Hello all, I have a question on crossover design.
I'm putting together a 3-way using the SB Acoustics SB15CAC as the mid and the Dayton RS225-8 as the woofer. I've measured the driver responses in the box, and have been modeling with VituixCAD to design a draft of the crossover design.
My question(s) - for the woofer/mid crossover, I need some hefty capacitors (from 80 to 120 uF). Can I construct the initial design of the crossover using nonpolarized electrolytics to make sure I have the frequency response correct (and then go to more pricey/better capacitors later)? I'd hate to drop $200 on caps and then find out I need a slightly different value...
I'm putting together a 3-way using the SB Acoustics SB15CAC as the mid and the Dayton RS225-8 as the woofer. I've measured the driver responses in the box, and have been modeling with VituixCAD to design a draft of the crossover design.
My question(s) - for the woofer/mid crossover, I need some hefty capacitors (from 80 to 120 uF). Can I construct the initial design of the crossover using nonpolarized electrolytics to make sure I have the frequency response correct (and then go to more pricey/better capacitors later)? I'd hate to drop $200 on caps and then find out I need a slightly different value...
EDIT: I misread!
NP electrolytics can, indeed, be used until you want to splash out on more expensive NP film caps.
NP electrolytics can, indeed, be used until you want to splash out on more expensive NP film caps.
Next question about nonpolarized electrolytics... How do they affect the sound as compared to better caps?
It's hard to describe what I'm hearing. Sounds a little like the mids are coming through a tunnel - kind of mushy, and not something that I think I can explain through the measured frequency response.
It's hard to describe what I'm hearing. Sounds a little like the mids are coming through a tunnel - kind of mushy, and not something that I think I can explain through the measured frequency response.
You should connect the negative leads of to identical electrolytics, this way you get a bipolar with half the capacitance of a single.
What you hear is subjective. Do not judge until you have compared to the same value of audio capacitors.
A good idea is to replace 10% of the bibolar with a film capacitor. In most cases possible negative effects disapear.
The best advice I can give you, do A-B testing with very short breaks if you want to stay objective. Once you have infected with the Voodoo virus you are lost.
What you hear is subjective. Do not judge until you have compared to the same value of audio capacitors.
A good idea is to replace 10% of the bibolar with a film capacitor. In most cases possible negative effects disapear.
The best advice I can give you, do A-B testing with very short breaks if you want to stay objective. Once you have infected with the Voodoo virus you are lost.
In my opinion, the "mushy" midrange sound you describe is not something that will eliminated simply by changing from electrolytic to film caps.
I would look to the crossover design rather than the capacitor type.
So, please let us see your crossover design.
I would look to the crossover design rather than the capacitor type.
So, please let us see your crossover design.
Thanks for that info. I'm wondering what directions I need to tweak the speakers next - crossover (and/or components), stuffing, maybe driver breakin...
I do have a little tweaking to do yet with the crossover design - the measured frequency response (of all three drivers w crossover) still has a couple of bumps, notably a dip near the woofer-mid crossover frequency and is a little "hot" above 3 kHz. The 40 and 110 Hz up & down are room modes; the 150 Hz suckout changes with how far the cabinets are in front of the wall. (I've tested the polarity of the drivers by reversing leads, and get a pretty steep dip at the crossover frequency, so I don't have a polarity issue.) The measured distortion with REW and a umik-1 is below the noise floor.
Here is the modeled response using the measured frequency responses of the drivers (my measurements of the drivers used, measured in the cabinet). The tweeter has some distortion below 1500 Hz, so I've used a 3rd order there to minimize that contribution, otherwise everything is 2nd order. I started with crossover components using an online tool then tweaked the values until I got a reasonably level response. My planned next step was to change components to adjust the tweeter level, the peak/trough at the mid/tweet crossover point, and the trough at the woofer/mid crossover point.
Last edited:
I started this journey when I fixed a friend's vintage turntable and, in connecting it to my system, discovered that I needed a new turntable because that vintage one was so much better. When my new turntable arrived, the albums sounded great, but that new turntable made it clear that the speakers needed upgrading. For all I know, right now, it might be that the new speakers are highlighting my crappy amp. 😉
If you want to use such a x-over simulation, all depends on the input. If your input is not the speakers response, but has the rooms fingerprint, it will not work.
Try to do measurements that have no room influence. Burn in will not change anything you can measure.
Try to do measurements that have no room influence. Burn in will not change anything you can measure.
Well, given the data of Fs=35 Hz and Xmax=5mm that's a midwoofer. The cone material and the size of the dustcap might imprint the 'mushi' sound.
I'd scale down the design and use it as the woofer. Candidate for the midrange: Visaton B80 ! Or....
I have used the SB15CAC as a mid range and I don't find it mushy. My guess is the crossover is the problem. I noticed that you a z offset of 2mm on the woofer and 0mm on the mid range. They should be a lot more, say 13mm for the mid and 24 mm on the woofer. I crossed the SB15CAC at 300hz and 3Khz. It looks like you are crossing at ~400hz and ~2.3khz? I would try giving the mid a larger pass band and lower the the padding resistor. The mid is rated at 86db and the woofer is rated at 86.8db, the 3.9 ohm resistor looks a little big to me. The reason for increasing the band pass is the rising shape of the mid range response (the green trace) in your sim, if you put the low pass and high pass filters too close together you start to see the hump that the high end of the response.
You fell for the usual trap: Developing a 3-way speaker. Something not as simple as it might seem. In practice a very complicated thing.
Maybe take another approach, develop a nice 2-way from the tweeter and mid woofer, then when you are satisfied, add a low cut and a sub woofer.
Then you find out that it is not even better sounding to go active for the sub, but even cheaper than buying appropriate passive components. Last, with an x-over of 200 Hz there is no need to have stereo sub's. So one sub amp will do.
The mid woofer will play just fine, at least down to 120 Hz.
Maybe take another approach, develop a nice 2-way from the tweeter and mid woofer, then when you are satisfied, add a low cut and a sub woofer.
Then you find out that it is not even better sounding to go active for the sub, but even cheaper than buying appropriate passive components. Last, with an x-over of 200 Hz there is no need to have stereo sub's. So one sub amp will do.
The mid woofer will play just fine, at least down to 120 Hz.