Frequency Response - Is this good ??

[mark.carline]

I think I am nearly there with the simulation in Speaker Workshop of a crossover and three drive units (tweeter, midrange and woofer) - All JM labs....

This is the final frequency response I get:

Is it any good or could I do better ???

Red - Woofer Output
Blue - Midrange Output
Pink - Tweeter Output
Black - All three above combined (but midrange out of phase)

[breguetphile]

If those are actual measured results of the final speaker thats a fairly impressive response. If its based on simulation you may want to use it as a starting point and then create the appropriate crossover and optimize the enclosure.

Good luck!

[PHilgeman]

Well, the range is 160dB so it is somewhat hard to tell, but it looks like there is a 8 or so dB lump around 200Hz, that will definitely be audible.

-Paul Hilgeman

[mark.carline]

well its just theory at the moment but it does include all the FRD and ZMA graphs from the actual driver units.....

Heres a close up.....

Whats a decent responce and how do I tell that I've got there ???

Could the box design do away with this hump - Its based on a sealed design with a 999Ltr (just for theory at the moment).

[5th element]

The graphs below show the speakers that ive just built, this is the Real frequency response measured after I built them. Ignor the stuff below about 1000hz its all meaningless as I used a very short window just to check the xover was working correctly.

What you want to do is end up with a reasonably flat frequency response say +-3dB. Then when you reverse the polarity of the midrange you get two big suckouts at the crossover frequencies to the bass and tweeter. Note that you may need to reverse the polarity of one or more of the drive units to get them to sum flat anyway. Just when you reverse with respect to each other you get the nulls.

For the measured response below the tweeter was out of phase electrically with the woofer for the first response. But wired electrically in phase to get the big null.

Either way you should strive for flat, then nulls in reverse.

The best way to do the speaker is to buy the drive units and then measure the frequency response and impedance of the units in the boxes. Then design the crossover around that.

One thing loudspeaker workshop doesnt simulate I dont think is the Z axis which IMO is much more important then the X or Y axis positioning of the drive units relative to one another.

Can you make targets with loudspeaker workshop?? ie you want to get 4th order linwitz slopes on all the drive units. So you make a target highpass at 3000hz and super impose it over the tweeters actual response then alter the crossover components to match the target.

LSwshop is not very intuitive to work with i know but I think the option to create target slopes is their sumwhere. This would make crossover design much easier.

Yes that peak in the bass looks not very nice.

[officeboy]

Where are you crossing over your woofer? that 100-200hz bump looks pretty bad. Maybe you can lower the crossover point to 50hz to help cut some of that hump away, or use a notch filter?

Also. you need to pick a box size first, then worry about your crossover. If you really are modeling with a 999L box, then well thats just not gonna work. Pick a box that gives you the lower end that you want, then massage the crossover network to smooth things out.

Also, recommended reading.
http://www.speakerbuilder.net/web_fi.../phat/hpif.htm

[5th element]

What focal units are these? if you want to send me the FRD and ZMA files for them all I'll be happy to give em a whirl in LSPcad and see what I can come up with.

Yes you need to get good bass response too from the woofer put it in a good box.

[mark.carline]

ok thanks!

If anyone wants me to email then the Speaker Workshop Files, ZMA and FRD files along with the driver unit specs then send me an email:

mark.carline@btinternet.com

thanks!!

[mark.carline]

p.s. I think the 200mhz bump is coming from the 10W6411 driver unit, see the response from that driver:

[Saurav]

One thing loudspeaker workshop doesnt simulate I dont think is the Z axis which IMO is much more important then the X or Y axis positioning of the drive units relative to one another.

It has a 'driver offset' parameter, which I assumed would be the Z axis, because I thought that's what was usually meant when one said 'driver offset'?

Can you make targets with loudspeaker workshop??

Yes, you can. You can also get SW to optimize your network to conform to the target over a specified frequency range. I've found this quite helpful when trying to get say a 3rd or 4th order acoustic slope by combining a 2nd order electrical network with the driver's natural rolloff. And like you said, you can superimpose different curves onto the same graph to be able to see what's happening more easily.

For a free piece of software, I think Speaker Workshop is pretty amazing. I've been messing with the modelling side of things. The next step is to use SW measure the drivers on the baffle and use those FRD/ZMA files instead of the ones I made from the printed specs.