is this 'power response' ok? *PIC*

[gordoncalder]

http://i234.photobucket.com/albums/ee202/gordoncalder/2.5wayextremishunzoomed.jpg

Provided Link: http://i234.photobucket.com/albums/ee202/gordoncalder/2.5wayextremishzoomed.jpg


I’m trying to model a 2.5-way version of my speaker (Adire extremis and seas 27tbfc). This is looking ok to me, but I’m always a little unclear on power response (the green line in the graph) and what I should be after. My translation of ‘power response’ is “build something with good dispersion” which may be wrong I don’t know.

Anyway, the power response here seems like a nice smooth line, but should it be lower than the summed response across all frequencies? I seem to remember it dropping off only in tweeter-territory in other designs. Did I do something wrong here, or is this alright?

The pic here is the “happy graph”, the link is for the “sad graph”. They’re both exactly the same, with different scale…you’ll see what I mean. :-)

thanks,

Gordon

[Pete Schumacher ®]

[cjd]

Power response as modeled looks good, but the transfer function peaking the way it does may cause some listener fatigue despite a smooth response.

C

[todd_o]

> I’m trying to model a 2.5-way version of my
> speaker (Adire extremis and seas 27tbfc).
> This is looking ok to me, but I’m always a
> little unclear on power response (the green
> line in the graph) and what I should be
> after. My translation of ‘power response’ is
> “build something with good dispersion” which
> may be wrong I don’t know.

> Anyway, the power response here seems like a
> nice smooth line, but should it be lower
> than the summed response across all
> frequencies? I seem to remember it dropping
> off only in tweeter-territory in other
> designs. Did I do something wrong here, or
> is this alright?

> The pic here is the “happy graph”, the link
> is for the “sad graph”. They’re both exactly
> the same, with different scale…you’ll see
> what I mean. :-)

> thanks,

> Gordon

A 5db scale would be better to determine how smooth it is.

[Pete Schumacher ®]

http://i12.photobucket.com/albums/a203/pete_schumacher/D25project.gif


> I’m trying to model a 2.5-way version of my
> speaker (Adire extremis and seas 27tbfc).
> This is looking ok to me, but I’m always a
> little unclear on power response (the green
> line in the graph) and what I should be
> after. My translation of ‘power response’ is
> “build something with good dispersion” which
> may be wrong I don’t know.

> Anyway, the power response here seems like a
> nice smooth line, but should it be lower
> than the summed response across all
> frequencies? I seem to remember it dropping
> off only in tweeter-territory in other
> designs. Did I do something wrong here, or
> is this alright?

> The pic here is the “happy graph”, the link
> is for the “sad graph”. They’re both exactly
> the same, with different scale…you’ll see
> what I mean. :-)

> thanks,

> Gordon

Your 2.5-way transfer function is not really indicative of a true 2.5-way. You've got some BSC on the main woofer, and ideally, that transfer function would be flat to the XO point. The peaking you've introduced to flatten the response might also present some problems.

I looked at the "zoomed" link at the bottom, and the power response is still quite good.

Check out the transfer function chart below. You ought to be shooting for something a little more like this.

[jkim]

Provided Link: My Speaker Building Pages (http://www.geocities.com/woove99/Spkrbldg/)


I have a comment on the peaking issue of your woofer net transfer function people talk about. An electrical third order can produce this kind of curve and adding a notch filter can exaggerate the shape. This can be used to remedy a 6.5" midwoofer's usual dip at around 1 kHz.

But don't go too wild in using this. The dip is usually called "surround dip" because it occurs when the cone's linear movement turns into breakup mode and is interfered by the driver's surround as the frequency increases. It is caused by some kind of phase interaction. Moving off axis the severity of this dip also changes. By making on-axis FR flat using a peaking transfer function, you can possibly overdo it---too much energy at the frequency where peaking occurs.

I used this type of filter for my Usher 2-way design. The Usher has a bit more severe dip than your Extremis. But I didn't completely fill up the dip for the above reason. Take a look at what I did. Find transfer functions in the page:
<A HREF="http://www.geocities.com/woove99/Spkrbldg/Usher_2way/Usher_2way.htm">http://www.geocities.com/woove99/Spkrbldg/Usher_2way/Usher_2way.htm</A>

- Jay

[gordoncalder]

> Your 2.5-way transfer function is not really
> indicative of a true 2.5-way. You've got
> some BSC on the main woofer, and ideally,
> that transfer function would be flat to the
> XO point. The peaking you've introduced to
> flatten the response might also present some
> problems.

> I looked at the "zoomed" link at
> the bottom, and the power response is still
> quite good.

> Check out the transfer function chart below.
> You ought to be shooting for something a
> little more like this.

hmm, ok, I'll try messing around a little more. I have a notch filter around 800-1200hz, and I think that's the cause of that transfer function peak. But without it there's a freq response peak (more of a broad hump than a peak)... I'll keep experimenting and come back with another sample.

thanks for the help,

Gordon