mboxler

I use this for capacitors http://www.pronine.ca/capimp.htm and this for inductors http://www.pronine.ca/indimp.htm Just enter the "resistor" value in the impedance field and it will calculate the -3db frequency.

If you are like me, you understand what capacitors are doing in a crossover, but may not know how they are doing it. I attached two videos that helped me. The first describes the capacitor itself. It helped me understand how an open circuit (the capacitor) can allow current to flow. The second places a capacitor in an AC circuit. It describes inductor as well, but that's another topic. If you replace the resistor with a speaker driver, or taps 0 - 5 in a heritage crossover, you will see why the current flows through them 90 degrees before the voltage across the capacitor changes. You will also see why more current flows at higher frequencies. I realize this is 20 minutes of your life that you will never get back, but I hope you find them educational. Mike https://www.youtube.com/watch?v=f_MZNsEqyQw https://www.youtube.com/watch?v=zO7RZZW0wSQ&list=PLkyBCj4JhHt-4PnnwpbG-ZKV_EjX03DX8&index=15

Yes. Actually, I believe Bob Crites measured a K-55 driver in a K400 horn and came up with closer to 14.5 ohms, but you get the point. If you mean the frequency at which the impedance of the capacitor is equal to the impedance of taps 0 - 5, yes. Using 14.5 ohms across taps 0 -4 as an example, the load across taps 0 - 5 would be double that, or 29 ohms. A 13 uf capacitor reaches 29 ohms around 422 hz. A 14.5 ohm resistor in parallel with the 14.5 ohm driver gets you a 7.25 ohm load. I realize the driver is not a pure resistive load, but... By moving these to taps 0 - 3, the reflected load across taps 0 - 5 is 4 times that load, or 29 ohms, right back to stock. No apology needed. You had me questioning myself. I'm not saying that the series resistor won't work, it just works differently. In my example, adding a 14.5 ohm resistor in series with 29 ohm impedance of taps 0 - 5 results in a 43.5 ohm load on the 13 uf cap. As mentioned before, this load is in parallel with the k-77 circuit, so that will change some as well. Again, if it sounds good, it is good. Mike

By across the squawker, we mean in parallel anywhere between taps 0 - 3 and the K-55. Placing the resistor across the + and - terminal block screws going to the K-55 is a convenient spot. A parallel resistor in and of itself will not change the voltage across the squawker. What it will do is draw more current from the amplifier. If the resistor is equal to the squawker impedance, the current drawn will double. Since the extra current passes "through" the 13uf capacitor, it charges twice as fast. This, in turn, doubles the corner frequency from 400 hz to 800 hz. We haven't reduced the voltage to the squawker, it just takes an octave longer to reach maximum voltage. By dropping the output taps to 0 - 3, we cut the current across taps 0 - 5 in half. Now the current through the 13uf capacitor is equal to the stock setup, and it's corner frequency is back to 400 hz. The maximum voltage across the squawker, however, is down 3 db at all frequencies. I assume a 14 - 16 ohm resistor was used??? L-pads work well because they drop the voltage across the squawker without changing the current drawn from the amplifier. The 13 uf cap charges at the same rate. Now let's place the so-called swamping resistor (usually 10 ohms) across taps 0 - 5. It has no effect on the voltage across taps 0 - 5, but it will draw a significant amount of current, and that current stays pretty much the same no matter what output taps are used. A 13 uf capacitor will charge way too fast. It needs to be replaced with, say, a 48 uf cap that will take longer to charge. I don't understand option two above. It would drop the voltage throughout the circuit, but it would also lower the corner frequency of the 13 uf capacitor because less current is "flowing" through it. At least I think that's right. No matter what, if the change sounds better to you, it is better! Mike

That's the easiest way to do it. This is an easy/inexpensive way to lower the voltage across the driver AND keep the current "through" the 13uf capacitor close to a stock AA crossover.

Did you end up soldering a 15 - 16 ohm resistor across the squawker and move the squawker to taps 0 - 3?

Although the video describes a transformer instead of an autotransformer, it will still work. If your primary inductance is 105.3 mH, then set your secondary inductance to 10.53 mH. The 10 db turns ratio is 3.165. 3.165 squared is 10. Since this a step down transformer, you divide the primary inductance (105.3) by 10 to get the secondary inductance (10.53). I haven't tried this, but I think I got the math right. If you want to model an autoformer, the inductor values would be 49.2 mH and 10.53 mH. Mike

I would be happy to send you my version of a 12 tap auto transformer. It seems to work pretty well. It's currently part of a Type A simulated circuit. I assume I can attach an .asc file in a private message. Mike

Looks good to me! You have now created two separate filter circuits. They will become one parallel circuit when the binding posts are strapped. They can also be paralleled when "strapped" at the amp...bi-wiring. They can also be run as separate circuits, each connected to its own amp...bi-amping. Mike

I like the colors! I believe you should move the common wire that connects to tap 0 from input L - to input H -. The autoformer is part of the high pass filter. Mike

This won't be too difficult... Relabel the input terminals as input low. Use the two empty terminal barrier strips as input high + and -. Move the wire going to the 13uf cap from the current input + to the new input high +. Cut the common wire between the output woofer - and the output mid -. If you cut it close to the woofer out - screw you may have enough to shift the remaining wire to the left. In the end, one common wire from input low - to woofer out -, one common wire from input high - to output tweeter - and output mid -. Now, connect your top binding posts to input high, and the bottom binding posts to input low. For single wire from your amp, you will need to strap your + binding posts together and your - binding posts together. For biamp/biwire, remove the straps. Actually, you made your speaker a little more versatile! I hope I visualized that right. Mike

Which part of Colorado?

Crap! I've seen that thread but never scrolled down far enough to see Mr. Warren's schematic of a T2A. Since the inductances of the various taps are listed on the T2A datasheet, he made it easier by replacing one four tap autoformer with four one tap autoformers. I, on the other hand, have been struggling with the one four tap math. The inductance of the T2A taps 0 and 5 must be 45.6 mH...4 times the 11.4 mH inductance between taps 0 and 3.

Coupling series inductors using the K command is the only way. While having a few beers with my brothers, I got an idea. I disconnected my 3619-ET from ALK crossover and measured the inductance from 0 - 5, 76 mH. Again I tried to measure the other taps, but the reading would always change each time I tried. I decided to calculate the inductances based on the turns ratios involved. I used 76 mH as my baseline inductance. I think it works. Maybe you can also try this and see if it's close. It's sure a lot cleaner. The response curve is similar, but the phase curve changed. The DCR values for inductors L1 thru L5 are .108 .076 .054 .038 and .093 ohms respectfully. Around .37 ohms total. Mike

Here's the response curve of the 20 X 10 mH inductor (mutually coupled) T2A, using tap 4 (14 inductors). Green is voltage across taps 0 and 5, red across 0 and 4. Not sure how close it is the the voltage across a real K-77.

This may be hard to see. I'd be happy to send you the .asc file if you want. Mike

I tried, but I seemed to get weird results on my DE-5000. Seems like the numbers would fluctuate each time I measured. More importantly, It seems that the measurement between taps 0 and 3 would not always equal 1/4 the value between taps 0 and 5. The issue, however, is that the output voltage of an autoformer is based on the winding ratio, not the inductance ratio. Inductance is based on several variables, and windings is only one of them. LTSpice let's you specify the inductance and DCR only. All other variables being equal, a 40 mH inductor should have twice the windings as a 10 mH inductor. My T2A LTSpice autoformer equals twenty 10 mH inductors in series and coupled. Tap 4 equals 14 of the inductors. Tap 3 equals 10 of the inductors. Tap 2 equals 7 of the inductors. Taps 1 equals 5 of the inductors. It comes out pretty close. My reasoning was that each of the 10 mH inductors would have the same number of windings. I can try to post my Type AA if you would like. Mike

Looking Good! I see the autoformers got a little roughed up. They were fine when I sent them. I should have mentioned that when I wrapped the common wire on the barrier strips, I intended them to be on the crossover side. It will work as you have it, but it will be easier to connect/disconnect the cables from the amp and to the drivers. If the amp is still on, as you connect/disconnect a + wire it will be too easy to touch it to the common wire below it. Mike

It wasn't pretty, but I got some numbers out of LTspice. I modified my test board and came up with 12.769uf and 2.077uf values. When I plugged these cap values into LTspice, I came up with voltages that were very close to what I measured with my Fluke meter. Back to LTspice... Again, 2.83 volts 400 hz, 14.2 ohm K-55, and 8 ohm K-77. Capacitors 13uf and 2uf. Classic Type A K-55 1.276 volts K-77 .074 volts 2uf cap moved to amplifier input K-55 1.38 volts K-77 .114 volts (3.6 db increase) 13uf and 2uf in series with K-77, no connection to autoformer K-77 .0985 volts My autoformer model is a work in progress. I'm curious what voltages others end up with. Mike

My apologies @John Warren . I've never used LTspice, and didn't understand the schematic. I've been using iCircuit and XSim, but can't simulate transformers/autotransformers. This link is interesting. Inductors are used along the a K-statement to mutually couple the inductors. Both transformers and auto transformers are discussed. I'm playing with LTspice now. Slow learner. https://www.analog.com/en/technical-articles/using-transformers-in-ltspice-switcher-cadiii.html

Looks like that autoformer simulation creates an electrical band pass across the 13 ohm resistor. Was that your intent? An accurate simulation would result in the voltage across the K-55 rising to 0db along with the K-77. I did find this, but I'm not smart enough to determine if it works. https://www.quora.com/What-is-the-equivalent-circuit-for-auto-transformers-impedances

Unless I'm doing something wrong, this comes very close to what I'm seeing if you assume the K-55 is a 16 ohm load. 13uf cap impedance + ((double the K-55 impedance // (2uf cap impedance + K-77 impedance))

Well, here's my results using a 2.83 volt 400 hz signal. My setup includes a 26.38uf mid cap, a 11.76uf tweeter cap, a 14.2 ohm resistor connected to tap 4 (tap 0 connected to ground), and an 8 ohm resistor. First, the two caps in series (or 8.136 ohms) results in .46 volts across the 8 ohm resistor. Nothing else connected during this test. Next, the classic Type A configuration with the tweeter cap connected to tap 5. .424 volts across the 8 ohm resistor 1.296 volts across the 14.6 ohm resistor Last, the new Type A configuration, with the tweeter cap connected to the amplifier input. .651 volts across the 8 ohm resistor 1.785 volts across the 14.6 ohm resistor The math required to reach the Classic Type A results ended up being for a parallel connection. It bothers me the the voltages the 8 ohm resistor are so close between the first and second tests, so I plan on running another one tomorrow at 800 hz. If anyone is interested in the math involved, I'd be happy to post it tomorrow as well. It's kinda nerdy. Right now my brain is fried. Mike .

Will do

Like I said in another thread, I'm not sure how the autoformer changes the circuit, but I would also think that the current flowing across taps 0 and 5 would have an affect. Before I send @Robbie010 the autoformers, I think I'll test this out. I don't have the 13uf/2uf caps to test with, but any two caps should do. I'll rig up the circuit with a 26uf cap to the autoformer and a 12uf cap to the tweeter. I have 14 ohm and 8 ohm resistors to use as loads. I'll pass, say, a 400 hz 2.83 volt signal and measure the voltages across each resistor. I'll then move the input to the 12uf cap and retest. If the caps are truly in series, the voltages across the resistors should not change, right? I've been wanting to test this for a while. Maybe it's about time. Mike