My New Crown XTi 1000s Are Here!!

[Erik Mandaville]

You can also just add a switch to toggle back and forth between the resistors -- you could include odd values, too.

Switches just make things a little easier. I think I would be inclined to use a parallel element across the driver after the series resistor, but the open-mindedness concerning this kind of attenuation is, for me, admittedly refreshing. The autoformer will be an interesting experiment, and you can install a switch for that as well.

Erik

[Deang]

Yeah, I'm back in the you mostly get what you pay for camp. Oh well.

Kudret -

"Would 12 ga litz be ok for the large ones?

If you can afford them, go for it. I really don't think you need air cores of any type in the LF section of the network, pretty much a waste of money, and I quit doing it myself years ago. If you're going to use air cores in the LF section, just use standard air cores. As for the HF section, Litz may or may not offer an audible improvement over standard air cores, I highly doubt it. You can actually use 14 AWG for all of them except one.

HF Section:

(4) L142.4 2.4 mH .56 22x45x89
(2) L141.8 1.8 mH .46 22x45x89

For the 6mH, the schematic calls out for less than .6 DCR. 12 AWG doesn't quite make it, you would have to go 10 AWG. Personally, I don't think .07 difference is that big of a deal. The coil I used is .9

(2) L126.2 6.2 mH .67 38x76x152
(2) L106.2 6.2 mH .47 38x76x152

LF section:

(2) S18.10 .10 mH .12 10x19x38
(2) S142.4 2.4 mH .36 22x45x89
(2) S141.8 1.8 mH .30 22x45x89

Your room is going to have a much greater impact on the FR than any small variations in the values you use. Bigger coils just mean bigger boards and more trouble figuring out where you're going to put them.

"You can also just add a switch to toggle back and forth between the resistors..."

Not a bad idea actually, especially since I only favor two values.

"I think I would be inclined to use a parallel element across the driver after the series resistor..."

You mean another resistor, make it an L-pad?

From my notes -- I have no idea where it came from.

"...try a single series resistor per tweeter of one Ohm up to two Ohms. When you need additional lowering you better use two resistors and assemble a dividing network. A single resistor of one ohm reduces the tweeter level about 2dB. A single resistor with a high value doesn`t add much additional attenuation but changes the crossover point of the tweeter. Adding resistors in series and parallel (L-pad) WILL alter the roll off characteristic in a significant way on top of attenuating dBs in a crossover network. In other words, you can add in one resistor in series with the driver which attenuates some dBs, or you can add one resistor in series and another one in parallel (L-Pad) to attenuate some dBs, but the frequency response of each of them is very different from the other...forget about the parallel resistor or L-Pad. In my network the parallel resistor causes the roll off to happen much earlier even though attenuation at high frequences remains the same, while a series resistor keeps the response curve similar..."

An L-pad works better if you flatten the impedance with zobel(s) or swamp the whole network back to something stable. It's pretty tough to calculate out for an L-pad based on x ohms when the in-network driver impedance is swinging between 6 and 20 ohms. Low value series resistors have an advantage here -- they attenuate while having less influence on the overall impedance, crossover point and FR.

[Kudret]

Thank you, Dean. That's been very helpful.

Kudret

[Erik Mandaville]

Dean:

I don't know why I was thinking that there was a dedicated tweeter as part of the equation.

It seems that a single series resistor would probably be best. You could use something like a single pole, six position switch for this. The crossover in the HF branch is of course before the attenuator (resistor or group of resistors), and would be connected to the single pole (contact in the center of most rotary type switches), with each output postion of the switch connected to one end of each of the resistors. I was looking at the board you made (again, very clean work), and there is space for a few more, with the same kind of layout. In other words, a total of even 6 resistors, maybe. So, with the crossover connected to the pole on the switch, and one end of each resistor connected to one of the positions on the switch, the other ends of those resistors can be joined at a single common point. From that common point, run a single lead to the positive connection for the driver. Even though the resistors share this common node, there will only be a single resistor in the circuit at a given time, so you can just reach back to the board and switch from one to the other easily. You could use, say a group consisting of 1.5ohms, 2ohms, 2.5ohms, 3ohms, 3.5ohms, and 4ohms, and that way get just slightly more flexibility in terms of what might sound the best. There probably won't be a big difference in the half-ohm steps, but maybe enough for slightly finer adjustment. The resistors can be practically touching to save space, and oriented just as the existing two are right now. 25watt jobs! Or put them anywhere on the board that works.

A piece of angle aluminum can be used to mount the switch, with two holes drilled where it connects to the board for greater stability than just a single hole. The other part is of course drilled for the shaft of the switch. If you did this, a shorting type switch might be a good idea. For the ends of the resistors, use a couple of 3 position solder terminal strips or just a barrier strip, and run a lead from each resistor connection point over to the corresponding position on the switch. I've been able to get 18AWG solid wire through the little holes on the switch positions, but they can't take too much more than that. 18AWG should be fine. 20AWG would work. You don't even have to crimp the daylights (aka crap) out of the connection on the switch, just push through the hole and quick touch with solder.

Erik

[DrWho]

No offense to Dean, but I've never understood how one could say "hey, nice networks" based on merely pictures alone. [6]

[Marvel]

No offense to Dean, but I've never understood how one could say "hey, nice networks" based on merely pictures alone.

I think that would be "NICE!" as in a real work of art, beautiful construction.

[Erik Mandaville]

Mike:

Dunno, I guess it's sorta like looking at a picture of a cool car, a cool motorcycle, or a bowl crammed full of vanilla ice cream and topped with a soft brownie and hot fudge.

I appreciate quality craftsmanship even if I may not happen to have a personal preference for the actual function of the product. Al K's ES crossovers border on sculpture and are wonderfully well-made, but they are more than likely NOT the best match for some of the very low power amplifiers I have. Dean and I don't always see eye-to-eye on issues (yeah, I know that's probably an understatement if there ever was one), but there is no doubt in my mind about the quality of the workmanship evident in those pictures. I will never hesitate to give credit where it's deserved. To the best of my knowledge, they were also designed by a very capable engineer; and that fact in conjunction with high quality craftsmanship is what helps me say, without reservation, "hey, nice networks" -- the only thing lacking in that quote being the exclamation point that belongs at the end of the statement. [pi]

...oops, I think I just hit the pizza icon. Kinda funny, though, so I think I'll just leave it!

Erik

[mike stehr]

Hey Dean! Nice Networks!

[Duke Spinner]

Top Notch ... !