BDS Questions

[weinstro]

Hi,

I have some BDS questions for the collective wisdom out there.

1. What does a good diffraction response curve look like? Something to shoot for as a target?

2. What does a bad one look like?

3. Is there some insight into fiddling with the baffle width and driver position to complement a given driver(s) response? Some thoughts on this would be appreciated....

4. How would I model a rounded baffle in one of the freeware tools? By rounded I mean that the entire baffle is a large radius, like a column or Sonotube.

Thank you,

Rob

[dlr]

Re: BDS Questions

Hi,

I have some BDS questions for the collective wisdom out there.

1. What does a good diffraction response curve look like? Something to shoot for as a target?

2. What does a bad one look like?

3. Is there some insight into fiddling with the baffle width and driver position to complement a given driver(s) response? Some thoughts on this would be appreciated....

4. How would I model a rounded baffle in one of the freeware tools? By rounded I mean that the entire baffle is a large radius, like a column or Sonotube.

Thank you,

Rob

The only question with a simple answer is for question one. An ideal response would have a step response with almost no "wiggles" in the response above the step. This would, however, include all off-axis responses simultaneously, not realistic for any but the rarest of systems.

A bad one? The more appropriate question is what is the polar pattern? That is, what is the response on- and off-axis? Most systems, even those with diffraction control to some degree, will have "bad" response in some off-axis areas, the exception being systems designed for controlled directivity that minimize off-axis variations. The question then morphs to what is the importance of a listening axis response vs. power response (sum of all axial responses). This gets worse when one considers that every speaker system will tend to sound different with room influences included, power response and short time-delay room reflections.

The only thing I can offer is a bit of what I used in a design I posted last year that only begins to touch on the topic. The writeup does touch on one area where I made specific use of baffle diffraction response in an attempt to complement a specific tweeter aberration.

The Chameleons

The system did get largely positive responses when auditioned at DIY NE last year, especially with vocals.

dlr

[weinstro]

Re: BDS Questions

Thank you for responding.

The only question with a simple answer is for question one. An ideal response would have a step response with almost no "wiggles" in the response above the step. This would, however, include all off-axis responses simultaneously, not realistic for any but the rarest of systems. dlr

I do try to look up the simple answers myself. I appreciate the target, though.

A bad one? The more appropriate question is what is the polar pattern? That is, what is the response on- and off-axis? Most systems, even those with diffraction control to some degree, will have "bad" response in some off-axis areas, the exception being systems designed for controlled directivity that minimize off-axis variations. The question then morphs to what is the importance of a listening axis response vs. power response (sum of all axial responses). This gets worse when one considers that every speaker system will tend to sound different with room influences included, power response and short time-delay room reflections.dlr

Hmmmm. If I understand correctly, now we're talking about the combined effect of diffraction response ripples and the driver(s) responses. And a power response, at that. This seems a bit complicated. Are there inexpensive tools like the FRD to help with this?

I don't have driver test data at every angle -- only what the factory provides. Unless, you're also suggesting that this is all determined empirically....

The only thing I can offer is a bit of what I used in a design I posted last year that only begins to touch on the topic. The writeup does touch on one area where I made specific use of baffle diffraction response in an attempt to complement a specific tweeter aberration.

The Chameleons
dlr

A subject search here led me pretty quickly to your website, and this design in particular. I also modelled your baffle on the BDS and came up with similar curves, thinking that yours would be a lot closer to "good" than "bad". I'm still giving myself some soak time with your write-up. I really like the article, as most folk don't spend a whole lot of time explaining their design choices and rationale.

I've also modeled ZaphAudio's L18 and L15 (before he retired it) designs on BDS as well. I've noticed that in Zaph's latest designs, he simply centers the drivers on the vertical axis.

BTW, how did you determine were to place the felt block on the Chameleons?

Regards,

Rob

[dlr]

Re: BDS Questions

Hmmmm. If I understand correctly, now we're talking about the combined effect of diffraction response ripples and the driver(s) responses. And a power response, at that. This seems a bit complicated. Are there inexpensive tools like the FRD to help with this?

The BDS is what I use for part of it. You can investigate pretty much any axis for diffraction. It does included a rough calculation of the driver's directionality as well, so it gives a reasonable approximation. You can also use "The Edge" and there are other tools, including Jeff Bagby's, but I prefer the BDS just for personal reasons as a long time user.

In case you haven't found them, look here.

don't have driver test data at every angle -- only what the factory provides. Unless, you're also suggesting that this is all determined empirically....

It can be modeled to a reasonable extent. Measurements are best, but even then most software tools can only use a limited number of actual measurements. The part most important for the power response influence of diffraction is the step area. The crossover also has an influence and that part can be modeled reasonably well. Again, Jeff's software does that as well. That's not part of a package such as the BDS, of course.

A subject search here led me pretty quickly to your website, and this design in particular. I also modelled your baffle on the BDS and came up with similar curves, thinking that yours would be a lot closer to "good" than "bad". I'm still giving myself some soak time with your write-up. I really like the article, as most folk don't spend a whole lot of time explaining their design choices and rationale.

I had thought for some time that the whole thought process would be good to present, so when offered the drivers for doing a design I decided that it they would be good candidates, since both midwoofer and tweeter have some issues that complicate the design process.

I've also modeled ZaphAudio's L18 and L15 (before he retired it) designs on BDS as well. I've noticed that in Zaph's latest designs, he simply centers the drivers on the vertical axis.

Some argue, with reason, that in-line isn't so bad. A small off-axis change can greatly reduce the direct response variations, not unusual for in-room positioning. That can also make the design a bit easier because it's often the case that an effort is made to reduce the diffraction dips/peaks on-axis that will likely negatively impact the power response. Outside of that, it reduces to opinion on the relative importance of the first (direct) response vs. in-room (power) response. It's not clear cut.

BTW, how did you determine were to place the felt block on the Chameleons?

Testing. The single bar between the drivers is largely a no-brainer, but the impact of the change must be measured to know the absolute improvement it makes. For typical dome tweeters there's not a single system that won't have a measurable improvement with a piece between drivers, sometimes surprisingly significant. The small pieces I used to make a triangle (I've yet to post pictures of that, I should add updates) were configured entirely empirically, but based on prior testing experience. In this case fairly small pieces were almost as effective as much larger ones, probably due to the small roundover built into the PE baffle and the baffle placement I chose for the tweeter.

dlr

[weinstro]

Re: BDS Questions

The BDS is what I use for part of it. You can investigate pretty much any axis for diffraction.

I use this one, as well. BTW, how does one get the Driver Group Sum Response chart to appear?

Testing. The single bar between the drivers is largely a no-brainer, but the impact of the change must be measured to know the absolute improvement it makes. For typical dome tweeters there's not a single system that won't have a measurable improvement with a piece between drivers, sometimes surprisingly significant. The small pieces I used to make a triangle (I've yet to post pictures of that, I should add updates) were configured entirely empirically, but based on prior testing experience. In this case fairly small pieces were almost as effective as much larger ones, probably due to the small roundover built into the PE baffle and the baffle placement I chose for the tweeter.

I also read through your article on natural felt blocks vs. synthetic felt rings.

Are you of the opinion that the shape of the ring has more of a detrimental contribution to diffraction than the type of material?

Thanks,

Rob

[dlr]

Re: BDS Questions

I use this one, as well. BTW, how does one get the Driver Group Sum Response chart to appear?

I forgot to check this at home this weekend. With Office 2007 here I cannot check, so I'll have to do that from home.

I also read through your article on natural felt blocks vs. synthetic felt rings.

Are you of the opinion that the shape of the ring has more of a detrimental contribution to diffraction than the type of material?

No, I would say (without absolute certainty) that the material is the most important. Shape has an impact, but for example I've not found any purely synthetic materials common available that work even modestly well, if at all. I've tried acoustic foams, but most are not dense enough to work for diffraction adequately, at least for dynamic drivers on a flat baffle.

dlr

[weinstro]

Before I cut some wood....

How does the attached BDS screen shot look?

This is my first solo attempt. I'd like to get some experienced feedback, as you can't really cut wood back together.

I've attached a GIF, but to fit the filesize limits this is probably so small as to be close to useless. So, if you have the old BDS tool (by P. Verdone), I've also attached the profile file. Just be sure to change the extension back to ".dlp". The profile is at 0 degree axis, but I've run the sim for 0 and 15 degrees.

Thanks in advance for the advice!

Regards,

Rob

[dlr]

Re: Before I cut some wood....

How does the attached BDS screen shot look?

This is my first solo attempt. I'd like to get some experienced feedback, as you can't really cut wood back together.

I've attached a GIF, but to fit the filesize limits this is probably so small as to be close to useless. So, if you have the old BDS tool (by P. Verdone), I've also attached the profile file. Just be sure to change the extension back to ".dlp". The profile is at 0 degree axis, but I've run the sim for 0 and 15 degrees.

Thanks in advance for the advice!

Regards,

Rob

I haven't had time to load it and check it out, but the graphs you posted look pretty good. My goal is usually to limit the magnitude of peaks/dips above the step area and pay a bit less attention to the step itself because you can't avoid having that in any case.

dlr