contrasseur - was this a question you were just thinking of or are you thinking this could be a possible verification of Xmax numbers? Or is there something else you're thinking. In pondering this thread, the thought of why and where else you might take this came to mind.

 

EDIT - Nevermind, your post #22 made it clear what you're trying to understand. Nice!

[Contrasseur]

Here I have shown a clean 10hz wave on the left, that same wave with distortions added as per the RFT19 at 10hz CEA2010 passing output on the right, as you can see the peak of the wave is over a db below what it should be, although the fundamental is still "there" and will read SPL as per the pure wave for 10hz content. When I measure the area under the curve for both the right side is actually about 7% higher, this correlates to the fundamentals output +harmonics, total combined sound pressure being higher, while each frequencies output can be measured independently the combined waveform indicates the actual tracking of the cone position over time required to produce all these frequencies at the same time, at the measured amplitudes.

 

How did you calculate these? I must have done it differently because even in my worst case scenario I could only get 11% variation at a single point, which is only 1dB.

[Contrasseur]

Contrasseur - was this a question you were just thinking of or are you thinking this could be a possible verification of Xmax numbers? Or is there something else you're thinking. In pondering this thread, the thought of why and where else you might take this came to mind.

 

EDIT - Nevermind, your post #22 made it clear what you're trying to understand. Nice!

I've been lurking this website for a long time because it has the most in-depth, accurate performance data. I've learned a lot from DIYaudio, BFM, AVS and others, but once you've read all the good threads the rest is just chaff of noobs parroting the one or two heuristics they've read somewhere. This site offers REAL RESULTS and thousands of data points that I'm constantly seeing new patterns in and learning more about what I'm doing. There's just this HUGE GLARING HOLE in my understanding that can't account for these variations!! So I'm having a whole lot of fun and learning a ton from crunching numbers in this thread and it's got me excited all over again!

 

Meanwhile I've got the girlfriend groaning "Come to beeeeedddd..." the past few days.

 

Thank God that woman think's I'm cute lol