Hello! Im not a big Hi Fi audio person, so sorry if these questions are stupid 
Anyways, I need to design a setup. The setup in question has a Woofer (4.5 ohms where I need it to crossover), a midrange speaker (10 ohms where I need it to cross over), and a tweeter (6.5 ohms where I need it to cross over). One of the speakers (the woofer) has a hump between 300-600 hz, so I need to begin my decent for the crossover before that. For the lower range of the mid, the crossover should be at 600 hz. I would like to use a 2nd order butterworth for all of this with a Q of 0.707.
the higher end is slightly more standard, however, I think i need to use an LPAD filter, because the efficiency of the tweeter is around 6db higher than that of the woofer.
Im starting to design the situation in spice. I have attached what I have so far.
I started to add the LPADs on the midrange speaker, but not the tweeter yet. The problem is that in the graph, the midrange speaker is slanted on the top and I have no idea why. Ive tried using calculators, but every one gives me a different answer. If anyone can help by explaining what I need to do or giving me equations so I can calculate it, that would be great! sorry this is so specific, it is for someone who is extremely particular in what they want. hopefully this doesn't sound too crazy and is possible to do.
Thanks so much and I look forward to hearing from you soon!
Anyways, I need to design a setup. The setup in question has a Woofer (4.5 ohms where I need it to crossover), a midrange speaker (10 ohms where I need it to cross over), and a tweeter (6.5 ohms where I need it to cross over). One of the speakers (the woofer) has a hump between 300-600 hz, so I need to begin my decent for the crossover before that. For the lower range of the mid, the crossover should be at 600 hz. I would like to use a 2nd order butterworth for all of this with a Q of 0.707.
the higher end is slightly more standard, however, I think i need to use an LPAD filter, because the efficiency of the tweeter is around 6db higher than that of the woofer.
Im starting to design the situation in spice. I have attached what I have so far.
I started to add the LPADs on the midrange speaker, but not the tweeter yet. The problem is that in the graph, the midrange speaker is slanted on the top and I have no idea why. Ive tried using calculators, but every one gives me a different answer. If anyone can help by explaining what I need to do or giving me equations so I can calculate it, that would be great! sorry this is so specific, it is for someone who is extremely particular in what they want. hopefully this doesn't sound too crazy and is possible to do.
Thanks so much and I look forward to hearing from you soon!
Looks to me like it's because your bandpass filter is relatively narrow and your inductor/capacitor choice for the low-pass is contributing also. Xsim is suggesting 0.787 mH and 4.02 µF for 2800 Hz, 2nd order, with Q of .707. That gives a more symmetrical shape. Bandpass calculations are often not that straightforward though, so it may need fine tuning in the end to meet a target curve.
Also, your model will be inaccurate if you base it on flat resistance (unless that's what your speaker has - which is typically only true of a planar unit or one that includes impedance equalizing circuits).
Also, your model will be inaccurate if you base it on flat resistance (unless that's what your speaker has - which is typically only true of a planar unit or one that includes impedance equalizing circuits).
Last edited:
In your original circuit, lowering L5 to 0.62 mH gets things closer to your target (in Xsim, anyway).
And no questions like these are stupid. Everyone starts at zero, and many of the details are not obvious, so some guidance is bound to be needed here and there no matter how diligent you try to be.
And no questions like these are stupid. Everyone starts at zero, and many of the details are not obvious, so some guidance is bound to be needed here and there no matter how diligent you try to be.
Last edited:
Xsim is a free crossover modeler by Bill Waslo. It includes CircuitBlocks for some standard crossover functions, which can be helpful for sanity checking basic component values.
https://www.libinst.com/
3D version, which adds off-axis modeling
https://www.diyaudio.com/community/threads/xsim-3d-development-i-could-use-some-math-help.317811/
And a short video showing its use
https://www.libinst.com/
3D version, which adds off-axis modeling
https://www.diyaudio.com/community/threads/xsim-3d-development-i-could-use-some-math-help.317811/
And a short video showing its use
Last edited:
In practice, designing only electrical filters without considering the true acoustic characteristics of each speaker will not give good results.
It is advisable to make acoustic measurements of each speaker, or at least of the midrange and tweeter, and to calculate the enclosure volume for the woofer by knowing its Thiele Small parameters.
And then simulate the filters in VituixCAD or XSim. Then the best results are guaranteed.
It is advisable to make acoustic measurements of each speaker, or at least of the midrange and tweeter, and to calculate the enclosure volume for the woofer by knowing its Thiele Small parameters.
And then simulate the filters in VituixCAD or XSim. Then the best results are guaranteed.
Hello,
I just downloaded Xsim, and I have one major question so far: how do I get the FRD response file and ZMA impedance file? Is this something I can create or are there existing ones that I can use?
And thank you for the tip. How do you advise making these acoustic measurements?
Thanks for all of your help so far!
I just downloaded Xsim, and I have one major question so far: how do I get the FRD response file and ZMA impedance file? Is this something I can create or are there existing ones that I can use?
And thank you for the tip. How do you advise making these acoustic measurements?
Thanks for all of your help so far!
by make acoustic measurements.
you can use REW for this. its free wonderful prog. also you need measurement mic and soundcard.
as for me XSim good but VituixCAD is the best
If the producer of the drivers dont publish them but do provide the graphics, you might want to try this: https://www.diyaudio.com/community/...-zma-file-from-graphic-response-curve.373693/
//
//
In addition to what the others have said above:
It depends on your level of commitment and what kind of performance you are targeting.
There are people that just use textbook values in crossovers and go on their merry way, oblivious of all the details they skipped. Some will add tuning crossovers by ear to this approach (there's a thread for that).
There are those that demand accurate measurements/modeling of all drivers in the actual enclosure, Z offsets, phase, diffraction, polar response, etc., etc.
And, of course, there's everything in between. Each person has their favorite hardware and software to do all this as well.
For measuring, I personally prefer CLIO or another packaged piece of hardware for generation and acquisition because they consistently perform without any real setup. Though this approach costs more, I've gotten a lot of use out of mine, and I have low tolerance for Windows/preamp/driver/sound card/etc. shenanigans. Many successfully use REW as a cheaper alternative.
Keep in mind that accurate acoustical measurement in a home environment is an endeavor in itself. It's not something you just casually pick up and run with. You have to understand the limitations of the methods and how to optimize them based on your environment to produce reasonably accurate results.
Hello again!! I just finished assembling my crossover on a board, and I'm having problems with getting a signal to come through it.
My arbitrary waveform generator is putting out a 3.5kHz sine wave. When I measure directly to the generator with an oscilloscope, I get the expected wave. But as soon as I even touch the input wire to the circuit to this direct connection between the generator and the oscilloscope, the wave immediately dies down. I thought it was a grounding problem; however, when I continuity-checked the input of the circuit to ground, as well as all of the outputs to ground, I get no continuity. My next thought was to disconnect all of the "first line" of components, meaning the first components the signal has to go through to get into the circuit. I did this in hopes of eliminating the possibility of components later in the circuit affecting the input. when I probed it again, I got an extremely fuzzy, sporadically moving sine wave. I then using an alligator clip put the "first line" components one by one back on the board, and the input didn't work for any of those either.
I think this behavior appears to be the signal coming into contact with ground; however, I don't know. There is a definite possibility that I am checking it wrong, but I have no idea. If any of you can advise me on how to troubleshoot this and/or continue, please let me know, your help is much appreciated.
Thanks!!
My arbitrary waveform generator is putting out a 3.5kHz sine wave. When I measure directly to the generator with an oscilloscope, I get the expected wave. But as soon as I even touch the input wire to the circuit to this direct connection between the generator and the oscilloscope, the wave immediately dies down. I thought it was a grounding problem; however, when I continuity-checked the input of the circuit to ground, as well as all of the outputs to ground, I get no continuity. My next thought was to disconnect all of the "first line" of components, meaning the first components the signal has to go through to get into the circuit. I did this in hopes of eliminating the possibility of components later in the circuit affecting the input. when I probed it again, I got an extremely fuzzy, sporadically moving sine wave. I then using an alligator clip put the "first line" components one by one back on the board, and the input didn't work for any of those either.
I think this behavior appears to be the signal coming into contact with ground; however, I don't know. There is a definite possibility that I am checking it wrong, but I have no idea. If any of you can advise me on how to troubleshoot this and/or continue, please let me know, your help is much appreciated.
Thanks!!
Not trying to be insulting, but are you using an amplifier also, or just the waveform generator? If just the waveform generator, what's its output impedance and max voltage? A normal configuration for this kind of test would include an amplifier to deal with the lowish impedance of the speaker.
Any chain that has a capacitor in series will not show continuity when tested with DC.
If none of this applies, a picture of your crossover (both sides) may help.
Any chain that has a capacitor in series will not show continuity when tested with DC.
If none of this applies, a picture of your crossover (both sides) may help.
Thanks for your reply. No, I am not currently using an amplifier. i currently have the amplitude set at 4v peak to peak (I just turned it on and that's what it was at so I didn't change it) and I have no idea what the impedance is, as there is no setting for impedance. I just pulled a lowpass filter off of the board and tested just it point-to-point using alligator clips, and it still doesn't work. There is signal going through the inductor, and I still get the waveform out the other side. But as soon as I attach the capacitor (going to ground), it goes flat. Is this because I don't have an amplifier? is one necessary to test my crossover system?
What signal generator are you using (manufacturer and model)?
Also, it seems like you are implying that you tested the inductor by itself. That gives you very high impedance for the signal generator to work into. If you connect the woofer also, you will likely pull the voltage down again. Similarly, attaching the capacitor does the same thing. I can show this in sims also, but I'm not sure it's worth the energy if it's not what's going on.
Also, it seems like you are implying that you tested the inductor by itself. That gives you very high impedance for the signal generator to work into. If you connect the woofer also, you will likely pull the voltage down again. Similarly, attaching the capacitor does the same thing. I can show this in sims also, but I'm not sure it's worth the energy if it's not what's going on.
Last edited:
I am using a BK precision 4053B signal generator and a Tektronix TBS 1000C Series digital oscilloscope.
Edit: Yes, I did test the inductor by itself, just to see if the inductor is what was causing the signal to disappear. Also, I think I understand why attaching the capacitor to the inductor would lower the voltage, but would it completely drain all of the signal? I'm really not sure. Again, I'm really new to this, and this is actually my first time doing a crossover design/build. Thanks so much for all of your help again, I really appreciate it.
Edit: Yes, I did test the inductor by itself, just to see if the inductor is what was causing the signal to disappear. Also, I think I understand why attaching the capacitor to the inductor would lower the voltage, but would it completely drain all of the signal? I'm really not sure. Again, I'm really new to this, and this is actually my first time doing a crossover design/build. Thanks so much for all of your help again, I really appreciate it.
Last edited:
From the 4053B manual: "For [CH1] and [CH2] connectors on the front panel, the generator has an output impedance of 50 Ω"
I would add an amplifier between your signal generator and the crossover/speaker. You're losing about 18 dB of signal across the internal impedance (at 50 ohms) into the full woofer+lowpass filter. Driving with that high an impedance also affects the filter's transfer function.
Also, if you're still at 3.5 kHz for the test signal, you'll only have about 0.03 volt getting through the woofer crossover with a generator output of about 4 V peak-to-peak and a 50 ohm source impedance (none of these are precise values, since this was just a quick and dirty check).
I would add an amplifier between your signal generator and the crossover/speaker. You're losing about 18 dB of signal across the internal impedance (at 50 ohms) into the full woofer+lowpass filter. Driving with that high an impedance also affects the filter's transfer function.
Also, if you're still at 3.5 kHz for the test signal, you'll only have about 0.03 volt getting through the woofer crossover with a generator output of about 4 V peak-to-peak and a 50 ohm source impedance (none of these are precise values, since this was just a quick and dirty check).
Last edited:
Why use my grandpa old technology when Columbus sailed to america . i think you can easily assemble a simple circuit from a few resistors to measure the impedance of a speaker. and use a microphone to measure its response. it only seems too complicated at the beginning to do this. But once you have done it about 2-3 times, you'll see how easy it is.
Then you will see the full picture of how the speakers behave at all frequencies, and you can build the most suitable crossover.
Realize that the impedance of any speaker is frequency-dependent, and measuring at a fixed frequency is not the way to build a proper crossover. Also, by measuring only the electrical signal you are not taking into consideration the acoustic behaviour of the speaker.
I highly recommend to learn and using the free REW for measurements and VituixCAD / XSim for crossover simulation.
. i think you can easily assemble a simple circuit from a few resistors to measure the impedance of a speaker. and use a microphone to measure its response. it only seems too complicated at the beginning to do this. But once you have done it about 2-3 times, you'll see how easy it is.Then you will see the full picture of how the speakers behave at all frequencies, and you can build the most suitable crossover.
Realize that the impedance of any speaker is frequency-dependent, and measuring at a fixed frequency is not the way to build a proper crossover. Also, by measuring only the electrical signal you are not taking into consideration the acoustic behaviour of the speaker.
I highly recommend to learn and using the free REW for measurements and VituixCAD / XSim for crossover simulation.