AR-7 and AR-17 Response Curves, etc.

[Rich W]

I've recently run a set of reverberant field frequency response tests on a pair of AR-7's and a pair of AR-17's. Methodology and location is identical to tests I ran on a pair of AR-3a's: http://www.classicspeakerpages.net/IP.Boar...?showtopic=5434

the only difference being the house curve, which rolls off at 1.5 db per octave above 2k vs. 3 db per octave for the AR-3a test. Curves were also processed with 1/6th octave smoothing. AR-17 tweeter set to "-3 db". AR-7 tweeter set to "Flat".

For reference, I superimposed last year's AR-3a curve for comparison:

A few observations:

- Although not designed for music listening specifically, I've been pretty fortunate to have a room that can do some justice to these vintage AR's. Note the similar bass responses down to about 60 hz, where the expected rolloffs on the AR-7 and AR-17 would occur. The room is giving them quite a bit of welcome help on the bottom octave though. Suckout at 70 hz, common to all three, is a room node artifact.

- The AR-17's (ADD series ca. 1977) are a recent restoration (new surrounds, Solen cap). As soon as I hooked them up, I knew how special they were. At my prime listening position, they measure an incredible +/- 2 db from 30 hz to 15k! They actually have a smoother response curve than my AR-3a's. The AR-17 uses the same drivers as the highly regarded AR-18's, but in a larger cabinet to further extend the low end. They would be compariable to the AR-6, which has the identical cabinet volume and similar woofer. Also, note the more extended high frequency response of the newer ferrofluid 1 1/4 inch tweeter vs. the older incarnation of the almost physically identical tweeter used in the AR-7 (and AR-6).

- The AR-7's, restored about 2 years ago, measure a respectable +/- 4 db from 40 hz to 9k. The elevated midrange (in particular the peak at 1.7k), while clearly audible on any program material, can be dealth with effectively with my one-third octave graphic equalizer. Incidentally, both my AR-7's exhibit the same midrange bump, even though one of the woofers came from an early vintage AR-18. I'm following with interest Speaker Dave's thread on enhancing the AR-4x crossover -his AR-4x's have an almost identical bump. I know the AR-7's should be performing better without any electronic assistance! Further testing on the individual drivers will be needed.

- In my particular room, It's difficult for me to admit that the AR-17's actually sound better in more listening locations than my AR-3a's. Specifically, the AR-3a's need to be heard at greater distances, at which they of course sound completely authoritative. However, at closer distances (closer than where these tests were run, but by no means in the "near field") the AR-17's have the edge, maintaining a more even mid-bass response and superior imaging /sound stage. The AR-3a's are definitely a large room speaker and need quite a bit of space to "breathe" to sound their best.

Rich W

[speaker dave]

I've recently run a set of reverberant field frequency response tests on a pair of AR-7's and a pair of AR-17's. Methodology and location is identical to tests I ran on a pair of AR-3a's.

Rich W

Hi Rich,

Great to see some comparative curves of more AR products. The AR17 is indeed amazingly flat. Glad to hear that you like its sound as well!

The AR7 curves look like the balance we've seen on most of the earlier systems with a climbing woofer response and a little downhill tilt to the tweeter. I'm sure some of the xover mods, similar to the 4x mods I'm doing would help out, although the component values would be particular to the system. If you are so motivated you could try the quasi second order woofer network to pull down the midrange some.

I'm jealous of your room which give an amazingly flat and extended response at LF. You are doing some spatial averaging, right? Are the individual curves nearly as flat?

David

[Rich W]

I'm jealous of your room which give an amazingly flat and extended response at LF. You are doing some spatial averaging, right? Are the individual curves nearly as flat?

David

Hi David,

Yes, I'm doing spatial averaging within an approximately 10 square ft. prime listening area. For the AR-17's, I ran pink noise through both speakers, set a very high number of averages in my RTA software, and moved the mic stand around within the test area. For the AR-7's, I ran individual tests at 10 different locations for each channel. The chart below shows 6 representative curves for the right channel, with the average curve for all 10 at the top. The individual curves are pretty decent. So you can see more clearly what's going on in the bass region, I limited the upper frequency to 2k. All the curves are in very close agreement above that.

With all the room augementation below 60 hz, it's really a mind-bender to hear a little pair of AR-7's reproduce (albeit at a reduced level) a 32 hz sub-contra C organ pedal. I actually ran tests where I got them to respond to a 27 hz sine wave - with substantial fundamental that you could feel. As you can see from my first set of graphs, the AR-17's bottom octave response falls right between the AR-7 and AR-3a, which is just about what you'd expect.

As far as the midrange peak on the AR-7's, I'm going to run some near-field tests to rule out any interaction with the room. I expect that varying the tweeter level control should tell me whether the woofer or the tweeter is the prime culprit contributing to that extra energy at 1.7k. My one-third octave equalizer can tame that peak pretty well, but it would be so much more satisfying to have the speaker respond more evenly through the midrange on its own. It's somewhat of a mystery to me that the AR-17's are so much smoother in the midrange, considering that one of my AR-7's has a later AR-17/18 type replacement woofer. As you probably know, the AR-7 has a "minimalist" crossover, the woofer is allowed to roll off with no inductor, and the tweeter network is first-order with a single 6 uf series cap. To me, that make the tweeter network in the 7's suspect. Maybe the tweeters themselves have aged to the point where they're starting to show anomalies near their resonant frequency. In any case, I can definitely hear that midrange coloration - from past test reports and AR's own response curves from back in the day, these things should be more neutral sounding. That said, they still sound incredible. In the same manner as you with your AR-4x's, I want to wring that last ounce of performance out of these little guys!

Best Regards,

Rich W