Have anyone experimented with a phase shield on a 5-6" midwoofer? I'm getting this driver 6″ SB17NBAC35-4 / Aluminum - Sbacoustics soon.. If I could knock down the on axis frequency peaks from 4.5kHz and up, this driver would be extremely good. I want it to match the 30degree response somewhat out to 7kHz. It would then be a lot easier to implement with a 2. order lowpass filter.
A 19mm shield would be 1/4 of the wavelength of 4.5kHz so it should cause some diffraction and effect the higher frequencies. If anyone have links to similiar products or experiments that would be great! I have REW and umic-1 so if this works in theory I could make some different sized shields and do some measurements.
A 19mm shield would be 1/4 of the wavelength of 4.5kHz so it should cause some diffraction and effect the higher frequencies. If anyone have links to similiar products or experiments that would be great! I have REW and umic-1 so if this works in theory I could make some different sized shields and do some measurements.
I think it's an interesting idea, well worth experimenting with. However i find that a simple notch filter takes down that peak nicely to allow for a 2nd-order acoustic crossover around 2khz. The woofers off-axis contribution at 7khz is quite insignificant at that point in my opinion.
Where do you want to cross over? And to what?
I have considered testing a shield for the surround. I suspect that could clean up the 1-2khz area.
Where do you want to cross over? And to what?
I have considered testing a shield for the surround. I suspect that could clean up the 1-2khz area.
Last edited:
My stock Seas Excel driver with a huge peak at 5kHz (came with a stock notch filter in the speaker) will be swapped out with this new driver. A notch can only do so much and since this SB driver doesnt have a good on and off axis respons in 5-7kHz region, a notch here will cause a dip in the off axis response. Se the graph on SB's site. I plan to cross over at 2.5-3kHz area btw.. Higher order filter will mess up the timing with the tweeter. Fixing this little problem with this driver would put in a class of its own. It measures extremely well besides this problem area.
Here is a better source of measurements:
SB Acoustics SB17NBAC35-8 | HiFiCompass
The response falls apart at about 5-6kHz due to a cone break-up. Not sure if it's redeemable with a phase-shield, i would guess not. But still, really interesting idea. Try it and post your results!
SB Acoustics SB17NBAC35-8 | HiFiCompass
The response falls apart at about 5-6kHz due to a cone break-up. Not sure if it's redeemable with a phase-shield, i would guess not. But still, really interesting idea. Try it and post your results!
Skogs, are talking about another driver? The SB looks fine to me in the 1-2kHz area, the red line is 60degrees off which I care little about. The 4 ohm version is a lot better than the one you linked to in 8 ohm.
I think there is some non pistonic movement causing the 30 degree roll off, and I suspect the on axis peak is coming from the dustcap area so it should be easy enough to cover. Also playing with the depth of the shield should give some interesting results. The lower frequencies should just ignore the phase shield and just wrap around it if I remember my highschool physics correctly 😉.
I think there is some non pistonic movement causing the 30 degree roll off, and I suspect the on axis peak is coming from the dustcap area so it should be easy enough to cover. Also playing with the depth of the shield should give some interesting results. The lower frequencies should just ignore the phase shield and just wrap around it if I remember my highschool physics correctly 😉.
You are right, they look suprisingly different in the upper region. I might suspect measurement error, but only because i don't understand why dispersion and break-up would differ between the two😛
Looking forward to seeing your results!
Looking forward to seeing your results!
Me neither.. but I've seen several measurements of the 4 ohms version and they all show the same.. the 4 ohms measurements should be accurate. Some mensions the cone/surround gives in which results in non pistonic movement in the outer area of the cone. If the measurements are right the brakeup will be a lot lower in level since its so much higher up in the frequency range.
I don't have good off-axis measurements at the moment, but here is a nearfield plot of the 8 ohm measured with a umik-1. I can't remember if its perfectly on axis or not. Looks like a few degrees of axis. But it shows a clear breakup at around 7600Hz. Very curious to see your measurements of the 4 ohm.
Attachments
Last edited:
A breakup that steep at 30 degrees will still be there after you do all you can mechanically.
Remove the dust cap, add a mushroom-cloud shaped phase plug, and add dead mass to the cone edge('edge treatment'). It will still be there, only very bad instead of extremely bad.
If you want to see for yourself, get a pair of a cheaper driver with the same problem, and try everything you can throw at it, good fun we can learn from!
Remove the dust cap, add a mushroom-cloud shaped phase plug, and add dead mass to the cone edge('edge treatment'). It will still be there, only very bad instead of extremely bad.
If you want to see for yourself, get a pair of a cheaper driver with the same problem, and try everything you can throw at it, good fun we can learn from!
