Hey all,
I want to design some crossovers for my DMLs. I am currently using a custom 4 channel amp with active crossovers (limited to HP/LP at 50-300Hz though) for panel testing and room measurements but I want to eventually build passive crossovers for 2 way DML systems, or 3 way FAST/WAW systems using DMLs for top end and an OB sub. I just finished the Udemy online course on crossover design and they have you use DATS to do an impedance graph and then make a ZMA file to load into XSIM with your far field measurements from REW, and then design your crossovers from there. However, Exciters don't exactly compute in DATS like a cone speaker because you only have Re, Le and Fs from the manufacturer. You don't have any of the Q measurements, Vas, etc because even when you mount the exciters to a substrate they still don't comply with any of those parameters, at least not in a conventional sense. How do you do about getting complete set of T/S specs to make a ZMA?
Am I overthinking this or something because I learned how to do it with conventional loudspeakers and the rules don't apply to exciters in the same ways (or at all)? I would rather come up with tailored crossovers rather than just throw some basic crossovers together at given frequencies based on my far field measurements and call it good. Or is that just the best way because they are Bipolar, open baffle and not really phase critical?
TIA
I want to design some crossovers for my DMLs. I am currently using a custom 4 channel amp with active crossovers (limited to HP/LP at 50-300Hz though) for panel testing and room measurements but I want to eventually build passive crossovers for 2 way DML systems, or 3 way FAST/WAW systems using DMLs for top end and an OB sub. I just finished the Udemy online course on crossover design and they have you use DATS to do an impedance graph and then make a ZMA file to load into XSIM with your far field measurements from REW, and then design your crossovers from there. However, Exciters don't exactly compute in DATS like a cone speaker because you only have Re, Le and Fs from the manufacturer. You don't have any of the Q measurements, Vas, etc because even when you mount the exciters to a substrate they still don't comply with any of those parameters, at least not in a conventional sense. How do you do about getting complete set of T/S specs to make a ZMA?
Am I overthinking this or something because I learned how to do it with conventional loudspeakers and the rules don't apply to exciters in the same ways (or at all)? I would rather come up with tailored crossovers rather than just throw some basic crossovers together at given frequencies based on my far field measurements and call it good. Or is that just the best way because they are Bipolar, open baffle and not really phase critical?
TIA
Last edited:
That course isn't going to give you what you need to cross a dml. I wouldn't worry too much about T/S parameters either.
What you need is to assess these from all angles to work out what they can do.
What you need is to assess these from all angles to work out what they can do.
Passive or DSP filters?
If DSP, I think you can focus on acoustical slopes. EQ your drives flat 2x (more is better) beyond the XO and then use the steepness you need. EQ to taste. Finished.
//
If DSP, I think you can focus on acoustical slopes. EQ your drives flat 2x (more is better) beyond the XO and then use the steepness you need. EQ to taste. Finished.
//
Amps are cheap but low freq crossovers are monsters. Any amp designed since 1950 will work fine.
B.
B.
And if you have a computer to spare, XO is also free...
CamillaDSP - Cross-platform IIR and FIR engine for crossovers, room correction etc.
//
CamillaDSP - Cross-platform IIR and FIR engine for crossovers, room correction etc.
//
T/S parameters are generally more about designing woofer enclosures, so not necessary for XSim to do its crossover job.
If you have clean impedance data and decent frequency response/phase data you've measured, that's all XSim needs. Frequency response is always baffle dependent if you're trying to do it accurately, so "textbook" data doesn't really get the job done. When you're talking about open baffle/DML stuff, even more so.
In the end, crossover design is about a lot of compromises and decisions. Many of those wind up being based on response at the intended listening distance and in-room response (especially if you're designing speakers just for yourself).
If you have clean impedance data and decent frequency response/phase data you've measured, that's all XSim needs. Frequency response is always baffle dependent if you're trying to do it accurately, so "textbook" data doesn't really get the job done. When you're talking about open baffle/DML stuff, even more so.
In the end, crossover design is about a lot of compromises and decisions. Many of those wind up being based on response at the intended listening distance and in-room response (especially if you're designing speakers just for yourself).
Thanks, Matt. This is very helpful, and in retrospect all makes perfect sense. I'm close to a 30 year stereophile but it's mostly been home theater and car audio. OB/DML stuff is a new interest for me and I have a tendency to think in old patterns of traditional enclosures and mechanics.