Statements: Modified frequency plot: *PIC*

[curt_c]

This is probably closer to the true on axis response, but still not completely accurate. For this compromised plot, I simply moved the mid response up 3dB. The actual response would be skewed, closer to the theoretical 4.8 dB at the low end, and less than that at the upper end, due to the absorption of the higher frequencies by the foam lining the transmission lines. Even though this plot shows something of a BBC dip, I think the actual acoustic response sounds much flatter. -As compared to a known flat monopole design. Rather then attempt to measure and model all the complexities involved, IMO, it may be easier to tweak dipole type designs by ear for the flattest acoustic response. This plot shows the crossover points much closer to the target 250 Hz and 4 kHz transfer function of the more easily modeled (monopole) woofer and tweeter, and consequently may be a more accurate portrayal of the actual response. By design, the path length between the front and the rear of the mids does put the addtional 6dB / octave dipole roll off into the stop band of the drivers, simplifying the crossover design.

C

[Jim Holtz]

> This is probably closer to the true on axis
> response, but still not completely accurate.
> For this compromised plot, I simply moved
> the mid response up 3dB. The actual response
> would be skewed, closer to the theoretical
> 4.8 dB at the low end, and less than that at
> the upper end, due to the absorption of the
> higher frequencies by the foam lining the
> transmission lines. Even though this plot
> shows something of a BBC dip, I think the
> actual acoustic response sounds much
> flatter. -As compared to a known flat
> monopole design. Rather then attempt to
> measure and model all the complexities
> involved, IMO, it may be easier to tweak
> dipole type designs by ear for the flattest
> acoustic response. This plot shows the
> crossover points much closer to the target
> 250 Hz and 4 kHz transfer function of the
> more easily modeled (monopole) woofer and
> tweeter, and consequently may be a more
> accurate portrayal of the actual response.
> By design, the path length between the front
> and the rear of the mids does put the
> addtional 6dB / octave dipole roll off into
> the stop band of the drivers, simplifying
> the crossover design.

> C

Thanks for the additional graph explaining why the FR dip is there with conventional measuring techniques. It's that di-pole thing causing it. :-)

Jim

[maynardg]

[dennismurphy]

Hi Curt I can't tell from the graphs which version lspCAD you're using. Have you switched over to 6.0? If so, what does the learning curve look like? Thanks.

[Jim Holtz]

Hi Maynard,

I'm not Curt but here is the URL to his website.

<A HREF="http://www.geocities.com/cc00541/index.html">http://www.geocities.com/cc00541/index.html</A>

The Statements project hasn't been posted yet. It'll be a while before that happens. For one thing, I have to create a drawing for the cabinets and I'm terrible at that! :-(

Jim

[curt_c]

> Hi Curt I can't tell from the graphs which
> version lspCAD you're using. Have you
> switched over to 6.0? If so, what does the
> learning curve look like? Thanks.

Hi Dennis,

I'm still using 5.25. I haven't really found a real need to update yet, so I cannot answer to the learning curve. FWIW, I do subscribe however, to both yahoo users groups for LspCAD and SE. I can certainly tell you one group is much more active than the other. This may be some indictation that the curve is not all that steep, relatively speaking.

C

[johnk]

> This is probably closer to the true on axis
> response, but still not completely accurate.
> For this compromised plot, I simply moved
> the mid response up 3dB. The actual response
> would be skewed, closer to the theoretical
> 4.8 dB at the low end, ....

Sorry Curt, but I'm not getting this at all. Designing a dipole for flay on axis response is no different than do so for a monopole: Measure on axis and the design the filters/compensation networks to yield flat response. The 4.8 dB thing is power, not on axis SPL.

Could you explain what you mean here a little more clearly?

[Pete Schumacher ®]

> Sorry Curt, but I'm not getting this at all.
> Designing a dipole for flay on axis response
> is no different than do so for a monopole:
> Measure on axis and the design the
> filters/compensation networks to yield flat
> response. The 4.8 dB thing is power, not on
> axis SPL.

> Could you explain what you mean here a
> little more clearly?

By gating, you don't get the sound from the back of the driver. If you ran frequency sweeps indoors, the reflected sound from the back of the driver would increase the SPL somewhat, but a short gate time wouldn't allow that to show up in the curves . . . at least that's how I heard him explain it.

[curt_c]

Hi John, I was hoping you would chime in here.

I took the measurements all on the tweeter axis at roughly 1 meter with 7ms gating IIRC. Everything looked reasonable and comparable to the MFG's posted plots.

I modeled in LspCAD for a normal flat response. but the result only sounded flat in-room when the mid was reduced to my first posted response plot. I can only assume the power response mandated the on-axis dip in response.

Your thoughts? The mids are in a foam lined untapered t-line of about 18 to 20 inches in length.

I got the 4.8 thing from one of your tech articles. No one has ever said I was a good reader.. ;^)

C

[curt_c]

Provided Link: http://www.pesupport.com/cgi-bin/config.pl?read=350600

[johnk]

> Your thoughts? The mids are in a foam lined
> untapered t-line of about 18 to 20 inches in
> length.

What is the cross sectional area, and where does the TL exit? What is the path length form the terminus to the front where the driver is? A 20" TL should have a 1/4 wave resonance no higher the 169 Hz and with the foam lining I would expect the higher frequencies to roll off. I don’t think you will see much dipole effect in the mid band. Just perhaps some additional rear radiation uncorrelated from front. But with the foam lining I would expect to see too much at the frequencies in question. Perhaps you could tale a near filed measurement of what is coming out of the TL terminus? It might be more the effect of the transition between the radiation pattern of the 4" mid and the ribbon that is causing your anomaly.

[curt_c]

Hi John,

IIRC, -(These were built by Jim Holtz) the mid lines are 5” square, and 18” long, centered, and running from front to back on a 12” wide x 60 something high baffle on an otherwise conventional rectangular enclosure. Prior listening to the terminus had indicated that the foam does absorb much of the higher frequency output. I had also thought of measuring the terminus, but only did an ungated response at 12 ft. before Jim picked up the speakers. –This was with the speaker and mic in the normal listening positions in the room. The plot below shows good correlation between the modeled (using previous individual measurements) and the ungated measurement. (Lower trace) I’m perplexed why this response sounds flat, and a flat response sounded bright. Thanks again for clarifying the power response issue. -Wishful thinking on my part...

C