Can sweeps damage tweeters?

[tony22]

I've heard from some folks that running sweeps, like the one in REW, can damage some tweeters - specifically the more fragile metal types like beryllium. Is there any way that the sweep in REW - even at its longest length - could do such a thing?

 

[John Mulcahy]

Sure. Tweeters are generally power dissipation limited, if you feed them signals at levels that are too high for too long they will suffer thermal damage. REW's sweeps have a pink spectrum, so the energy falls at 10 dB per decade as frequency increases. The time the sweep spends going from 10 kHz to 20 kHz is the same as the time it spends going from 10 Hz to 20 Hz, so the high frequency part of the sweep is traversed quickly, but if you use high levels and long sweeps you can cause damage. There is no need to use high levels for acoustic measurements.

 

[tony22]

I've never gone above the -12dB level shown on the Measure panel. Hmm, I might have it -9dB once. How high is too high? In the past I'd always use the longest sweep. Is -12dB at the longest sweep potentially dangerous?

 

[John Mulcahy]

What matters is the SPL that is produced and the length of the sweep. You can't tell the SPL from the sweep level setting as it depends on the interface and amplifier gains, the system volume setting and the speaker sensitivity. If you have calibrated the SPL reading or you are using a calibrated USB mic you can use Check levels to see the corresponding SPL, or look at your graphs. 75 dB SPL is sufficient for acoustic measurements and is very unlikely to damage any speaker, even those with limited power handling. The longer the sweep, the more time it spends in the frequency range of the speaker and the more power the tweeter has to dissipate. I have made thousands of measurements with 256k sweeps at around 75 dB SPL and never seen any damage at all to any speaker I have measured.

 

[tony22]

If you have calibrated the SPL reading or you are using a calibrated USB mic you can use Check levels to see the corresponding SPL, or look at your graphs. 75 dB SPL is sufficient for acoustic measurements and is very unlikely to damage any speaker, even those with limited power handling.

Yep, I do. And I never go above 75dB SPL. What if I was using the longest sweep length available?

 

[John Mulcahy]

A 256k sweep at 48 kHz sample rate lasts about 5.5 seconds. A full range sweep spans about 10 octaves and the tweeter would typically be active during the last 3 octaves, so for a little less than 2 seconds. It would take a very loud signal to overheat a tweeter in less than 2 seconds. A 4M sweep lasts 16 times as long, so the tweeter would be active for about 30 seconds. A loud signal could overheat a tweeter in that time. REW's longer sweeps are provided to facilitate high signal-to-noise ratio measurements of electronic equipment, particularly at very high sample rates. I don't know why anyone would use them for speaker measurements.

 

[tony22]

Is 4M new? I thought 1M was the longest sample rate; that's what I'd been using.

Here's what's got me scratching my head. After reading about this I decided to see if my tweeters could reproduce out to 20KHz. My speakers are well designed, with a FR spec'd out to -2dB at 20KHz. I set up my UMIK-1 with the capsule inches in front of the tweeter, facing it. I used the RTA tool but wasn't sure if I should run Pink Noise to do this test. Is this a legitimate test? Or, maybe I should ask, how would I do a test like this? The tweeters on both speakers sound fine, with no distortion or obvious problems, but I'd like to figure out how to test them just for my own sanity.

 

[John Mulcahy]

Just use a sweep. The mic should be very close, a fraction of an inch. The sweep level will need to be much lower than for a listening position measurement due to the mic's proximity. Mic calibration may be inaccurate at the highest frequencies.

 

[tony22]

If I use a calibrated SPL meter and run it right up to the midrange, say, just to get an easier way to read SPLs, what should it measure right up at the speaker to do this?

 

[High Concept]

Is there some theory floating around that the measurements must be made at high SPLs? That a sweep at 75 dBA SPL is “better” (more accurate, yields more data, etc.) than one at 70, or even 65? Run six sweeps, 5 dBA apart, 50 to 75 dBA at the mic position, and plot them on the same page. The curves are identical in the frequency axis and proportional in the amplitude axis. This won’t change until the background noise in the room becomes a factor (low S/N ratio). Vis-a-vis tuning or evaluating your room, all the response curves are equally valid.

Assuming a satisfactory S/N ratio, the only need I can see for a particular sweep SPL is to answer the question, “how flat (in dB) is my room?” – which has no answer – to “how flat is my room in the presence of a stimulus at X dBA?” Or, of course, to be consistent over time.

 

[tony22]

I think I understand your rhetoricalness. So maybe the question is - how much over the ambient SPL is enough for the sweep SPL to be effective?

 

[High Concept]

I think I understand your rhetoricalness. So maybe the question is - how much over the ambient SPL is enough for the sweep SPL to be effective?

Actually, my rhetoricalness was aimed at the world in general, not directly at you! But it's an interesting, if subjective, question.

What I do know is that my afternoon is blown. After lunch, I am going to set this up in my recording studio (noise floor: 27 dBA) and repeat this from 80 dbA down to mud. My tweeters are beryllium (Focal Trio6 be) but I have factory spares in a drawer. Ah, the life of a geek!

Steve Johnson
Redwood City, CA

 

[tony22]

My tweeters are beryllium (Focal Trio6 be) but I have factory spares in a drawer. Ah, the life of a geek!

Good luck!

 

[John Mulcahy]

If I use a calibrated SPL meter and run it right up to the midrange, say, just to get an easier way to read SPLs, what should it measure right up at the speaker to do this?

Level shouldn't be a concern, you'll be measuring right at the tweeter. Drop the sweep level 20 dB from where you normally set it and see what you get.

 

[tony22]

Level shouldn't be a concern, you'll be measuring right at the tweeter. Drop the sweep level 20 dB from where you normally set it and see what you get.

I set up a short sweep and positioned the capsule of the mic right in front of the tweeter (trickier than I expected!). Both tweets measure out essentially flat to 20KHz. All good. Thanks to all who helped.