ATC SM75-150S Dome Mid Measurement Data

[ShinOBIWAN]

There's always plenty of heated debate regarding ATC's dome mid so I finally decided to put together measurement data and post it here to perhaps remove some of the subjectivity and also the bashing for the sake of bashing.

I should also thank Ralph who's FR and impedance data I have borrowed. Ralphs measurements followed my own very closely so thats reasurring regarding accuracy, it also saves me much time capturing, editing, converting and uploading that data.

Just a couple of very quick notes:

The distortion test was performed outside and at 90dB average measured at the mic which was exactly 2m from the driver. The driver itself was mounted in a Perceive Satallite enclosure with no crossover in place for both distortion and CSD tests.

Cumulative Spectral Decay and distortion measurements were taken in ARTA and STEPS software(v1.1.0) using a Behringer ECM8000 mic, both ARTA and STEPS were calibrated using a true RMS DMM, the mic sensitivity was taken as per manufacturer specs.

On Axis:


The amplitude response is not as smooth or as flat as might be expected for driver that costs so much or is held in such reverence. However the phase response is remarkably smooth between 300Hz-4kHz. The above plot is a 'nearfield' LF response merged at 700Hz with a 'farfield' high frequency response taken at 0.5m. The first major breakup occurs at about 4.5kHz and is also evident in the impedance plot below. Below 1kHz the response rolls off smoothly.

As mentioned, in order to use this driver in a system requires considerable EQ. This isn't easily achieved with a passive crossover, and the driver benefits from an active implementation.

Off-axis:


The off-axis response is where the driver really shines. It's easy to eqalise a driver to be flat on axis, but not so easy to equalise it for a flat power response as well. Only a driver with a consistent power response can be equalised to achieve both.

Below is an example of the ATC's consistent off-axis performance. EQ has been applied to flatten the response somewhat, and the consistency in amplitude response as the measurement microphone is moved off-axis is easily seen; at 45 degrees the amplitude response remains within +/-1dB between 400Hz and 2kHz. Only above 2kHz does the driver start to beam and the power response fall away consistent with size of the driver's diaphragm.

Note that the vertical scale is 1dB/division. By augmenting the ATC mid with a controlled directivity HF driver, a smooth roll off in the system power response can be achieved. By 4.5kHz things are getting nasty, so to use the driver up to around 3.5kHz as is found in ATC's own speaker systems suggests quite steep crossover filters are needed.

Impedance of 8ohm version


The impedance plot is well behaved with just a small glitch at 700Hz and a larger one at 4.5kHz. The driver resonates at 320Hz which is only marginally outside the claimed operating range of 380Hz to 3.5kHz. Normally a midrange driver is not used close to it's resonance frequency. But in the case of the ATC mid, it works very well down low, sounding better crossed around 300-400Hz than higher up. Of course, when used below 400Hz, a steep filter is recommended or power needs to be limited not exceed Xmax.

Distortion:


CSD:


Hope that helps demystify the fog of hype/BS that surrounds this driver. Make of the measurements what you will and please note the test conditions - it makes all the difference.

If anyone is interested in other measurements/conditions I'll should be able to provide them given time.

There also seems to be a distinct lack of measurements for the Scan R2904 ring radiator, if anyone wants those I do have them.

EDIT: Uploaded wrong file for distortion measurements, now fixed.

[Garmt]

Good stuff! I would be very interested in measurements of the ring radiator as well. Do you also have data for the normal 'dome' version of the Scanspeak Revelator for some comparison?

[dunkyboy]

Great work. Correct me if I'm wrong (and I'm a total noob when it comes to this sort of thing!) but those distortion measurements look very good! Nearly entirely all below 1% within its operating range, and majorit 2nd order. The decay chart also looks very well behaved within its operating range.

Dunc

[Tenson]

Its not a bad measuring midrange. Until we have measurements of an alternative midrange though, it does not say a lot. Stand up PMC/Volt!!

It is about what I would expect, but since it costs such a lot of money maybe it should be doing better.

I think Ant is on the money with his comments about crossing lower than ATC do in their implementation. There are tweeters which perform a lot better in the upper midrange where they overlap. The RAAL tweeter can cross as low as 1.5KHz and has a decay that is all over and done in less than 0.5ms with distortion all below 1%.

[Stereo Mic]

It's better crossed over at 500hz IME and also if it's not used above 3khz.

[ShinOBIWAN]

Originally Posted by brizonbiovizier

So you do care!

At last - the reason ATC SCM100s makes my ears hurt has been spelt out clearly and unequivocably.

Of course those measurements are without crossover.

Go design a loudspeaker and you'll realise that actually measurements aren't as important as you'd think, implementation and room are of course everything.

You should also know that the PMC version is inferior to the ATC. Its actually a stripped version and more inline with the ATC standard version. Pop it out take a look.

[Tenson]

Originally Posted by ShinOBIWAN

You should also know that the PMC version is inferior to the ATC. Its actually a stripped version and more inline with the ATC standard version. Pop it out take a look.

Without measurements you can not really jump to that conclusion and implementation is everything anyway.

[ShinOBIWAN]

Originally Posted by Tenson

Its not a bad measuring midrange. Until we have measurements of an alternative midrange though, it does not say a lot. Stand up PMC/Volt!!

It is about what I would expect, but since it costs such a lot of money maybe it should be doing better.

I think Ant is on the money with his comments about crossing lower than ATC do in their implementation. There are tweeters which perform a lot better in the upper midrange where they overlap. The RAAL tweeter can cross as low as 1.5KHz and has a decay that is all over and done in less than 0.5ms with distortion all below 1%.

OK show me another driver that does 0.06% to 0.3% THD between 500hz to 3Khz and at real world levels of 90dB at 2m(96dB/1m) measured outside and to put those distortion figures into perspective - they're approaching the THD of solid state amps. At the same time show me a driver which offers the same or better off-axis performance and overall power response at those SPL and distortion levels. Simple answer is you can't.

CSD plot has a -35dB floor as opposed to the marketing BS you see with 25 and even 20dB + screwy gating. In fact if you raise the CSD floor to -25dB the mid decays in 1.5ms @ 500hz. Again show me another mid or mid/bass that does that and with the attributes I mentioned above.

The real deal breakers are sky rocketing distortion below 500hz that simply put is piss poor, and a series of resonances starting at 4.5Khz lending a hard and forward sound to the lower treble. Its easy to see why ATC aren't using this driver correctly with 380hz and 3.5Khz XO points along with 4th order slopes.

I've had these measurements for months and months now and the more drivers I test the more I'm impressed with the ATC's performance. There's also the otherside of the story - subjectives and I've still yet to hear better mids on a loudspeaker at any price than what I've got with my DIY speakers.

Final words: Anyone who judges the ATC mids on the performance of ATC's implementation is an idiot basically, doubly so given the data I've provided here. Its like saying you don't like a car because someone stuck a 2 speed gearbox in there.

[ShinOBIWAN]

Originally Posted by Tenson

Without measurements you can not really jump to that conclusion and implementation is everything anyway.

I was speaking with Billy down at ATC earlier this year and he gave me the low down on the PMC dome mid after I mistakenly assumed they used the super. Its lower sensitivity, less BL lower build and no copper in the motor.

Those are important attributes to a well designed loudspeaker I have no shame when stating that they're inferior to the ATC supers.

[ShinOBIWAN]

Originally Posted by Stereo Mic

It's better crossed over at 500hz IME and also if it's not used above 3khz.

I'm currently investigating a new(to me) type of crossover slope: 2nd order Neville/Thiele which offers some interesting properties including brickwall filtering after an initial and traditional roll off characteristic. By using linear phase subtrative crossover generation and combining the slopes the ringing is cancelled out.

The real problem with using such steep slopes is that you need to very accurately match the tonal qualities of each.

[dunkyboy]

Any idea why ATC might be inclined to use such obviously wrong crossover points for the drivers they designed and [presumably] know more about than anyone..? You'd think after 20-odd years they would've noticed their mistake and shifted them slightly?

Dunc

[darrylfunk]

could some one point out there reasons for assuming atc have their crossover points wrong as distortion is still fairly low at 500 hz , are there crossovers not tuned so the drivers stay voiced more naturally 380hz is quite hi for a bass driver to reach so surely its just a judgement they have come to.
also dont pmc crossover the mb2 a at 380hz as well !!!
seems a bit harsh to critic the designs to me.

[bottleneck]

I assume setting a crossover point isn't exclusively a decision that you make to avoid a drive unit's points of distortion.

Other points - such as avoiding crossing over in the vocal speech area, or perhaps to integrate into a speaker design parameters, to meet the best crossover point of another alternate driver must all be concerns I'm sure.

I wouldn't personally begin to second guess the engineers at ATC. Whether I like their products or not is immaterial. The fact that they employ more qualified audio engineers in the field of loudspeaker design than you will typically find in an audio forum on the internet has to be a given.

[Tenson]

Ant you sound a little stressed

If you notice, I was saying you were right, crossing a little lower than ATC do from the mid to the tweeter seems a good idea. The RAAL tweeter may have its CSD gated at -25dB but even if you take that into account from 2.5KHz which is where I might choose to Xover it will have a FAR quicker decay rate. The distortion measurements are done at 85dB/1m so I'll give you that one, but they are still very low from 2.5KHz upwards. THD is not such an important factor IMO anyway if it is mainly even order. The horizontal dispersion is also EXCELLANT from 2.5KHz upwards.

Thus I am saying that there are tweeters which perform better than the ATC mid from about 2KHz upwards where it would be a wise move to Xover above.

I would be very interested to have similar measurements from a high quality compression driver and CD horn.

Personally I have never heard fully convincing vocals from a speaker with a mid dome. I don't know why.

[fox]

It could be because you approach the midrange dome with a preconception and thus you predispose yourself up to not be convinced?

I like those wavy pattern things earlier on BTW. Looks like serious stuff. I'm sure that's why so many CMI fairlights were sold in the 80s, you can't fail to look kick ass with one of those...



Well almost...