Hi-freq Active Crossover Needed

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Any thoughts on a good active x-over I can build to control a set of hi-freq drivers. It needs to have a variable x-over point as I don't yet know where I want to x-over to occur. I going to build a seperate set of hi-freq boxes so I can tri-amp my setup. I'm going to try a Phase Linear 400 for the hi frequencies to see how it sounds. If there's a good schematic I'd be interested in building it.
 
I would suggest a Linkwitz-Riley crossover. This is a 4th order crossover with good phase characteristics at the crossover region, which is important with high frequency since this contributes to perception of imaging. I'd suggest looking at Rod Elliott's Project 9. Also check out this reader's experience.

As for adjustability, you can experiment with different crossover frequencies by placing the resistors in DIP sockets. For a 4th order stereo 2-way crossover, you need to adjust 16 resistors simutaneously to adjust the crossover frequency. You could use 16-pin DIP sockets holding 8 resistors each, one set for each channel. Solder one DIP socket to the board, and insert the resistors into another socket (to create a resistor pack). Then you can easily swap out different resistor packs. Once you decide on one to keep, either just leave it there or unsolder the sockets and solder the resistors in place. You would do this with the resistors rather than the caps because they are cheaper and come in a broader range of values than caps.
 
*keep in mind that at high frequencies, driver spacing will significantly affect the phase across the crossover point.

that said, i've had good luck with the LR4 filter in the past.

here is a report i did in the past:

http://www.ee.siue.edu/~cstahl/e592b.pdf

it shows two variations on the sallen-key type filters. using a gain of 1 allows you to use equal value resistors, making it ideal for use with ganged potentiometers. using a gain of 2 allows you to use equal value capacitors.

IIRC for the Q of the LR2/BW2 blocks, you end up with caps of C and 2C, so that works out. i guess i'd go with the gain of 1 like on the other websites.
 
Thanks for the responses! I'm considering the Ashly XR 2001. This has balanced input/outputs. I have a Phase Linear 4000 preamp, two Phase Linear 400 Series II power amps, and two tube mono-blocks. Currently I'm using a Paradigm X-30 crossover for the subs. Will the Ashly crossover with the balanced inputs work with my OLD gear? The X-30 has RCA connectors and is compatible, the Ashly crossover uses 1/4" connectors.

Do you know of any 3-way crossovers that work with the old gear? Does this create a problem with the Rod Elliot Project? I'd like to buy a crossover if I can, otherwise I will build one.
 
Hi,
I just wanted to chime in.
Just because the electronic crossover is a Linkwitz-Reiley 24dB/oct doesn't mean that this will be the end result of your speaker.
See, speakers have their own passband- on either end of their limit they roll off- usually at 12dB/oct. If you dial in your electronic crossover close to the natural roll-off point of the driver, you might end up with 36dB/oct final slope. If you dial it in where the driver's response is very flat, you might get lucky and end up with response close to LR-4.
The problem is that midrange and tweeter might not overlap enough for the electronic crossover to be the dominant source of the crossover curves. Without EQ on each crossover leg you will likely end up with something less than you expect.
Always consider the drivers' natural response as it dictates the choice of the electric curves.
A good electronic crossover should let you control the slopes on either side of the crossover point independently. For example you should be able to dial in a 2000Hz tweeter High Pass and 1500Hz midrange Low Pass. With these points on a single pot it's very hard to find a good crossover point where the drivers sum properly.
Remember to play with the polarity. Should you end up with 6th order acoustic crossover your tweeter and midrange will be 180 degrees out of phase at the crossover point.

Good luck.

PS: you might want to download a demo of lspCAD and simulate in software what I just described using real world driver's measurements. It's really eye opening how textbook electric crossover curves don't really work with real world drivers.
 
Jan, thank you for all the information. Unfortunately I'm going to need a bigger brain to understand it!

For the low end I have a pair of ported subs I designed and built. for mid-range I have a pair of JBL L100s. I have yet to settle on the high freq drivers/cabinets (I will build them). Currently I use a Paradigm X-30 to crossover the subs at 80 hz. The subs are powered by a Phase Linear 400 and a pair of tube mono-blocks drive the JBLs. I want to split off the hi-freq before the mono-blocks and route it to a second Phase Linear 400 to power the hi-freq drivers. My choices are to keep the X-30 and run a seperate crossover for the mid/high split, or to come up with a three-way crossover and replace the the X-30.

What has me confused is the input connectors/signal. I believe this old gear with RCA connects won't work with the new crossovers that have balanced input/outputs. At this point I don't know what crossovers will be compatible with my old gear.

Don
 
If you are looking for a DIY solution to your active crossover requirement, have a look Here To properly de3sign your crossover you'll need to absorb Jan's post, but the group buy board gives you the freedom to do as he suggests.

I have a primer on active crossover design with some examples and a spreadsheet to save number crunching on my website (button below). Once you think you understand what I've written, visit www.linkwitzlabs.com to get the REAL in depth information.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.