AR9,90,91,92 replacement UMR's from AB tech

[Guest radkrisdoc]

They are all one and the same. Present day AB tech replacement number 200044. I asked for clarification and I was told that all drivers used in this series had the same characteristics. He said the crossover compensation/characteristics may be different in each speaker, but the drivers were essentially the same. I reconfirmed this as many times as I could...."what? say that again? Oh okay". Then he said that they have a few of the older drivers left (not for sale, as I understood) and that the 200044's are modeled after those characteristics. The only difference between these drivers is the brown and black flange.

I read this other older thread where Tom Tyson states that 200032 and 200028 are different mids. Im still scratching my head, cant fit the pieces together. Help! Someone please clarify!

[Guest Bret]

>Someone please clarify!<

I can't. But I can muddy things further. Here's the reason: I had the original pair of drivers from my 9s, an original one driver from a 92, and a pair of the 200044s.

Swapping drivers around I found that the 200044s were louder than either my 28s or the 32s in the AR-9. In the 92s they sounded as though they were "right" for the speaker. In the AR-9 they sounded a little "loud." Probably due to a crossover difference in the 9 and the 92, but I don't know that.

Now, the question is if my old ones are damaged (doubtful) or aged (definitely) and if the 200044 sounds like these used to. Can't say.

Somewhere I wrote down the DC resistance and I think I might have posted a message with those figures. Essentially they weren't meaningful.

To get the new ones to sound "right" in the 9s I had to attenuate them with the switches.

While you guys are wishing, here's what I wish. I wish we could get a functioning AR-9, 90, 11, 10pi, 3a, LST, LST II (name your favorites) and get all the replacement drivers that are available for them, then get someone who is a lot smarter than I am to look at the new drivers' differences with the originals and design crossover "corrections" for the drivers we can actually GET. (referencing the recent thread about the inductor differences between the foam and cloth versions of the 12" in the AR-3a)

The reason for that is that I'm finding quite a bit of difference in performance in the used drivers I'm buying, regardless of model number, and would *really* like to be able to buy *new* drivers to go into the old cabinets and tweek the crossover to put the whole thing back in balance. The Tonegen 12", for instance, aren't really all that close to the originals. I'm betting that they are way more efficient and have a much higher Fs. If we had the information I'm wishing-for you could rebuild a pair of speakers that sounded alike AND sounded like a pair of ARs.

It's the same problem with replacing the capacitors. I don't know of any source for a data sheet on the originals, so it's kind-of hard to "match" them with any precision.

We might be able to figure the UMR situation out if we just knew the real differences between the 28 and 32 and 44. The 28 and 32 carried different model numbers for some reason.

You'd think someone in the business (like a driver manufacturer or retailer) might be interested enough in the information to pony-up a buck or two for some lab time. But maybe there just isn't enough money in old speakers to make it worthwhile.

If I knew how, or was near a lab that could/would do it, I'd just do it myself. . . you'd think that all these specs and parameters would be in some file cabinet somewhere.

Bret

[Guest radkrisdoc]

>Swapping drivers around I found that the 200044s were louder

>than either my 28s or the 32s in the AR-9. In the 92s they

>sounded as though they were "right" for the speaker. In the

>AR-9 they sounded a little "loud." Probably due to a

>crossover difference in the 9 and the 92, but I don't know

>that.

>To get the new ones to sound "right" in the 9s I had to

>attenuate them with the switches.

Exactly what I felt....the difference between the 200032 and the 200028 is that the 32 is "louder", needs to be tamed with the attenuation switch; not only that, when I was hooking up the AR's the first time, I had a Polk LSi9 on the other channel. My first impression was the midrange had a "telephone-like" quality and I was quite puzzled at the time. Now I realize that the speaker with the original 200028 driver does not sound like that. The 200032 with the AR-90 crossover is the problem, it needs to be replaced with an original. With such an audible difference, what is one present day replacement driver (200044) going to do, it cannot change characteristics like a chameleon depending on whether it is installed in an AR-90, 91 or 92.

>We might be able to figure the UMR situation out if we just

>knew the real differences between the 28 and 32 and 44. The

>28 and 32 carried different model numbers for some reason.

>If I knew how, or was near a lab that could/would do it, I'd

>just do it myself. . . you'd think that all these specs and

>parameters would be in some file cabinet somewhere.

Hopefully wont be a problem if someone here identifies a lab. I am gonna try doing that. If cost is reasonable, we should be able to do it. Obtaining drivers....if I succeed in obtaining a lab's service, I already have 2 different upper mids. Just have to see how it goes.

And oh by the way, Tonegen. I am surprised that people here still refer to that Japanese oem manufacturer. The company long merged with Foster in 1986 and now is virtually non-existent.

http://www.foster.co.jp/annai_e/history_e.htm

I guess some other company is still making those Tonegen drivers.

Maybe all of us who need to buy drivers should get together and contact AB Tech to confirm driver discrepancies.

[Guest Bret]

>I guess some other company is still making those Tonegen drivers.<

I think we still refer to them as that because they were first built by Tonegen and made in the same plant after Tonegen was bought. I could be wrong.

Bret

[Pete B]

The 200044 probably uses a new and stronger magnet as we're seeing with the ceramic magnet woofers:

http://www.dextermag.com/Page82.aspx

I'd estimate between 1.5 and 2 dB more output, that's a lot over the wide bandwidth that the midrange covers. Does this driver also use ferro fluid?

It would help to get as much physical data as possible about the drivers, especially the motors. Gap height, coil height, DC reistance, inductance, most of what I listed for the tweeter. Full measurements are even better.

I can help with a resistive pad design but I need at least the DC resistance of these drivers. If people can post DCR data it would help. Anyone who has an old and new driver can veryify the correct pad by using something as simple as the Radio Shack SPL meter since even if the response is not perfectly flat (they're reasonably good in the midrange) it does not change. We don't need absolute calibration since all we need to know is the relative difference. I can give more details if anyone is interested in trying this.

It may seem like a patch to use a resistive pad, however, it is actually better than the original driver because the more efficient driver with the resistive pad will run cooler with less compression and a higher thermal limit. The best thing would be to leave out the pads and use all of the more efficient drivers. A solution would be needed for the overdamped bass response such as line level EQ, or a Linkwitz transform.

Pete B.

>Swapping drivers around I found that the 200044s were louder

>than either my 28s or the 32s in the AR-9. In the 92s they

>sounded as though they were "right" for the speaker. In the

>AR-9 they sounded a little "loud." Probably due to a

>crossover difference in the 9 and the 92, but I don't know

>that.

[Guest Bret]

>Full measurements are even better.<

Do you know of a way we could obtain that?

My hypothesis is this:

Either we figure something out or our speakers are on their very, very last legs except as disused collector's items.

Why can't we get all the original parameters on the old drivers? Sure, they may be "in the bottom of a locked filing cabinet in a disused lavoratory with a sign outside that says, "Beware of the leopard," but they must be somewhere.

Bret

[Pete B]

It's a lot of work to do acoustical measurements and I don't think they're required to solve this problem. A pad on the new driver should do it, if the voicing is still not quite right we can go from there. It's good that you have another 10pi as a reference. It's interesting that most of AR's published FR curves show the individual drivers which means the acoustical response can be reverse engineered in a program such as CALSOD. But the basic driver information is still needed.

I personally would probably put more effort into full restoration of something like the AR-9s since the work could also lead to an improved reissue that might be of value.

>>Full measurements are even better.<

>

>Do you know of a way we could obtain that?

>

>My hypothesis is this:

>

>Either we figure something out or our speakers are on their

>very, very last legs except as disused collector's items.

>

>Why can't we get all the original parameters on the old

>drivers? Sure, they may be "in the bottom of a locked filing

>cabinet in a disused lavoratory with a sign outside that says,

>"Beware of the leopard," but they must be somewhere.

>

>Bret

[Pete B]

I've worked with dome midrange drivers and from my experience a 1.5" dome driver is going to have a very hard time meeting up with a 12" woofer from a large signal perspective and the high crossover point supports this observation. There does not seem to be as much competition in dome mids and I don't think I've seen a higher Xmax than .5 mm which combined with the dome size does not offer much volume displacement (VD). There are few if any good quality 1.5" mids on the market, there is a Scan 1.5" but they call it a tweeter and I don't know if the Fs is low enough.

Most 3 to 5" cone mids have at least 1 mm Xmax, often much more, and obviously more radiating area. Cone drivers typically have very good off axis response since the dust cap acts as the radiator at high fequencies and dispersion is often better than a large dome.

I have for many years preferred something like the upper midrange in the NHT 3.3 which I believe is a polycone SEAS 4 or 5" driver. I don't believe the exact driver is available anymore but the SEAS MCA11FC or MP14RCY/P might be a good choice. This is a .pdf by the way:

http://madisound.com/seas/h522.pdf

The old Peerless 821615 is not one of my favorites but might also be a good fit, would have to check Fs.

[Pete B]

Just noticed this data from the AR literature:

1.5" Mid Fs:

AR-10pi 400 Hz

AR-9 800

AR-11 400

1.5" Mid:

AR-10pi 13,000 gauss

AR-9 13,000

AR-11 13,000

The higher Fs for the 9 makes sense since it has a higher crossover frequency and a higher Fs reduces excursion due to out of band energy.