ATC/Volt midrange drivers and 118 dB SPL, continuous?

[audiothings]

I'm looking at building a 3 way system similar to the ATC SCM200ASL Pro or the Quested Q212D.

Both these systems are capable of playing at 118 dB, continuously.

The former uses the 's' version the ATC SCM75-150. The non 's' version has a sensitivity of 94 dB (1W/1m) with a power handling of 75 watts leading to a maximum continuous output of 112.7 dB. In ATC's professional monitors, the midrange units are crossed over at 380 Hz and 3.5 KHz.

The Quested appears to be using some version of the Volt VM752. This driver has the same 94 dB sensitivity, and a power handling capability of 100 watts, leading to a maximum continuous output of 113.9 dB. In Quested's professional monitors, the midrange units are crossed over at 450 Hz and 4500 Hz with LR24 filters.

Will I be able to generate the 118 dB continuous I'm going for, with either of these two drivers, under real world circumstances, with LR24 filters on both ends? Or is the ability to generate such levels limited to the custom versions of the drivers?

 

[Tytte71]

You must define continuous sound pressure level. I reckon you speak of peak sound pressure level when playing music. I can't see any reason why you should use 118dB steady state (sinusoidal) signal.

The power sensitivity of the standard ATC SM75-150 is ~91dB/W. The S version is ~94dB/W. 500W is needed to drive the standard version at 118dB peak and 225W for the S-version. IOW a lot higher than the maximum input power recommended by ATC

The problem gets worse when you look at the voice coil excursion with the standard driver crossed 450Hz/LR24. It takes ~240W to reach xmax at 450Hz (accounted for the -6dB at crossover frequency). You get ~112dB/1m SPL at xmax. Beyond xmax the driver rapidly becomes severely nonlinear (height of gap is 5mm and voice coil length is 3,5mm).

http://test.atcloudspeakers.co.uk/wp-content/uploads/2014/01/ATC-Driver-Specs.pdf
http://studio-hifi.com/images/ATC75-150S_JeffBagby.pdf

You need another driver than the standard version to reach 118dB SPL.

 

[xrk971]

You need a midrange cone in a FLH (tractrix or constant directivity horn). Tractrix gets 105dB to 110dB at 1w and 1m.

http://wp.volvotreter.de/projects/pics_plans/traxtrix-midbass-fronthorns/

 

[audiothings]

You must define continuous sound pressure level. I reckon you speak of peak sound pressure level when playing music. I can't see any reason why you should use 118dB steady state (sinusoidal) signal.

Genelec gives the clearest representation I can find right now, of what I am trying to achieve.

For their 1238A monitor, their specs are 'Maximum RMS acoustic output > 116 dB @ 1 m' and 'Maximum peak output per pair with music material > 124 dB.

I have heard the 1238 and I have heard speakers that deliver more SPL, and in this case, I'm going for the latter.

I am (perhaps falsely) under the impression that the RMS tests are performed with pink noise.

 

[Charles Darwin]

I'm not sure that the quoted max continuous SPL for the whole speaker system actually implies that the mid can do the same on its own and over its range.

Chances are the test uses pink noise which rolls off at -3dB/oct.
So a speaker rated at 120dBspl produces that at say 50Hz but at 100Hz it would only need 117dB, 114 at 200Hz and 111 at 400 to comply with the rating.

I might be wrong...

 

[Tytte71]

Genelec's RMS and peak data is maybe useful for the purpose of convincing some buyers, but is IMO worthless as basis for your own design. Starting by scaling your system for a specific max SPL is a good thing to do. As an alternative to your "continuous output" design parameter I would use mean sound pressure level as a basis. Music is not constant an the dynamic range varies with the program material. If you start by e.g. 95dB mean sound pressure level (which is very loud and certainly not recommended for long term listening) and add 25dB (Crest) to allow for the majority of recordings to be played effortlessly at 95dB SPL, then your speakers should be able to play at 120dB peak within its linear range. You will soon find that this kind of performance is pretty challenging to achieve using traditional hifi loudspeaker drivers, especially if you are looking for a low distortion system.

I can understand you opinion about the Genelecs' qualities. The combination of controlled directivity, high capacity and good system integration is revealing. It is also possible to make better speakers than the Genelec 1238.

 

[xrk971]

I would ask what are the distortion n figures at those highs SPLs for Genelecs? Actually even at. Modest 90dB what is distortion? Below 200Hz is not pretty.

Just aiming for THX certified speakers is a very challenging task for most hifi speakers.

please correct me if I am wrong but - I think THX requirement is to reach 85dB reference level at listening position (2 to 3m) and still have 20dB headroom. That's 110dB peak at 2m or 116dB at 1m peak and 91dB reference at 1m.

I have tried designing to this before. It's very challenging and to to keep it lower distortion. You pretty much need pro audio setups.

 

[Charles Darwin]

Genelec states <1% 50-200Hz and <0.5% above 200Hz at 95dB/1m.

 

[andy19191]

Genelec gives the clearest representation I can find right now, of what I am trying to achieve.

In the Technical Data section (on RHS) for the KH 420 the specified maximum SPL at 1m varies from 109 dB to 122.4 dB depending on the conditions. In the Measurements section the maximum SPL at 1m for 1%, 3% and 10% distortion is plotted against frequency and this is perhaps the most informative piece of data. The 3" dome is a recent Neumann design using neo magnets and is waveguide loaded so some interpretation required. I think the old O500C used an ATC 3" dome but, again, waveguide loaded.

 

[5th element]

Of course it's going to be using some kind of noise stimulus otherwise the tweeters in these speakers would fry within seconds.

 

[torgeirs]

So you want a 3 inch speaker to play 2 dB louder than the Genelec 5 inch speaker at 450 hz?
Genelec 5 inch mid. 116 dB
You want 3 inch mid. 118 dB?

I doubt the Genelec is less than 92 dB sensitivity. And it has probably larger x-lin

So either Genelec is using a cheap driver or you are trying to defy physics🙂

 

[torgeirs]

Hm, Genelec says:
Maximum short term sine wave acoustic output on axis in half space, averaged from 100 Hz to 3 kHz @ 1 m @ 121 dB @ 250 watts. Assuming flat response.

250 watts is 8 times doubling of 1 watt => 3 dB x 8 = 24 dB gain over sensitivity => 121dB - 24 dB = 97dB midrange sensitivity
Hard to beat with 2 dBs using a 3 inch without a horn. (Specially if you can only use 100watts)

 

[audiothings]

So you want a 3 inch speaker to play 2 dB louder than the Genelec 5 inch speaker at 450 hz?
Genelec 5 inch mid. 116 dB
You want 3 inch mid. 118 dB?

I doubt the Genelec is less than 92 dB sensitivity. And it has probably larger x-lin

So either Genelec is using a cheap driver or you are trying to defy physics🙂

I am by no means an expert on the matter. But it seems to me that a 3" midrange driver is putting out 121 dB continuous in the ATC SCM300ASL Pro. The 3" midrange driver in the Quested Q412D is putting out 130 dB peak.

I have asked the question here, about the ATC drivers "blowing up" - and the answer seems to be an emphatic "no!". The loudspeakers do seem to be able to deliver to their specs.

 

[xrk971]

The ATC SCM300ASL has a huge magnet (~15lbs IIRC) and large voicecoil - more akin to a pro woofer. It's not surprising that it can handle a large thermal load under continuous use. Also note that it has a small waveguide which gives the sensitivity somewhat of a boost (probably 1.5 to 2dB at least) so that helps somewhat.

An externally hosted image should be here but it was not working when we last tested it.

 

[torgeirs]

I guess they tailor the short time noise to the max output of the drivers.
So 135 dB from the bass section that handles 4 octaves. And 114 - 118 dB from the rest. Total 130 dB for whole speaker.

Actually that is not that far away from realworld music stimulus (scaled down a bit🙂

I would say Genelec is more transparent in their spec.

 

[AndrewT]

How much heating effect is there in a 100µs single shot transient?
100us is a half period of a 5kHz sinewave.
25us would be the half period of a 20ishkHz transient.
Now repeat them for a 2% duty cycle. No wonder treble drivers with really effective HP filtering can survive enormous transients.

 

[5th element]

How much heating effect is there in a 100µs single shot transient?
100us is a half period of a 5kHz sinewave.
25us would be the half period of a 20ishkHz transient.
Now repeat them for a 2% duty cycle. No wonder treble drivers with really effective HP filtering can survive enormous transients.

Exactly and it really does make the whole 'max SPL' wars of loudspeakers like this somewhat meaningless.

Of course the power rating of loudspeaker drive units has been rather meaningless for years anyway, at least with regards to how manufacturers choose to display it.

You could theoretically test a mid/bass, or similar, with sine waves and come up with a figure for how many watts its coil can take before thermal failure but even this has limitations.

The drivers orientation, frequency of the test signal and enclosure all pay a large roll in determining how much power a driver can actually handle. So straight off the the bat, trying to use a sine wave as a stimulus has already thrown up a number of variables that will affect the end result. So is sine wave testing pointless? Well not entirely, but it's not as useful as you'd actually think...

Next you have to consider that music isn't steady-state, it's transient in nature, so the sine wave power handling figure has almost no meaning in the real world anyway, but what could be considered a useful figure?

The answer is nothing. No matter what figure a manufacturer attributes to its drivers it wont really mean very much to the end consumer.

Scanspeak and SEAS typically quote a tweeter power handling in watts, with a noise stimulus and with details provided on the filter used...this isn't particularly useful either. All it really does is allow them to claim a power handling that sounds useful to someone perusing their data sheets.

So X model tweeter can handle 150 watts with X stimulus and Y filter in place. YIPPEE! Well I'm not listening to noise, I'm listening to music (which can actually vary quite widely) and my crossover is completely different to the one they used...will my tweeter survive?

That doesn't really answer the question but manufacturers of such active loudspeakers are required to provide maximum, continuous, SPL figures. Clearly these are going to be derived when using broadband noise because any single frequency sine wave stimulus isn't going to allow for the largest figure. They want to have all drive units (and amplifiers) contributing so as to spread the thermal load and to arrive at the largest figure possible.

 

[xrk971]

I know Weltersys did some destructive testing of a TC9FD full range driver in a FLH to see how much power it takes to let the magic smoke out. These drivers are rated by Vifa as 30w rms and 60w max. IIRC it was able to handle some insane number like 400w. Also, with a steep high pass filter, dynamic bursts of perhaps 1 second at ~kW was applied and the driver survived - I think that work was done by Barleywater.

So, really, max music power ratings are meaningless. Sine wave RMS numbers do show how good heat dissipation in the voicecoil is though - but also depends on speaker box topology and access to air flow internally.

 

[torgeirs]

Yes, but as tytte writes in post #2, the x-max is the limiting factor at low frequency.
As you know, doubling the freqency gives 1/4 the x. So at maybe 1000? Hz and up the thermal capacity sets the limits of max SPL

 

[torgeirs]

Just for fun:
x-max @ 450hz @ 240w @ 112 dB
=> x-max @ 900hz @ 960w @ 124 dB
=> x-max @ 1800hz @ 3840watts @ 136 dB
=> x-max @ 3600hz @ 15k watts @ 148 dB

Disregarding -6 dB LR filter. So maybe the top octave can be over 130 dB with unlimited watts and short burst stimuli