Frequency response of AR dome midrange?

[ligs]

I find the following graph very intriguing as it appears with the quote below from http://www.ebay.com/itm/PAIR-ACOUSTIC-RESEARCH-AR3-AR3a-MIDRANGES-/291289667620?pt=Vintage_Electronics_R2&hash=item43d2379824

"These are fully functioning midranges, but not perfectly matched as depicted in the frequency response plot in pic three using a sweep signal through a 25 uF capacitor. The level of the outputs are identical as the signal strength and mike placement were the same. The red response is for the older driver and the green response is for the 200010-1. For those not familiar with this nearfield response, the flat response is typical. It's only when the mike is located some distance from the driver that wall effects produce the typical jagged response.

The driver with the wire leads is probably from an AR3 or LST and dates to 1970-3. The other driver with the connecting posts underneath is labeled 20010-1 and probably dates to 1974-6."

I don't know much about close mic measurements but they are generally used only for woofers also both midranges look like 1.5" to me.

[Carlspeak]

The seller's measurements are somewhat of a surprise to me considering the age of the drivers and also the original AR published measurements which can be found in the CSP library. What is most puzzeling is how far up the FR those mids stayed flat before rolling off. The sellers tests shows them rolling off above 10 kHz, well above AR's ~ 4 kHz that did not employ a HP xover component. If those mids really do roll off that high, a mod to the xover would be needed to roll them off around 4 kHz. Otherwise a nasty response hump would emerge in the 5-10 kHz range due to the summing of the mid and tweeter's response.

Secondly, the low end roll off of the green mid is too high for it to work properly with the woofer and the original xover design. The typical xover point for the 3a and LST applications was in the 500-600 hz range. The 700-800 roll off point would cause a dip in the response in the 600-700 hz range due to the woofer's roll off around 400 hz.

A last comment: The dB level of the sellers test also seem very high. However, he does point out the tests were done close mic'd. So, he must have played them pretty loud to start with and the close proximity of the mic amplified the net SPL readings.

[tysontom]

I find the following graph very intriguing as it appears with the quote below from http://www.ebay.com/itm/PAIR-ACOUSTIC-RESEARCH-AR3-AR3a-MIDRANGES-/291289667620?pt=Vintage_Electronics_R2&hash=item43d2379824

"These are fully functioning midranges, but not perfectly matched as depicted in the frequency response plot in pic three using a sweep signal through a 25 uF capacitor. The level of the outputs are identical as the signal strength and mike placement were the same. The red response is for the older driver and the green response is for the 200010-1. For those not familiar with this nearfield response, the flat response is typical. It's only when the mike is located some distance from the driver that wall effects produce the typical jagged response.

The driver with the wire leads is probably from an AR3 or LST and dates to 1970-3. The other driver with the connecting posts underneath is labeled 20010-1 and probably dates to 1974-6."

I don't know much about close mic measurements but they are generally used only for woofers also both midranges look like 1.5" to me.

These eBay measurements are not meaningful. Close-mic measurements work fairly well for woofers being measured down below the frequency of ultimate radiation impedance, but this doesn't work very well for high-frequency drivers. Those curves don't even resemble the true output from the AR 1-1/2-inch mid-range driver, so it's hard to tell what he was getting, or even if these response curves emanated from the drivers themselves. The vertical resolution is heavily smoothed. The only standard (and acceptable) way to effectively measure an individual mid-range driver or tweeter is into anechoic space (or with gaited measurements) at 1 meter, on- and off-axis, usually with crossover in place. These test methods date back to the American Standards Association [ASA C16] and the RETMA 103 Standard and so forth, and these methods are widely accepted in the industry and are repeatable.

--Tom Tyson

[ligs]

Tom,

Thanks so much for the insight. I just began to realize how critical the midrange driver is. In the case of AR3a it covers about 10 octaves including the region our ears are most sensitive!

George

[tysontom]

Tom,

Thanks so much for the insight. I just began to realize how critical the midrange driver is. In the case of AR3a it covers about 10 octaves including the region our ears are most sensitive!

George

You bet, George!

You are right about the critical range of frequencies covered by the AR-3a midrange driver. In the case of the AR-3a, for example, the 1-1/2-inch driver determines much of the character of the AR-3a, its perceived sonic accuracy and smoothness, etc. Fortunately, this driver is exceptionally smooth and uniform -- one of the best dome drivers ever -- and it exhibits very little coloration to the sound. I don't know of any other midrange driver from any manufacturer that was smoother or more uniform within the same operating range -- if subjected to the same level of objective testing in an anechoic chamber under the same conditions. The response of the AR-3 midrange was very much the same as the 3a -- equally smooth and uniform -- but lacked the off-axis output of the latter.

There were some complaints early-on from Consumer Reports about a "thickness" in the lower-midrange, upper-bass frequencies that the organization attributed to the woofer and the midrange, but I think the issue had to do with a crossover choke that was used in the AR-3a around the time CU commented on the speaker. Roy Allison mentioned to me that the value was indeed changed once they discovered a small (2 dB) rise in output around the crossover frequency, and this likely accounted for the "thickness" mentioned by Consumer Reports. I don't think there was any fault in the AR-3a driver, and any test would show that this AR-3a mid-range driver (as I showed in the anechoic-chamber test of the original driver) exhibits extremely uniform, clean and smooth output, probably unsurpassed by any driver of this caliber. Indeed, Consumer Reports later tested the AR-LST (but did not include it in a standard test report) and stated that it was, by far, the most accurate speaker they had ever tested, and that it had an accuracy rating of 95%. Since the AR-LST and the AR-3a have the exact same drivers and same crossover topology, the differences are slight except for spectral balance. By the way, properly working AR-3 2-inch midrange drivers are also extremely smooth and uniform in its operating range as well.

The important thing about the AR-3a midrange driver is its durability over time. There is very little to go wrong with it; early versions of this driver, prior to 1977 or so, did not use Ferrofluid in the gap, and these earlier drivers work essentially the same way as when new, and this is good considering the the level of deterioration that has occurred on some examples of the 3/4-inch tweeter. In the midrange, there is no urethane foam under the dome or in the gap, etc.

--Tom Tyson