MakeMineVinyl
Well-Known Member
In my quest to endlessly tweak, and remembering how good my Altec A-7-500s sounded when I had a 3rd-order electronic crossover many, many years ago, I decided to rip open my crossover and re-think the whole thing.
I decided to really go out on a limb and go ultra minimalist - a single pole, 6dB per octave crossover between the woofer (15") and the HF horn. Building something like this is simplicity in the extreme; just a single resistor and capacitor in front of the LF and HF power amplifiers for low pass and high pass respectively. The entire network can be passive, so its an electronic crossover without the active electronics (although I do use a single buffer IC before the power amp inputs to give a more predictable load impedance for the RC filters).
1st order crossovers have a big advantage in phase response - they're the only type which can successfully reproduce a square wave. All other types mangle a square wave into something almost unrecognizable. Do we actually listen to square waves? Heck yes; a clarinet's waveform is essentially a square wave (in fact square waves are used as the basis of synthesizing a clarinet voice on an old-skool analog synthesizer). Theoretically, the advantage should be audible as more accurate reproduction of transients etc.
1st order crossovers do have one dis-advantage though; a lot more low frequency information is passed to the tweeter, so one must be careful. In my case, since I'm running a maximum of 5 watts into the tweeter, and since anything over a half of a watt never happens at even extreme listening levels because of the efficiency of my horns, I'm pretty safe. However, I did take this opportunity to change the basic crossover frequency to a minimum of 1.2kHz, which was the optimum frequency arrived at by Altec Lansing for the A-7 series while I worked there (this was when they were still located in Anaheim, across the street from Disneyland).
The transformation was pretty unreal, in a very good way. Everything seemed to open up, and the soundstage widened to well beyond the physical edges of the speaker cabinets (even with program material which was not processed to give head-related transfer function cues for false surround type effects). Additionally, the speakers became much more "invisible" in that the sound never seems to be actually be coming from the speakers themselves with well recorded stereo material - that's a huge feat with speakers the size of a refrigerator.
After much experimentation, I arrived at a very unconventional location for the crossover points, 1.2kHz for the woofer and 4kHz for the HF horn. In theory, this would produce a huge suckout, but the HF horn has excess energy in the range below about 2kHz which was very effectively smoothed out by shifting its crossover point much higher. The magenta curve for the HF horn below shows the response with the crossover point raised. The green curve for the woofer is its response with no filtering at all for reference.
Here are the response curves; the magenta curve is the HF horn only, the green curve is the woofer only, and the yellow curve is the final listening position response. I have to use a link since for some reason the forum cannot directly post pictures from a Google drive account.
https://drive.google.com/file/d/0B7DJENUU7T6RMDFleW4xazlXT1E/view?usp=sharing
For those with an experimental streak, this is a very simple and potentially beneficial project. Just be very careful and keep the listening levels low and monitor the amount of energy getting to the tweeter to avoid overloading it until it can be determined that proceeding is safe. Raising the crossover frequency above stock is a way to mitigate this to a certain extent, but you still need to be careful.
Single pole crossovers can, and have been implemented in passive crossovers in manufactured speakers, including some high end ones made today, but here I'm talking about low level passive filtering before the power amps.
I decided to really go out on a limb and go ultra minimalist - a single pole, 6dB per octave crossover between the woofer (15") and the HF horn. Building something like this is simplicity in the extreme; just a single resistor and capacitor in front of the LF and HF power amplifiers for low pass and high pass respectively. The entire network can be passive, so its an electronic crossover without the active electronics (although I do use a single buffer IC before the power amp inputs to give a more predictable load impedance for the RC filters).
1st order crossovers have a big advantage in phase response - they're the only type which can successfully reproduce a square wave. All other types mangle a square wave into something almost unrecognizable. Do we actually listen to square waves? Heck yes; a clarinet's waveform is essentially a square wave (in fact square waves are used as the basis of synthesizing a clarinet voice on an old-skool analog synthesizer). Theoretically, the advantage should be audible as more accurate reproduction of transients etc.
1st order crossovers do have one dis-advantage though; a lot more low frequency information is passed to the tweeter, so one must be careful. In my case, since I'm running a maximum of 5 watts into the tweeter, and since anything over a half of a watt never happens at even extreme listening levels because of the efficiency of my horns, I'm pretty safe. However, I did take this opportunity to change the basic crossover frequency to a minimum of 1.2kHz, which was the optimum frequency arrived at by Altec Lansing for the A-7 series while I worked there (this was when they were still located in Anaheim, across the street from Disneyland).
The transformation was pretty unreal, in a very good way. Everything seemed to open up, and the soundstage widened to well beyond the physical edges of the speaker cabinets (even with program material which was not processed to give head-related transfer function cues for false surround type effects). Additionally, the speakers became much more "invisible" in that the sound never seems to be actually be coming from the speakers themselves with well recorded stereo material - that's a huge feat with speakers the size of a refrigerator.
After much experimentation, I arrived at a very unconventional location for the crossover points, 1.2kHz for the woofer and 4kHz for the HF horn. In theory, this would produce a huge suckout, but the HF horn has excess energy in the range below about 2kHz which was very effectively smoothed out by shifting its crossover point much higher. The magenta curve for the HF horn below shows the response with the crossover point raised. The green curve for the woofer is its response with no filtering at all for reference.
Here are the response curves; the magenta curve is the HF horn only, the green curve is the woofer only, and the yellow curve is the final listening position response. I have to use a link since for some reason the forum cannot directly post pictures from a Google drive account.
https://drive.google.com/file/d/0B7DJENUU7T6RMDFleW4xazlXT1E/view?usp=sharing
For those with an experimental streak, this is a very simple and potentially beneficial project. Just be very careful and keep the listening levels low and monitor the amount of energy getting to the tweeter to avoid overloading it until it can be determined that proceeding is safe. Raising the crossover frequency above stock is a way to mitigate this to a certain extent, but you still need to be careful.
Single pole crossovers can, and have been implemented in passive crossovers in manufactured speakers, including some high end ones made today, but here I'm talking about low level passive filtering before the power amps.
