Hello
I got two pairs of Hertz Dieci DSK 130.3 to install in my Renault Laguna II. Front and back doors allow only for 13cm speakers so I have to live with this.
After I installed them I must say I really like the mids that come out of this thing but the tweeters are killing me.
The crossover that comes with the set is total crap. A second order passive crossover only for the tweeter. The crossover box has both outputs for tweeter and mid/woofer but the mid/woofer connection is just a straight Wire from the input. The whole thing is a joke, using a ferrite core inductor and an electrolytic cap.
Now, the datasheet does state that the mid/woofer has a SPL rating of 93 and the tweeter of 94 but there's absolutely no attenuation/l-pad for the tweeter. As you can imagine, since the mid/woofer also reproduces the high frequency there's a highs nightmare in my car.
Now I took it upon myself to design a proper crossover for this thing, and I went for a second order Linkwitz-Riley alignment for both mid/woofers and tweeters.
I used the information presented here by Rod Elliott (great guy!):
Passive Crossover Network Design
I will use air-core inductors from Jantzen (cheap and good) and a mix of film caps. For the capacitors that are inline with the signal for the tweeter I am going with a polypropylene cap - Jantzen Cross Cap (cheap and I buy them from the same place where I get the inductors).
For the caps in the mids section I'll go with Wima polyester caps (real cheap).
I decided to make an L-pad for the tweeters so I can tame them a bit. I haven't decided on the values yet as even if it's just a 1dB difference between the units, either the mid/woofer Speaker is rated higher or the tweeter is rated lower. (yes, they don't say if the SPL is rated at 1W/1meter or 2.83V/1 meter. even so, doesn't matter as either way there's 1dB difference between them).
I did measure the tweeter and I found that it has a reasonably flat impedance. Wasn't expecting that but for sure it doesn't need a correction network.
The mid/woofer does need a Zobel network and it's calculated at about 3.1-3.3ohms/21-22uF. I did measure it's impedance and +3db frequency and the excel sheet spits out these values.
Since the original crossover has 3.5KHz crossing point I chose the same value for mine.
With the Zobel network in place the mid/woofer has an impedance of 3.25ohms at 3.5KHz and the tweeter has a nominal impedance of 3.97 at the same frequency (pretty accurate).
Now, with these values and the crossover frequency the components values came out like this:
For a 12dB/octave filter:
Tweeter: 5.73uF (5.6uF+0.1uF should do)
361uH (370uH inductor with one turn unwounded).
Mid/Woofer: 7uF(6.8uF+0.22uF should do)
295uH (300uH should do)
For the L-pad I will get 3 sets of simulated values.
For the nominal 1dB difference (with mid/woofer inductor losses accounted for):
Series resistance: 0.57ohms (0.56R available)
Parallel resistance: 18.74ohms (18R available)
For 2dB difference:
SR - 0.86R
PR - 11.04R
For 3dB difference:
SR - 1.12R
PR - 7.55R.
I will get the next standard values I find and make some tests and see how much I need to kill the highs.
Even with a separate amplifier there will be no need for higher than 5W rating resistors (for Lpad/Zobel).
The L-pad must be after the filter on the tweeter so there's no extra heat dissipated for no reason.
What I don't know for sure is...should the Zobel be before or after the filter on mid/woofer?
I attached the schematic for both filters and a picture of the stock filter.
Since on the front the mids are in doors and tweeters on the corner of the dash I need to make separate units for each driver. Hope I get to post some pictures with the finished product.
Hope Hertz answers my email where I asked them to provide graphs for impedance and frequency response of the drivers. Also to clarify if the stated SPL is at 1W/1meter or 2.83V/1meter.
I need to design the pcbs now.
I got two pairs of Hertz Dieci DSK 130.3 to install in my Renault Laguna II. Front and back doors allow only for 13cm speakers so I have to live with this.
After I installed them I must say I really like the mids that come out of this thing but the tweeters are killing me.
The crossover that comes with the set is total crap. A second order passive crossover only for the tweeter. The crossover box has both outputs for tweeter and mid/woofer but the mid/woofer connection is just a straight Wire from the input. The whole thing is a joke, using a ferrite core inductor and an electrolytic cap.
Now, the datasheet does state that the mid/woofer has a SPL rating of 93 and the tweeter of 94 but there's absolutely no attenuation/l-pad for the tweeter. As you can imagine, since the mid/woofer also reproduces the high frequency there's a highs nightmare in my car.
Now I took it upon myself to design a proper crossover for this thing, and I went for a second order Linkwitz-Riley alignment for both mid/woofers and tweeters.
I used the information presented here by Rod Elliott (great guy!):
Passive Crossover Network Design
I will use air-core inductors from Jantzen (cheap and good) and a mix of film caps. For the capacitors that are inline with the signal for the tweeter I am going with a polypropylene cap - Jantzen Cross Cap (cheap and I buy them from the same place where I get the inductors).
For the caps in the mids section I'll go with Wima polyester caps (real cheap).
I decided to make an L-pad for the tweeters so I can tame them a bit. I haven't decided on the values yet as even if it's just a 1dB difference between the units, either the mid/woofer Speaker is rated higher or the tweeter is rated lower. (yes, they don't say if the SPL is rated at 1W/1meter or 2.83V/1 meter. even so, doesn't matter as either way there's 1dB difference between them).
I did measure the tweeter and I found that it has a reasonably flat impedance. Wasn't expecting that but for sure it doesn't need a correction network.
The mid/woofer does need a Zobel network and it's calculated at about 3.1-3.3ohms/21-22uF. I did measure it's impedance and +3db frequency and the excel sheet spits out these values.
Since the original crossover has 3.5KHz crossing point I chose the same value for mine.
With the Zobel network in place the mid/woofer has an impedance of 3.25ohms at 3.5KHz and the tweeter has a nominal impedance of 3.97 at the same frequency (pretty accurate).
Now, with these values and the crossover frequency the components values came out like this:
For a 12dB/octave filter:
Tweeter: 5.73uF (5.6uF+0.1uF should do)
361uH (370uH inductor with one turn unwounded).
Mid/Woofer: 7uF(6.8uF+0.22uF should do)
295uH (300uH should do)
For the L-pad I will get 3 sets of simulated values.
For the nominal 1dB difference (with mid/woofer inductor losses accounted for):
Series resistance: 0.57ohms (0.56R available)
Parallel resistance: 18.74ohms (18R available)
For 2dB difference:
SR - 0.86R
PR - 11.04R
For 3dB difference:
SR - 1.12R
PR - 7.55R.
I will get the next standard values I find and make some tests and see how much I need to kill the highs.
Even with a separate amplifier there will be no need for higher than 5W rating resistors (for Lpad/Zobel).
The L-pad must be after the filter on the tweeter so there's no extra heat dissipated for no reason.
What I don't know for sure is...should the Zobel be before or after the filter on mid/woofer?
I attached the schematic for both filters and a picture of the stock filter.
Since on the front the mids are in doors and tweeters on the corner of the dash I need to make separate units for each driver. Hope I get to post some pictures with the finished product.
Hope Hertz answers my email where I asked them to provide graphs for impedance and frequency response of the drivers. Also to clarify if the stated SPL is at 1W/1meter or 2.83V/1meter.
I need to design the pcbs now.
Attachments
Forgot to mention that I will be taking out everything before 125Hz from the head-unit. It uses an active crossover. I still have separate bass outputs on the back (line level) that I'll feed to a subwoofer that I'm going to get in the future.
I am thinking of a sealed active subwoofer. So far I am oscillating between Alpine SWE-815 and SWE-3200.
The SWE-815 is advertised as a 100W RMS unit but after some german testing website took the amp apart they found a 2x28W amplifier inside. It's bigger brother is 150W but true to the spec (and twice as expensive here).
I am thinking of a sealed active subwoofer. So far I am oscillating between Alpine SWE-815 and SWE-3200.
The SWE-815 is advertised as a 100W RMS unit but after some german testing website took the amp apart they found a 2x28W amplifier inside. It's bigger brother is 150W but true to the spec (and twice as expensive here).
Hi,
Your over complicating it. Just L-pad the tweeter after the original x/o.
Sadly Rod E. knows a lot more about electronics than he does
loudspeakers, and the advice is over-simplistically wrong, whilst
being true in theory, its simply not the way to go about it.
rgds, sreten.
Your over complicating it. Just L-pad the tweeter after the original x/o.
Sadly Rod E. knows a lot more about electronics than he does
loudspeakers, and the advice is over-simplistically wrong, whilst
being true in theory, its simply not the way to go about it.
rgds, sreten.
Last edited:
Zobel goes next to the driver, so that the filter sees a flatter load impedance. But a parallel resistor as part of an L pad or attenuator should also make the load more resistive, which might be good enough. Best to breadboard and listen, and measure if you can.
I already ordered the parts so I might as well go ahead and do the crossover.
How come the more expensive components kits have a proper crossover? Some of them have a see-through plastic case and I can see two inductors, two caps and a few resistors for L-pad.
These speakers are not exactly bottom line ones (in the grand scheme of things) and I thought a proper filter might improve their sound.
In the future I will also add a 4x60W or so amp and use the Head-Unit for low level signal only.
Thanks for the info.
I attached some pictures with some of Hertz's crossovers (for more expensive models). They do seem to make them, they shaved some cost on mine.
Attachments
Hi,
To design a proper x/o you need to take driver responses into account.
An on paper ideal electrical x/o with zobels is rarely the right solution.
rgds, sreten.
To design a proper x/o you need to take driver responses into account.
An on paper ideal electrical x/o with zobels is rarely the right solution.
rgds, sreten.
Well I did measure the speakers.
Used the 10R series resistor mentioned in that sheet. Then I measured the impedance at the crossover point. Then I added the Zobel and measured the impedance again.
On the scope it had a pretty flat impedance from 100Hz to 10000Hz. Had a bump at Fe but yeah.
The tweeter had a pretty flat impedance without a correction network and I just left it as it is.
Used the 10R series resistor mentioned in that sheet. Then I measured the impedance at the crossover point. Then I added the Zobel and measured the impedance again.
On the scope it had a pretty flat impedance from 100Hz to 10000Hz. Had a bump at Fe but yeah.
The tweeter had a pretty flat impedance without a correction network and I just left it as it is.
Can you share some ideas on this matter? I would study a bit. I'm curious as to what else can be done with a quantifiable improvement.
The problem is that filters designed on paper don't take into account that de frequency response of a loudspeaker is not linear.
In your case, the midrange speaker will roll of at a certain frequency. That is why the manufactor didn't bother to ad a lowpass part on the crossover for that midrange.
Ofcourse you can add a nice filter to make that roll over at a lower frequency or to make that roll off smoother. And yes, that can make that speaker sound better.
But a good designed crossover has to take into account the frequency response of the connected speaker for the combined output of the speaker and the crossover to be right.
For example i once designed a passive crossover for a focal composet. I had to make a first order lowpass filter. Together with the speakers own roll off the result was a perfect 2nd order lowpass that matched the 2nd order highpass for the tweeter.
Just start with the lpad to tame those tweeters a bit and go from there. But don't be fooled
In your case, the midrange speaker will roll of at a certain frequency. That is why the manufactor didn't bother to ad a lowpass part on the crossover for that midrange.
Ofcourse you can add a nice filter to make that roll over at a lower frequency or to make that roll off smoother. And yes, that can make that speaker sound better.
But a good designed crossover has to take into account the frequency response of the connected speaker for the combined output of the speaker and the crossover to be right.
For example i once designed a passive crossover for a focal composet. I had to make a first order lowpass filter. Together with the speakers own roll off the result was a perfect 2nd order lowpass that matched the 2nd order highpass for the tweeter.
Just start with the lpad to tame those tweeters a bit and go from there. But don't be fooled
Me think it's because the speakers are mounted very off-setted so there's the natural off-axys response roll-off 🙄
The mids have a pretty high frequency response. On my bench the ear said it was pretty up in volume at 10khz.
I just received the frequency response and impedance graphs for the tweeter from Hertz directly. Pretty nice of them.
I'm still waiting for the same graphs for the mid/woofer.
Attachments
Last edited:
Seems Hertz came through and delivered the frequency response of both drivers. Didn't deliver the impedance graph for the woofer but it's still good enough.
What I find weird is that the graph shows a crossing point of 5.8KHz, but the spec sheet says it's actually 3.5KHz with the stock crossover.
I will still use 3.5KHz in my design.
edit:
looking better at the graph I think I should choose a 5KHz crossing point for a more linear response.
What do think?
What I find weird is that the graph shows a crossing point of 5.8KHz, but the spec sheet says it's actually 3.5KHz with the stock crossover.
I will still use 3.5KHz in my design.
edit:
looking better at the graph I think I should choose a 5KHz crossing point for a more linear response.
What do think?
Attachments
Last edited:
I think it's better to keep the 3.5KHz crossover point as it's a bit lower than the mid/woofer peak.
How much should I attenuate the tweeter by? I see the mid/woofer is at an average of 90dB and the tweeter goes at 98dB or so. Should I go for 7-8dB?
How much should I attenuate the tweeter by? I see the mid/woofer is at an average of 90dB and the tweeter goes at 98dB or so. Should I go for 7-8dB?
Since I already ordered the coils for a 3.5KHz crossover point, I only replaced the caps values.
I decided in the end to go for 4KHz. The Fs of the tweeter is 2.8KHz, and the peak of the mid/woofer is at 5KHz, so I decided in the middle is best. That's 4KHz. Not the best thing but it's the best in this situation.
I still need to decide for the tweeter attenuation. If no one has any input, I'll go for 7-8dB.
I need to unwind the coils when I get them. Their value needs now to be lower for 4KHz vs 3.5KHz. Better to unwind than to need a larger value 🙂
After I get all the parts, I need to test and then make the pcb.
I decided in the end to go for 4KHz. The Fs of the tweeter is 2.8KHz, and the peak of the mid/woofer is at 5KHz, so I decided in the middle is best. That's 4KHz. Not the best thing but it's the best in this situation.
I still need to decide for the tweeter attenuation. If no one has any input, I'll go for 7-8dB.
I need to unwind the coils when I get them. Their value needs now to be lower for 4KHz vs 3.5KHz. Better to unwind than to need a larger value 🙂
After I get all the parts, I need to test and then make the pcb.
- Status
- Not open for further replies.