That looks OK provided your data is good, and measuring a null at the crossover point on the listening axis with one driver reversed will suggest this. The quality of your mic is not that important for this type of measurement.
You've also changed the crossover point which will make small changes to the character of the speaker beyond what this kind of data is capable of showing, although this is not a reason not to continue as you are. What it does mean is that you should be wary of drawing conclusions based on what you might hear.
You've also changed the crossover point which will make small changes to the character of the speaker beyond what this kind of data is capable of showing, although this is not a reason not to continue as you are. What it does mean is that you should be wary of drawing conclusions based on what you might hear.
Chris, that last picture of the crossover is asking for a notch at 3/4K you might fix it not that important only for a better transition, also the tweeter might be too bright.
These high spots are all on axis, here is the woofer and tweeter FR plots, the peaks drop off listening off axis.
Attachments
You can get better phase tracking than that, but yes that crossover would work more normally. You can check the alignment another way by clicking the "reverse polarity/normal polarity" button on the tweeter section. Then move the listening/measurement height up and down until you get the deepest null. Note the angle (next to the height you are putting in) - this is your forward lobe angle, where you would usually be listening around. If it's no good, get the phase aligned on a different axis.
Are you still doing it with zero z-offset, though? You're almost sure to need around -0.015 to -0.03 or more on the woofer, and this will affect your design quite a bit. Also, it looks like you need to figure out where the reference level is for the overall response. Some of that rolloff below 1kHz is from the test setup, but some of it is probably real rising response that you'll need to deal with along with your baffle step. I guess you can investigate when you get your mic. I'm guessing you'll want to be around 85-86dB for the system sensitivity.
Are you still doing it with zero z-offset, though? You're almost sure to need around -0.015 to -0.03 or more on the woofer, and this will affect your design quite a bit. Also, it looks like you need to figure out where the reference level is for the overall response. Some of that rolloff below 1kHz is from the test setup, but some of it is probably real rising response that you'll need to deal with along with your baffle step. I guess you can investigate when you get your mic. I'm guessing you'll want to be around 85-86dB for the system sensitivity.
Just to butt in here and add my 2 cents worth of thinking here. When we listen to our sound systems we do not generally move in the vertical plane. We are sitting in a chair or on a couch and if anything we may move side to side but not often up and down. So I would look at your design with the thinking of how high the devices are going to be relative to your actual sitting ear height. If you are looking for a perfect solution while looking at the vertical axis and moving up and down you will find no workable solution baring a point source speaker, a coaxial or full range speaker is the only solution to that situation. If that is one of your criteria all multi-device solutions will fail including mtm and every other perturbation of design. Just my thoughts on this factor in the design of the crossover and speaker implementation.
If you look at actual polar measurements of carefully-designed small 2-way speakers compared to coaxials and full-rangers, I think you'll see that isn't as nearly as true as you think. Perfection doesn't happen in either case. Regardless, with a 2-way, including coaxials, it's up to the crossover design how it works out, thus the need to work with phase as in the topic of this thread.
dumptruck,
Don't get me wrong, I completely agree that phase response is very important, actually more important than most would regard it. But I am only talking about in the vertical plane, not the horizontal plane. I think that time/phase response in a loudspeaker combination is much more important than just at the crossover point. I just make it a point to work on the global response of the loudspeaker and not that specific point only. You can not look at this factor in a bubble, you need to look at all group delay aspects to get the sound right, just getting things correct at the crossover point does not make for a great sounding loudspeaker.
Don't get me wrong, I completely agree that phase response is very important, actually more important than most would regard it. But I am only talking about in the vertical plane, not the horizontal plane. I think that time/phase response in a loudspeaker combination is much more important than just at the crossover point. I just make it a point to work on the global response of the loudspeaker and not that specific point only. You can not look at this factor in a bubble, you need to look at all group delay aspects to get the sound right, just getting things correct at the crossover point does not make for a great sounding loudspeaker.
Sure, sure, and I mentioned some other larger picture aspects a minute ago. As far as this thread, the OP correctly identified an aspect (phase tracking and time offset) that they did not know how to work with, so might as well consider it in a bubble for a minute, get a handle on what affects what, and then go back to the juggling act that is designing a complete system.
The woofer plot you show here reveals cone breakup where different parts of the cone get out of synch with each other, shown where the axial plots shoot off in different directions.
Looks like everyone is telling you the same thing, bart-dood, but you're not listening. 😀
Uploaded with ImageShack.us
Read my lips. The lively paper/reed Peerless 830875 has some cone-breakup issues around 4-6kHz that also mess up rolloff and phase badly. It needs treating to match it up to a tweeter on an overall 4th order acoustic. 😎
http://www.diyaudio.com/forums/multi-way/228143-help-veeper-tweeker-2.html
I haven't particularly examined your choice of tweeter, but have given you a start point here. This is better than the usual steep LC bass circuit with unreasonably low crossover point.
http://www.diyaudio.com/forums/multi-way/228143-help-veeper-tweeker.html#post3334347
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
Read my lips. The lively paper/reed Peerless 830875 has some cone-breakup issues around 4-6kHz that also mess up rolloff and phase badly. It needs treating to match it up to a tweeter on an overall 4th order acoustic. 😎
http://www.diyaudio.com/forums/multi-way/228143-help-veeper-tweeker-2.html
I haven't particularly examined your choice of tweeter, but have given you a start point here. This is better than the usual steep LC bass circuit with unreasonably low crossover point.
http://www.diyaudio.com/forums/multi-way/228143-help-veeper-tweeker.html#post3334347
- Status
- Not open for further replies.