I have been a little dissatisfied with the limited sweet spot of my speakers and the mediocre stereo imaging I get.
My area of focus on my 3-way speaker is with the midrange and the tweeter.
Midrange is the Audax PR170M0
Tweeter is the Morel MDT-37
Crossover Point 2.7 kHz
My first thought was to check the off-axis response of the system and I have the following plot:
Each line represents a 15° change in microphone position from zero (on-axis) to 60°.
Due to the room size the mic was 31" from the front of the speaker, aimed at the intersecting point between the Audax PR170M0 midrange and the Morel MDT-37 tweeter.
The driver pair are mounted so that they are 5.275" (134 mm) center to center. A small portion of the tweeter is carved out so that it fits close to the Audax midrange. The baffle size for the two drivers is 8" by 12", with the sides sloping backwards at about 30° to reduce the baffle diffraction step. The sloped sides appear to do just that and leave no ill traces of of a step.
My Observations:
The off-axis response looks pretty disappointing to me and may, in part, explain the sensitive sweet spot I get.
In the vertical direction I have these two plots that are in 7.5° increments along the azimuth axis and are plotted at the same distance of 31".
This is going upwards in azimuth...
This plot is going downwards in azimuth.
I would like to see a better performance from the Audax and Morel as far as off-axis response, but there seems little that I can do to coach better dispersion with these drivers.
The Audax has a lot of ripple on the frequency response, but that may be just the way it is with a paper cone. Most mids in this SPL range have this syndrome.
Your Opinions?
The question remains what can I do to improve the performance of the imaging and widen the sweet spot, if anything?
Errata:
The following is just information that some may find interesting or necessary...
Crossover Details
Now, the crossover was a design performed by Madisound and is available here: LEAP.pdf
For those of you with a sharp eye, the tweeter's polarity, if wired according to that schematic, produced a huge suckout at the upper crossover point at 2.7 kHz. It looked like this:
So, before I began any actual measurements I reversed the tweeter polarity and every other plot here has the tweeter reversed from the schematic.
Cone Breakup
The gyrations in the plots from about 300 Hz to 1100 Hz (for those with questions) appear to be a combination of the woofer's cone break up and the Audax midrange. The crossover point is 400 Hz.
This is the JBL 2235H:
This is the Audax PR170M0
My area of focus on my 3-way speaker is with the midrange and the tweeter.
Midrange is the Audax PR170M0
Tweeter is the Morel MDT-37
Crossover Point 2.7 kHz
My first thought was to check the off-axis response of the system and I have the following plot:
An externally hosted image should be here but it was not working when we last tested it.
Each line represents a 15° change in microphone position from zero (on-axis) to 60°.
Due to the room size the mic was 31" from the front of the speaker, aimed at the intersecting point between the Audax PR170M0 midrange and the Morel MDT-37 tweeter.
The driver pair are mounted so that they are 5.275" (134 mm) center to center. A small portion of the tweeter is carved out so that it fits close to the Audax midrange. The baffle size for the two drivers is 8" by 12", with the sides sloping backwards at about 30° to reduce the baffle diffraction step. The sloped sides appear to do just that and leave no ill traces of of a step.
My Observations:
The off-axis response looks pretty disappointing to me and may, in part, explain the sensitive sweet spot I get.
In the vertical direction I have these two plots that are in 7.5° increments along the azimuth axis and are plotted at the same distance of 31".
An externally hosted image should be here but it was not working when we last tested it.
This is going upwards in azimuth...
An externally hosted image should be here but it was not working when we last tested it.
This plot is going downwards in azimuth.
I would like to see a better performance from the Audax and Morel as far as off-axis response, but there seems little that I can do to coach better dispersion with these drivers.
The Audax has a lot of ripple on the frequency response, but that may be just the way it is with a paper cone. Most mids in this SPL range have this syndrome.
Your Opinions?
The question remains what can I do to improve the performance of the imaging and widen the sweet spot, if anything?
Errata:
The following is just information that some may find interesting or necessary...
Crossover Details
Now, the crossover was a design performed by Madisound and is available here: LEAP.pdf
For those of you with a sharp eye, the tweeter's polarity, if wired according to that schematic, produced a huge suckout at the upper crossover point at 2.7 kHz. It looked like this:
An externally hosted image should be here but it was not working when we last tested it.
So, before I began any actual measurements I reversed the tweeter polarity and every other plot here has the tweeter reversed from the schematic.
Cone Breakup
The gyrations in the plots from about 300 Hz to 1100 Hz (for those with questions) appear to be a combination of the woofer's cone break up and the Audax midrange. The crossover point is 400 Hz.
An externally hosted image should be here but it was not working when we last tested it.
This is the JBL 2235H:
An externally hosted image should be here but it was not working when we last tested it.
This is the Audax PR170M0
If you want more regular dispersion, you have to go with smaller drivers, and or drivers that can be crossed over lower. But looking at your first FR, it doesn't look too bad - the thing I look for in horizontal polars is that the lines are basically parallel, which they seem to be. It would be a pretty good idea to take your measurements outside where you can get longer gate times - you're pretty much only looking at the treble range, but the mids are almost more important - not in terms of polar response, but more for the overall FR of your system. I would highy recommend measureing at a greater distance too - like minimally 6 feet off the ground, and a speaker-mic distance of 5 feet. I do my best measurements 11 feet off the ground, and a mic distance of 6 feet.
What is your baffle dimensions/layout?
There isn't much you can do about your vertical polar response, unfortunately, besides get the drivers closer and cross lower.
Do you feel the treble is rolling off too soon, leading to a small sweetspot? Your tweeter is slightly on the large side, but not by much...
I'd guess that stereo imaging problems are more a function of differences between the two speakers FR, and their placement in your room. How much room do you leave between the speakers and large objects - I consider 3 feet about minimal.
Hopefully others will chime in too.
What is your baffle dimensions/layout?
There isn't much you can do about your vertical polar response, unfortunately, besides get the drivers closer and cross lower.
Do you feel the treble is rolling off too soon, leading to a small sweetspot? Your tweeter is slightly on the large side, but not by much...
I'd guess that stereo imaging problems are more a function of differences between the two speakers FR, and their placement in your room. How much room do you leave between the speakers and large objects - I consider 3 feet about minimal.
Hopefully others will chime in too.
This is what the cabinet looks like:
An externally hosted image should be here but it was not working when we last tested it.
Front baffle layout:
An externally hosted image should be here but it was not working when we last tested it.
Note that the front baffle is raked backward at an angle of 3.8° as illustrated in this isometric view:
An externally hosted image should be here but it was not working when we last tested it.
Measurements were taken with the cabinet in the room center. Normally one speaker is closer to the wall, but the room layout will change in the future, so I am trying to get all the gremlins out of the beast first, then apply veneer.
Moving my head a few inches from the listening position does show some treble fall-off. Overall imaging is a little vague in my opinion. If you get into the sweet spot it is better. I am sure the room has enough to do with it, too.
Try switching to CD type waveguide or horn.
Rob
Do you mean a full horn and compression driver?
I came up with this idea a few months back, but my girlfriend could not get the idea out of her mind that she was looking at two people mooning us all of the time.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Something along those lines. There are option besides a 4430 clone
. You could go with 1" or 1.5" PT waveguides or grab a set of Earls Waveguides. Just have to work out the compensation and driver integration if you strike out on your own.With the improvements you seem to be looking for it's a good bet you could end up happy trying it out. One thing is for sure using horn/waveguide similar to a 2344 you don't have to worry all that much about a small sweet spot. Just toe them in a bit and go.
Rob
Room. I saw you later posted that the room could have something to do with it.
I noticed hard tile floors in your pic, not your listening room??
Anyway, I have never been an 'imaging freak', other things float my boat (or maybe I was not because I did not particularly have it??). Got bored one day (and had taken my speakers as far as they will go) and so just hung raw f/glass batts in my room....room treatment was always on my list and it just started that day.
Just to give you an idea of the rough and ready experiment!! They have been there for a year now.
My room is large and very solid (two foot solid brick walls), but that also made it very ringy.
I was staggered at the improvement getting rid of the slap echo made. Many improvements, but NOW I had imaging like never before. In fact, it rivalled the imaging I had only ever heard once before, and funnily enough that guy also had a well treated room.
Only a survey of two, but it seemed pretty conclusive to me!
Anyway, all I want to get across is that for an experiment (as you can see) it is very simple and cheap to do, and you can always put the batts in the ceiling after your experiment.
I noticed hard tile floors in your pic, not your listening room??
Anyway, I have never been an 'imaging freak', other things float my boat (or maybe I was not because I did not particularly have it??). Got bored one day (and had taken my speakers as far as they will go) and so just hung raw f/glass batts in my room....room treatment was always on my list and it just started that day.
Just to give you an idea of the rough and ready experiment!!
They have been there for a year now.My room is large and very solid (two foot solid brick walls), but that also made it very ringy.
I was staggered at the improvement getting rid of the slap echo made. Many improvements, but NOW I had imaging like never before. In fact, it rivalled the imaging I had only ever heard once before, and funnily enough that guy also had a well treated room.
Only a survey of two, but it seemed pretty conclusive to me!
Anyway, all I want to get across is that for an experiment (as you can see) it is very simple and cheap to do, and you can always put the batts in the ceiling after your experiment.
Yes, it is, but things will change at some point in a fairly major way.
My goal is to complete these speakers, but I have some things to do before I can veneer them. My list is:
1. Finish crossover tweaks and any other overt flaws in the system. So, I am trying to dial in the sound the best I can.
2. Fix cabinet resonances - there are a lot. Leaning toward constrained layer with green glue.
3. Fix flaws in the cabinets (nicks, dents, irregularities) so they are ready for veneer.
4. Room renovations. The room gets a new layout and the system moved or turned 90° so that the speakers fire down the long path of the room. This not only improves the seated listening position, but will play right into the attached kitchen. So casual listening will benefit as well.
5. Room treatments where possible.
Even with the room changes and treatments I expect that the room will be less than ideal, but I have limits there. I can accept that, but at least I want to make sure the bow and ribbon are done on the speakers so that they are not a source of inferiority.
Not a great idea having them uncovered... you will have glass fibres floating around in the air.
Looking at their shape and their arrangement near the wall edges made me wonder, would they work better if they were arranged in an ENABL pattern? (Just joking)
In my experience, better off-axis response don't necessarily lead to better stereo imaging, but it will help the overall tonal balance. What you need to get control over is the *level* of the early room reflections, as these do affect stereo imaging a lot. Treating primary reflections points in the room with absorbtion or diffusion is very effective.
I agree that early reflections may be your problem, especially since one speaker is closer to the wall. Also, matching the FR between the two loudspeakers is important. The three ways to manage this problem (that I know of) are: 1) add acoustic treatment to the near side wall, 2) use drivers with narrow directivity for mid and tweet and 3) move the speakers away from acoustically large objects (the wall) or otherwise ensure that those acoustically large objects are placed symmetrically to left and right of the centre of your speakers.
I like your speaker box design, but I think that the ledge to left and right of the midrange is an anomally wrt the care you took elsewhere in trying to manage diffraction. Perhaps you can fill in those corners, somehow, and smoothen the transition?
I like your speaker box design, but I think that the ledge to left and right of the midrange is an anomally wrt the care you took elsewhere in trying to manage diffraction. Perhaps you can fill in those corners, somehow, and smoothen the transition?
Thanks, all. Like I said, I know there are problems that I have with the existing room layout. That will change in the future and I can address room deficiencies at that point.
I just wanted to clean up the speakers in the meantime and get them in order.
I guess the operative question is, how does the off-axis response I have measured compare with other loudspeaker systems:
Bad
Poor
Mediocre
OK
Excellent
I am thinking that Excellent and Bad are probably not applicable, but it would be nice to get opinions from those a little more experienced on the subject.
I'll try to find a calm day that I can wheel one of these speakers outside for a better sweep of the low and midrange areas. I would like to better understand the ripple I am getting.
Lastly, the boxes have a lot of resonances that I need to address. I am thinking of ordering some green glue and some baltic birch (1/2" or 3/8" thick) and line the interior panels with a layer of green glue and the baltic birch. The cabinet volume is 6.7 cubic feet and I had designed the woofer box for something closer to 6 cubic feet.
As far as the left and right shelves go, I could add a a small panel to make that transition smoother and see how that impacts the diffraction.
I just wanted to clean up the speakers in the meantime and get them in order.
I guess the operative question is, how does the off-axis response I have measured compare with other loudspeaker systems:
Bad
Poor
Mediocre
OK
Excellent
I am thinking that Excellent and Bad are probably not applicable, but it would be nice to get opinions from those a little more experienced on the subject.
I'll try to find a calm day that I can wheel one of these speakers outside for a better sweep of the low and midrange areas. I would like to better understand the ripple I am getting.
Lastly, the boxes have a lot of resonances that I need to address. I am thinking of ordering some green glue and some baltic birch (1/2" or 3/8" thick) and line the interior panels with a layer of green glue and the baltic birch. The cabinet volume is 6.7 cubic feet and I had designed the woofer box for something closer to 6 cubic feet.
As far as the left and right shelves go, I could add a a small panel to make that transition smoother and see how that impacts the diffraction.
A couple of hints:
If you want to confirm that the source of the ripple is the cones, do a high resolution impedance measurement using MLS or FFT to get both magnitude and phase. If you plot it on the complex plane (real and imaginary parts of Z), cone breakups and other mechanical resonances will show up as little unexpected loops.
The cabinet may very well be a significant source of ripple and lobing. The little shelves and their associated spandrels at the bottom of the midrange are suspect.
Really, the off-axis response isn't too bad at all, considering how ripply the direct response is.
If you want to confirm that the source of the ripple is the cones, do a high resolution impedance measurement using MLS or FFT to get both magnitude and phase. If you plot it on the complex plane (real and imaginary parts of Z), cone breakups and other mechanical resonances will show up as little unexpected loops.
The cabinet may very well be a significant source of ripple and lobing. The little shelves and their associated spandrels at the bottom of the midrange are suspect.
Really, the off-axis response isn't too bad at all, considering how ripply the direct response is.
Apart from the ripples, it looks very good! This is the advantage of a 3-way: you have more control over off-axis response. An equivalent 2-way would have had a dip in the midrange off-axis.
Are you aware of the Harman papers that discuss off-axis (power) response? They have some nice graphs for comparison.
Are you aware of the Harman papers that discuss off-axis (power) response? They have some nice graphs for comparison.
Last edited:
I think I need some help understanding what you are talking about with the MLS/FFT impedance measurement.
I just ordered Joseph D'Appolito's Testing Loudspeakers from Parts Express. It will be here Tuesday.
However, I am trying to jam my brain full of as much knowledge as possible. Is there a link or can you provide me an explanation of the process so I can run that test?
No, I am not aware. Most of my knowledge comes from Vance Dickason's book, which is good, but only goes so far.
I'll do a search and see if I can find what you cited.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.