I modelled and merged the Baffle Response and Woofer Response. Added the new overall response to the model.
It definitely did have an effect on the lower end. I tried to compensate like you suggested, but by adding two notch filters.
Now we have quite a few components.
How am I able to simplify this? As far as I understand, I should aim to have as little components as necessary.
Also, the impedance is pretty low but doesn't go below 4 ohms, which I guess is okay.
This is probably the reszlt of your (parallel) notch filters.
I suggest you remove those and also the 2,2 ohm series resistor from the woofer, unless that is the resistance of your coil (would be too high) or you explicitly want to massively increase qts of your woofer. You will also have to reduce the tweeter resistor. This results in a clean tuning for the midwoofer and a better efficiency.
Then try to get the desired response by increasing the woofer series coil.
As you are planning to make a very ambitious and maybe expensive enclosure you might also consider using better drivers. Tcp115 is very cheap but has quite high distortion levels. rs125 midwoofer is highly regarded, with maybe a bit less bass extension. Just my 0.02€
Cheap drivers are cool 😉 . I don't pay much attention to distortion normally.. unless it has something to say. I don't think it's very important with all else being good.
I am stuggling with removing a trough right around the crossover freq. Which makes me think, maybe these aren't the best match.
However, I am just a student and trying to keep the cost down. Basically, below 300USD for everything.
Casting the cabinet will not be super expensive. 3D printing the internal support and building a cheap latex mould. I was actually more concerned with the tweeter, but I like the looks of it. hahah
Reverse polarity on one of the drivers (tweeter, usually). This is normal for 2nd order crossovers and might solve the problem.
Looking forward to your progress with casting concrete!
Last edited:
You won't actually be able to get rid of that dip between 3-4kHz because it is inherent to the woofer and because your tweeter choice must be crossed over at or above 4kHz. Your last version shows good phase alignment in the xo region so reversing polarity won't be the answer to the problem. The only way to fix it is to choose a different tweeter that is comfortable crossing at about 3kHz or lower. And that is something I would seriously consider if you haven't purchased the drivers yet.
The x, y and z coordinates are what tell the program the location of your drivers on the baffle. X is the horizontal relationship and should be 0 when drivers are vertically aligned. All good.
Z is the relative difference in acoustic centers and is best measured but can be guestimated sometimes from the spec sheet drawings as the point where the cone/dome meets the voice coil. With the planar tweeter, it'll be the location of ribbon itself and it doesn't look like the ribbon is set back from the frame to any large degree and with the woofer, somewhere in the neighborhood of 25-30mm looks about right albeit the drawing is not really very informative. So the woofer z set to about 25 or 26mm would be my choice but 29mm won't make a huge difference.
Y is the horizontal relationship usually from the center to center of the 2 drivers. With a planar I think you can use the bottom of the ribbon to the center of woofer as your guide (someone correct me on that one if I'm wrong). You have your's set to 0 which is telling the program that the tweeter is directly in the middle of the woofer, also called a coaxial which is not what you've got here. There is also a rule of thumb to more or less try to get your tweeter and woofer vertically as close together as possible especially so when you have a higher xo frequency. Your post #1 drawing has the tweeter and woofer unnecessarily quite a distance apart.
You've learned now how to add in the baffle diffraction to the woofer. You should also do the same for the tweeter for more accurate results.
Lastly, usually you would add in/splice/merge the woofer's LF response due to the box alignment and likewise change up the LF of the woofer impedance response for the same reason, but when working with a 2-way with a fairly high xo point, it won't really make any difference to the xo sim so that one is actually ok to skip in your particular situation.
Have you figured out how to anchor any necessary screws into the cement btw? That can actually be more than a little problematic.
The x, y and z coordinates are what tell the program the location of your drivers on the baffle. X is the horizontal relationship and should be 0 when drivers are vertically aligned. All good.
Z is the relative difference in acoustic centers and is best measured but can be guestimated sometimes from the spec sheet drawings as the point where the cone/dome meets the voice coil. With the planar tweeter, it'll be the location of ribbon itself and it doesn't look like the ribbon is set back from the frame to any large degree and with the woofer, somewhere in the neighborhood of 25-30mm looks about right albeit the drawing is not really very informative. So the woofer z set to about 25 or 26mm would be my choice but 29mm won't make a huge difference.
Y is the horizontal relationship usually from the center to center of the 2 drivers. With a planar I think you can use the bottom of the ribbon to the center of woofer as your guide (someone correct me on that one if I'm wrong). You have your's set to 0 which is telling the program that the tweeter is directly in the middle of the woofer, also called a coaxial which is not what you've got here. There is also a rule of thumb to more or less try to get your tweeter and woofer vertically as close together as possible especially so when you have a higher xo frequency. Your post #1 drawing has the tweeter and woofer unnecessarily quite a distance apart.
You've learned now how to add in the baffle diffraction to the woofer. You should also do the same for the tweeter for more accurate results.
Lastly, usually you would add in/splice/merge the woofer's LF response due to the box alignment and likewise change up the LF of the woofer impedance response for the same reason, but when working with a 2-way with a fairly high xo point, it won't really make any difference to the xo sim so that one is actually ok to skip in your particular situation.
Have you figured out how to anchor any necessary screws into the cement btw? That can actually be more than a little problematic.
Sanity check: Still ridiculous values on the inductors: 33mH is far too large, even for a woofer circuit, (probably should be in the order of 0.33mH or lower for a tweeter); and 470uH is too small to be effective on a woofer, should be in the order of 4.7mH... the problem with simulations is that sooner or later, they meet reality...
try costing a 33mH inductor, and take note of the dimensions
try costing a 33mH inductor, and take note of the dimensions
Agreed, 33mH is far too large just about anywhere. But I think so is 4.7mH for the woofer inductor here when the xo frequency is up around 4kHz. Somewhere between .5-1mH strikes me as in the proper ballpark but also with the appropriate shunt values --> 15ohm with 66uF going to ground are the ones on the woofer that are both kind of in the ridiculous region. No offense intended to the samthebaam - you only learn by doing and getting feedback.
It may help a little to know some electronics here. Any connection to ground without resistance before the driver is essentially a short. Very small inductor and very large cap values going to ground are each letting almost all of the signal through (ie. with very little resistance) which may make your amp very, very unhappy. It may depend on the size of the inductor or the cap in series before the shunt leg as that will determine how much of the signal (ie. current) that each shunt leg may or may not be passing through to start off with.
So try a larger cap and a smaller inductor on the tweeter and maybe increase the woofer inductor slowly after you drop the shunt cap and even get rid of the shunt resistor to start off with. A clear indication as you play with the combination of values that the shunt values are too strong (too small an inductor on the tweeter or too large a cap on the woofer) is that the 'knee' at the rolloff starts to increase in SPL. When that happens stop or pull back the shunt value a little and go back to the series component and decrease it in the case of the tweeter cap and increase it in the case of the woofer.
It may help a little to know some electronics here. Any connection to ground without resistance before the driver is essentially a short. Very small inductor and very large cap values going to ground are each letting almost all of the signal through (ie. with very little resistance) which may make your amp very, very unhappy. It may depend on the size of the inductor or the cap in series before the shunt leg as that will determine how much of the signal (ie. current) that each shunt leg may or may not be passing through to start off with.
So try a larger cap and a smaller inductor on the tweeter and maybe increase the woofer inductor slowly after you drop the shunt cap and even get rid of the shunt resistor to start off with. A clear indication as you play with the combination of values that the shunt values are too strong (too small an inductor on the tweeter or too large a cap on the woofer) is that the 'knee' at the rolloff starts to increase in SPL. When that happens stop or pull back the shunt value a little and go back to the series component and decrease it in the case of the tweeter cap and increase it in the case of the woofer.
Haven't purchased anything yet. Still messing around with it all. I am a total newbie, so all the input I'm getting here is really appreciated.
jReave: Do you have a suggestion for another Tweeter. I was aware it's not the best, but I really liked the rectangular look and it isn't too pricey.
I was also eying the Dayton Audio AMT Mini-8 but when I look at its frequency graph, I don't think it would be any better
So any further recommendations are welcome.
Yes, as I am 3D printing the internals (meaning the internal box that will support the cement walls) I was intending on modelling some mounting brackets right into it.
Mmh I think I will have to look around and maybe start over. Starting with different components. These will however influence the enclosure (required internal volume) and then of course the whole crossover but I am considering it.
Definitely learning a lot and enjoying myself.
Thanks for all the input!
You are correct. The AMT Mini-8 would not work well either. What you need is a tweeter with a resonant frequency, Fs (or a FR that stays more or less flat in the case of a planar/ribbon) at least 1 octave below your expected xo frequency, but preferably more than that. Another way of saying that is that when the xo is in place, you want the Fs to be a minimum of about 24dB down below the fundamental, which means a tweeter with a lower Fs can have a shallower xo slope and one with a higher Fs needs a steeper one.
If you wanted to stick with a planar, perhaps the least expensive one I'd choose is the GRS 3 1/2", but the price doesn't really make sense compared with your woofer cost and it's N/A right now anyways.
Sticking with more traditional tweeters, these 2 don't look too bad in this application:
https://www.parts-express.com/Dayton-Audio-TD25F-4-1-Soft-Dome-Tweeter-4-Ohm-275-022
https://www.parts-express.com/Dayton-Audio-ND25FA-4-1-Soft-Dome-Neodymium-Tweeter-275-059
And SB Acoustics is also known for very good drivers:
https://meniscusaudio.com/product/sb-sb26st-c000-5/
If you wanted to stick with a planar, perhaps the least expensive one I'd choose is the GRS 3 1/2", but the price doesn't really make sense compared with your woofer cost and it's N/A right now anyways.
Sticking with more traditional tweeters, these 2 don't look too bad in this application:
https://www.parts-express.com/Dayton-Audio-TD25F-4-1-Soft-Dome-Tweeter-4-Ohm-275-022
https://www.parts-express.com/Dayton-Audio-ND25FA-4-1-Soft-Dome-Neodymium-Tweeter-275-059
And SB Acoustics is also known for very good drivers:
https://meniscusaudio.com/product/sb-sb26st-c000-5/