I'm experimenting simulating the conversion of 3 ways to 3.5 ways speakers. The objective of this experiment is to arrange the FR in a concept of integrated subwoofer speakers. I started by having a 3 ways speakers with typical 3 drivers; one tweeter, one midrange, and one woofer. Then, I'm going to add a second woofer to the system where the second woofer has its own cabinet.
The problem is I don't know the Target Function of the 3.5 ways speakers since I've never seen it before. I don't know what the ideal FR of the goal should be. So, I'd like to ask experts here for the criteria that would be recommended.
However, I have randomly tweaked the circuit myself. For the 3.5 ways conversion, I doubled the value of the inductor and halved the capacitor of the first stage crossover, and, doubled the second stage inductor again referred to the first stage.
Here is the original 3 ways system:
Here is the modified 3.5 ways:
Here is the simulated impedance plots of the two cases for comparison; blue dotted line = 3 ways, black solid line = 3.5 ways:
I'm quite uncertain if it would work. Therefore, please advice me whether I did it correct or not.
PS. I know it should be done using actual measurements, but this is for research purposes.
The problem is I don't know the Target Function of the 3.5 ways speakers since I've never seen it before. I don't know what the ideal FR of the goal should be. So, I'd like to ask experts here for the criteria that would be recommended.
However, I have randomly tweaked the circuit myself. For the 3.5 ways conversion, I doubled the value of the inductor and halved the capacitor of the first stage crossover, and, doubled the second stage inductor again referred to the first stage.
Here is the original 3 ways system:
Here is the modified 3.5 ways:
Here is the simulated impedance plots of the two cases for comparison; blue dotted line = 3 ways, black solid line = 3.5 ways:
I'm quite uncertain if it would work. Therefore, please advice me whether I did it correct or not.
PS. I know it should be done using actual measurements, but this is for research purposes.
The sim indicates what you would have without room or cabinet effects depending on how you have manipulate the frequency data.
You will also have room gain at these frequencies assuming you want to use it in an enclosed space. You have not included the low frequency roll off of the original 3 way bass driver driver in either sim. How that behaves will affect the low frequency data in the graphs i think.
I imagine that you understand that all your drivers are at the same location in space X, Y and Z, consequently real world data from a few hundred hertz and above are going to have real world effects for phase, driver dispersion and cabinet refraction causing things to be not so perfect.
As you are concerned about Low frequency aspects it might be worth noting that most decent subwoofers have a phase control to allow better integration if people care to use it. Possibly if you were to build and measure this getting the low frequency phase to look reasonable could be a challenge.
The sim is also showing that you have 7mH and 14mH in the bass section of the crossover. I would be considering how to make these myself as they would be expensive in terms of copper and shipping charges for me in UK.
As to a target function, I do not know. I do not work with the low frequency and put up with what the driver gives me. Big bass is not a big issue for me. As your sim data for the other three drivers has nice symmetrical shapes with the roll on and roll of being parallel, I would expect that might be a sensible approach for now and as others have said they like to use a sealed driver bass mid if they are going to integrate with a sub woofer often with an aim of somewhere around 80 Hz crossover point.
Maybe both of us should consider more time in the sub woofer forum 🙂
You will also have room gain at these frequencies assuming you want to use it in an enclosed space. You have not included the low frequency roll off of the original 3 way bass driver driver in either sim. How that behaves will affect the low frequency data in the graphs i think.
I imagine that you understand that all your drivers are at the same location in space X, Y and Z, consequently real world data from a few hundred hertz and above are going to have real world effects for phase, driver dispersion and cabinet refraction causing things to be not so perfect.
As you are concerned about Low frequency aspects it might be worth noting that most decent subwoofers have a phase control to allow better integration if people care to use it. Possibly if you were to build and measure this getting the low frequency phase to look reasonable could be a challenge.
The sim is also showing that you have 7mH and 14mH in the bass section of the crossover. I would be considering how to make these myself as they would be expensive in terms of copper and shipping charges for me in UK.
As to a target function, I do not know. I do not work with the low frequency and put up with what the driver gives me. Big bass is not a big issue for me. As your sim data for the other three drivers has nice symmetrical shapes with the roll on and roll of being parallel, I would expect that might be a sensible approach for now and as others have said they like to use a sealed driver bass mid if they are going to integrate with a sub woofer often with an aim of somewhere around 80 Hz crossover point.
Maybe both of us should consider more time in the sub woofer forum 🙂
Last edited:
The target function is what you want it to be. You define it.
Most people go for a flat response with possibly a shallow roll off above 10kHz or so.
Well-know recording engineer Bob Katz recommends rolling off from 10kHz to 20kHz of 3 or 4dB.
Try one, and if you don't like it, change it, listen again.
Jan
Most people go for a flat response with possibly a shallow roll off above 10kHz or so.
Well-know recording engineer Bob Katz recommends rolling off from 10kHz to 20kHz of 3 or 4dB.
Try one, and if you don't like it, change it, listen again.
Jan
+1
One of the mantas Zaph had when he was active, which is also how I approach multiway design, is worth repeating for the OP, as it echoes what you put in your final line above, Jan:
One of the mantas Zaph had when he was active, which is also how I approach multiway design, is worth repeating for the OP, as it echoes what you put in your final line above, Jan:
As has been said, the only 'target transfer functions' in practice are the ones you create or decide to track for a given system yourself. However, purely electrical assumptions are for the birds anyway, as they go straight out of the window the second you introduce real world drive units, with non-flat frequency, phase responses, frequency varying impedance loads, and are spread over a large baffle area rather than being coincident on all three axis.
That is why I use Acourate, which allows me first to characterise/correct the drivers, then do a sensible xover and finaly correct for the room. All those issues you mention are largely gone. For $ 400, less than a couple of boutique caps.
Jan
Jan
The target for the 3.5 way is simply the baffle step (inverted). Sure you also add that to your preferred voicing, but that's not the question you asked today 😉
The usual way to do 3.5 way in theory like this, is to make a normal flat 3 way, and then add the other woofer with an inductor so it suits the baffle step.
The usual way to do 3.5 way in theory like this, is to make a normal flat 3 way, and then add the other woofer with an inductor so it suits the baffle step.
True, but they ideally need to be addressed anyway, whatever approach to filtering is preferred, since for quality results you really need to work with the on-baffle frequency, phase, impedance responses of the drivers rather than simple fixed resistive loads as shown by the OP.