Open Baffle musings

[fdieck]

Re: Open Baffle musings

http://techtalk.parts-express.com/sh...53#post1870053

A list of proof concept low $ drivers

[charlielaub]

Re: Open Baffle musings

Plugged the drivers into ABC Dipole. Looks like the woofer needs some help

Thanks for posting this earlier on in this thread. You are right, the original driver that I was thinking of using was not going to get anywhere near 100Hz!

Today I finally got around to modeling my MWOB system drivers in ABC Dipole. I forgot how useful this software is (I last used it in 2008 I think!). I finally "re-learned" how to use it and here are screen shots:

For the midwoofer (Exodus Anarchy):


For the tweeter (Tang Band W2-800SL):


With the midwoofer crossed over around 100 Hz and the tweeter around 1k Hz I can get 105-110 dB SPL (Xmax limited). The power required for that level of output is pretty high because of the boost needed for the drivers. With moderate power levels of 100W into 8 ohms, I can get over 100 dB.

Also note that the dispersion pattern is quite washed out for the midwoofer, and is not anything like the equivalent circular baffle. When you move off axis, however, the basic behavior of a dipole is followed. Even at 75 degrees off axis, the response does not show a null in the passband between 100Hz and 1k Hz. This is as close to constant directivity as I am going to get, and it's the goal that I was chasing after, so I am happy to see this result. Of course it's only a model...

For the tweeter, the model also seems to confirm my suspicion that the dipole peak is not that peak that I am seeing in my measurement (see post # 100) around 2k-3 kHz. I only see a little of the null around 8k Hz, so the driver's own directivity must be starting to kick in at this point (it has a 1.5" diameter cone, so that is possible).

-Charlie

[fdieck]

Re: Open Baffle musings

Thanks for posting this earlier on in this thread. You are right, the original driver that I was thinking of using was not going to get anywhere near 100Hz!

Today I finally got around to modeling my MWOB system drivers in ABC Dipole. I forgot how useful this software is (I last used it in 2008 I think!). I finally "re-learned" how to use it and here are screen shots:

For the midwoofer (Exodus Anarchy):
[ATTACH=CONFIG]30980[/ATTACH]

For the tweeter (Tang Band W2-800SL):
[ATTACH=CONFIG]30981[/ATTACH]

With the midwoofer crossed over around 100 Hz and the tweeter around 1k Hz I can get 105-110 dB SPL (Xmax limited). The power required for that level of output is pretty high because of the boost needed for the drivers. With moderate power levels of 100W into 8 ohms, I can get over 100 dB.

Also note that the dispersion pattern is quite washed out for the midwoofer, and is not anything like the equivalent circular baffle. When you move off axis, however, the basic behavior of a dipole is followed. Even at 75 degrees off axis, the response does not show a null in the passband between 100Hz and 1k Hz. This is as close to constant directivity as I am going to get, and it's the goal that I was chasing after, so I am happy to see this result. Of course it's only a model...

For the tweeter, the model also seems to confirm my suspicion that the dipole peak is not that peak that I am seeing in my measurement (see post # 100) around 2k-3 kHz. I only see a little of the null around 8k Hz, so the driver's own directivity must be starting to kick in at this point (it has a 1.5" diameter cone, so that is possible).

-Charlie

I pluged into Edge and I think it is a dipole peak.

http://www.tolvan.com/index.php?page=/edge/edge.php

PM me and I will send my file the softwhere is silly easy to use.

[charlielaub]

Re: Open Baffle musings

I pluged into Edge and I think it is a dipole peak.

http://www.tolvan.com/index.php?page=/edge/edge.php

PM me and I will send my file the softwhere is silly easy to use.

No. I am using the driver "nude" so there is only the frame to provide a "baffle". Try modeling this with the Edge:
baffle = 50x50mm (this is the frame OD of the driver)
speaker size = 40mm (round)

You will see the first dipole peak is at approximately 5k Hz.

-Charlie

[jhollander]

Re: Open Baffle musings

I can’t help myself…a narrower baffle is really going to help the mid to tweeter x-o, sorry.

The bump you’re seeing could be reflections. Could be wrap around contribution at your x-o pint. If you’re using PCD flip the mid or tweeter polarity and look for a matching bump. Could be baffle related.

When I get something unexpected I get out my painters scaffold and measure outside…individual drivers on baffle… and then do it again… and measure a few off axis… some with no x-o…derive my z offsets, well you get the idea.

[fdieck]

Re: Open Baffle musings

No. I am using the driver "nude" so there is only the frame to provide a "baffle". Try modeling this with the Edge:
baffle = 50x50mm (this is the frame OD of the driver)
speaker size = 40mm (round)

You will see the first dipole peak is at approximately 5k Hz.


-Charlie

We are talking about your driver siting on the edge of a 2 foot baffle are we not? I want to make sure I am on the same page as you are what you are measuring.....

[charlielaub]

Re: Open Baffle musings

I can’t help myself…a narrower baffle is really going to help the mid to tweeter x-o, sorry.

All that is going to do is depress the low end response of the midwoofer. It won't change the tweeter response. I am not going to make it so narrow that I can operate below the midwoofer's first dipole peak... Maybe you can expand on "help the mid to tweeter x-o"??? I'm just not sure what you mean, specifically.

The bump you’re seeing could be reflections. Could be wrap around contribution at your x-o pint. If you’re using PCD flip the mid or tweeter polarity and look for a matching bump. Could be baffle related.

When I get something unexpected I get out my painters scaffold and measure outside…individual drivers on baffle… and then do it again… and measure a few off axis… some with no x-o…derive my z offsets, well you get the idea.

I think I will move the tweeter away from the baffle so that it is in "free space" and remeasure to see if the "bumps" are still there. First I have to figure out just how to do that... suspend on wires maybe...

I'm surprised that you use a scaffold outdoors for measurements. There is absolutely no reason to do that unless you need to measure the frequency response below 200 Hz or so. Above that you can easily collect an impulse response in room, gate off the portion that contains the reflections from the walls, and then invert it into a frequency response.

Also, there is no need for "operation scaffold" to get driver offsets. Set up the mic. Do the gated impulse measurement for each driver. Then measure with both drivers. The interference between the driver output makes unambiguous determination of the acoustic offset pretty straightforward, using a software that can do it. PCD can (which I mostly don't use), as can my ACD software (which I use).

-Charlie

[jhollander]

Re: Open Baffle musings

For your mid shown in post #107 the first dipole null is 976 Hz , not sure what the measured FR looks like there but assume we would want the o-x at this point. Change the baffle width to 10 inches first dipole null is 1,635 Hz.

Yes your w-m itegration point moves higher but where do you want to compromise?

The scaffold & measuring outside only insures no reflections. This happened to be one of causes of my mid range bump issues. Your results may vary.

[Deward Hastings]

Re: Open Baffle musings

but where do you want to compromise?

That is alwaysthe question. And to get over a decade (of frequency response) out of a driver always involves significant compromise of something . . . the same problem SL had with ORION. Flat on axis, flat power response, constant directivity . . . you can always get one, sometimes (maybe) get two, and never get all three (with the drivers stretched that far). The original ORION got the first two at the expense of the third . . . in some ways I regard it as the best of the family. All the later versions sacrificed something in frequency response to get some approximation of directivity control. Always compromises . . .

[charlielaub]

Re: Open Baffle musings

For your mid shown in post #107 the first dipole null is 976 Hz

No, I'm afraid not. Let's look at the response model again:


The frequency you mention (976 Hz) is the position of the first null for a point source on an equivalent ROUND baffle (shown in yellow). It has very little to do with the response from my non-point-source driver mounted quite eccentrically on my rectangular baffle according to the model (in blue), and confirmed by my on-axis measurement (in plot below, measurement is above 200Hz. This is blended with a model of the response below 200Hz):


The dip at 2200Hz, peak at 4300 Hz, and subsequent roll off above are all part of the driver's response and not from the baffle. The 976 Hz dip is not there, because the diffraction by the baffle edges mostly cancels itself out, very unlike the situation for a circular baffle. I invite you to plug my numbers into your own copy of ABC Dipole and model all the off axis response angles. You will see that the off-axis responses essentially follow the on-axis response without peaks and dips. This is possible because of the position of the driver on the baffle, and the size of the baffle surface. It's still a dipole, just not the "classic" pattern, so I don't need to restrict myself to operating below the "classic" first dipole peak frequency (some people call this the "linear" response regime). What is more restrictive for the crossover point to the tweeter is the off-axis response of the midwoofer, but that is not significantly down at or below 1k Hz. You can see this for yourself in the off-axis responses in ABC Dipole.

not sure what the measured FR looks like there but assume we would want the o-x at this point. Change the baffle width to 10 inches first dipole null is 1,635 Hz.

Yes your w-m itegration point moves higher but where do you want to compromise?

The scaffold & measuring outside only insures no reflections. This happened to be one of causes of my mid range bump issues. Your results may vary.

Use a gated impulse technique and do in room measurements. You'll find it to be much more convenient than the scaffold approach! ;) Here's a tutorial that I wrote about using ARTA to make these kind of measurements:
audio.claub.net/tutorials/FR measurement using ARTA.pdf

-Charlie

[jhollander]

Re: Open Baffle musings

So I guess my explanation of why a narrow baffle helps with tweeter integration got lost on the statement “assume you want the x-o point at the first dipole null” and how “I” solve a mid range issue with outdoor measurements.

I'm sorry you did not find these suggestions helpful.

[charlielaub]

Re: Open Baffle musings

I guess my explanation about why there is no first dipole null for the midwoofer was lost on you. There is no need to cling to your dogma, because it does not apply to my system.

[charlielaub]

Re: Open Baffle musings

Thanks for commenting on the possibility of reflections... I looked at the setup again and thought maybe there could be a reflection off of the top of the baffle, since the tweeter is sitting directly on top.

Today I put a second W2-800SL in a bar clamp and positioned it well above the baffle, as shown here:


Then I did a FR measurement on it, and found the following:



The bump at 2k-3k is still there, so it's not a reflection from the baffle. The only other origin that I could think of was a reflection off of the motor back into the cone. All my sims of the driver show that the +6dB baffle peak should be showing up around 4k Hz. The peak around 2k-3k is too narrow to be the baffle peak anyway... so at this point I think it is inherent to the driver although I have to admit that I am not 100% convinced.

-Charlie

[charlielaub]

Re: Open Baffle musings

While I was set up, I took another on-axis measurement of the current config:


I'm liking this a lot. I hope to set up some subs today, crossed over around 100 Hz, so that I can listen to the system with some bass.

-Charlie

[charlielaub]

Re: Open Baffle musings

I finally had time today, while the measurement gear was out, to take a bunch of off-axis FR measurements in the horizontal plane of the MWOB with crossover version 2. The mic is elevated to be about equal to the top of the baffle (48" above the floor) for all of these measurements, and they are valid above 200 Hz.



-Charlie