mboxler

Or is it... As I understand it, Klipsch developed the AK-3 to correct a phase issue with the AK-2. I assume the 5mh value was chosen so that it, in parallel with the inductance across taps 0-3 of a T4A, corrected the K-55M portion of that issue. My 3636's measure around 19mh across taps 0-3. This, in parallel with a 5mh inductor equates to almost 4mh across the K-55M. I don't know the inductance across taps 0-3 on the T4A. If it's 10mh, then the parallel inductance would be 3.3mh. If the inductance across taps 0-3 on the T4A is around 10mh, then perhaps I should use a 4mh inductor across taps 0-3 of the 3636 to get to the same 3.3mh parallel inductance?

I'm building new AK-3's to replace my original AK-2 crossovers. Willing to spend up to $70 shipped. If none available in a week or two I'll just go with the Crite 3636's I already have. Thanks, Mike

I don't know what to say. My 80uf cap, at 100hz, measures 81.27uf and .017 DF. 81.27 at 100hz = 19.57 ohms. 19.57 * .017 = .33269 ohm ESR. At 1000hz, measures 78.80uf and .037 DF. 78.80 at 1000hz = 2.02 ohms. 2.02 * .037 = .07474 ohm ESR. This is on a DE-500 LCR meter. May try to find manufacturer's datasheet.

2-5 ohms? At what frequency? My 80uf electrolytic measures .08 ohm ESR @ 1khz. A capacitor with 2 - 5 ohms ESR would definitely change the frequency response. Mike

You and me both. Throw in an autoformer and things get really interesting! It's all about resistors, capacitors, and inductors. Once you understand each component's impedance and phase at a given frequency, then it's all about complex math and ohm's law. I'd go into much more detail, but I'm afraid I'd be going way off the rails. Perhaps one should start a thread on the basics of passive filters? For example. 2.5mh inductor and a 6 ohm driver. 382hz 2.83 volt signal. How can the voltage drop across both components be 2 volts? There may be others that find the answer interesting. Then again, maybe not. Mike

Not in Dean's league, but... The parallel capacitor will always create a 2nd order electrical filter. The size of the capacitor, however, dictates the frequency at which the slope starts to drop 12db per octave. For a visual, I've attached the two different capacitors across the K33, simulated by a 6 ohm resistor. Again, this is with the 2.5mh series inductor. The top green plot is an 80uf capacitor, the middle a 40uf capacitor, and the bottom no capacitor. As you can see, an 80uf capacitor creates an underdamped filter...the voltage across the K33 actually rises around 1.6db before falling. Acoustically things will be different, but I thought this might help clarify how parallel capacitors affect a filter. Mike

My theory, it's all about real estate. A 4mh inductor takes up maybe 20% more space compared to a 2mh. It seems as though Klipsch likes 50uf capacitors max, which are almost as large as the inductors. Without stacking you are doubling the capacitor space occupied when you double the capacitance. Again...only a theory.

If one assumes the impedance of the K55 is around 15 ohms, then I would say the AK-4 filter was designed for a 30 ohm load, and the voltage drop across the K55 would be around 6db. If one were to design the same filter with a 15 ohm L-Pad, then you simply double all the capacitors and halve all the inductors to get a similar frequency curve. I'd be happy to simulate the above, just to see if I'm close. Oh...I never said adding a series resistor would not decrease the current through the circuit. I fully realize that, with a properly designed L-Pad, all other component values can remain unchanged. Edit...should have said 6db Mike

My wife says I'm a good mediator 🙂 Here are the high pass circuits of the AK-4 Mike

I believe the AL-2 and AL-3 are identical, except a poly switch was added to the AL-3. Are you referring to the AL-4 perhaps? Mike

It is interesting that Klipsch decided to not add a notch filter to the woofer circuit of the AL-4 as they did to the AK-4.

😎

It's actually two 50uf capacitors paralleled for a total of 100uf 🙂

I couldn't find an LTspice simulation of the AK-4, so here's what I came up with. The AK-4 high pass to the tweeter in green and the AA in red. If I did it correctly, seems as though EQ is reversed between the two in that frequency range(?). Mike

I have so many components sitting around, I just might. Curious how it would compare to the ALK Universal kit I built a long time ago.

A little off topic, but... I have designed, but not built or tested, a crossover loosely based on Al's Universal, so we'll call it the Universal 4500. The new components not only change the electrical cross between the mid driver and tweeter to 4500hz, but also allow you to change the attenuations of either/both of them. This was accomplished by the combination of the familiar "swamping" resistor, and the moving/modification of the mid driver low pass components to the output side of the autoformer. Since this is just a third order to the tweeter, I'm not sure it's steep enough for a K77 at that lower crossover point. No idea how it would sound, but would be an interesting experiment! Mike

Interesting! I guess I was misremembering. Thanks, Mike

Thanks for the clarification. I always thought that the K-77M could run lower with a steeper filter (voltage wise), but it's SPL was too low below 6000hz to sound right. Or I could be misremembering 🙄

I use REW. I don't have an AL-X crossover, but the high pass of the ALK ES5800 is almost identical. Here is a plot of the voltage across a real K-77M. Still can't figure out why REW gets cranky at the lower SPL.

I believe the AL-4 is 4500hz, due to the K-77F(?).

Pretty sure the high pass filter to the K-77 was identical on the AL, AL-2, and AL-3 crossovers...around 6000Hz. Mike

Here are my LTspice and REW plots for the high pass of the ALK ES5800t crossover. Can someone tell me what's going on with REW below the 70db level? Is there a measure setting I'm getting wrong, or is this noise on the TRS connection from the crossover to the Line In jack? Thanks, Mike

Please don't get me wrong. If you are satisfied with your results then that's all that matters. I assume you meant 44.8 total inductance (2.8 * 16) 🙂. Your tests appear to confirm my math. For example, tap 4 represents 75% of the total windings, therefore the output voltage will be 75% of the input voltage, or -2.5db. I hope I read your plot correctly. Likewise, tap 2 represents 37.5% of the total windings, resulting in a -8.52db drop in voltage. I guess I'm too much of a perfectionist, and was trying to get your model closer to -3db and -9db. I was also reminding all that on the T2A, tap 3 is actually -3.35db. I often wonder, regarding a new Crites Type AA crossover, that the combination of the 3636 autoformer on the -3db tap and the lower ESR Sonicaps might be the reason it seems brighter than a T2A/polyester cap combination. Mike

Sorry I'm not much help. If I understand you correctly, you are saying 1/16 of the total autoformer inductance equates to 3db (?). I'm saying that it's purely the ratio of input turns to output turns that dictates voltage change. In your example (4 2.8mh inductors connected in series and mutually coupled), each inductor would have the same number of turns. Therefore, if you connect the input to the top of the first and the bottom of the fourth, and connect the output between the first/second inductor and the bottom of the fourth inductor, your total output turns will be 75% of the input turns. As a result, your output voltage will also be 75% of your input voltage, or -2.5db. The next tap, 50%, or -6db, and the last tap, 25%, or -12db. If, in your model, you replace the 13uf capacitor with a wire, you should get the same input to output voltage changes as I just described (I think). Mike