AR9 woofer crossover design

[soundminded]

I finally got around to reading Tim Holl's technical description of the design of the AR9 and had a couple of questions about the woofer crossover design for anyone who might know.

On pages 3 and 4 (images timholl4 and timholl5) the explanation for the crossover design says that an LC network has effectively zero impedence at the loudspeaker resonant frequency where the speaker impedence is highest and is shunted by a resistor which transitions as a series element into the woofer drive circuit as you move away from the resonant frequency to maintain impedence and protect the amplifier. But in the crossover schematic, this resistor is replaced by a 10 mh inductor. Is the DC resistance of the wire of this inductor used as the referenced resistor???? I'm puzzled by this apparant contradiction. BTW, in this design, any significant air leak would change the resonant frequency, radically alter the impedence curve and present a serious danger to some amplifiers.

Also I assume that the series inductor to the right of the series parallel LC combination and not given any value constitutes a first order low pass filter to restrict the woofer to signals below 200 hz. Is that right and does anyone know the value of that inductor? Would it be around 3mh?

Since the woofers are in parallel and the overall impedence of the system is nominally 4 ohms, confirmed by impedence curve in the user's manual, are the woofer voice coils wound for 8 ohms each? This would make them substantially different from all of the other 12 inch AR woofers and non interchangable.

[Guest Bret]

If the woofers can be 4 ohms in their cabinet, and the upper drivers can be 4 ohms in their cabinet, and the two cabinets being parallelled together are 4 ohms, then why can't two 4 ohm drivers in parallel be 4 ohms, too?

I finally gave-up trying to understand the whole thing because I couldn't seem to make anyone understand my question. I kept getting really good answers to all sorts of questions I hadn't asked, but nobody ever even got close to answering my question.

I hope someone will answer your question because it will answer mine.

BTW - I understand that one 12"AR woofer is much like (not identical, guys don't jump me) another 12" AR woofer.

I'm really looking forward to having Ken explain some of the tests he's run because I'd like to see if there was any meaningful difference in the performance of the old 10pi cone material compared to the stiffer 9 material. As it is, I'm not even sure I can tell what effect I might be able to chaulk-up to the cones.

Bret

[soundminded]

"I assume that the series inductor to the right of the series parallel LC combination and not given any value constitutes a first order low pass filter to restrict the woofer to signals below 200 hz."

Obviously, this is in combination with the 470 microfarad shunt capacitor forming a 12 db/octave low pass filter. Butterworth?

[soundminded]

I feel your lack of understanding is due to other people's inability including mine to explain it clearly enough.

While the drivers are rated 4 ohms nominally, when combined with their respective crossover elements they are only 4 ohms at some frequencies, the frequencies over each of their intended operation. At other frequencies they are much higher because the crossover elements in series with them are higher at those frequencies. When a low impedence say 4 ohms, is in parallel with high impedences say 25 ohms, the result is still only about 4 ohms or slightly less. So at say 50 hz, the woofer circuit (including the woofer itself and its crossover elements) may be 4 ohms plus a little with its crossover elements but the midrange and tweeter circuits may be 25 ohms each because of the impedences of their crossover elements (in this case the series capacitors.) at 1000 hz, the midrange may be 4 ohms but the tweeter and woofer circuits may be 25 ohms. And at 10 Khz, the tweeter circuit may be 4 ohms but the woofer and midrange circuit may be 25 ohms (in this case due to the series inductors.)

In the case of the AR9, you have 2 woofers in parallel operating at the same frequencies 28 hz to 200 hz. If each is 4 ohms, their combined impedence is 2 ohms. If they are 8 ohms, the combination yields 4 ohms. However, the crossover elements in series will increase the impedence. slightly. What Tim Holl correctly said is that the impedence is maximum at resonance, about 28 hz. The specified impedence of a driver is nominal and can vary all over the place. With the two woofers in parallel, at some frequencies, the combination will be much less. The added element is there to add series impedence but only at some frequencies within the woofer's operating range to protect the amplifier from seeing too low a load impedence. At resonance, it is bypassed by what appears to be a bandpass filter. BTW, it is also clear that to cut costs, in the woofer circuit, the two drivers share the same crossover elements so both woofers must be connected and operating and they must be the nearly identical or the crossover characteristic curve will be radically changed.

[Guest Bret]

I appreciate very much your willingness and stick-to-it-ness in explaining these crossover mysteries to me. I fear that I am so ignorant that I don't know how to ask the right question.

I completely understand what I have been told, thus far, concerning single frequencies. I didn't actually put pen to paper, but I thought the suggestion that I draw the resistance curves of the various drivers and overlay them caused a particularly effective image.

I see, very well, very completely, I believe, that the inclusion of high impedance everywhere except where a driver is being driven, has the effect of changing the overall impedance very slightly.

What's blowing my mind is that the 9 is essentially two, completely separate, cabinets; two crossovers that don't "touch" anywhere. Therefore the crossovers cannot "interact."

The upper cabinet operates at about 4 ohms, and the lower cabinet operates at about 4 ohms. Since either cabinet's impedance is VERY high outside its operating range and since a high impedance when added to a low impedance still gets you something close to the low impedance in total; I get how playing a 100Hz signal into the combined cabinet would yield a 4 ohm (or thereabouts) impedance; or into the separate cabinet would yield about 4 ohms..

If I play pink noise into the upper cabinet I'm going to have 4 ohms, about, overall. I can grab onto that.

If I play pink noise into the lower cabinet, I'm going to have 4 ohms, about, overall. I got it.

So far so good.

I lose all ability to think rationally when I strap those two cabinets together in parallel and STILL get 4 ohms. You might as well be telling me that 4+4=4 or that the impedance is kept in check by the "good woofer fairies" because it's all beyond me.

Could it be that the impedance is a sometimes thing and that both cabinets showing 4 ohms at the same time simply doesn't happen?

You, yourself, were explaining that impedance might dip well below 4 ohms at some frequencies in the woofer. Well, if I unstrap the cabinet on an AR-9 for bi-amping I don't read anywhere that cautions me to use an amplifier that can withstand a 2 ohm load on the woofers nor do I see anything in the sweep graph that shows an impedance that low.

I'm thinking you might as well give up and that I ought to keep my day job, give my VOM away, and forget everything I thought I knew about calculating parallel resistence.

Bret

[tysontom]

>I finally got around to reading Tim Holl's technical

>description of the design of the AR9 and had a couple of

>questions about the woofer crossover design for anyone who

>might know.

>

>On pages 3 and 4 (images timholl4 and timholl5) the

>explanation for the crossover design says that an LC network

>has effectively zero impedence at the loudspeaker resonant

>frequency where the speaker impedence is highest and is

>shunted by a resistor which transitions as a series element

>into the woofer drive circuit as you move away from the

>resonant frequency to maintain impedence and protect the

>amplifier. But in the crossover schematic, this resistor is

>replaced by a 10 mh inductor. Is the DC resistance of the

>wire of this inductor used as the referenced resistor???? I'm

>puzzled by this apparant contradiction. BTW, in this design,

>any significant air leak would change the resonant frequency,

>radically alter the impedence curve and present a serious

>danger to some amplifiers.

>

>Also I assume that the series inductor to the right of the

>series parallel LC combination and not given any value

>constitutes a first order low pass filter to restrict the

>woofer to signals below 200 hz. Is that right and does anyone

>know the value of that inductor? Would it be around 3mh?

>

>Since the woofers are in parallel and the overall impedence of

>the system is nominally 4 ohms, confirmed by impedence curve

>in the user's manual, are the woofer voice coils wound for 8

>ohms each? This would make them substantially different from

>all of the other 12 inch AR woofers and non interchangable.

The woofers in the AR-9 are the standard 12-inch, 4-ohm design. There were no 8-ohm 12-inch woofers. I think (without having a crossover schematic handy) that there were series-parallel elements present to keep the impedance correct, for the most part. Also, it would take a serious air leak to cause a resonance problem and subsequent amplifier-loading problem. That's an unlikely condition unless someone mounted the woofers with screws missing or without a gasket or something.

--Tom Tyson

[tysontom]

>BTW - I understand that one 12"AR woofer is much like (not

>identical, guys don't jump me) another 12" AR woofer.

This is basically true. All of the AR 12-inch woofers use the same voice coil, same number of turns, same voice-coil overhang in the gap, same magnetic force in the structure, and so forth. There were some minor differences in the weights and construction and density of the cones over time, but these differences were slight and had very little effect on sound quality. AR went about improving the cones going forward -- primarily in the upper range of response -- and they also began to stiffen the suspension elements to increase reliability. From the first highly-compliant, 12-inch ceramic woofer in the 1969 AR-3a up through the AR-9/AR-9LS versions, the compliance differences are quite noticeable, but the audible differences were slight. I think there were response improvements in the upper bass with the later woofers. The original AR-3a design is probably more a "pure" acoustic-suspension design than the later woofers -- especially if you use the classic 90/10 model -- in that it relied more on the air in the enclosure for restoring force than the mechanical springs of the woofer's mechanism. But this is purely academic and probably inconsequential. The later woofers, on the other hand, were superior with respect to power handling and resistance to "bottoming."

>I'm really looking forward to having Ken explain some of the

>tests he's run because I'd like to see if there was any

>meaningful difference in the performance of the old 10pi cone

>material compared to the stiffer 9 material. As it is, I'm

>not even sure I can tell what effect I might be able to

>chaulk-up to the cones.

>

I think that Ken's tests are going to reveal not so much the differences in the various iterations of the 12W's cone material but rather the effect of retrofitting scivers, voice coils and spider assemblies -- which I assume to be the case with some of the woofers he is testing. It will be most helpful knowing what effect rebuilding the woofer will have on its performance. How does it compare with the original? In the end, I think we want to restore performance of a given AR speaker as close to its original value as possible. Improvement over the original design would be fantastic, but arguably unrealistic.

--Tom Tyson

[harry398]

This is another thread that went quiet that I find fascinating 

 

More info needed.

 

What baffles me is the differences in woofer circuits between the ar9 and the ar9lsi. (Yes I know the 10 inch isn't 4 ohm,but still).    

I wish some of the original designers of this phenomenal speaker would arrive and Talk.