Measurement and Testing of Acoustic Research Drivers

Measurement and Testing of Acoustic Research Drivers



  1. RobHolt
    As mentioned elsewhere I'm rebuilding a pair of AR4xa but using the later drivers from 1979/80 as used in the AR17/18.

    In the process I've been taking lots of measurements and thought I'd share them here along with a few observations.

    Effectively I have the original 4xa coated cloth surround woofer dated Nov 1972, a 1980 build of the 'universal' made in USA woofer as used in the AR18 (200001-1), and a 1982 version of the outwardly similar unit as used in the 18S (200037-0). The latter two units re-foamed.




    First up the two versions of the outwardly very similar universal drivers.

    Measured in the 4xa cabinet pulled away from boundaries and at 0.5m on axis:


    Observations: The 1982 build unit as used in the 's' versions is around 3dB hotter over much of the range.

    Now lets look at the 1980 universal against the cloth 4xa driver:



    Observation: The blue trace is the older cloth surround driver and is clearly more sensitive through most of the range, again by around 3dB. Also we can see that it rolls off earlier than the later version. Now lets equalise the mid points for these two drivers:



    Observations: Now it gets interesting. The newer driver (black trace) not only has better LF extension but also produces greater output. Worth having for a small AS cabinet.


    Now lets see what happens when we take the 1980 universal woofer and compare the loading of the 4xa 19ltr cabinet against the 9ltr AR18 cabinet:



    Observations: As expected the 4xa cabinet (blue trace) allows better LF extension and less peaking - all good stuff.

    Now lets look at the Fs for the old and new drivers. The 4xa specification gives the driver free air resonance at 35Hz, and thats exactly what we get from this 40 year old driver. However the new drivers both come in at 33Hz which is near as dammit the same for our purposes. Again all good:




    Observations: Nothing other than the older driver is slightly higher impedance overall, and given it also plays around 3dB louder it likely employs a lower moving mass, slightly trading LF extension against mid sensitivity.

    Finally the tweeters.

    Here we have the 'pointy dome' 1979/80 doped version from the AR18 (200014-3) against the flatter undoped version from the 18S (200038-0). Fed via a 10uf cap to protect from LF:


    Observations: The later 'S' version is clearly more sensitive to the tune of around 3dB and rather more peaky.

    Impedance for the tweeters was extremely similar and here is the 1979/80 version plot:


    Observations: Nice benign load and very well damped fundamental resonance at 1.8khz.

    Lots to wade through but hopefully it will be of use to folk when they come to replace drivers.
    What's apparent is that clear similarities run through the entire line from the 70s into the 80s drivers and that AR adjusted LF extension against sensitivity with different models. Certainly the 'universal' woofer and tweeter from the late 70s and (80) are both around 3dB less sensitive than the next version as used in the 'S' models, and care should be exercised when matching these universal drivers - they do come as pairs.

    Note: The dip around 140hz in all woofer traces is a room effect - likely a speaker-floor cancellation.