Jump to content

Anyone Have AR-3a Measurements, System, Drivers


Pete B

Recommended Posts

Looking for test reports or any other source for measured AR-3a system input impedance. Individual driver impedance would be helpful also. Mid and Tweeter would be helpful. Didn't find any of this in the library, did I miss it?

Pete B.

Link to comment
Share on other sites

I have the full fold out technical analysis of the AR-3a. I would scan it, but it folds out to be huge, making it impossible to scan. Let me see if I can find it, in my mess of AR stuff.

In the mean time, if anyone else has this document, feel free to post the data that Pete is looking for. Its the fold out document with the red printing.

Link to comment
Share on other sites

Thanks Joe, I'm not sure what your referring to exactly, but scans of the system and driver input impedances would be helpful. I got frequency response data from the library, but it's not very clear, better scans of this would help also. I'm working on a full model of the AR-3a and have a rough first cut. I used the woofer data from Ken's work. I'm doing this to determine the characteristics of the mid and tweeter in support of the AR-LST discussion.

Pete B.

>I have the full fold out technical analysis of the AR-3a. I

>would scan it, but it folds out to be huge, making it

>impossible to scan. Let me see if I can find it, in my mess of

>AR stuff.

>

>In the mean time, if anyone else has this document, feel free

>to post the data that Pete is looking for. Its the fold out

>document with the red printing.

Link to comment
Share on other sites

Sounds like a cool project! I have been designing an AR-4x type system for a while now. I used off the shelf tweeters, but the woofer, I designed from the ground up. I just need to get around to having the woofer made!

My collection of AR documents is at home back in New York. In case you did not know, I away at College in Chicago. I will have the documets mailed to me, and then I will try to find a wide scanner, and scan them.

Link to comment
Share on other sites

>Looking for test reports or any other source for measured

>AR-3a system input impedance. Individual driver impedance

>would be helpful also. Mid and Tweeter would be helpful.

>Didn't find any of this in the library, did I miss it?

>

>Pete B.

Pete,

You may already have this, but this information was on the AR-3a Technical Data Sheet #L-1621 and other places.

http://www.classicspeakerpages.net/dc/user_files/443.jpg

Fig. 19 shows AR-3a impedance with level controls set on the "normal" position; Fig. 20 shows AR-3a with controls set at mimimum; Fig. 21 shows AR-3a with controls set at "maximum."

One important thing to note is that the impedance swings below four ohms at some frequencies when the controls are set to maximum, but not by much. The AR-3 went below three ohms in places.

Individual-driver impedance curves were not published. You could easily measure them yourself if you have an AR-3a at your disposal and are willing to disconnect the drivers.

--Tom Tyson

Link to comment
Share on other sites

>You may already have this, but this information was on the

>AR-3a Technical Data Sheet #L-1621 and other places.

>

>

>http://www.classicspeakerpages.net/dc/user_files/443.jpg

>

>Fig. 19 shows AR-3a impedance with level controls set on the

>"normal" position; Fig. 20 shows AR-3a with controls set at

>mimimum; Fig. 21 shows AR-3a with controls set at "maximum."

>

>One important thing to note is that the impedance swings below

>four ohms at some frequencies when the controls are set to

>maximum, but not by much. The AR-3 went below three ohms in

>places.

>

>Individual-driver impedance curves were not published. You

>could easily measure them yourself if you have an AR-3a at

>your disposal and are willing to disconnect the drivers.

>

>--Tom Tyson

Thanks very much Tom, this is exactly what I was looking for and it is very interesting. I don't have an AR-3a to test but expect probably in the near future. Testing drivers will be helpful but we don't know how they've aged so old data is helpful also. Would you happen to have the AR-3a woofer input impedance directly driven without the crossover network? AR-3a woofer frequency response data without the crossover network would also be helpful.

I'm seeing some interesting results from this simulation. The ESR of the 150 uF cap strongly influences the input impedance dip around 450 Hz due to slight peaking in the filter's amplitude response with very low ESR. This is common, I've seen it before. The capacitor ESR strongly influences the Q of the HF rolloff of the woofer XO. A very low ESR cap causes more peaking in the amplitude response and a lower impedance dip in the input impedance. Here the ESR of the cap lowers the Q of the filter and flattens the amplitude response. Using the cap's ESR is not the best way to control Q since it then becomes dependent on capacitor type, but it is the most economical way, use a low cost cap and eliminate the need for a series resistor. However, now more power is dissipated in the cap and it will heat and age faster. One forum member reported the 470 uF cap in the same location in an AR-9 exploding with a very high power amplifier.

I'm also finding that the 2.85 mH woofer inductor over compensates for baffle loss resulting in a depressed response around 200 to 500 Hz when simulating with DUT 5/6 woofers (low Bl). The response with the 1.9 mH inductor looks much closer to the published response curves. The 2.85 mH coil might compensate for the newer higher Bl woofers, by reducing the upper bass (200 to 500 Hz) output and in a relative sense increasing the below 100 Hz output. It may have also been changed to give the speaker a better balanced sound in room. I'm not exactly sure just offering some possibilities.

The input impedance dip to below 3 ohms at 900 Hz in the above figure 20 is very interesting in that the controls are all the way down offering a simple 16 ohm resistive load to the crossover network. This is caused by the 50 uF cap resonating with the .88 mH inductor with a higher Q when the pot is all the way down. There is very little damping in the circuit. The .5 ohm series resistor was an attempt by the designers to reduce this impedance dip effect. "Upgrades" to a lower resistance inductor will also make this dip lower and therefore should not be used. The Q of this midrange high pass filter is high and there is peaking in the amplitude response but it is also strongly effected by the ESR of the 50 uF XO cap. Low ESR caps will raise the Q even higher and also raise the associated amplitude peaking. The .5 ohm resistor and inductor resistance provide at least some loss/damping.

I'll try to estimate the ESR of the 150 and 50 uF caps based on this input impedance data. Thanks again Tom!

Pete B.

Link to comment
Share on other sites

Simulation of Fig 20 yields some interesting results. It is valuable because the unknown mid and tweeter input impedance are removed from the analysis with the pots down. The woofer inductor is 1.9 mH. The 50 uF has fairly high ESR, and an effective value closer to 40 uF near the crossover frequency. I think the 50 uF was two paralleled paper caps? The 150 uF seems to have very low ESR, was this typically a Compulytic?

I used estimated values for the DC resistance of the AR inductors, anyone have measured values for the .044, .88, and 1.9 mH?

Pete B.

Link to comment
Share on other sites

>Simulation of Fig 20 yields some interesting results. It is

>valuable because the unknown mid and tweeter input impedance

>are removed from the analysis with the pots down. The woofer

>inductor is 1.9 mH. The 50 uF has fairly high ESR, and an

>effective value closer to 40 uF near the crossover frequency.

>I think the 50 uF was two paralleled paper caps? The 150 uF

>seems to have very low ESR, was this typically a Compulytic?

>

>I used estimated values for the DC resistance of the AR

>inductors, anyone have measured values for the .044, .88, and

>1.9 mH?

>

>Pete B.

Pete,

I don't have an impedance plot for the unmounted AR-12W, but I do have the FR plot. Measurements made outdoors, speaker mounted flush in the ground (2π Steradians) and a calibrated microphone set at approximately 1.5 meters (the old RETMA standard) above the speaker and rotated through different solid angles of 30 and 60 degrees off axis. The Q of the original AR-3a woofer system is close to 1.0. Close-miking is much easier and gives almost identical on-axis response results, but as you know you cannot measure off-axis that way. The woofer coil was in the circuit I believe.

http://www.classicspeakerpages.net/dc/user_files/448.jpg

--Tom Tyson

Link to comment
Share on other sites

Pete,

The DC resistance values I have are:

AR coil #4/.88mh= .4+ohms

#7/1.9mh= .7+ohms

#9/2.85mh=.9+ohms

The small coil .044mh is .2+ohms...

In earlier versions the woofer(150uF) and mid(50uF) caps were combined in a big (heavy!) plastic encapsulated/Chicago Industrial "block" with a common input wire. Later versions have single large can types (Chicago Industrial "Royalitics") for the woofer and mid. The tweeter went from a single 6uF "block" to a blue metal Sprague. The latest version I have seen dating to late 1974 has a small modern stye NPE on the mid, the larger blue Sprague NPE on the tweeter and the big Royalitic can on the woofer.

I don't believe the mid was ever served by paralled caps in the '3a.

Roy

Link to comment
Share on other sites

Thanks again Tom,

I've seen similar plots like this but never with such an expanded scale, this is helpful. It matches simulation fairly well with the 1.9 mH inductor. About 1 dB of the peaking around 450 Hz comes from the woofer XO and seems to be deliberate, however it is the opposite of what's needed for baffle step compensation, it was probably designed for use near room boundaries where BSC is not needed. This was probably also done to raise the system's voltage sensitivity slightly. The low frequency response does looks to be nearly text book with a Q of about .9 or very close to 1 as you stated. I was interested in the woofer input impedance in box without the crossover but it's not too important. The larger 2.85 mH inductor does provide much more BSC compensation, this is typically how it's done - make the woofer inductor larger.

Pete B.

Link to comment
Share on other sites

>Pete,

>The DC resistance values I have are:

>AR coil #4/.88mh= .4+ohms

> #7/1.9mh= .7+ohms

> #9/2.85mh=.9+ohms

>The small coil .044mh is .2+ohms...

>

>In earlier versions the woofer(150uF) and mid(50uF) caps were

>combined in a big (heavy!) plastic encapsulated/Chicago

>Industrial "block" with a common input wire. Later versions

>have single large can types (Chicago Industrial "Royalitics")

>for the woofer and mid. The tweeter went from a single 6uF

>"block" to a blue metal Sprague. The latest version I have

>seen dating to late 1974 has a small modern stye NPE on the

>mid, the larger blue Sprague NPE on the tweeter and the big

>Royalitic can on the woofer.

>

>I don't believe the mid was ever served by paralled caps in

>the '3a.

>

>Roy

Thanks Roy,

I'd guess that Royalitic is Chicago Industrial's spin on Compulytic. Probably a low ESR type. I think there's a schematic in the library that shows two caps in parallel for the 50 uF, I'll trust your input.

Pete B.

Link to comment
Share on other sites

>I think there's a

>schematic in the library that shows two caps in parallel for

>the 50 uF, I'll trust your input.

Hi Pete,

That schematic is one of a Layne Audio rebuild.

Along with the parallel mid caps it shows l-pads instead of rheostats and advises connecting the mid out of phase with the woofer. Its very misleading for folks looking for the original configuration. I have not seen a '3a from 1969 through 1974 that has any of the drivers wired out of phase with the others. The mid and tweeter crossover components are installed on the negative side of the circuit. On the "front-of-the-cabinet-wired" versions, a common yellow wire connects the positive sides of the tweeter and mid with the positive cabinet "T" terminal via the connected #2 pot terminals. In the back-wired versions there are 2 yellow wires, one from each pot.

I forgot to mention that the inductors are all 17 ga wire.

Roy

Link to comment
Share on other sites

Thanks Roy,

Yes that's the one, didn't make much sense when I read it.

Pete B.

Link to comment
Share on other sites

Bret posted pictures of the AR-9 crossover where there was a clear view of the 470 uF/50VNP and patent number 3842325. Here's a link to the patent:

http://patft.uspto.gov/netacgi/nph-Parser?...5&RS=PN/3842325

This 470 uF was the one that one forum member reported "exploding" I commented that a high ESR cap would dissipate power but this is a very low ESR design. I believe that it was over voltaged, exposed to over current, or had aged so that the ESR was much higher.

The patent shows how special construction methods are used to provide low ESR. Most NPEs are not constructed in this way as r_laski's picture shows the typical single contact point:

http://www.classicspeakerpages.net/dc/user_files/433.jpg

Pete B.

Link to comment
Share on other sites

Thanks also to rjr513 for the figures. Now I'm scratching my head looking at these. Tom's figure of the AR-3a woofer radiating into 2pi or half space suggests that there is no BSC compensation. However, the woofer response in rjr513's figure is into 360 degree radiation and suggests that there *is* BSC compensation, since it is reasonably flat. In fact, it is *over* compensated since 100 Hz is slightly higher than 300 to 500 Hz. I think the system in this plot used the 2.85 mH inductor. Is this much later literature from the time when the 2.85 mH was commonly used? Systems with the 2.85 mH inductor and the low Bl woofers will have much more bass output:

http://www.classicspeakerpages.net/dc/user_files/440.jpg

Almost forgot to comment on this plot below, this is in between, less peaking at 450 Hz than Tom's plot, but flatter than the one above. It says that the woofer was measured out doors and all drivers were into 360 degrees? Can anyone confirm how the woofer was measured, half space or full space?

http://www.classicspeakerpages.net/dc/user_files/436.jpg

Link to comment
Share on other sites

The figure below is from Allison's AES article "The Influence of Room Boundaries on Loudspeaker Power Output", June 1974. It is an AR-3a (woofer only) measured under 2pi and 4pi conditions. The agreement with theory is excellent, note that it is heavily smoothed and that he uses PWL rather than SPL on the Y axis, PWL dB values should be doubled when converting to SPL. SPL is analogous to voltage and PWL to power. Curve B is very similar to Tom's graph above and shows a system without BSC, curve A further supports this since it shows baffle loss under 4pi conditions. I believe that the early AR-3a did not have BSC and it was incorporated some time in the 1970s with the 2.85 mH inductor:

http://members.aol.com/basconsultants/2pi4pi.jpg

Link to comment
Share on other sites

The full-page AR-3a literature was from the 1973-74 time frame. Note the 10 American Drive address in Norwood MA.

This is a great series of literature, which spans the entire set of late-Classic AR's: LST, LST-2, 3a, 5, 2ax, 8, 6, 4xa and 7. It even has the incorrect photo of the grilleless 4xa that AR continually used at that time--they show the 7!--but only the sharpest-eyed AR aficionado EVER notices that.

The other curves (with the slight rise at 450Hz) were made during the Thorndike Street Cambridge MA era, prior to 1973.

Differences between the data are likely caused because the measurements were made at different times (a few years apart, when both the measuring equipment and conditions were slightly different), slight changes to the components of the 3a itself (both intentional and unintentional), and marketing "creativity"--which is an ever-present reality despite our unwavering faith to the contrary.

Steve F.

Link to comment
Share on other sites

Tom,

I just noticed that this FR plot is identical, as far as I can tell, to one given in the AR-LST brochure from the library in file arlsttwo.zip and the XO point is shown as 575 Hz. The annotations are even the same. I would not be surprised if AR used plots from the 3a just to avoid having to redo them for LST literature but I can't say for sure.

I also notice that the LST brochure from library file arlstone.zip shows a downward sloping response similar to the AR-3a response given in the "10 AMERIAN DRIVE" brochure, and states the XO point as 525 Hz.

Pete B.

>Pete,

>

>I don't have an impedance plot for the unmounted AR-12W, but I

>do have the FR plot. Measurements made outdoors, speaker

>mounted flush in the ground (2π Steradians) and a

>calibrated microphone set at approximately 1.5 meters (the old

>RETMA standard) above the speaker and rotated through

>different solid angles of 30 and 60 degrees off axis. The Q

>of the original AR-3a woofer system is close to 1.0.

>Close-miking is much easier and gives almost identical on-axis

>response results, but as you know you cannot measure off-axis

>that way. The woofer coil was in the circuit I believe.

>

>http://www.classicspeakerpages.net/dc/user_files/448.jpg

>

>

>

>--Tom Tyson

Link to comment
Share on other sites

  • 2 months later...

Steve F I think your over simplifying my analysis of what's going on with the inductor change. I'm reposting this here since it applies:

I covered some aspects of your question about baffle step compensation in another thread but did not state it outright:

http://www.classicspeakerpages.net/dc/dcbo..._id=6180&page=4

The center of the baffle step as can be seen in Allison's measurements is about 400 to 500 Hz. A crude form of baffle step EQ could have been obtained by setting the mid and tweeter controls about -3 dB from flat in an early AR-3a with the 1.9 mH inductor. It would be more abrupt, and not the full 6 dB but the full amount is not usually required in any case. What I was showing in that thread without stating it outright was that the newer design with the 2.85 mH inductor seems to have modern baffle step compensation since in post #6200 we see that the woofer is flat into half space as is stated in the figure, probably with the 1.9 mH inductor:

http://www.classicspeakerpages.net/dc/dcbo...id=&page=4#6198

Whereas here in post #6227 we see that the woofer is flat, actually overcompensated since the response slopes slightly down, rather than up, where it is stated as being measured into a 360 degree solid angle or hemisphere. This probably has the 2.85 mH inductor:

http://www.classicspeakerpages.net/dc/dcbo...id=&page=4#6198

The two cases should look like A and B in Allison's AES paper if there was no difference in the speaker system:

http://www.classicspeakerpages.net/dc/dcbo...id=&page=4#6198

In fact the speaker in Allison's paper probably had the 1.9 mH inductor.

I believe that the early systems with the 1.9 mH inductor were designed to be flat into half space and the later systems with the 2.85 mH were designed with baffle step compensation. It would be interesting to hear both types of AR-3a's side by side, they should sound very different.

Pete B.

Wally asks from another thread:

>BTW, would AR speakers benefit from Baffle Step equalization?

>When would manufacturers have stsrted to design speakers with

>Baffle Steo Equalization built in? I have some well reviewed

>speakers like Paradigms and Celestion DL10 that are much

>bassier than AR's when against a wall for instance. The bass

>on these sounds about the same as my AR's when these (newer!)

>speakers are away from room boundaries.

Link to comment
Share on other sites

It’s questionable whether the 3a ever incorporated BSC into its design. If it did, AR never mentioned it in any of their post-1973 literature. Roy Allison’s independent research in 1972-93 identified and quantified many of these issues for the first time in the industry. The 3a never was presented to the public by AR as addressing these issues.

Of course, it’s possible that the 2.85 mH inductor change to the 3a was meant as some sort of EQ for that effect. AR’s marketing during that time period was so incredibly inept that it is not unthinkable that the engineers could have incorporated such a change into the design but the Sales/Marketing departments didn’t know how to publicize it properly.

Personally, I doubt it. I still think that the 3a’s basic design approach was unchanged, and that any component/measurement changes over the years were simply the result of the normal migration of parts used and assembly methods employed—as always happens with any item that is sourced and manufactured for a long time at multiple locations. But it’s possible—I certainly have no proof either way, only a hunch.

Perhaps Tom might have some internal strategy documents that would shed some light on this.

Steve F.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...