For most tweeters, a round over of 3 or 4 inches is probably not enough to ameliorate the diffraction in its passband.
In limited simulations, I have found little improvement in moving from a 1 inch to 4 inch round over, there is an improvement, but the resultant change in baffle loading screws up the frequency response quite seriously, requiring a total crossover rethink, adding EQ elements to bring it back to flat again. Much like waveguiding a tweeter.
It's far easier to pick a tweeter with a flat response and suitably narrow pattern, in the first place.
At midrange frequencies the round overs required are even larger, and less practical
In limited simulations, I have found little improvement in moving from a 1 inch to 4 inch round over, there is an improvement, but the resultant change in baffle loading screws up the frequency response quite seriously, requiring a total crossover rethink, adding EQ elements to bring it back to flat again. Much like waveguiding a tweeter.
It's far easier to pick a tweeter with a flat response and suitably narrow pattern, in the first place.
At midrange frequencies the round overs required are even larger, and less practical
Or tall....
Well, doing the maths, anything less than 12" is in the midrange...isnt it?
I mention is because, there are plenty of proponents of the wide baffle idea, for reducing (rather moving) the problem
Going the opposite way, making the baffle dimensions as small as possible, just means more correction required for step loss, and a larger affected bandwidth.
FWIW Scott, I appreciate the counter argument, I only wish you could be bothered to back it up with some thinking and maths to support the idea.
Well, doing the maths, anything less than 12" is in the midrange...isnt it?
I mention is because, there are plenty of proponents of the wide baffle idea, for reducing (rather moving) the problem
Going the opposite way, making the baffle dimensions as small as possible, just means more correction required for step loss, and a larger affected bandwidth.
FWIW Scott, I appreciate the counter argument, I only wish you could be bothered to back it up with some thinking and maths to support the idea.
Last edited:
If by useful you mean not already known and understood, then no, it's common knowledge. Nothing in this thread is "useful".
There's a change in impedance when the wave from the tweeter hits the felt, so some of the energy is diffracted/reflected off the edge of the felt instead of passing through it. If too much of it happens at the same distance, it can accumulate and become more noticeable at certain angles.
It's why xrk has the nautilus-shaped baffle in one of his projects; each "angular section" sees a different distance to the baffle and so baffle diffraction is smeared across a larger angle/frequency.
It's why xrk has the nautilus-shaped baffle in one of his projects; each "angular section" sees a different distance to the baffle and so baffle diffraction is smeared across a larger angle/frequency.
I wouldn't be using a response plot to design the crossover in such a case.
1" radius is minimally sufficient to make an impact on diffraction at the more sensitive region of a few kHz, so it can be worthwhile if there is a problem.
Using a larger driver to control directivity to the same frequency might also introduce cone breakup into the band.
Allen,
"A Few kHz" is not the most sensitive region, perhaps 1kHz to 6kHz is the most sensitive region.
A lot (too much), scientific assumption going on, I fear.
Sadly, in my sims, a 1" round over has almost zero affect on these frequencies.
I totally agree with the post above yours, the baffle needs to be dealt with in a far more aggressive manner, than a simple and minimally effective 1" roundover can provide.
That being said, pigs will fly before I am tempted to use foam core for anything, and I'm quite happy to investigate angled, or trapezoidal baffles and absorbent. Its clearly still a matter of opinion, argued with science on both sides, still....I trust my ears, mostly😀
"A Few kHz" is not the most sensitive region, perhaps 1kHz to 6kHz is the most sensitive region.
A lot (too much), scientific assumption going on, I fear.
Sadly, in my sims, a 1" round over has almost zero affect on these frequencies.
I totally agree with the post above yours, the baffle needs to be dealt with in a far more aggressive manner, than a simple and minimally effective 1" roundover can provide.
That being said, pigs will fly before I am tempted to use foam core for anything, and I'm quite happy to investigate angled, or trapezoidal baffles and absorbent. Its clearly still a matter of opinion, argued with science on both sides, still....I trust my ears, mostly😀
Last edited:
The nautilus concept goes a little beyond diffraction, so may be ambiguous without some explanation.
As I alluded to in my last post, are you looking at a response plot when making this judgement, and if so, how can you be sure what's really going on?
Oops. I meant this one, not the (Nautaloss) back-loaded horn. Open baffle Nautilus: the Fibonacci solution to edge diffraction
Sorry for for the late reply, Allen. What I can say again is that due to the intense interaction between sound wave and damping material, it is quite easy to dampen high frequency waves, which occurs mainly by absorption through a combination of conversion to heat, dispersion, a decrease in amplitude and an increase in wavelength. OK?
Lol I love the way the Seasonal Spirit wears off in time for Boxing Day!
Today is my xmas #2, so I'm happy to ignore trolls!
Today is my xmas #2, so I'm happy to ignore trolls!
There are plenty of articles, tests results, and forum posts in the past on diffraction. Imaging is also improved when diffraction is addressed.
One useful link below.
Diffraction Doesn't Have to be a Problem
One useful link below.
Diffraction Doesn't Have to be a Problem