Question for Curt

[johnpastuck]

I measured the 2-ti100 at my design axis which is maybe 15 to 25 degrees off axis on the 2 drivers and I'm getting a15db spike at 10k. All I changed was r1011 from 30 to 15 ohms and L 1021 from .3 to .18 to account for the impedance change due to the 2 drivers in paralelle. Measurements taken at 1 meter. Is this expected off axis? Does it lessen as distance increases?
Did I screw it up? Any Idea's?

TIA

John

[curt_c]

Hi John,

Is this a question about the design you brought to Iowa DIY? If so, read on...

I modeled 2 W41337's in parallel with your crossover modifications, and assumed a 2 foot distance between the mids. I found about a 5 dB peak at around 15 K when measured at infinite distance along the vertical axis equidistant from each driver, however this results in a flat modeled on axis response at 2 meters, indicating this was your design listening distance. At 1 meter the model indicated comb filtering starting around 14K, so there was no peak, but a significant suckout.
-Unfortunately, all this rambling doesn't explain your measured response peak.

I'd guess the shunt notch filter is not functioning properly. This would result in a 15 dB peak around 14K, according to the model.

C

> I measured the 2-ti100 at my design axis
> which is maybe 15 to 25 degrees off axis on
> the 2 drivers and I'm getting a15db spike at
> 10k. All I changed was r1011 from 30 to 15
> ohms and L 1021 from .3 to .18 to account
> for the impedance change due to the 2
> drivers in paralelle. Measurements taken at
> 1 meter. Is this expected off axis? Does it
> lessen as distance increases?
> Did I screw it up? Any Idea's?

> TIA

> John

[johnpastuck]

> Hi John,

> Is this a question about the design you
> brought to Iowa DIY? If so, read on...

> I modeled 2 W41337's in parallel with your
> crossover modifications, and assumed a 2
> foot distance between the mids. I found
> about a 5 dB peak at around 15 K when
> measured at infinite distance along the
> vertical axis equidistant from each driver,
> however this results in a flat modeled on
> axis response at 2 meters, indicating this
> was your design listening distance. At 1
> meter the model indicated comb filtering
> starting around 14K, so there was no peak,
> but a significant suckout.
> -Unfortunately, all this rambling doesn't
> explain your measured response peak.

> I'd guess the shunt notch filter is not
> functioning properly. This would result in a
> 15 dB peak around 14K, according to the
> model.

> C
Thanks Curt,
So the lowerd load impedance should not affect r1041? Wouldn't the resistance of the shunt need to be lower than the load at the feq in question ? The peak looks closer to 12k, maybe increase the value of c1041? Sorry for all the dumb questions but I just didn't have time to measure and model everything and it sounds great in my living room but it's still a work in progress.

Thanks again

John

[curt_c]

> Thanks Curt,
> So the lowerd load impedance should not
> affect r1041? Wouldn't the resistance of the
> shunt need to be lower than the load at the
> feq in question ?

R1041 just sets the Q of the notch and is not particularly sensitive to changes in driver load.

> The peak looks closer to
> 12k, maybe increase the value of c1041?

Try paralleling another .1 ufd to C1041. That should move it down to around 12K

C

[johnpastuck]

> R1041 just sets the Q of the notch and is
> not particularly sensitive to changes in
> driver load.

> Try paralleling another .1 ufd to C1041.
> That should move it down to around 12K

> C
That dropped it to 10 db, and it is 20db down.
Funny thing is the esl has a peak around 11k, they must cancel a bit though, when combined its 3 or 4 db. I'm starting to see why the Wilson WAMM uses an equalizer. What software do you use that allows you to model at the different distance?

Thanks

John

[curt_c]

I use LspCAD, and I'm sure SE will show the summation at different distances as well. I don't know of any others...

C