Hello. I'm, looking for advice on new 3-way speaker build.

Jump to Latest
J

Jonny English

Member
Joined 2020
#1
Hi - completely new to the forum and speaker building. So much info and help on the forum it is unbelievable. Long story short I want to design and build a 3-way speaker. After much reading about drivers and 'playing' with xsim, REW I came up with the following, trying to keep costs low-ish! Dayton Audio Woofer DSA270-8 10", and Mid-bass RS150P-4A 6" and SB Acoustics Tweeter SB29RDNC

The theoretical crossover looks like the attached. I know that the enclosure will change everything, but intend to work on a ported or TL design floor stander.
3way guess.JPG
I'd really like some opinions on whether this is destined for failure, or has a chance of a decent speaker.

Thanks in advance for any advice.
 
#4
Hi Johny, at a quick glance it is looking good, but the impedance curve is looking off to me.. I'm wondering a bit about the impedance due to the 120uF cap, that could definitely be an issue, but it doesn't look too bad from your posted picture.

The other thing that seems a little unusual is that the tweeter cap is 10uF which seems high for the crossover frequency... also the bump in the impedance at a bit over 3Khz seems out of place too (though I haven't looked at the impedance curve for your tweeter..

Are you sure your impedance files you are using are right?

Tony.
 
#5
Hi,

few thoughts while zipping afternoon coffee 🙂
Summary: I'd lower the crossoverpoints to ~200Hz and ~2kHz



Wouldn't a bit smaller mid driver work better since it is crossed over so high? A 6" driver starts narrowing already at ~2k whereas a ~4" driver would meet the tweeter better at the 3k crossover point dispersion wise?

The other way around would be to lower the mid-hi crossover point, I don't know how low the tweeter can be crossed? Closer to 2k would be better for the 6" mid I think, if the tweeter still sound fine with such low crossover point.

Another is the c-2-c spacing with the woofers to the mid driver. Ideal case seems to be 1/4 wavelength distance at crossover, which would be at 500Hz only < 20cm, one 10" would be ok but the other one is outside of this. So, with the 6" mid and two 10" woofers you could lower the crossover frequency from 500Hz to somewhere closer to 200Hz.
 
Last edited:
#6
I took another look at your circuit as you posted it on another thread, and I can see something I overlooked. So I did the maths and your woofer filter is not as expected.

Could you post your woofer response and impedance without a filter?
 
#9
wintermute said:
Hi Johny, at a quick glance it is looking good, but the impedance curve is looking off to me.. I'm wondering a bit about the impedance due to the 120uF cap, that could definitely be an issue, but it doesn't look too bad from your posted picture.

The other thing that seems a little unusual is that the tweeter cap is 10uF which seems high for the crossover frequency... also the bump in the impedance at a bit over 3Khz seems out of place too (though I haven't looked at the impedance curve for your tweeter..

Are you sure your impedance files you are using are right?

Tony.
Click to expand...
Very good question @wintermute - I will check. The impedance subject is one I need to do some more reading on! Also some suggestions by others about changing the cross over points and the size of the mid driver to 4 or 5.25 in. I actually realised that the mid is 40w RMS which isn't well matched to the other two drivers at 80 - 100 W. So I will have a look at replacing the mid and redo the xover.
 
#10
tmuikku said:
Hi,

few thoughts while zipping afternoon coffee 🙂
Summary: I'd lower the crossoverpoints to ~200Hz and ~2kHz



Wouldn't a bit smaller mid driver work better since it is crossed over so high? A 6" driver starts narrowing already at ~2k whereas a ~4" driver would meet the tweeter better at the 3k crossover point dispersion wise?

The other way around would be to lower the mid-hi crossover point, I don't know how low the tweeter can be crossed? Closer to 2k would be better for the 6" mid I think, if the tweeter still sound fine with such low crossover point.

Another is the c-2-c spacing with the woofers to the mid driver. Ideal case seems to be 1/4 wavelength distance at crossover, which would be at 500Hz only < 20cm, one 10" would be ok but the other one is outside of this. So, with the 6" mid and two 10" woofers you could lower the crossover frequency from 500Hz to somewhere closer to 200Hz.
Click to expand...
Many thanks will have a look at this now
 
#12
I'd be trying something like this on the low pass, it should give you a similar roll off (note I only eyeballed it didn't do a direct comparison) but friendlier impedance...

Tony.
 

Attachments

  • low_pass.png
    low_pass.png
    6.4 KB · Views: 226
#16
I was just trying to work out what your acoustic target was for the woofer to mid crossover. The closest fit I can get to the woofer rollof with the 1.3mH and 120uF is 2nd order bessel at about 400 Hz.

Note to get close to that, the previous suggestion I had won't do it, the 56uF needs to raise to 82uF. The impedance is a bit lower, but still a bit better than with the 120uF..

Tony.
 
#17
Jonny,

You've used on-axis reference files for all drivers. In the actual speaker you will not be on-axis to any driver except the tweeter.

You have also not programmed Z-offsets for the cone drivers. The mid is slightly behind the plane of tweeter and the woofers further back. These will need to be adjusted in your simulation as well.

With two woofers the situation is even more complicated specially with large woofers crossed so high up. Each will have a different Z-offsets, and separate response curves as the listening angles will be different.

There is also zero baffle step - with two large woofers you can use a very small amount of it, maybe 2dB but you still need at least a little bit. On a 12-13" wide baffle you'll have a peak in the middle of the midrange passband, this is also something you need to account for in the crossover.

A design this complex needs measurements of the actual build. In the absence of that you have to at least get the Z-offsets right, factor in baffle effects and get the correct response curves for each driver depending on its 'tilt' away from the listening axis.
 
#18
wintermute said:
I'd be trying something like this on the low pass, it should give you a similar roll off (note I only eyeballed it didn't do a direct comparison) but friendlier impedance...

Tony.
Click to expand...

Thanks Tony. I realised that the frd & ZMA files weren't correct. This threw up some issues with the Mid and tweeter when I used correct ones (using SPL Trace in VituixCAD). Also taking into account have two woofers was probably overkill and in the interests of keeping costs down I've changed the design to:
Scanspeak 2604-832
FaitalPro 6FE200 6"
Dayton DSA270 10"
Resulting in the following.
xSim v2.JPG

Crossovers now at 473 and 1712 . Having modelled the Cab in BoxyCAD the two XO values W-M were 428 and including floor bounce 544. So in the right ball park. I assuming the next steps are to get the drivers in a test baffle in the room and do some actual measurements.

Can you advise if I risk damaging the drivers if connected directly with no xover doing a sine sweep ? Is there a thread or link that explains the best way to measure in order to finalise the xover?

Many thanks again for everyone's comments - I am extremely grateful.

wintermute said:
I was just trying to work out what your acoustic target was for the woofer to mid crossover. The closest fit I can get to the woofer rollof with the 1.3mH and 120uF is 2nd order bessel at about 400 Hz.

Note to get close to that, the previous suggestion I had won't do it, the 56uF needs to raise to 82uF. The impedance is a bit lower, but still a bit better than with the 120uF..

Tony.
Click to expand...
Thanks Tony - the new xover is much closer to 400 now at 473, albeit with a different mid

Sangram said:
Jonny,

You've used on-axis reference files for all drivers. In the actual speaker you will not be on-axis to any driver except the tweeter.

You have also not programmed Z-offsets for the cone drivers. The mid is slightly behind the plane of tweeter and the woofers further back. These will need to be adjusted in your simulation as well.

With two woofers the situation is even more complicated specially with large woofers crossed so high up. Each will have a different Z-offsets, and separate response curves as the listening angles will be different.

There is also zero baffle step - with two large woofers you can use a very small amount of it, maybe 2dB but you still need at least a little bit. On a 12-13" wide baffle you'll have a peak in the middle of the midrange passband, this is also something you need to account for in the crossover.

A design this complex needs measurements of the actual build. In the absence of that you have to at least get the Z-offsets right, factor in baffle effects and get the correct response curves for each driver depending on its 'tilt' away from the listening axis.
Click to expand...
Thanks Sangram - I'm still learning and not sure about how or what needs to be adjusted. I understand the concept about time alignment, but don't want the 'look' of the baffle to have a step in it. By making adjustments to the xover will this compensate so drivers dont need to be physically aligned?
 
#20
Jonny English said:
Thanks Sangram - I'm still learning and not sure about how or what needs to be adjusted. I understand the concept about time alignment, but don't want the 'look' of the baffle to have a step in it. By making adjustments to the xover will this compensate so drivers dont need to be physically aligned?
Click to expand...

Xsim allows you to enter offsets. We usually take the top of the voice coil as the offset distance, assuming tweeter is listening axis. For woofer on top alignments we use 0 offset if woofer is the listening axis.
 
You must log in or register to reply here.
Top Bottom