Hello..
My name is Ryan from Indonesia.
First of all, i have soooo little experience about diy loudspeaker, i'm not a professional and i have limited tools and resources.. and my english is not good either
I'm Currently building floorstanding speaker with the woofer SB16PFCR25-08 And tweeter SB26ADC-C0004
The drivers are chosen because it is affordable and produced in my country indonesia, so it is an advantage for me..
I built the cabinet with MDF
The cabinet front baffle width is 20cm, with 35cm depth and 85cm tall.
The baffle thickness is 36mm (Double 18mm MDF)
I have measure the raw frequency response of the drivers, measurements done with omni mic (not calibrated) i'm aware that it is not 100% accurate, but i think i may get the idea with it.
The measurements are 4ms gated and is done in my living room, the speakers was on top of a chair
The pink one is the woofer
The green one is the tweeter
Without any crossover
The issue is 3khz dip on the tweeter response..
What cause this?
Is this because of the tweeter diffraction? The tweeter is not flush mounted, but the side baffle are rounded..
Is the baffle thickness of the tweeter affect this? because it is 36mm i have no idea whether it is too thick or not.. and the tweeter cabinet is separated from the woofer.
Thank you for responding. I really need help because i don't have anyone near me to be ask this kind of questions...
My name is Ryan from Indonesia.
First of all, i have soooo little experience about diy loudspeaker, i'm not a professional and i have limited tools and resources.. and my english is not good either
I'm Currently building floorstanding speaker with the woofer SB16PFCR25-08 And tweeter SB26ADC-C0004
The drivers are chosen because it is affordable and produced in my country indonesia, so it is an advantage for me..
I built the cabinet with MDF
The cabinet front baffle width is 20cm, with 35cm depth and 85cm tall.
The baffle thickness is 36mm (Double 18mm MDF)
I have measure the raw frequency response of the drivers, measurements done with omni mic (not calibrated) i'm aware that it is not 100% accurate, but i think i may get the idea with it.
The measurements are 4ms gated and is done in my living room, the speakers was on top of a chair
The pink one is the woofer
The green one is the tweeter
Without any crossover
The issue is 3khz dip on the tweeter response..
What cause this?
Is this because of the tweeter diffraction? The tweeter is not flush mounted, but the side baffle are rounded..
Is the baffle thickness of the tweeter affect this? because it is 36mm i have no idea whether it is too thick or not.. and the tweeter cabinet is separated from the woofer.
Thank you for responding. I really need help because i don't have anyone near me to be ask this kind of questions...
Attachments
I am not sure what is happening here, the woofer looks reasonable, the tweeter not so good.
How does the woofer measurement agree with the manufacturers data above 1KHz? if its a reasonable match then you know you are ok on the woofer side.
At what distance did you do the measurement's at and the woofer and tweeter are wired completely separate form each other?
The tweeter looks wrong, in that it should be declining below 1KHz your measurement goes down to 800hz or so ?
Maybe try it out of the box on its own put a 3uf in series with it and measure it to see if it now behaves on its own when measured some 50 or 100cm away.
If it doesn't behave has the phase corrector plate been knocked or has the tweeter been dropped.
The cabinet look nicely finished, with a thick baffle it is common to relieve the internal wood around the back of the woofer with router bit to give a 45 degree chamfer to let the woofer breath, did you do this?
With some further work it is possible to get you a good preforming speaker with a crossover in the 2.0 to 2.5KHz zone
How does the woofer measurement agree with the manufacturers data above 1KHz? if its a reasonable match then you know you are ok on the woofer side.
At what distance did you do the measurement's at and the woofer and tweeter are wired completely separate form each other?
The tweeter looks wrong, in that it should be declining below 1KHz your measurement goes down to 800hz or so ?
Maybe try it out of the box on its own put a 3uf in series with it and measure it to see if it now behaves on its own when measured some 50 or 100cm away.
If it doesn't behave has the phase corrector plate been knocked or has the tweeter been dropped.
The cabinet look nicely finished, with a thick baffle it is common to relieve the internal wood around the back of the woofer with router bit to give a 45 degree chamfer to let the woofer breath, did you do this?
With some further work it is possible to get you a good preforming speaker with a crossover in the 2.0 to 2.5KHz zone
Tweeters, especially with thick front plates don't like to be not flush mount.
Assuming your tweeter measurement is ONLY your tweeter, the speed of sound / diameter of your tweeter flange is around 3.4 kHz, so that would be in the right neighborhood to be an issue with not being flush mounted.
If the green line is Tweeter + Midwoofer and crossover, that dip would be caused by bad phase matching due to not taking the acoustic offsets into account.
If the green line is Tweeter + Midwoofer and crossover, that dip would be caused by bad phase matching due to not taking the acoustic offsets into account.
Diffraction can definitely affect things in such a way. Remember you've got the combined diffraction effects of the tweeters faceplate not being flush mounted and the edges of the baffle only having a very small round over.
You can see the effects of flush mounting woofers and tweeters, to varying degrees, here. Although this is only valid for the tweeter and woofers tested it shows how much of an impact diffraction can have.
http://www.zaphaudio.com/mtg-surface.html
You can see the effects of flush mounting woofers and tweeters, to varying degrees, here. Although this is only valid for the tweeter and woofers tested it shows how much of an impact diffraction can have.
http://www.zaphaudio.com/mtg-surface.html
I've guessed your baffle size and tweeter placement.
Having a bigger edge radius may help you (blue = 30mm, orange = 0 mm = square edge). You probably have something in the middle. Also - flush mount all drivers... so the tweeter doesn't "See" any edges (including woofer).
This is why we build test boxes first 🙂
Having a bigger edge radius may help you (blue = 30mm, orange = 0 mm = square edge). You probably have something in the middle. Also - flush mount all drivers... so the tweeter doesn't "See" any edges (including woofer).
This is why we build test boxes first 🙂
Hello..
I really appreciate all the replies..
Thank you for your reply...
The measurements done from around 100cm distance..
The woofer and tweeter was measured separately. Also wired separately..
SB16PFCR25-08
Above is the Woofer Frequency response from the manufacture.. I'm aware of this was measured with IEC baffle..
And below is the Tweeter
SB26ADC-C0004
And i also measured the tweeter outside the box. I just put it on top of the cabinet
The Purple one is the tweeter outside the box
For the cabinet construction.. I made a mistake, forgot to make 45 chamfer.. Already had in mind but somehow i forgot...
For the phase corrector, may i know what and where is it located? i don't know much about this thing..
Does this tweeter has some anomaly from it looks?
Thank you..
I really appreciate all the replies..
Hello..
Thank you for your reply...
The measurements done from around 100cm distance..
The woofer and tweeter was measured separately. Also wired separately..
SB16PFCR25-08
Above is the Woofer Frequency response from the manufacture.. I'm aware of this was measured with IEC baffle..
And below is the Tweeter
SB26ADC-C0004
And i also measured the tweeter outside the box. I just put it on top of the cabinet
The Purple one is the tweeter outside the box
For the cabinet construction.. I made a mistake, forgot to make 45 chamfer.. Already had in mind but somehow i forgot...
For the phase corrector, may i know what and where is it located? i don't know much about this thing..
Does this tweeter has some anomaly from it looks?
Thank you..
Hello..
So would flush mounting the tweeter help with the issue? at least a bit..
The green line is the tweeter only.. I don't know but the tweeter response looks very strange, even with diffraction factors.. The dip is way too deep..
How do you think?
Thank you..
Hello..
Is my tweeter diffraction looks normal? Because it dipped like 6db. Looks like a hole..
The purple line is the tweeter outside the box. I put it on top of the cabinet.
The green one is the tweeter inside the box, surface mounted..
Hello..
Do flush mount and round over the edge with bigger radius help with the issue?
Because i think the dip is way too deep.. And i have no idea if this normal or not..
DIY is hard but super fun to me. But at this point it is frustrating...
Well.. i think i'm going to mod the cabinet if necessary, because i want the maximum outcome from this build.. i mean not perfect but the best it can do..
Also remember, you are focussing on one axis - on axis. A dip on axis may be filled in off axis as the relative baffle edges shift in relation to the mic (one closer, one further away). This effectively smooths the effect, especially for centred tweeters where an on axis - both sides are diffracting "in unison".
For example, assuming I've done this correctly, baffle diffraction sim - orange = on axis, blue = 45 degrees off axis.
This is another reason we design with on AND off axis measurements. If you optimise for one axis, you may be making it worse for the others and the in room response may be better (or worse) as a result
For example, assuming I've done this correctly, baffle diffraction sim - orange = on axis, blue = 45 degrees off axis.
This is another reason we design with on AND off axis measurements. If you optimise for one axis, you may be making it worse for the others and the in room response may be better (or worse) as a result
In terms of shape options and impact, this old reference can be useful. Although it only shows a cube (equidistant from all 4 edges), a rectangle lessens the peaks / dips as we remove the number of equidistant edges.
If big roundovers are a problem, then large chamfers seem to work quite well. Either 45 degree or 2 * 22.5 degree
If big roundovers are a problem, then large chamfers seem to work quite well. Either 45 degree or 2 * 22.5 degree
If it helps, here's my waveguide mounted SB26ADC tweeter on a trapezoid double chamfered (22.5 degree) baffle:
And baffle in question
I'm pretty sure the SB26ADC is quite smooth through that lower treble region. Therefore the baffle shape, relatively square edges (roundover must be large to have significant effect), symmetrical position and not flush mounting... are all contributing to your baffle diffraction response.
And baffle in question
I'm pretty sure the SB26ADC is quite smooth through that lower treble region. Therefore the baffle shape, relatively square edges (roundover must be large to have significant effect), symmetrical position and not flush mounting... are all contributing to your baffle diffraction response.
You can try close mic (5mm) measurements, they will ignore the box and let you compare to 1m measurements
Last edited:
The dip of the tweeter response is without a doubt because of not being flush mounted.
Absolutely essential.
When estimating your volume for the bass driver I end up with 35-37 liters.
That is too much for SB16PFCR-08.
I'm building with that driver right now and have both in simulation and in real life ended with 16 liter tuned at 38 Hz.
In your bass measurement you also can see baffle step, which needs to be compensated for to get a linear frequency response in bass and midrange.
If I where you, and if my estimation of around 35 liter cabinet volume I would suggest another bass driver in parallell placed at 60% of the hight between floor and current driver and make the design a 2,5 way design and compensate for the baffle step by rolling of the lower bass/midrange earlier than the upper driver.
With your baffle width, baffle step occurs between approx. 150-650 Hz and are in the range of 4dB.
Since SB16PFCR-08 don't have a copper cap on the pole piece I would aim at giving +1dB higher amplitude than bass range and compensate baffle step with 3dB. Full baffle step compensation might end up at bit "dark" in midrange because of no copper cap, cabinet volume and the drivers size.
35-37 liter of volume is also way too much for midrange reproduction.
Two drivers in separate 15-16 liter bass reflex boxes would be better, but for really good midrange reproduction a cabinet volume of 4-8 liters is recommended. Not necessary, but recommended. As a 2-way design bass depth and linearity also is important.
I played around with the tweeter and you should not expect perfectly linear response from SB26ADC. It's not as flat as manufacturers curves indicates.
But it's fairly good, have extremely low distortion and sounds very good. Have some colouration because of underdamped chamber resonances.
Absolutely essential.
When estimating your volume for the bass driver I end up with 35-37 liters.
That is too much for SB16PFCR-08.
I'm building with that driver right now and have both in simulation and in real life ended with 16 liter tuned at 38 Hz.
In your bass measurement you also can see baffle step, which needs to be compensated for to get a linear frequency response in bass and midrange.
If I where you, and if my estimation of around 35 liter cabinet volume I would suggest another bass driver in parallell placed at 60% of the hight between floor and current driver and make the design a 2,5 way design and compensate for the baffle step by rolling of the lower bass/midrange earlier than the upper driver.
With your baffle width, baffle step occurs between approx. 150-650 Hz and are in the range of 4dB.
Since SB16PFCR-08 don't have a copper cap on the pole piece I would aim at giving +1dB higher amplitude than bass range and compensate baffle step with 3dB. Full baffle step compensation might end up at bit "dark" in midrange because of no copper cap, cabinet volume and the drivers size.
35-37 liter of volume is also way too much for midrange reproduction.
Two drivers in separate 15-16 liter bass reflex boxes would be better, but for really good midrange reproduction a cabinet volume of 4-8 liters is recommended. Not necessary, but recommended. As a 2-way design bass depth and linearity also is important.
I played around with the tweeter and you should not expect perfectly linear response from SB26ADC. It's not as flat as manufacturers curves indicates.
But it's fairly good, have extremely low distortion and sounds very good. Have some colouration because of underdamped chamber resonances.
Last edited:
The tweeter do not like to be placed too close to a sharp edge like your top edge.
Distance to top edge is similar to tweeters outer diameter and is probably also contributing to the dip.
If possible I would change position and place tweeter below bas/midrange driver.
Dip at approx. 3000 Hz.
340m/s / 3000 Hz = 11,3 cm. Outer diameter of tweeter mounting frame?
Also note the peak at approx. 6kHz because of 1/2 wave length diffraction.
Both will be reduced quite much with flush mounting, but getting away from the top sharp edge will help even more.
For a test, add 20cm of baffle on top by placing a 20x20cm piece of the MDF you used for the the cabinet.
Place exactly at the edge which will extend your baffle 20 cm upwards.
Measure again and see what happends.
Distance to top edge is similar to tweeters outer diameter and is probably also contributing to the dip.
If possible I would change position and place tweeter below bas/midrange driver.
Dip at approx. 3000 Hz.
340m/s / 3000 Hz = 11,3 cm. Outer diameter of tweeter mounting frame?
Also note the peak at approx. 6kHz because of 1/2 wave length diffraction.
Both will be reduced quite much with flush mounting, but getting away from the top sharp edge will help even more.
For a test, add 20cm of baffle on top by placing a 20x20cm piece of the MDF you used for the the cabinet.
Place exactly at the edge which will extend your baffle 20 cm upwards.
Measure again and see what happends.
Last edited:
Hi.. Thank you for reply..
Very nice looking speakers especially the waveguide for the tweeter.. 😍
Going to try flush mounting the tweeter..
Hello there..
Your reply are really insightful for me..
I think i made too many error with the build...
Now i'm thinking of building new boxes, but its gonna cost a lot and time consuming..
For now i think flush mounting is the best solution i got.. Going to try it first and see the results after..
2.5 way speakers are absolutely be my next project after this..
But i think the crossover trial end error will be much more harder? Matching the phase and get the flattest response...
Thank you so much!