Peak limiting for driver protection

toastyghost wrote:   Despite what Elliot says, a well-qualified model is quite accurate for predicting a subwoofer's excursion and such, so long as the loudspeaker is driven in such a way that it is linear time-invariant. That generally means small to medium drive voltage signals, which avoid significant nonlinear behaviour. Most of these things can be calculated to get you started, but when it comes to measurements all of the functionality of Smaart (& software tools which are way more appropriate for loudspeaker design) can be had for free, using REW, VituixCAD, OpenSoundMeter, or ARTA.

Elliot Thompson wrote: toastyghost wrote:   Despite what Elliot says, a well-qualified model is quite accurate for predicting a subwoofer's excursion and such, so long as the loudspeaker is driven in such a way that it is linear time-invariant. That generally means small to medium drive voltage signals, which avoid significant nonlinear behaviour. Most of these things can be calculated to get you started, but when it comes to measurements all of the functionality of Smaart (& software tools which are way more appropriate for loudspeaker design) can be had for free, using REW, VituixCAD, OpenSoundMeter, or ARTA. <p ="western" style="margin-bottom: 0in">Which concludes you need to literally Measure. I used many of programs listed above and put some down throughout the years. Horn Response is a Simulator. It predicts things in the ideal situation. All simulators can only predict. They cannot guarantee the results are accurate. Measurement software can show different results among-st one another. However, the results are still a better representation of real world conditions than any simulator in a virtual world can offer. <p ="western" style="margin-bottom: 0in"> <p ="western" style="margin-bottom: 0in">There comes a time when sound professionals needs to step up to the next level. Searching for articles that may offer a keyword based on their dilemma, is not productive (nor needed) once you have the proper measurement tools at your disposal. <p ="western" style="margin-bottom: 0in"> <p ="western" style="margin-bottom: 0in">Best Regards, <p ="western" style="margin-bottom: 0in">

“Make a measurement” - of what, exactly? How? What stimulus? What distance? Where should the mic be placed? How should the room be excluded, or not? Your suggestion is vague and frankly, unhelpful.

Measurements are useless if the person making them doesn’t know how to do them properly, or understand what they are looking at, the limitations of the processing and electro-acoustic environment that was used to make the measurement, & what the various elements of the results are caused by.

Essentially, to make a measurement useful, you need to know what should appear on the screen before making the sweep. The computer is just answering a question, it's down to the operator to know how to ask the right question and assess whether the answer is correct.

This can be as simple as knowing in advance that your ported subwoofer should have a natural fourth-order roll off at the low corner, without any electrical filtering in place. If the plot produced from measurement doesn't show that, then you can try to figure out why: maybe room modes, reflections, driver unloaded, air leak, etc etc etc

The vast majority of useful measurements when designing a loudspeaker are also small-signal ones. That’s because you want the speaker to be linear time invariant for the maths to work, both in convolution of the measurement stimulus and to qualify it to electro-acoustic concepts. Funnily enough, most models also work on the same principles.

“No models are perfect, but some are useful.”

Measurements of non-linear behaviours are absolutely an advanced stage of making measurements, especially in real-world environments. I’d argue that many folks don’t even know how to do them in such a way that the data is even remotely representative of the device under test alone.

You know a really great way for people to bridge these gaps in knowledge and experience?

Using simulation tools.

Same for building what they sim - a huge problem for lots of people - and then measuring that, and qualifying it back to the simulation.

Any discrepancy then can be investigated, and used to refine either the model, the build, or the measurement process. Often, there need to be tweaks to all three, and several repeats before the person begins to really understand what's happening.

All of the above caveats about knowing what you're looking at still apply, every step of the way.

Each of the links I’ve posted are to in-depth articles from my own personal reference manager/research archived, written by hugely respected industry engineers who've worked for many major brands, on a literal professional sound training website...

One could argue the benefits in approach to helping people learn the what & the why for themselves by pointing them in the right direction and offering qualified, detailed advice with examples rather than just posting vague suggestions, but that probably wouldn't be very much in the spirit of the season

Edited by toastyghost - 28 December 2022 at 12:47am

toastyghost wrote:   “Make a measurement” - of what, exactly? How? What stimulus? What distance? Where should the mic be placed? How should the room be excluded, or not? Your suggestion is vague and frankly, unhelpful.

Edited by Elliot Thompson - 28 December 2022 at 10:15am

Elliot Thompson wrote: Augusts wrote: I have just one simple question - does the value that reaches Xmax would be considered Program power or AES? Your scenario was exceeding the Xmechanical limit. If you want to know how much watts you are encountering, during and prior to the Xmechanical limit, connect a Multi-meter on the output of the amplifier. Multiply the volts times volts and divide it by your impedance load. That will give you the peak wattage your speakers are receiving.  For Example: 40 x 40 / 4 = 400 watts You can then, look at the wattage figures of Program Power and/or AES to determine which wattage your figure falls under. Best Regards,

The question is not about where to limit. The question is how to do that correctly

Edited by Augusts - 28 December 2022 at 10:56am

I am telling you guys, the speaker hit the backplate because, the Vpeak was 141Volts for a 4ohm driver. that is 6000w. and from what I concluded that a single rms limiter does not limit the speakers excursion all together , but that there is a relation between the AES( that is used to calculate the RMS limiters treshold), +3db Program Power (which is the dynamic music material and not pure sine waves) and +3db for Peak power which is as I understand a very short peak that cannot do a mechanical damage to a subwoofer because it is very short.

Edited by Augusts - 28 December 2022 at 11:12am