Extracting phase.....

[lasse]

[Chris Roemer]

I believe it's the dotted line (which I think JB refers to as the "tail").

[dynamo]

I recall reading as well that the dotted line is an accurate prediction of the true roll off, much more so than the measurement that low, unless anechoic.

Might as well go to 20hz with the dotted line for your sims. I have found with some programs that if I don't go to 20hz that the resulting response graph just becomes a horizontal line over to 20 from the point I left off, which is no good either..

[jhollander]

Yes, use the dotted line, that will give you a more accurate representation of the phase.

[ErikSquires]

You do not have to go down to 20 Hz. You do however need much better measurements.

Everything below 500 Hz appears to be noise. You probably took the measurements using too low of a volume setting. Usually you want to get to at least the -30 dB level to say it's good enough. Your phase won't be right below your cut-off, but at -30 it will hardly have any effect on simulations.

To get absolute levels, check your gain to ensure a reference signal at 60Hz produces 2.83V. That's the equivalent of 1 Watt. Measure at 3' on axis In OmniMic we can select test tones with bass excluded to ensure we don't damage the tweets.

Best,


Erik

[Pete Schumacher]

You do not have to go down to 20 Hz. You do however need much better measurements.

Everything below 500 Hz appears to be noise. You probably took the measurements using too low of a volume setting. Usually you want to get to at least the -30 dB level to say it's good enough. Your phase won't be right below your cut-off, but at -30 it will hardly have any effect on simulations.

To get absolute levels, check your gain to ensure a reference signal at 60Hz produces 2.83V. That's the equivalent of 1 Watt. Measure at 3' on axis In OmniMic we can select test tones with bass excluded to ensure we don't damage the tweets.

Best,


Erik

Erik, the noise floor of a measurement room is a tough nut to overcome. In this case, doing what has been suggested, applying a tail to the measurement that continues to drop at a 2nd order rate is exactly what is needed. Measuring at a higher volume won't help with the noise floor. It will always be there. And including the noise floor is a sure way to mess up the absolute phase of the tweeter response.

[lasse]

Hi all, and thanks for the feedback.

Erik has a valid point in that I most definitely could improve my measuring skills but I feel that the weaker point for me comes next. To come up with a sound XO-topology for the system.

Thanks Pete for your kind remark that I probably have done what is possible to do in a small room in the basement

I include a picture, summing the measurements taken at 120 cm distance at a offset-angle of 10 degrees at tweeter height.

Your comments and suggestions would be most welcome, both sanity-check-wise and for choise of alignment.

Best regards//lasse

[jhollander]

I'm not sure I understand your note. The tweeter line looks fine, but what it the blue line? If that is your summation with another driver, that looks OK but there not much to go on as we can't see the summation of the drivers.

[Pete Schumacher]

I'm not sure I understand your note. The tweeter line looks fine, but what it the blue line? If that is your summation with another driver, that looks OK but there not much to go on as we can't see the summation of the drivers.

I'm guessing the blue is the woofer he's trying to combine it with.

[ErikSquires]

Um, my point was just, raise the volume and re-measure. You'll end up with the correct chart, but no, it won't go to 20 Hz. :-)

If this really is the best you can measure, do a nearfield measurement instead. Put the mic 1/4" away from the center of the dome.

On the other hand, you almost reached the -30 dB point... so you don't need that much better measurements.

[lasse]

Hi,

Yes, I forgot to tellyou.

The red Lines Are the tweeter response and phase, and the blue ones are the corresponding woofer info. The woofer SPL is combined far Field and nearfield, as mentioned by Erik.

Regards// lasse

[Pete Schumacher]

Um, my point was just, raise the volume and re-measure. You'll end up with the correct chart, but no, it won't go to 20 Hz. :-)

If this really is the best you can measure, do a nearfield measurement instead. Put the mic 1/4" away from the center of the dome.

On the other hand, you almost reached the -30 dB point... so you don't need that much better measurements.

Doing what you suggest with the near field measurement won't measure the effects of the baffle which is very important to get proper summation.

The 12dB/octave tail is the better option than even higher measurement levels that have the potential to damage the tweeter.

[dlr]

Doing what you suggest with the near field measurement won't measure the effects of the baffle which is very important to get proper summation.

The 12dB/octave tail is the better option than even higher measurement levels that have the potential to damage the tweeter.

+1

A tweeter is simply a closed box system that exhibits a 12db/octave slope asymptote below Fs on an infinite baffle and the same below the baffle step peak on a baffle.There's no need to measure below 300-500 Hz for a tweeter.

I have one comment on the woofer. If the response is summed using a near-field measurement, it needs to have the baffle diffraction included that appears to be missing in the response shown. The near-field response alone lacks that as it approximates an infinite baffle response. It should be modified using one of the diffraction tools, such as the BDS or the one by Jeff Bagby prior to designing the system crossover.

dlr

[lasse]

+1

A tweeter is simply a closed box system that exhibits a 12db/octave slope asymptote below Fs on an infinite baffle and the same below the baffle step peak on a baffle.There's no need to measure below 300-500 Hz for a tweeter.

I have one comment on the woofer. If the response is summed using a near-field measurement, it needs to have the baffle diffraction included that appears to be missing in the response shown. The near-field response alone lacks that as it approximates an infinite baffle response. It should be modified using one of the diffraction tools, such as the BDS or the one by Jeff Bagby prior to designing the system crossover.

dlr

Hi,

Forgive me again,

The speakers, in this case, have a somewhat odd placement. See the attached Picture. I thought that the need for bafflestep compensation was minor in this situation.

Am I right?

Best regards//lasse

[dynamo]

If that is the speaker on top of the cabinet, I think that you will still have some diffraction effect as the wave rounds the baffle and steps back to the wall, but I don't think you need any BSC for that application. But I also am no expert..