how to model baffle effects for OB mids?

[Max_Andrews]

I want to run two PA130 drivers open baffle, but I want to use a wrapped baffle (so the enclosure has a front and sides but no back). I can simulate the response in FRM by setting the baffle width as if it were "unfolded," but then the peak is in the wrong place because the front baffle is actually a lot smaller. So if the front baffle is 10" wide i'd get a 2dB peak around 700hz, with the normal baffle step loss profile following down until the "full" width of the baffle including the sides is reached, where it would fall off a lot faster because of the rear wave cancellations. Sow how do I model both of these in the same simulation?

[Pete Schumacher ®]

> I want to run two PA130 drivers open baffle,
> but I want to use a wrapped baffle (so the
> enclosure has a front and sides but no
> back). I can simulate the response in FRM by
> setting the baffle width as if it were
> "unfolded," but then the peak is
> in the wrong place because the front baffle
> is actually a lot smaller. So if the front
> baffle is 10" wide i'd get a 2dB peak
> around 700hz, with the normal baffle step
> loss profile following down until the
> "full" width of the baffle
> including the sides is reached, where it
> would fall off a lot faster because of the
> rear wave cancellations. Sow how do I model
> both of these in the same simulation?

The reason is, that so much more mid energy is going into the room, but delayed in arrival. If you were to assume anechoic, you may find that a "flat" on axis response may sound a bit forward in a typical room, due to all that extra energy into the room. OB mids can sound really nice, but they are much more room/placement dependent, and a typical application will see a "saddle" response in the mids to overcome that extra energy.

[Max_Andrews]

>Yeah the PA130 is pretty "shouty" as it is, but that's part of why I like it, sounds great for a lot of vocal stuff. I'm using the SD1.1 tweet and it has a rising top end response, my current XO has about 1.5dB more in the higher registers than the mids and I've tamed the bump around 2khz a bit, haven't chosen a woofer yet, probabably dual 8" classics. They have a bit of emphasis at 100hz descending to flatter around 350hz where they will be crossed, so hopefully the rising response of the SD1.1 and the bump in the woofer will give the "saddle" response.

For modeling, should I use the profile from the 10" wide baffle or the 24" wide "imaginary" baffle?

[terryo]

> For modeling, should I use the profile from
> the 10" wide baffle or the 24"
> wide "imaginary" baffle?

I'd just use a 10" wide "test" panel of plywood or MDF and add cardboard wings attached with duct tape. In that manner, you'll have a much better appreciation of what you see with your various models. What difference are you seeing with a 10" vs. 24" baffle? With a test panel and a spl meter (Radio Shack) you can compare the results as well as just listen to music. As you've mentioned, "flat" isn't necessarily where it's at.

Best Regards,
TerryO

[Paul Carmody]

To do the Aethers, I ran the mids in response modeler with the actual baffle width, but as "open baffle"--even though they had 4" sides. Yes, it's true that the back of the drivers will see something like a 15" wide baffle, but I'll get to that in a minute.

First, the reason I believe it's important to use the actual front baffle width is to get the true diffraction characteristics of the front waves of the drivers.

Second, the reason I believe it's important to sim it as "open baffle" is that you are getting almost no help from a box to give you any sort of low end extension. The side "flaps" help a bit, but they're nothing compared to an actual closed or vented box.

However, I would also agree with Pete in regards to rear energy (though there are ways to tame it, and its SPL to the listener pales in comparison to the front wave). With open-back designs, the room becomes a part of the sound. My solution to this was to use simulation to get me close, then do a LOT of listening to familiar recordings and tweaking. Truth be told, though: these speakers will sound different in every room. And there just isn't a simulation for that. But I can still give you a graph of what the computer thinks it will sound like ;-)

[markk]

I always hate when folks say this. It's not as simple as that "extra energy from the rear."

As you drop below the baffle step in a box, the box becomes omnidirectional and "all that energy" is going to the rear too. Just thinking that the rear wave in a dipole is somehow "extra" is inaccurately simplifying the situation.

[Paul Carmody]

This is what I did on mids on the Aethers. Then I pressed "Save Baffle Diffraction Curve to BDS Register Above." Saved the new FR curve, and extracted minimum phase.

Note: there is no reason to do a box model for impedance on open mids, for obvious reasons.

[markk]

In a box, the rear wave is only partially absorbed, then reradiated anyway. Granted, as you go up higher in frequency, the absorption improves in the box. Still, though, I don't think folks should think about the rear wave in a dipole as "extra." It's way oversimplifying what is going on.

[Pete Schumacher ®]

> In a box, the rear wave is only partially
> absorbed, then reradiated anyway. Granted,
> as you go up higher in frequency, the
> absorption improves in the box. Still,
> though, I don't think folks should think
> about the rear wave in a dipole as
> "extra." It's way oversimplifying
> what is going on.

It's true that the boxed mid will still have energy "wrapping" around the enclosure, but the amount between the OB and sealed box is definitely different, and as you go up in frequency, the difference is even greater. I'm not saying that you can't model for a flat response on axis, and likely pretty accurately, but the applications I've seen people put forth seem to have a recessed voicing in the on-axis response.