Or simply not flat
and any ground plane or mic phase issue would further influence the dip.
Or simply flat
But impossible to do with small baffles, to achieve un equal distance to edges.
or something achievable real world with wider baffle
and offset mounting, so no edge is equal distance.
Typical 1.3 ratio or 1.6 ratio aka close enough to golden ratio.
And all the way up to 1.7 ratio actually still works, and more likely
because of physical driver sizes and mounting edges.
and any ground plane or mic phase issue would further influence the dip.
Or simply flat
But impossible to do with small baffles, to achieve un equal distance to edges.
or something achievable real world with wider baffle
and offset mounting, so no edge is equal distance.
Typical 1.3 ratio or 1.6 ratio aka close enough to golden ratio.
And all the way up to 1.7 ratio actually still works, and more likely
because of physical driver sizes and mounting edges.
Since you already picked double 18mm for the baffle, go for big chamfers like described here. Scroll to the bottom. You might consider adding a third layer of mdf. That will optimize high frequency radiation pattern. It will not necessarily improve midbass-tweeter integration, a waveguide would be better for that.
Waveguide just makes center to center farther.
Also makes off axis poor.
horrible vertical, horizontal drops like a rock.
Crossing to a 6" driver.
You never ever need a waveguide
because crossover is not low enough to remotely care.
Specially since its obvious offsetting the driver creates
a low end rising response if the crossover even needed it.
What it would do is probably create the same old annoying
30 ohm crossover peak to remove all the response not needed.
Added to the woofer baffle step, why stop at 30 ohms.
Crank it up to 60 ohms
Also makes off axis poor.
horrible vertical, horizontal drops like a rock.
Crossing to a 6" driver.
You never ever need a waveguide
because crossover is not low enough to remotely care.
Specially since its obvious offsetting the driver creates
a low end rising response if the crossover even needed it.
What it would do is probably create the same old annoying
30 ohm crossover peak to remove all the response not needed.
Added to the woofer baffle step, why stop at 30 ohms.
Crank it up to 60 ohms
I do, the most magical waveguide ever used for
a small woofer or midrange.
= None
So easy to make magic lol
a small woofer or midrange.
= None
So easy to make magic lol
Every kind of baffle is a waveguide. You only have to optimize the dispersion of the total system from time to time. Waveguide-like horns can and do help out with that.
There ain't much magic... It's mostly a matter of preference of dispersion and design goals 😉
I happen to love waveguides and hugely prefer them over direct radiating drivers. But I can still admit that everything has its merit 👍
Re mic calibration, I personally think you should have some sort of calibration, otherwise you will be targeting problems that may not exist. For example, here is mine (UMIK1). This is the MiniDSP supplied calibration with the mic. I am not asserting it is true / accurate (that is - I haven't had the mic independently calibrated):
It's hard to calibrate a mic "full range" due to room reflections. It should be possible to create a self calibration using farfield (gated) measurements if you use a known tweeter measured on a known baffle and create a "difference" between that and measuring the same tweeter on the same baffle at the same distance. you should be able to do this with the SB26ADC given you know the SB Acoustics measurement conditions for the tweeter (and they seem to be consistently produced to measure very similarly). Note: the bass part of the SB Acoustics measurement cannot be used as it includes their room modes / reflections.
Below the room reflection - you could splice in the tail end of the above (or other) omni-mic files as these tend to be more similar down low. The reason to try and get the low end response in your calibration file is not because of crossover design (for a 2 way this is going to be >= 1.5KHz in many cases), it's to try and get baffle step right.
It's hard to calibrate a mic "full range" due to room reflections. It should be possible to create a self calibration using farfield (gated) measurements if you use a known tweeter measured on a known baffle and create a "difference" between that and measuring the same tweeter on the same baffle at the same distance. you should be able to do this with the SB26ADC given you know the SB Acoustics measurement conditions for the tweeter (and they seem to be consistently produced to measure very similarly). Note: the bass part of the SB Acoustics measurement cannot be used as it includes their room modes / reflections.
Below the room reflection - you could splice in the tail end of the above (or other) omni-mic files as these tend to be more similar down low. The reason to try and get the low end response in your calibration file is not because of crossover design (for a 2 way this is going to be >= 1.5KHz in many cases), it's to try and get baffle step right.
Hello everyone.. I got new mic yesterday. Re run the measurement
I think the boost from my previous measurement is from the microphone..
But the dip still there. Should i accept the dip and go on?
Going to make 1 inch round over for the last try..
Or should i make new baffle?
Thank you all. Really appreciate it..
I think the boost from my previous measurement is from the microphone..
But the dip still there. Should i accept the dip and go on?
Going to make 1 inch round over for the last try..
Or should i make new baffle?
Thank you all. Really appreciate it..
Good results with the new mike. I already gave a clue on baffle modification 😉. An 1” roundover won’t be enough, I’d predict.
No, you should not be satisfied with that result.
Is this a measurement of both drivers and crossover?
From this discussion i realized that
cabinet has a big impact on diffraction..
But how commercial speaker (ex like sonus faber lumina) or dynaudio etc, they made some models with no round over chamfer or any diffraction counter. How they deal with this?
I mean would the measurement of the speakers shows some dip and peak as well? Because on google the measured response of those speakers didn't show any severe diffraction..
cabinet has a big impact on diffraction..
But how commercial speaker (ex like sonus faber lumina) or dynaudio etc, they made some models with no round over chamfer or any diffraction counter. How they deal with this?
I mean would the measurement of the speakers shows some dip and peak as well? Because on google the measured response of those speakers didn't show any severe diffraction..
Please could you give a little bit more information to tell us exactly what was done to give the measurement shown in post 50? Mic distance, on tweeter axis, Response of the loudspeaker and crossover combined?
Does the new mic come with a cal file and how did you apply that?
Also some details of the crossover used even if you just mention second or third order and corner frequency, 3Kiloherz maybe. I am not sure hat caused the little wiggle you have there, Possibly reversing phase to one of them could make the dip better or worse.
Maybe as you get to the final phase you will need some attenuation on the tweeter.
Does the new mic come with a cal file and how did you apply that?
Also some details of the crossover used even if you just mention second or third order and corner frequency, 3Kiloherz maybe. I am not sure hat caused the little wiggle you have there, Possibly reversing phase to one of them could make the dip better or worse.
Maybe as you get to the final phase you will need some attenuation on the tweeter.
Well, often: yes. See these or these. And probably quite some more. A lot with mediocre baffle step correction, a lot of others with diffraction-induced irregularities on-axis. But: if you take care of your baffle and pick everything right, you get this (as an example).
[Edit] There is a funny little wrinkle in the bespoke dip in your response curve though. And it appears in all of your far-field measurements. That should be examined, could it be some fault in your measurement setup, like something causing reflections near your speaker or microphone?
Last edited:
He did say he was measuring at 100cm. Seems a little far. Mic only needs to be at a distance greater than twice the baffle width to be far enough to include baffle data.
Centered baffle is about 8", and wavelength of 3k is 4.5". Is anything protruding or bumped out at about 101mm from center of dome? Since it has not been suggested, place heavy felt between tweeter and woofer and see if the dip goes away.
- measure without anything around the speaker to cause a reflection.
- placing a thick blanket on the floor in front of speaker can help.
- I think he should measure closer to 40 to 60cm distance.
- if the speaker being measured is not equidistant from floor and ceiling, usually this can help.
- if the speaker is operating parallel to walls, it should be rotated to be more orthogonal or non parallel.
- if the measurements are not being taken directly on the tweeter axis, this is also important, unless the listening axis is to not be this axis.
Centered baffle is about 8", and wavelength of 3k is 4.5". Is anything protruding or bumped out at about 101mm from center of dome? Since it has not been suggested, place heavy felt between tweeter and woofer and see if the dip goes away.
Here is the notes of the measurements
1. The mic is behringer ECM8000, Is not specifically calibrated mic, or has specific calibration file, just generic cal file, I'm aware that this mic is far from 100% accurate, but hopefully is still on the acceptable tolerance
2. The measurement i posted here is just the tweeter only, without any crossover, But from REW the sweep start from around 400hz (not sure whether it is safe or not) and gated around 4 to 5ms
3. The mic position is at the tweeter level, around 100cm distance,
4. The speaker is in the middle of the room (not exactly middle, and standing on top of chair) to minimize reflection.
1. The mic is behringer ECM8000, Is not specifically calibrated mic, or has specific calibration file, just generic cal file, I'm aware that this mic is far from 100% accurate, but hopefully is still on the acceptable tolerance
2. The measurement i posted here is just the tweeter only, without any crossover, But from REW the sweep start from around 400hz (not sure whether it is safe or not) and gated around 4 to 5ms
3. The mic position is at the tweeter level, around 100cm distance,
4. The speaker is in the middle of the room (not exactly middle, and standing on top of chair) to minimize reflection.
The concept, that frequency response is dependend on baffle and placement on baffle, is not accepted by most beginners, as they can not imagine it. "Yes, but that last bit is not that important to me" is what you hear from most. It is just as hard to remove such ideas from a DIYS person as it is to convice them that an universal crossover doesn't work. When they get their first microphone and finally measure something usefull, that resulting mess usualy is an ugly surprise.
There is a cheap and simple, uncomplicated way of testing actual baffle response, even before you build a speaker from wood.
Get some panels of XPS insulation. You can cut them into any shape, just with a knife. Measure. So simple, so easy and cheap. Even no complicated to clean dirt. No more "yes, but..." If you ask at a building site, in most cases you can carry cut off XPS pices away for free, as it is expensive to dispose.
If even that is too much work, do simulations or buy a finished speaker and never measure it.
There is a cheap and simple, uncomplicated way of testing actual baffle response, even before you build a speaker from wood.
Get some panels of XPS insulation. You can cut them into any shape, just with a knife. Measure. So simple, so easy and cheap. Even no complicated to clean dirt. No more "yes, but..." If you ask at a building site, in most cases you can carry cut off XPS pices away for free, as it is expensive to dispose.
If even that is too much work, do simulations or buy a finished speaker and never measure it.