need some x-over advise before I build... (1 Viewer)

[Jon Hancock]

Ron,
This looks a *lot* better than your original graphs- one question for you- are you (or the program) taking into account the difference in time offset, assuming you're mouting the tweeter and midwoofers on the same baffle? It *looks* like possibly you are, since the woofer corner and roll off is softer and slower, which would introduce some phase shift to compensate for the time offset- and you're summing pretty nicely. Impedance curve looks fairly realistic, assuming a box tuning of about 32-33 Hz, and that that is a bit above the actual driver free air resonance (lower hump is taller...).
You can also do a check by reversing the phase on the tweeter, (in simulation) and seeing how deep a null you get. If you get a 24 dB or better null at the crossover frequency, you're about as close as you're likely to get with simulation without your own measured curves.
The only real question I'd have to raise is that I think you've implemented the best curve possible for these drivers, using this approach- but note that the 2 K peak on the midwoofer is only down about 8 dB. That may still produce some audible affects. I've found it's helpful (when you get to the hardware stage) to listen to the filtered driver output; if you hear any "crud" in the rolled off response, (with the tweeter disconnected), you're going to still wind up with some grunge and edge in the presense region, though with the tweeter playing you'll be harder put to identify the source (masking).
Take a look at this measured test box curve with a HiVi M8a- the 2.8 kHz peak of the driver is down 40 dB. When you listen to the driver and crossover by itself, it's rolled off, but it's also "clean"- no crud in the output from the driver cone modes.

Good work- you've definitely made some progress with the tools.
Best regards,
Jon

 

[Danny Richie]

Good feedback again from Jon.

I agree with him on all accounts.

A couple of other notes is that the crossover point is to the lower limit of that tweeter but I think the XT25 will handle it fine if not driven to hard.

One thing I have implemented a few times that has worked well in applications like this is instead of using the typical impedance equalization on the tweeter that uses a resistor in line with a cap. You can use a resistor in line with a larger value inductor.

This really won't effect the slope much until below the crossover point. It is there that the impedance will begin to rise on the tweeter and power will be further by-passed through the LR network.

It will steepen the tweeter slope below the crossover point much like that of a higher order network and still without putting more reactive components in series with the driver or even reactive within the pass-band it is playing.

You might can also play with the values in the woofer EQ circuit to see if you can minimize the peak at 2.1kHz.

This is when actual measuring tools come in real handy.

It is one thing to work from simulations. It is another to work from actual measurements.

If you do this often you might consider a Clio lite or something.

 

[Ron D Core]

So what value inductor would you recommend for that? Unfortunately my program won't let me put an inductor in the cap's place. I will end up wanting to pump 100-150 watts through them at some point, so that may be a handy thing to have. Regarding the equipment investment, I don't plan on going to crazy with all this to make it worth it. maybe if I was able to sell my designs, but I don't see that happening unless they want to pay alot for my hours of hard work. Man I can't wait to get going on this, I think I'll work on some renders for people to enjoy today Everyone loves alittle eye candy once in awhile.

 

[Patrick Sun]

Eye candy is nice, but the make-or-break is in the crossover design.

 

[Mark Hayenga]

Aren't there a few guys from Tucson on the bass list / Mad forum? Might be worth hooking up with some of those guys if they have measurement equipment - you get better data, they get to play with your drivers for an hour Patrick was able to find a guy with Laud in his area, you might try doing the same in AZ.
JH said:

The better caps sound better because they insert less noise.

How is the cap inserting any significant noise? I mean, I know there's stuff like Johnson noise in resistors and other passive components, but the induced voltage from that stuff is on the order of micro or nanovolts and only causes problems in active systems that I was aware of.

 

[Danny Richie]

Ron,
It would be hard for me to recommend an inductor value. Typically I design off measurements and would need to measure it to know for sure.
Mark,
Yes, The better caps sound better because they insert less noise.
If they measure they same in every way what do you think makes them sound different?
Take a look at Jack Bybee's site to read a little about noise and noise filtering. http://www.bybeetech.com/
If you want the technical cap talk call Jeff at Sonic Craft, http://www.soniccraft.com/ he'll talk your ear off about caps.

 

[Ron D Core]

Ok guys, thanks alot for the help guys, I came up with a 3rd order Butterworth that roles the 2.2kHz area down to about -30 dB. The phases seem to look better as well. I think I'm going to put it together and see how it works out. I'll look into a digital camera before I start anything of course