My Current Project...(Teasers...)

[rpb]

I'm puzzled. My sim ends up at 90dB. Johns is at 85dB. In PCD, the response of Johns filter looks much different below 1k. The transfer function shows 1k to be down 12dB. It looks like he was correcting for baffle step, but that is already in the measurement. Or maybe I'm doing something wrong here.

[mattp]

I would still likely place an LC across the woofer and see if you can't get to -50dB on the breakup. As low HD and resolving these should be, I would remove all doubt and kill that blemish. This also might allow you to shift the xover higher, as I think Xmax could be your issue, not HD. I would try some higher output with xover and see, but 1500+ w/o waveguide is probably recommended. I have a set of TL, but not that model.
Wolf

Thanks Ben. What I put in bold were exactly my thoughts for this design. The design I came up with was very similar to his first which was a surprise to me, the main difference being the values for the woofer filter since he accounted for baffle step. He sent me a 4th order with notch and a XO around 3k that I think I will try next.

Otherwise I will always have that mild panic thought in the back of my head that these will just shatter one day.

And a cap of about 1uF across your padding resistor will lift the top.
Wolf

This is very useful info. I might try it to see if I hear any real difference, My ears are only good to about 14k but it's cheap enough to try. And according to my sim you nailed that 1uF value BTW. I have this sneaking suspicion you have done this once or twice...

I'm puzzled. My sim ends up at 90dB. Johns is at 85dB. In PCD, the response of Johns filter looks much different below 1k. The transfer function shows 1k to be down 12dB. It looks like he was correcting for baffle step, but that is already in the measurement. Or maybe I'm doing something wrong here.

I know he accounted for baffle step but he did use his own corrected/blended response using a NF measurement I sent him for everything below probably 600-700 Hz. I did not correct the file I posted earlier in the thread so the phase was probably not accurate. Hopefully John will chime in and shed some light on his reasoning for this. When it comes to proper XO design I don't know what I don't know - so I am all about learning from you veterans.

[rpb]

Here's a steep 2k filter that hammers the 5k spike.

[jhollander]

The original files did not have a baffle step blended with the nearfield. Attached are the latest files

[rpb]

The original files did not have a baffle step blended with the nearfield. Attached are the latest files

Wouldn't the on axis, in box measurement already capture baffle step?

[rpb]

Here's a 4th order version.

[rpb]

The original files did not have a baffle step blended with the nearfield. Attached are the latest files

The woofer frd is wrong. It's a tweeter frd. The tweeter frd is also a tweeter frd, but with a steeper roll-off. I didn't look at the others.

[jhollander]

Nope. Maybe you are not use to seeing woofer with baffle step?

[jhollander]

The original files had blended response with gating to 1K

[rpb]

Nope. Maybe you are not use to seeing woofer with baffle step?

Look again. Here's your posted woofer, and tweeter frd files.

[Wolf]

I initially questioned the 4.5mH coil on the woofer as well. That is a pretty large value for a 5" midbass, even when comping BSC with a full 6dB. Mind you, for this type of design, I'd shoot for a Super-Q laminate or a large gauge air-core. I've not played with the files, but I might be...
Later,
Wolf

[rpb]

Thanks Ben. What I put in bold were exactly my thoughts for this design. The design I came up with was very similar to his first which was a surprise to me, the main difference being the values for the woofer filter since he accounted for baffle step. He sent me a 4th order with notch and a XO around 3k that I think I will try next.

Otherwise I will always have that mild panic thought in the back of my head that these will just shatter one day.



This is very useful info. I might try it to see if I hear any real difference, My ears are only good to about 14k but it's cheap enough to try. And according to my sim you nailed that 1uF value BTW. I have this sneaking suspicion you have done this once or twice...



I know he accounted for baffle step but he did use his own corrected/blended response using a NF measurement I sent him for everything below probably 600-700 Hz. I did not correct the file I posted earlier in the thread so the phase was probably not accurate. Hopefully John will chime in and shed some light on his reasoning for this. When it comes to proper XO design I don't know what I don't know - so I am all about learning from you veterans.

Do you have gated frd files from this measurement, that have not been altered, spliced, corrected, etc.?

[Wolf]

I'm hoping the files Matt posted initially were both on the tweeter axis. That's how I roll...
Wolf

[rpb]

My post was just deleted as spam, so I'll ask again if Matt has gated frd files for this measurement that have not been corrected, tweaked, modified, in any way.

[rpb]

In addition, I'd replace the foam blocks under the speaker with a speaker stand that sits on some blocks. That foam will probably affect the baffle step slightly. That's how I roll!