Jump to content
The Classic Speaker Pages Discussion Forums

Howard Ferstler

  • Content Count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About Howard Ferstler

  • Rank

Contact Methods

  • Website URL
  • ICQ

Recent Profile Visitors

5,319 profile views
  1. I first suggest that you absolutely make sure that you located each speaker in the same place and fed it the same signals from the same channel. That way, you eliminate any boundary-related issues and any issues with your amp, preamp, or source generator. Assuming you measured correctly, this situation is most likely a problem with driver aging, but the crossover might be having problems, too. A multiple malfunction, so to speak. The pink and white noise broad-spectrum level differences would probably not be related to capacitor deterioration. I suggest you swap the tweeters and see if the differences change places above 2 kHz. If so, then one tweeter might be having problems (or perhaps even both tweeters, given their respective ages). You could also do the same thing with the woofers. That will cull out any anomalies with the crossover. Howard Ferstler
  2. The Model Four crossover was the most complex one Roy designed for a two-way speaker, excepting perhaps the still later Model Four that he configured for the Kentucky outfit that brought the speaker back to life. Most of those he used with other two-way models simply installed ballast resistors in series with the tweeter. With the Model Four, not only did he use those attenuating ballast resistors, but he also installed a tapped connection to the choke, which allowed it to better dovetail when the tweeters were adjusted in level. Because of this, the level control is more complex than usual. Howard Ferstler
  3. Some CD-6 models retained the first-order tweeter filtering of the earlier Model 6 version. Many later CD-6 models had the same crossover network as the CD-7. The earlier version of the CD-7, namely the Model 7, had that same simpler network. Howard Ferstler
  4. I need to footnote my own commentary. One other mod I did recently was to insert an 8.2 mdf capacitor in parallel with each of the tweeter ballast resistors in all four networks (4 ohms in the inner panels and 6 ohm in the outer panels), as well as tweeter 4-ohm ballast resistor in the network within the center system I built. Thpse ballast resistors basically pad down the output of the tweeters (similarly to what we get with the tweeters in the Model One), but if one inserts a proper capacitor in parallel with that tweeter, as the frequency climbs the resistor begins to become bypassed and so the current flowing to the tweeter increases. This ramps up its output as its native output tends to roll off. My mod increased the output gradually above 8 kHz, reaching a maximum of +2 dB at 10 kHz and +4 dB at 12.5 kHz. Now, this may sound like a crank move on my part, but Roy actually did something like that with the resurrected Model Four he re-designed for the Kentucky outfit. He inserted a 12 mfd capacitor in parallel with that tweeter network's 2-ohm ballast resistor in order to get those tweeters to be a bit more impressive in the top octave. I used the 8.2 ohm cap because the tweeters seemed to be rolling off above 8 kHz more than they did when they were newer. I frankly think that both his and my "updates" would have zero impact for old guys like me, but it would certainly impress product reviewers who measured those systems and marveled at how they were potently flat right out to hypersonic frequencies. I did it on a lark, and would do it again. Howard
  5. The IC feature malfunctioned right out of the box. One or the other speaker would suddenly switch a panel on or off, and nothing I did (or what Roy suggested I do) fixed things. I even taped over the sensors on the fronts of the cabinets. I checked the schematic and there was one 1500 pfd capacitor in there that had me suspicious. (Its sole function, being hooked between the two G terminals in the IC-switching network, was to keep the 12-volt feed to that network from negatively interacting with, I believe, the tweeters.) If one looks at the schematic it can be seen that its insertion was an afterthought that resulted in some post-production problems. Whether or not it was the problem (probably not, since each system had one and the chances of both being defective were slim), there was a problem and after a while, since the omni mode was my prime listening requirement at that time, I figured that I could just hard wire things and stop worrying about sudden imaging shifts in the middle of a listening session. Prior to opting for my latest dual-network solution I spent some time with the systems hard wired to the focus mode, and it did not take long for me to realize that I wanted the switching functions back. Hence, those much more extensive mods done recently. Incidentally, Bill, I still have one of the IC network wall-wart power connectors for the systems, and it is possible that somewhere in my collection of old audio gear I have the remote switching hand-set too. If you hunger for either, let me know. I can at least surely supply the wall wart. All bets are off with that controller, however. Howard
  6. Yes, in the focus mode the attenuated panels are playing so low in level that it is impossible to accurately measure their contributions. The wrap-around sound from the louder panels, cabinet shading notwithstanding, simply overwhelms them. Their output is basically inconsequential as best I can tell. With my new, dual-crossover arrangement the outer panels can be (1) turned completely off, (2) attenuated by one to two dB, or (3) attenuated by perhaps 6 dB. The second assumption is based upon what Roy got with his Model One attenuations. The third is basically a guestimate on my part, since, as I noted, it is impossible to accurately measure the level of the attenuated panel without the output of the louder panel skewing the results. Up close you can definitely tell the differences in the panel outputs, but things get dicey as you move out further and get to the prime listening locations. I have listened to the speakers in the "stereo" mode of my system (surrounds and center turned off) and the difference between the now pseudo-omni mode (outer panels down 1 or 2 dB) and either of the other modes (completely off or down about 6 dB, as per my estimate with the latter) is surprisingly slight. You get a tad better detail and focus with the focus mode (needless to say, and this will depend upon room size/shape and sidewall materials), but somewhat, although not overwhelmingly so, better spaciousness with the outer panels running nearly full tilt (with room size/shape and sidewall materials again having an impact). The 6-dB down setting splits the subtle differences. However, once the surround and center channels are engaged all bets about stereo performance, soundstaging and imaging and clarity are off. the differences, especially involving spaciousness, become slight, indeed, and even the detail and focus differences become less apparent, due to the impact of the MTTM vertical array on the center speaker. The more speakers you have in a system the more things tend to blend together into a systematic whole. I rather like it that way. Howard Howard
  7. Hey, I am back. Here is some poop on those IC-20s of mine. The IC circuit switching acted up almost from the beginning and I finally decided to hard-wire the things into the omni mode and leave it at that. I gutted the IC network, and removed the sensors and lights on the front, covering the holes with a small brass plaque that says "Allison." This was my first mod and I could pull it off, because I had the crossover and IC schematics, courtesy of Roy. The second mod, done some years later, was a complete rebuild of the crossover network, following the factory specs. This involved replacing the old, back-located crossover mount with standard 5-way binding posts and rebuilding the network (new caps, new resistors, same chokes) on a triangular-shaped board that was screwed to the interior bottom of the cabinet. While the original cap arrangement had some poly jobs for the mid and tweeter sections and one standard non-polarized job for the woofers, the replacement caps I used were all poly types, including a physically huge one for the woofers. Because the new network used up some space at the cabinet bottom, the original reversed woofer tube would no longer fit, and so I re-configured the woofers both facing forward, just like with the Model One. Because I use outboard subwoofers for the frequencies below 80 Hz the modest increase in even-order distortion caused by this change would be meaningless. More on the subwoofers, below. However, now I have even more substantially modified my IC-20s, basically installing TWO crossover networks into each. The original network had to kind of split the difference in terms of choke and capacitor values, because it had to sometimes handle both four tweeters and four mids (omni mode) and at other times handle just two of each, with only a modest amount of feedback from the greatly attenuated other-panel drivers (focus mode). This works OK, because of the wide coverage capable from the mid and tweeter drivers, but it actually was not as perfected an arrangement as Roy got with either the Model One and Model Two units, or the Model 9 and CD-9. Those have the crossover chokes and caps precisely aligned with the fixed driver loads. Yep, two separate networks in each system. The one for the inner panels is similar to the version in the Model One: two tweeters, two mids, two woofers; and with some subtle changes (slightly different series choke and capacitor values) to compensate for driver aging with the midranges, plus going to second-order filtering for the tweeters to protect them better than what was possible with the first-order filtering of the Model One. (Note that the factory IC-20 also used second-order filtering for its four tweeters, as do all of the CD and AL series systems.) The outer panels each have their own crossovers (mounted on still another board and screwed to the interior rear of the cabinet, just above the binding posts), with the parameters similar to the mid/tweeter sections of the inner panels. The mid-tweeter panel networks are wired in parallel (each is nominally 8 ohms, like a Model One, with parallel wiring delivering nominally 4 ohms), and the outer panels can be turned off with 20-amp-rated switches. No changes were made to the networks for the woofers, which run together full time. Both panels have their driver complements fully protected by a total of FIVE polyfuses, each sized for the loads involved. The original IC-20 had three. Further mods involved some additional switch work. The inner panels have permanent resistors installed in their networks and those resistors partially mimic the "concert-hall slope" Roy made possible with the Model One's own switching option. The slope chosen for the inner panels was somewhat between Roy's full slope and the less aggressive middle slope. The outer panels use slightly larger contouring resistors, given them the full-slope contour. This means that the outer panels play one or two dB less loud than the inner panels when they are turned on. An additional switch for each outer-panel section allows the option to drop the output of the turned-on outer panels about 6 dB lower than the inner panels, kind of splitting the difference between the focus and omni modes. Finally, two additional switches allow the user to decrease the output of any of the panels in operation several dB in two stages. Because of impedance differences seen by those switches when either the inner or outer panels are in operation they offer up different attenuation amounts. With just the inner panels working each switch uniformly drops the mid/tweeter output by roughly 2 dB. With both panels in operation each switch uniformly drops the id/tweeter output by roughly 4 dB. These attenuations do not change the slopes of the contouring; they just change the mid/tweeter section outputs in relation to the woofer outputs below 350 Hz. The switches used for this additional contouring and shaping are also 20-amp rated jobs, and I do not expect them to develop oxidation problems. Most members of the Allison group I belong to know that I also employ flanking "ambiance" speakers, with a pair of side-wall mounted Model Fours each 30 inches away from the front wall and sitting on 5-foot high bookcases (these Fours are also modified to pretty much mimic what Roy did with the Model Four he designed for the Kentucky people) and a pair of Model Eights (modified to mimic the network behavior of the later CD-8) at the same height and on identical bookcases flanking the listening couch further back into the room. The output levels of these four surround speakers are MUCH lower in level than what we get with the standard Dolby set-up parameters, because it is their job to add a comfortable amount of simulated hall ambiance to the classical and baroque music I listen to the most. In my case they were initially set and then modified by the program in my mid-grade Pioneer receiver. I use the Dolby "height" mode to deal with those more forward surround speakers. I can leave the receiver settings fixed, because this installation is used for music listening only. This package also has a single, floor-standing center speaker (38 inches tall, with an angled-back slightly front panel) that uses Allison mids and tweeters in an MTTM vertical array, with a single Model Four woofer. The crossover network for the mids and tweeters is identical to that for the inner-angled panels of the IC-20s, with the woofer network mimicking what we have with the CD-9. I built the cabinet myself out of thick, solid cedar and MDF. The grill is a cut-down IC-20 grill. It has no level controls, and it is set to operate about 2 dB below the standard Dolby set-up level. The installation also has two subwoofers: cylinder types, 68 inches tall, 14 inches in diameter, with Dayton Reference drivers, and with power coming from a Crown XLS1000, 350 wpc power amp. An ART equalizer contours the sub outputs below 80 Hz. This arrangement with the IC-20s allows me to both shape the radiation pattern of each system from wide (but favoring the inner angles) to the usual, dual-panel super-wide pattern, and to adjust brightness levels when faced with recordings that are a bit too brittle sounding for my taste. Most of the time I run just the inner panels, plus the center and surrounds, with the slope adjustments turned off. This is a lengthy description of what I have done, following parameters set up by Roy long ago and not wildly experimental, but with some changes that are related to driver aging. If anyone is interested I can post some photos, at least once I figure out how to do it. Howard Ferstler
  8. This is my last visit and last post. I leave the field to Zilch, speakerdave, Ken, genek, and so forth. I move on to enjoying my throwback audio systems and my recording and book collections. Incidentally, I have removed all of the attachments that I previously installed in posts (pictures, graphs, PDF article reprints, etc.), because if I leave this place my stuff leaves, too. I only wish I could eliminate all of my posted messages. Those still itching to view graphs of any kind need only scope those supplied by Zilch, but don't expect to see any photos of his listening room. PS: at least, when I was here I used my real name, which is more than I can say for some of you people. Howard Ferstler
  9. If one cares to, they can go to the "kitchen" section and check on one of my posts dealing with the "Roy Allison interview in 1992 issue of The Audio Critic, David Ranada was the interviewer." Some here may be interested, because that particular post by me (which shows up on page 3 of the series) has an attached review of the AR-303 loudspeaker done by Julian Hirsch back in 1995 (Stereo Review magazine). What may interest some people here even more is that Hirsch does A/B listening and measuring comparisons between the 303 and the AR-3a. I would post the attachement here, but I have only a limited amount of attachment space left. Howard Ferstler
  10. Attached is a copy of Julian Hirsch's review of the AR-303 in the June, 1995 issue of Stereo Review. I have to assume you were happy with his assessment, too. In it, he does a direct measuring and listening-session comparison between that newer speaker and the AR-3a. Note that he got the crossover specs wrong with the comments on the AR-3a, saying they were 550 and 6.5 kHz, when they were actually 575 and 5 kHz. Well, nobody's perfect. Interesting that the 303 woofer is run up to a somewhat higher (and more directional) frequency than that of the 3a. The important thing will be his listening/comparing work, and everybody here who owns either speaker will certainly be interested. Howard Ferstler
  11. Two points. First, it is interesting that Allison's article on imaging and directivity that I posted on the "Library Additions and Corrections" page has managed to get a total of 80 hits, while the Kitchen thread that "debated" the issue managed to get over 2500. How in blazes does anybody debate (or read debates even if they are not posting any comments of their own) about what Allison said about dispersion and NOT go and read Roy's paper? OK, OK, I know that some here went into and out of the kitchen debate numerous times and probably only accessed Roy's paper once, but did each of those 80 article readers then go into the kitchen debate over 30 times? Second, yes, the Kitchen "debate" series was finally shut down by the webmaster, no doubt because things were getting pretty heated. (I am glad to note that the final bunch of caustic messages in that lengthy thread were at least not posted by me and that most of my postings were lengthy, fairly decent, informative, and not obnoxious defenses of my postions.) However, in addition to getting heated they were also, at times, getting pretty interesting (and informative), and shutting the whole thread down in order to calm certain people down (possibly the webmaster needed more calming than anybody else) smacks of an almost theological concept of protection. Hey, if things get REALLY bad, send a warning to the participants, and if they continue to be REALLY bad, then shut the thread down. However, I do not think that particular thread was headed towards perdition at the time it was terminated, and the fact that it scored over 2500 visits shows that it was a manifestly popular session. Howard Ferstler
  12. I recently posted an additional article about Roy Allison in the Library additions and corrections section. This one is an interview Roy gave to Ranada as part of a series in the Spring/Summer, 1992 edition of The Audio Critic. In the article Roy not only discusses the Allison boundary effect (something he discovered as part of his research on the AR-3a in typical home-listening rooms), but also offered comments upon the series of live vs recorded concerts Edgar Villchur presented when both were at Acoustic Research. Virtually all AR fans should be interested in what Roy had to say about those concerts, since they relate to the real-world quality of the AR-3 speakers used in the demos. I bracketed that section in pencil marks in the draft when I first read it years ago. Howard Ferstler
  13. In the "Library additions and corrections" section I just posted a PDF copy of a draft that Roy Allison presented at the 99th AES convention in October of 1995 . It deals with loudspeaker directivity. Howard Ferstler
  14. I just posted a PDF copy of a draft that Roy Allison presented at the 99th AES convention in October of 1995 in the "Library additions and corrections" section. It deals with loudspeaker directivity. Howard Ferstler
  15. Obviously, any components other than wire will have some kind of impact on the sound of a crossover network. After all, it is the job of such components to do just that. Assuming equal resistance values, no resistor will sound different from any other. There is no esoteric characteristic that would make one sound different from another. What might matter is power-dissipation ability, however. I suggest using at least ten-watt rated versions for AR speakers of this vintage. Howard Ferstler
  • Create New...