Review by James J. Jewell
"So there I was, sitting in the comfort of my living room, amongst the detritus of major home renovation (Fig. 1). I was flipping through a hard-print DIY audio magazine (I will let you guess which one) trying to ignore the adds in the margins for products I can neither afford nor rationalize, when a little something tugged at my subconscious. My eyes went over to the left margin where I was confronted with the question "How do you improve the World's Best speakers". That's a pretty bold statement on many levels! Upon further inspection, it became apparent that this was an ad for some kind of a crossover upgrade for Klipsch horn products. Fortunately it came with a web link, which I was connecting to faster than you can say "crappy 56K dial-up account".
The big (and I mean BIG!!) fully horn-loaded Klipsch products are the Klipschorn, the LaScalla, and the Belle Klipsch. For those of you unfamiliar with the Klipsch Heritage horn line-up or horns in general, they possess a quality critical to low-powered tube amps and desirable to others: High sensitivity. They run at approximately 104dB/W/m at one watt, measured at one meter. Assuming an 'average' speaker sensitivity of 87dBm/W/m, the Klipsch horns are about five times more sensitive, or efficient, than many common speakers. This means that I can fill my living room with 'loud' music at a fraction of a watt, and still have nearly 30 watts of clean "Class A" hot-rodded Dynaco power in reserve for transients and percussion. The downside of this is that any noise in your system is going to stick out like a fart in church. The LaScalla is a three-way horn system with the mid-range "skwaker" and the tweeter being true horns and the bass being a 15" cone speaker fitted to a horn-loaded cabinet. The skwaker is actually more efficient than the other two drivers, and typically has to be padded down a bit to form a smooth three-way musical mix. Apparently the horn 'sound' is not universally enjoyed. Metal ones tend to ring unless properly damped, and they sound edgy and aggressive to some listeners. On the plus side, they respond to dynamics and subtleties in the music far better than most other drivers, due to the tiny low-mass diaphragm that's driven by the amp. I find they produce a very 'live' or 'front-row' sound which I like.
So, back to the web page... I went to www.alkeng.com and discovered a clear and unassuming site with a variety of paragraphs on filter design. A few more clicks, and I discovered some basic information and some photos of the Klipsch crossover and its variants. Still, I did not see any cost information so I continued to scroll down. At the very bottom of the page I came across what amounts to every DIYerselfers dream: The full schematics zipped for free download!!! Hot Damn! This kept getting better by the minute. After a few minutes of flipping through the documentation, those famous words started to float through my head. You know what I'm talking about. Lets say them all together now: "Aw, I can do this". Ah, like a crack junkie trying to kick the habit, I rationalized that I could take on just one more project. I mean what harm could one more itty-bitty little project cause? Needless to say, I did not consult my girlfriend prior to this endeavor. You see, she has this crazy notion that I should spend my time working on the whole-house renovation that I started a few years ago so that we can get married and live in a dream house together. I say "four walls and a roof, baby!" "Aw, I can do this" got me to buy this *^#% fixer-upper house in the first place. "Aw, I can do this" convinced me to try to fix that mechanical Chelsea ship's clock that sits in a box in my basement in 600-some odd pieces. On the positive side, "Aw, I can do this" got me to build a fantastic sounding pre-amplifier (a future article?), "Aw, I can do this" got me to join the United States military and survive Officer Training School and "Aw, I can do this" got me to build this fantastic crossover. Hopefully "Aw, I can do this" will get you to do more than just listen to your Hi-Fi system.
Al's balancing network (he does not call it a cross-over too often) improves on the original in a few areas. First, there is just the component quality upgrade. By using audiophile-grade components, one could expect an improvement with even the original design. The original is loaded with iron-core coils and motor caps (typically used in air-conditioners and can sound OK in audio, although one usually sees them in high-tension B+ supplies rather than in signal path coupling position). But Al's big contribution is designing a network that provides a nearly constant impedance to the amplifier. This is particularly important with tube-amps! But what really gilds my lily is that his design allows the user to adjust the level of attenuation that the skwaker sees without affecting the impedance matching. You can choose from approximately 5, 6, 8, 11 and 14 dB of attenuation. Like your female singers to sit in your lap? Keep the midrange a little hot. Mid-range 'horn-sound' grind on your nerve's a little? Pad that sucker down a little more! And if you buy Al's product, or wire your own properly, you can adjust the padding without any tools!
So after looking over the well-written plans and schematics, I decided to build this bad-boy on the installment plan, i.e. piecemeal. I had some HDF lying around the house from a kitchen cabinet project as well as a bunch of terminal strips. The now-defunct Parts Connection had some great sales on, well, on everything, so I ordered the Solen inductors. While Al specifies Litz-wire in the smaller two coils and solid 14-guage for the woofer inductor, I bought all three in Litz because of the sale. The bases from the HDF were made slightly smaller dimensions than the plans call for. Not for any particular reason other than I apparently can't read a schematic properly. Because there was originally just enough spare HDF to make the two bases, and did not want to have to buy another 2'x4' piece to correct for my mistake, I set out to come up with a new and more compact component layout. My prime goal was to minimize magnetic coupling between any of the four inductive components (three coils and an autotransformer). Al arranges the three coils orthogonaly to minimize coupling, which I also did, however I managed to place mine a little farther apart by keeping them close to opposite corners. By specifying air-cored torroids, the ALK crossover should have few problems with magnetic interaction, since the fields of torroids tend to stay close to the physical outline of the coil. The other (and probably the main) reason for using these coils is the low DC resistance and the lack of a ferric core to saturate. There is, of course, a size penalty to be paid, but c'mon, these speakers are freakin' huge to begin with, and there is plenty of room inside to shove all of these components. The penalty paid for moving the magnetic components farther apart, and not following Al's component layout is that I had to use longer wiring. However there was no way of quantifying any sonic penalty this may have. I suspect that the inches of wire added are so small relative to the length of wire in all of the coils and the transformer that any negative contribution would be imperceptible.
A few more liberties were taken with the design. I had lots of terminal strips lying around the house, so a descision was made to use those to mount capacitors side-by-side. I had plenty of tightly twisted pairs of 18-gauge wire from another project in the basement, and decided to use a three-way binding post at the input to the x-over. I stuck to his component list, with the exception of two capacitors. While staying with the recommended values, I could not bring myself to pay $20+ for the Hoviland Musicaps. Therefore a substitute of a film-and-foil cap that I believed to offer equal performance at about a quarter the price was in order. Aside from these minor tweaks, Al's design was followed to the letter.
So back to the chronology... The next component I ordered was the transformer. This is a custom jobbie by the good folks a Universal Transformer Corporation in Wylie Texas. No, this is not the same UTC that made all of those famous tube-era pieces of iron that us tube rollers try to get our hands on, but rather a small hand-built operation making solid pieces of equipment. The price was right, at about $25, and you'll need one per channel. The folks there are friendly and my only complaint is that they don't have e-mail, or accept credit cards, and as a red-blooded American I expect nothing less than instant gratification. So why am I willing to pay $25 for an x-former but not for a cap? I guess my subconscious does some kind of pound-per-dollar calculation, which finds a roll of film-and-foil lacking, but can rationalize a hunk of iron and copper. On top of that, the transformer is, in my mind, the key element that makes Al's design so special.
Once the transformer and capacitors arrived, I mounted them in a fashion that I thought made optimal use of the remaining real estate on the base plate. With that came the point-to-point wiring. I've taken to calling the crossover 'Medusa' or "the spaghetti incident", and the reason for which should be obvious after looking at the photos. While my layout may have been optimized from a component size-and-shape standpoint, it certainly was not from a wiring point of view. I had these long twisted pairs of white wire running all over the place (Fig. 2). Al's layout looks much more elegant.
The stock Klipsch design uses a barrier strip as a central hub for routing all of the various wires to and from the actual x-over components as well as the three horns. This allows for a good deal of flexibility for bi-wiring (snake oil if you ask me, but you probably didn't) and bi-amplifying (scientifically legitimate), so it was wise for Al to keep it in his design. However it is also supposed to be the location where you actually hook up the length of speaker wire from your amps, and those barrier strips are pretty darn narrow, and don't allow you use a very nice termination on your speaker cable. In fact, you are basically limited to using a 0.25" or smaller spade terminal, of the kind that you can get for a few pennies a piece at your local RatShack. This makes it almost impossible to hook up large gauge speaker cable, since finding a spade terminal large enough to crimp onto 10-gauge wire and yet small enough to fit onto that pesky barrier strip is a futile search based upon mutually exclusive criteria. So I opted to take a pair of color coded, gold-plated three-way binding posts and mount them to a special block added to the base. A small length of wire went from them to the terminal strip, using the el-cheapo spade that was berated a moment ago. Now I had a x-over with a proper gold binding post.
Al recommended a particular attenuation setting for my model of the LaScalla. To change this setting all one needs to do is switch settings on the autotransformer (Fig. 3), and change the value of a power resistor that parallels the autotransformer. I bought enough resistor values to allow for a variety of settings. I attached female connectors onto the ends of the resistors (Fig. 4) and onto the wires that run up to the autotransformer. This allows an easy-to-change non-solder connection.
After putting in the protective fuse, I took the units and slid them into the back of cabinets, on the opposite side of the original x-over. It was time to perform the smoke test... Not only was there no smoke, there was an immediate improvement in clarity even without any component break-in.