Read more and sign up Let me give you a quick example on how to transform 2. Then, we need to look at the impedance: So, for the mid-bass driver we have But since the driver is 8 ohms, it has the same efficiency value The tweeter has the efficiency quoted for both types, and the one we are interested in is So in this case, the tweeter has a lower output compared to the mid-bass.
In case the tweeter is louder than the mid-bass, you need to put a resistor in series with the tweeter to tame it down. Actual measurement of the speakers When we actually measure them, we can see that the tweeter is actually louder by a couple of decibels. This can happen for several reasons. First of all, you have to take the parameters quoted by the manufacturer with a grain of salt. Secondly, you have to take into account the acoustic centers of the speakers, which is considered to be in the place of the voice coil.
While both speakers are mounted flush on the baffle, the acoustic center of the mid-bass is around the magnet area, which is behind the tweeter. Therefore, the tweeter is actually in front acoustically , compared to the mid-bass. So, we made a correct assumption just by judging the spec sheet. Anyway, you can always listen to the finished speaker, and subjectively tell if the tweeter is too harsh. If so, adding a resistor in series with the tweeter will tame it down.
You can see a dip in the response between 2 kHz and 4 kHz, for both the mid-bass and the tweeter. This is actually some baffle step issue.
You can fix it in the crossover design. Also, you can off-set the tweeter. Placing it closer to the side of the baffle not in the center , will mitigate this problem. However, we are not aware of this issue as this article assumes we have no measuring tools. As I said in the beginning, compromises will be made, and we shall continue. There are some speakers which are designed with 1st order filters in mind.
However, even those have some issues at certain frequencies particularly the resonant frequency of the driver. Many times, when this solution is selected, a single capacitor is used for the tweeter and nothing else for the mid-bass. However, we are going to test a complete filter, with a high pass filter for the tweeter and a low pass filter for the mid-bass. Being a 1st order filter, I suggest a crossover frequency of at least 6 kHz.
So, the circuit diagram should look like this: R1 is actually the resistance of the inductor not a separate resistor. You can tell that from the icon in the circuit diagram. After hooking all the components up, here is the frequency response: The frequency variation is about 8 dB, which is pretty bad, but forgivable for a crossover which was made blindfolded.
The problem here is that protection for the tweeter is not good. If you have some music rich in frequencies close to tweeter Fs, and you turn the volume up, you might permanently damage the tweeter.
It provides adequate protection for the tweeter, the responses sum up flat and it has only 4 components. Lesser components means less complications in the audio signal path, and of course, cheaper filter.
First of all, we need to choose the crossover frequency. Unlike the 1st order filter, we can go lower in frequency. The resonant frequency of the tweeter is around Hz. Which is insane. Go at least 1 octave above that, so 2 kHz. But just to be safe, I picked the crossover frequency at Hz. Now to calculate the components: Here is the circuit diagram: After setting up this filter, here is the frequency response: We can observe that the mid-bass response is much more linear now.
However, the tweeter is significantly louder. Revised LR2 Now you have to trust your ears since we presumed that no measuring tools are used to tell you if the tweeter is louder. In this case, the tweeter dominance is quite noticeable. Therefore, a series resistor is in order. You can play with different resistor values and listen to how it sounds. In a perfect world it would double the resistance of the tweeter and halve the input power, so the SPL would go down by -3 dB.
I will use a 3. Here is the circuit diagram with the extra resistor: And, again, here is how the frequency response looks like: The response is much more linear, except for the dip at 2. That dip is actually the baffle step effect of both tweeter and mid-bass stacking together, creating that dip. Placing the tweeter offset to the side will reduce this issue. Again, we have no way to tell. However, the response is pretty decent.
Excluding the dip at 2. Flattening the impedance curve We know that when you measure the impedance of a speaker in free-air, you will have a spike at the resonance frequency, and after that, the impedance will progressively go up as frequency increases.
To make the impedance have a more steady value across the frequency range, we will use 2 tricks: Mid-bass Fortunately, there is a separate article I made about the Zobel network , and I suggest you check it out.
The same speaker was used so there is no point in doing this again. In conclusion, the values for the components are as follows: Since the impedance is now flattened for the high frequencies at least, where the filter will take effect , we will not use the nominal impedance 8 Ohms to calculate the crossover components.
Tweeter The tweeter has a big spike at resonance frequency, and we need to flatten that out. Again, there is a separate article concerning the notch filter , which uses this exact same speaker. The values for the components of the notch filter are as follows: Once again, since we flattened the impedance curve of the tweeter, we are going to recalculate the values of the components from the tweeter section. Tweeter attenuation Again we are going to add a 3.
The response looks a bit more linear now. So, here is the diagram again: The dip in the response at 2. You can definitely play with the position of the resistor and see how it sounds. Conclusion and subjective opinion Making passive crossovers for speakers is really an art. A difficult one, I might add. At least having the ability to measure the frequency response is definitely a plus. Looking at the last frequency response, you might be contempt with the result.
However, let me give you the facts. The bass response is not bad, by itself. If you push the volume up, you will get good bass response. But if you do that, the mid-range is so loud, that you need to turn it down.
So, you see, linearity is very important. Having a variance of more than 3 dB across the frequency range is a problem which cannot be ignored. You could use these speakers with a subwoofer. It would be a great combo.
Or you can get yourself some measuring tools and a crossover modeling software and start building crossovers with minimal flaws.
Related Software Links: How Do You Get EPLAN Electric? | Latest Rosetta Stone - Learn Hebrew (Level 1, 2 & 3 Set) Version