Refreshing AAs

[Guest " "]

I think you have the makings here for an intresting DIY offering to sell via e-bay. A kit to modernize your AA's. 20 bucks, well done.

Thank you for your encouragement and comments.

My intention was simply encourage folks to do something about there old crossovers, rather than getting caught up in "thinking" about the various options (and doing nothing).

No kit needed. Simply go to PartsExpress.com

You will need the following Solen capacitors

4 @ 2.0 uF caps (027-534)

2 @ 6.2 uF caps (027-558)

2 @ 6.8 uFcaps (027-560)

this totals $ 21.12 in parts

Some cable ties, silicone adhesive or caulking, and solder lugs are helpful. The 6.2 & 6.8 caps are paralleled to replace the original 13 uF cap. It is also a good time to clean up any oxidized contacts and attach some pipe foam for a snug fit into the corners.

The investment is small. There will be some improvement and you can always do something more ambitious in the future. But at least do this.

Good Luck,

-Tom

 

Great. I share you views on over complicating a project like this. Problem defined, problem solved.

I was going to ask you to post the parts list, but did not what to change the direction of your thread.

This is high quality DIY work.

Sure, there are pros and cons for doing "other" or "different", but what you have done, and shared with us is more than adequate, fun, and certainly DIY.

[Al Klappenberger]

Guys,

It is true that the woofer will dictate the impedance seen by the amp at low frequency. In the case of the K33 that is 6 Ohms (not 8 and not 4). It is in series with a 1 mHy voice coil that must be tuned out by the filter. The network determines the impedance above the lowest crossover. In the case of the AA, it goes up to over 30 Ohms at mid frequency! This is becasue of the use of the transformer without compensation. It's a dirty trick to get away with a small series cap (13 uF) where a larger one is really necessary for an 8 Ohms system. Anyhow, my original ALK "universal" network is 6 Ohms at extreme low frequency and gradually moves to 8 Ohms from there up. It stays 8 Ohms within about +-1 Ohm. It's also a pure resisitive load. It looks like an 8 resistor, not like a cap or inductor at verius frequency points like the AA.

Eric: You can't just stick a swamping resistor into a network unless it was designed to be there. This is probably why you had to remove it. If you removed it from one of my network designs I'm sure you would be putting it back in a hurry!

Tom,

What bothers me about these posts on low cost upgrades to the AA is that most people who are not experineced filter designers have no way to know how bad a design the AA is. Becasue it was designed by PWK, everybody assumes it must be winderfull! It takes sombody like me who is a filter specialest to realize how much better it could be. The truth is the AA is simply an "A" with a quick fix to the tweeter filter and some zener diodes stuck in to keep from blowing out tweeters!

Al K.

[PrestonTom]

Tom, The improvement going to new caps you saw when you rebuilt your late model type AA crossovers is probably less than most see when when rebuilding older crossovers. The little black caps you replaced usually read around 0.25 ohms each for ESR. There are exceptions to that with some of them reading close to an ohm ESR. Most of the older metal can caps in the AAs test at around 1 ohm ESR. With that reading, the difference is much greater after replacing the caps. New good caps will have an ESR of around 0.02 ohms. One more point on the Type AA rebuild is to make sure that Klipsch used the right type screw in the 245 uH inductor. Lots of times you will find a steel screw in it and it should be brass. If it has a steel screw, the value of that inductor will be increased a lot. I have not been able to find the correct brass screws, but they are readily available in stainless and that is what I use. You can just check the screw with a magnet. If it is attracted by the magnet, you should replace it. Bob Crites

Bob,

Thanks for the info on your measures of the ESR on old caps. In spite of all the equipment we have in the lab at work, I had nothing to measure the ESR on the old caps. Actually, I could not even find a LCR meter.

Yes, I forgot to mention the screw that holds down the inductor. In my case, it was probably brass, but I did not have a magnet handy. Easy fix: take it out completely and put a dab of adhesive underneath

Good Luck,

-Tom

[PrestonTom]

..........

Tom,

What bothers me about these posts on low cost upgrades to the AA is that most people who are not experineced filter designers have no way to know how bad a design the AA is. Becasue it was designed by PWK, everybody assumes it must be winderfull! It takes sombody like me who is a filter specialest to realize how much better it could be. The truth is the AA is simply an "A" with a quick fix to the tweeter filter and some zener diodes stuck in to keep from blowing out tweeters!

Al K.

Al,

Yes, I understand your concern. There is a good deal of hero worship around PWK and of course filter design, an entire topic in itself, has improved over the years.

Yes, the A & AA networks are pretty simple; a single element to low pass the woofer, a single element to high pass the mid (and let it roll off on its own for the upper cutoff) and a single element (now 3 element + diodes) for the tweeter. Obviosly there is much room for improvement. I have only seen bits & pieces regarding the later AK series, although it looks like the design became much more sophisticated.

As I see it , and I suspect we are in some agreement, the biggest difference in sound you will get is in terms of changing the topology, and you have championed this position. A big difference will also be gained when a dereriorated part is replaced (refurbishment) and certainly Bob Crites has championed that position (and effectively also).

Unfortunately many threads are focused on the comparatively smaller issues (details perhaps). If you want a big difference tackle the worn out parts, or bigger still, tackle the basic topology. If you want to spend an even greater amount of money, then further chase down esoteric differences and focus on the "flavor" of the capacitors etc. My concern is this perspective sends folks down a very expensive path. The further improvements gained when you follow this path will be much smaller in comparison. If you have sufficient funds, then certainly do all three: update the worn out parts, pursue a better topology (while you are at it), and use "premier" components. However with all due respect, it is a reasonable question to ask whether the additional expense results in an additional (and appreciable) improvement? It is also reasonable for me to cringe when I realize that many are just waiting and waiting and doing nothing.

I'll be a broken record, since I believe it is an important point. Whether my suggestion is a complete solution or not, is not the issue. All I am suggesting is that for $21 & 2 hours of effort you will get an improvement. I have been honest about the nature & magnitude of the improvement (although I was not able to do an actual A/B test) and think that others should give it a try also. I am interested in hearing about their experiences.

Good Luck,

-Tom

[Erik Mandaville]

Al:

"Eric: You can't just stick a swamping resistor into a network unless it was designed to be there. This is probably why you had to remove it. If you removed it from one of my network designs I'm sure you would be putting it back in a hurry!"

I'm sure you're right. I have some ideas about those in general, and have used them on my Lowthers, which of course don't use crossovers at all.

I very much to prefer to leave them out -- at least on this type A. I built a higher order network a couple of years ago, tried the same thing then, and took it out then, too.

Thanks,

Erik

[Guest " "]

Guys,

It is true that the woofer will dictate the impedance seen by the amp at low frequency. In the case of the K33 that is 6 Ohms (not 8 and not 4). It is in series with a 1 mHy voice coil that must be tuned out by the filter. The network determines the impedance above the lowest crossover. In the case of the AA, it goes up to over 30 Ohms at mid frequency! This is becasue of the use of the transformer without compensation. It's a dirty trick to get away with a small series cap (13 uF) where a larger one is really necessary for an 8 Ohms system. Anyhow, my original ALK "universal" network is 6 Ohms at extreme low frequency and gradually moves to 8 Ohms from there up. It stays 8 Ohms within about +-1 Ohm. It's also a pure resisitive load. It looks like an 8 resistor, not like a cap or inductor at verius frequency points like the AA.

Eric: You can't just stick a swamping resistor into a network unless it was designed to be there. This is probably why you had to remove it. If you removed it from one of my network designs I'm sure you would be putting it back in a hurry!

Tom,

What bothers me about these posts on low cost upgrades to the AA is that most people who are not experineced filter designers have no way to know how bad a design the AA is. Becasue it was designed by PWK, everybody assumes it must be winderfull! It takes sombody like me who is a filter specialest to realize how much better it could be. The truth is the AA is simply an "A" with a quick fix to the tweeter filter and some zener diodes stuck in to keep from blowing out tweeters!

Al K.

 

Al, thanks for the summary on the issues the universal network is designed to correct. The specifics about the autofomer kinda relates to discussions going on in other threads.

Tom, I have to agree with you on the costs benefit analisys.

[Hardhead]

PrestonTom--

Concerning your question: "t is a reasonable question to ask whether the additional expense results in an additional (and appreciable) improvement?" The answer is a resounding (pardon the pun) "Yes."

If you do some research in the archives regarding the ALK crossovers, you will see that the comments of his customers are overwhelmingly positive. Anyone who takes the time and effort to look in the archives--including those too "paralyzed" by their lack of knowledge to make well-informed decisions regarding their crossovers--will find a tremendous amout of well-considered and informative postings.

[Al Klappenberger]

Eric,

If you make this comment:

I very much to prefer to leave them out -- at least on this type A. I built a higher order network a couple of years ago, tried the same thing then, and took it out then, too.

You do NOT understand the basic concept of the swamping resistor! You CAN NOT take a swamping resistor in or out on a whim! The choice to use it or not to use it TOTALLY CHANGES THE ENTIRE NETWORK ELEMENT VALUES depending on if you choose to use it or not use it.

Al K.

[Deang]

That's why I asked him what cap value he was using at the primary position. Based on his preference with the resistor out -- I suspected he didn't scale the parts to reflect the stepped down impedance. If he put the resistor in, using a 13uf cap at the primary -- the crossover point between the bass bin and squawker would be about 1500Hz.

[Al Klappenberger]

It is quite obvious that many people do not understand what the "swamping resistor" is actually doing.

The 13 uF cap in the AA fixes the low crossover to 400 Hz for a 32 Ohm (roughly) load. The swamping resistor drops the load to 8 ohms. That is a ratio of 32 / 8 = 4. If you design the filter (1st order = 1 cap) with the swamping resistor the load impedance is now 8 Ohms so the filter must be "impedance scaled" by a factor of 4. This means the 13 uF cap must be changed to 13 x 4 = 52 uFd. The two must be changed TOGETHER! The performace will be the same EXCAPT, the 52 uF and swamping resistor option will look like 8 ohms to the amp. With the 13 uF cap alone the amps sees 32 Ohms! Again: This is simply a dirty trick done in the days when a 52 uF cap was the size of Texas. There is no blasted reason on earth to stick with 1960 compromises in 2006! Look at the specks on your amp and see what it is rated to operate into. 4 Ohms, 8 Ohms, maybe even 16 Ohms if it's a few years old. Have you ever seen a spec for a 32 ohms load on a home stereo amp? Then why load one with 32 Ohms? This is silly! The only execption to this is amps used for 70V distribution systems. That's another story though.

Al K.

[PrestonTom]

PrestonTom--

Concerning your question: "t is a reasonable question to ask whether the additional expense results in an additional (and appreciable) improvement?" The answer is a resounding (pardon the pun) "Yes."

If you do some research in the archives regarding the ALK crossovers, you will see that the comments of his customers are overwhelmingly positive. Anyone who takes the time and effort to look in the archives--including those too "paralyzed" by their lack of knowledge to make well-informed decisions regarding their crossovers--will find a tremendous amout of well-considered and informative postings.

Hardhead,

I am afraid that you have misunderstod me, I certainly did not suggest that ALK crossovers would not provide an improvement. Certainly it is the case that replacing worn out parts & re-designing the topology can improve the sound.

I also hope that my comments are not viewed as an attack on the ALK network. That was also not my intention. My comments were much more general in nature and were made after reading a large number of threads on these topics. Actually there is a wide variety of options available and I still believe it is confusing topic.

Good Luck,

-Tom

[Erik Mandaville]

Al:

As I said above, I'm sure you're right. The network I made a couple of years ago was actually quite similar to your Universal type 'A'. Maybe I need to go back to that, and rebuild it EAXACTLY to your specifications, using Hovland caps (or would Solens work in the HF branch?) and try again. I seem to sense a sort of 'lossy' quality that may have just been my own aging hearing. Maybe I just prefer the sound of very simple, low order networks. Who knows, but again, I'm sure your right.

That said, you indicated above:

"What bothers me about these posts on low cost upgrades to the AA is that most people who are not experineced filter designers have no way to know how bad a design the AA is. Becasue it was designed by PWK, everybody assumes it must be winderfull! It takes sombody like me who is a filter specialest to realize how much better it could be. The truth is the AA is simply an "A" with a quick fix to the tweeter filter and some zener diodes stuck in to keep from blowing out tweeters!"

In other words, Paul Klipsch was a human being who, like all of us, made mistakes -- or at least in this case a crossover network with room for improvement? So he designed the really 'bad' (your words) type AA. He is also the individual whose legacy has brought to us some wonderful things, not the least of which has become a very significant part of the foundation of your own work. What we have the luxury of knowing today, Al, very often comes from the groundwork done by those before us. I want to submit a couple of things: 1) that what is better to you may not always and without failure be 'better' for someone else -- just as the type AA network was at the time of its conception and construction thought highly enough of to put into production. And 2) that you have tools of technology, convenience, and speed at your disposal that the designers of these earlier Heritage crossovers perhaps did not.

In their day, the AA or infamous type AL (which I actually also liked for awhile) may have been well thought-of. As bad as you say it is, I found the AA to be a decent network, and similar in many ways to the very simple type 'A'-- the more obvious difference to me being the lower tweeter output of the AA (the aspect I didn't care for) It seems others have been satisfied with it, too, even with refreshed capacitors of the same kind originally used in the Heritage line. Or the cheap black Solens, which are actually very good capacitors. This would indicate that, as poor a network as the AA is, its deficiencies are not so great that people are throwing them into the trash can without looking back or thinking twice.

Erik

[Deang]

You and me are so often times on the opposite ends of things -- but that was a pretty good post. The AA doesn't appeal to me, but I enjoy the sound from a nicely built set of Type A's very much.

[Al Klappenberger]

Guys,

The reason people do not hear now bad the AA network is is simply becasue the inharent problems in a mechanical transducers like a horn and driver are far worse than those of even the worst electroinc filter! Even the very best of them! To hear the difference between a first rate network and a realy band one you must isten through all the distortion, time delay and bad frequency response of a mechanical transducer to hear it! This is why it's so easy for me to see how bad the AA is. I'm looking at it through computer simulation, electronic instruments and 30 years of experinece designing them at microwave frequency where the requitements are FAR more stringent becasue the loads are perfect resistors and not mechanical transducers!

Al K.

[Erik Mandaville]

Dean:

The type 'A' may be a case where less really is more. I once made a pair using cobbled parts from very old networks, which was really the wrong way to go about it. They sounded like absolute......well.....just very bad. Brand new, or at least known 'good' caps, and decent air-core inductors were an entirely different situation. They didn't cost very much, and work pretty good for me.

Off and on over the past few years, I've considered the possibility that passive, speaker-level crossovers may have some intrinsic drawbacks, but just have not had the motivation to take the other road. I'm still learning about surround sound, but maybe one of these days I'll start working on it. I have the amps I would need, but for some reason not the mental drive.

Al:

"This is why it's so easy for me to see how bad the AA is. I'm looking at it through computer simulation, electronic instruments and 30 years of experinece designing them at microwave frequency where the requitements are FAR more stringent becasue the loads are perfect resistors and not mechanical transducers!"

I understand this. And again, and with all due respect....

If I brought a microscope and wet-mount slides to the public swimming pool I'm using for physical therapy after spinal fusion surgery, I'm sure I would see more than I would want in terms of the zillions of microscopic, scary-looking earthlings, paddling around with strange fins or tiny water-jet propulsion systems, looking for unsuspecting creatures like me. I doubt I would ever go swimming again! I don't see those invisible critters -- either in the pool or water I drink, so I continue on with a healthy fun swim and cold glass of water when I get home. Hmmmm......now I'm wondering If I should really use that pool! It's doing the job as well as the fitness club right down the road, though, and is very inexpensive. As long as I'm staying healthy and getting the exercise I need, maybe it's just as good as the more expensive place!?

In other words, the imperfections you speak of, discovered through testing that takes place at microwave frequencies and with perfectly resitive loads, may sound pretty cool on the one hand, but may not necessarily be directly observable when applied to the real-world and obviously much more crude example of developing effective frequency dividing circuits for people and the mechanical devices they use to play music in their homes.

Sure, I understand that such tight design tolerances can do nothing except improve the much more crude behavior of woofers, squawkers, and tweeters (maybe). However, in terms of what can actually be detected by the ear, the quality of which being as subjective as it is, the fact remains that there are some who are pretty happy with what they have, even if what they have happens to be the scoundrel type AA.

Erik

[Al Klappenberger]

Eroc:

In other words, the imperfections you speak of, discovered through testing that takes place at microwave frequencies and with perfectly resitive loads, may sound pretty cool on the one hand, but may not necessarily be directly observable when applied to the real-world and obviously much more crude example of developing effective frequency dividing circuits for people and the mechanical devices they use to play music in their homes.

In yet other words: The crossover network isn't worth upgrading becasue the dirivers and horns are the weak link.

In that case why do YOU (Eric) bother to build networks? To the extremes, since the drivers are the weak link we all might as well hook our stock Khorns to WallMart boom boxes!

No! Every component in the stream must be optimized in the end. There is no end to how far you can go. I simply scream about networks becasue this is what I know. I'll let guys like Bruce Edgar work on the horns. BTW: horn drivers are not really very bad loads to a filter. They just look a bit inductive. That is not hard to deal with.

AL K.

[Erik Mandaville]

Al:

I understand that this is a sensitive subject. You're also somewhat missing the point I'm trying to make.

Walmart boomboxes usually have loudspeakers built into them, but I've read about people hooking up k-horns to Sony Walkpeople with pretty good results.

The fact that the type AA network in your estimation measures poorly does not necessarily make it an inferior network. To you it does, since you are basing your qualification on its measured response. However, some users may use it over another because they simply like the way it sounds, without respect to its response profile. I've read the writing of well-known audio reviewers who have indicated both that horn speakers are by their very nature horribly colored and inaccurate. Based on that, I suppose I should have carted mine to the nearest dumpster and then sold my house in order to try to qualify for a mortgage on a new pair of monoliths from Wilson Audio. Same is true for the now ancient debate over low powered single-ended triode amplification.

Why do I build networks? It's fun, and because my speakers need them in order to function as properly as my ears are able to distinguish. This is just for the Klipschorns, though. The Lowthers don't need crossovers, but that comes with a host of other tradeoffs. Nothing is free, I suppose, but I do know that it's possible to make very good sounding networks out of the two types we have been discussing.

And Al: I respect the hard work that you do, and have publicly commented on your very high level of craftsmanship and construction. That can't be denied. Nor can the fact that MANY people love what you have created for these old but still very capable speakers.

However, if someone is happy with his or her A, AA, AL or whatever crossover, or for that matter the type of amplifier or preamp or source component they use, I am uncomfortable telling them what a bad choice they have made. I have wanted to build a pair of ES networks for my klipschorns, but they are extremely expensive to build to your specs. I'm also not convinced I will like them better than this pair of type 'As' we're listening to right now. As I've said, I have made a few higher order networks in the past, and always seem to prefer more simple designs.

Erik

[Erik Mandaville]

I need to correct something: I have continually referred to the network we're now using as the type 'A'. In fact it's only part 'A.' The small inductor on the squaker is the change I made in order to make more effective use of Bob's new tweeter, which also worked fine with the stock 'A.' I just like what the inductor brings to the overall sound. The horn responsible for sqauwking seems happier on the job, too.

Erik

[Deang]

"The fact that the type AA network in your estimation measures poorly does not necessarily make it an inferior network."

Sure it does, and it also explains why 99 out 100 prefer the sound of the constant impedance designs. Most stay with the old networks simply because they've either never heard the alternatives to know the difference, or have convinced themselves that what they are using is pretty much as good as it gets, and making the change isn't likely to net any worth while gain -- a nonsensical position propagated by many around here who themselves haven't heard the alternatives.

As far as variations on the A types go, I said I enjoy mine very much -- but I only use them for low volume listening. They certianly wouldn't be my networks of choice if I could only have one set.

[BEC]

I like being in that top one percent.

Bob Crites