Trying my luck, and learning, to design a crossover for some drivers that I ordered and going to install in my car. I have chosen the Dayton Audio AN25F--4 and RS100-4.
I want to design a passiv crossover to get the most out of the install and tried my luck in Xsim. Pretty steep learningcurve and a lot of fun trying to adjust, to a point 😉
Maid a screenshot of what I have done so far. But not sure why the overall-line is lower than the original? Also tried to set a crossover between 3000-4000 Hz, but when I want to lower the dB on the tweeter with a resistor it goes to sh**, and it makes a massive spike below 2000 Hz which results in a massive peak og dip in the overall-curve.
Anyone go any tips for a good starting-point?
And, oh yes I am a total rookie!😀
I want to design a passiv crossover to get the most out of the install and tried my luck in Xsim. Pretty steep learningcurve and a lot of fun trying to adjust, to a point 😉
Maid a screenshot of what I have done so far. But not sure why the overall-line is lower than the original? Also tried to set a crossover between 3000-4000 Hz, but when I want to lower the dB on the tweeter with a resistor it goes to sh**, and it makes a massive spike below 2000 Hz which results in a massive peak og dip in the overall-curve.
Anyone go any tips for a good starting-point?
And, oh yes I am a total rookie!😀
your trying to pad at the wrong spot...and this seems a rather complex x-over, i'd start over and try to use the minimum number of components until you make a better correlation as to what is happening when changes are made.
this is the second thread you've opened on the same topic.
this is the second thread you've opened on the same topic.
Well, my last post was not about the design it self but about the software.
I will make a more simple version. But what du you mean by pad the wrong spot?
I will make a more simple version. But what du you mean by pad the wrong spot?
To my understanding, a resistor in series lowers the "volume" on the whole signal. So resistor R2 lowers the dB on the whole, and R1 lowers it on the signal in the interval set by C4 and L3.
Keeper in mind, I am very new to crossover and trying to learn.
Keeper in mind, I am very new to crossover and trying to learn.
i'm likely not the best teacher but doesn't that change your source impedance?
pads can be series or parallel.
pads can be series or parallel.
Yes, it increases the resistanse and therefor the dB on the driver become lower. Like restraining the level of the sound.
The tweeter plays much louder than the midrange, so the tweeter-level has to be reduced.
The tweeter plays much louder than the midrange, so the tweeter-level has to be reduced.
well i'm not sure what to say, your system response looks anything but flat, i had a look at the drivers published responses and wonder why you even need to add a tweeter to the RS 100.
trying to design a x-over when there is no measurement of the driver installed in a box or in this case your car is only going to lead you farther off course but that's just my opinion.
trying to design a x-over when there is no measurement of the driver installed in a box or in this case your car is only going to lead you farther off course but that's just my opinion.
This is what your filter is doing....meaning, if you had a perfectly flat tweeter and put your filter in place, your tweeter response would look like this. Decreasing the 3 ohm resistor in the series notch will deepen that valley.
Adding series resistance does increase the source impedance as seen by the driver, and thus decreases the damping factor. Typically, that is more of a concern for a woofer, not a tweeter.
I have had success with a series resistor being the first component of the tweeter filter. I also have had success using an L-pad on the tweeter side. Neither approach is wrong. Heck, for the Elac UBR62 speakers Andrew Jones, a highly respected speaker designer, put a series resistor between the two capacitors of a third order filter on the tweeter circuit.
Last edited:
I agree.
Mount the drivers, then take frequency response and impedance measurements, then design the crossover.
Having said that, with the cost of quality passive crossover components continually getting more expensive, good performance amplfiers and DSP getting less expensive (at least in home audio it is), the tuning power of DSP, and the performance tradeoffs of using inductors, the case for skipping the passive crossovers altogether and going all active keeps getting more compelling. Indeed, I don't foresee myself ever again using passive crossovers.
Thanks for very good feedback 🙂
I actually have a DSP amplifier in the car, but it does not have enough channels to "DSP" every speaker. It is all about comprimises, unfortnately.
So I thought I would have some passive crossover on the front speakers to kind of "control" more of the frequency and then use the power of DSP. Maybe I am better of just haveing a 1-way crossover going to the tweeters 🙂
But, good fun learning some new stuff 😉 And DSP is not cheap here in Norway, at least I think.
It's a BMW F20 1-series and a Matchup 7BMW DSP amplifier by the way.
I actually have a DSP amplifier in the car, but it does not have enough channels to "DSP" every speaker. It is all about comprimises, unfortnately.
So I thought I would have some passive crossover on the front speakers to kind of "control" more of the frequency and then use the power of DSP. Maybe I am better of just haveing a 1-way crossover going to the tweeters 🙂
But, good fun learning some new stuff 😉 And DSP is not cheap here in Norway, at least I think.
It's a BMW F20 1-series and a Matchup 7BMW DSP amplifier by the way.
It is a good idea.
How will the mid and tweeter be mounted, or what will the distances apart be.
Ideally they should be mounted as close as possible to each other.
Normally in simulation we need to account for the center to center spacing
If vertical mounting, and of course also horizontal positions as well.
So the overall phase of the drivers is somewhat realistic.
Also the " baffle" or likely the door which mounted too would be simulated.
Normally we model in free space or full space because the baffle provides half space loading.
So the responses will be much much different than a factory FRD
Depending on the size of the baffle and the diffraction from the baffle.
As far as the initial issue, the massive spike can result from inductor values
And padding methods. Usually a parallel resistor is also added with the driver.
So it behaves more like a typical L pad / Divider.
The amount of padding is much like a typical L pad / voltage divider.
But you can also control the impedance.
Not much different than using a 4 ohm or 8 ohm or even 16 ohm L pad to change the impedance.
It is possibly to pad before or after the filter, of course the series resistor will change the behavior of the filter.
Sometimes it is beneficial to pad before or after the filter with a series resistor. But there is additional impedance
control by adding the second parallel resistor
The massive response dip is likely a phase issue at the crossover point.
Depending on the filter orders you can either flip the polarity of the tweeter or woofer.
Or change the crossover point. So their is no phase cancellation.
This is why as mentioned in the beginning, The driver positions to each other need to be defined.
So you are working with realistic phase relations.
How will the mid and tweeter be mounted, or what will the distances apart be.
Ideally they should be mounted as close as possible to each other.
Normally in simulation we need to account for the center to center spacing
If vertical mounting, and of course also horizontal positions as well.
So the overall phase of the drivers is somewhat realistic.
Also the " baffle" or likely the door which mounted too would be simulated.
Normally we model in free space or full space because the baffle provides half space loading.
So the responses will be much much different than a factory FRD
Depending on the size of the baffle and the diffraction from the baffle.
As far as the initial issue, the massive spike can result from inductor values
And padding methods. Usually a parallel resistor is also added with the driver.
So it behaves more like a typical L pad / Divider.
The amount of padding is much like a typical L pad / voltage divider.
But you can also control the impedance.
Not much different than using a 4 ohm or 8 ohm or even 16 ohm L pad to change the impedance.
It is possibly to pad before or after the filter, of course the series resistor will change the behavior of the filter.
Sometimes it is beneficial to pad before or after the filter with a series resistor. But there is additional impedance
control by adding the second parallel resistor
The massive response dip is likely a phase issue at the crossover point.
Depending on the filter orders you can either flip the polarity of the tweeter or woofer.
Or change the crossover point. So their is no phase cancellation.
This is why as mentioned in the beginning, The driver positions to each other need to be defined.
So you are working with realistic phase relations.
Last edited:
I will take measurements when I get the drivers in the car and see how they do 🙂
All mounted in the factory location because I like the originale style, see pictures (not my car, just copy frim the the internetto show location).
And what does it mean when you say "pad"?
All mounted in the factory location because I like the originale style, see pictures (not my car, just copy frim the the internetto show location).
And what does it mean when you say "pad"?
Agreed. This has motivated me to run some comparative simulations - https://www.diyaudio.com/community/...overs-without-measurement.189847/post-7938256
So, what the best advise us actually to use a one way crossover to protect the tweeter only?
The reason I was thinking a two way is that it is easier to EQ the drivers because each driver get the frequencyrange that is best for the placement and performance. The mids in the doors is at an hard angle from the listening position and I guess the high frequency drops off radically anyway, so better to let the tweeters play the high frequency for a better soundstage. Not that they are 100% on-axis but at least a lot better than the mids in the doors.
The reason I was thinking a two way is that it is easier to EQ the drivers because each driver get the frequencyrange that is best for the placement and performance. The mids in the doors is at an hard angle from the listening position and I guess the high frequency drops off radically anyway, so better to let the tweeters play the high frequency for a better soundstage. Not that they are 100% on-axis but at least a lot better than the mids in the doors.
Yeah, hi, I think that the off-axys loss of the treble might lead to factory settings, without the Crossover.
The tweeter might go well with that crossover, omitting the LCR, and yes, try 3, 4 and 5 kHz...also equal C in the CLC high pass usually do the trick
The tweeter might go well with that crossover, omitting the LCR, and yes, try 3, 4 and 5 kHz...also equal C in the CLC high pass usually do the trick
It has been decades, but the last time I used a series resistor on the tweeter side, I also used a resistor in parallel with the tweeter (I refer to the two resistors together as an L-pad). The goal was to, in addition to bringing down the overall tweeter output, reduce (to a small extent) deviations in the tweeter's impedance as seen by the input side. Note that when the impedance of the tweeter and L-pad system are changed, for example by changing one or both of the resistance values, the other filter component values need to be changed if the same general filter shape is desired.
I wouldn't. Here is why, in my opinion:
1. You want a filter on the woofer because you don't want the woofer playing up into the frequency range where it begins breaking up. Woofers typically cannot be run full range without an audibly negative affect due to cone breakup. Personally, I have not heard any that can.
There there are full range drivers, but they usually are used without a tweeter. GR Research has a full range driver that is supposed to be pretty good, but even if its cone breakup is not audible, it still is going to suffer from beaming at higher frequencies (see below).
2. You don't want the woofer playing up into the frequency range where it starts beeming. This may be especially problematic in a car where the ears of people located in different seats are at different angles with respect to the direction the woofer is pointing. If you optimize the sound for the driver's seat, the sound in the passenger seat will not be very good.
Now, there are controlled dispersion designs where the frequency response off axis stays pretty neutral, and mosly just the SPL changes as you move off axis. This can be an advantage in a car where you always are sitting closer to once side or the other. You are more off-axis to the closest woofer/tweeter than to the woofer/tweeter on the opposite side of the car, so this balances out the perceived SPL. But, that is a whole new rabbit hole to go down, and I personally have not gotten into it.