Jump to content

Crossover help


Lynx4

Recommended Posts



3 hours ago, VanArn said:

I looked up the basic specifications for the speakers used in this project and it would seem that the crossover circuit  by Lynx4  is on the weird side. Perhaps starting with a conventional  Butterworth or Linkwitz/Riley  second order filter would be a better approach.  For a 3kHz crossover frequency with a four ohm impedance, the inductance value is 0.3mH for both coils and the capacitors are 9.4 mFd in the case of the Butterworth  filter. An impedance correction network is required for the  low frequency section and using the impedance and phase  derived from the D.A.T.s  graph from the  Parts Express site ( and trusting that these are accurate ) a series network consisting of a  resistor of 4.5 ohm  /10 watt and a 36 mFd  capacitor placed in parallel across the bass speaker', would suit.  A "L" pad  is required to be fitted to the tweeter to allow the sensitivity to match that of the bass speaker and I would suggest a series resistor of  1.3 ohm and a parallel resistor of 10 ohm would be suitable values to use. The inductors should be wound with a heavy gauge wire, say 1.4mm, to avoid a high d,c. resistance loss.

Awesome thank you. I'll sketch that out unless you already have a copy of the circuit diagram? 

Thanks 

Link to comment
Share on other sites

18 hours ago, Lynx4 said:

if I did something wrong or not. 

Not exactly (and not as "totally unworkable" as others have commented) ;) 

 

The design of your crossover will heat this resistor, so hot, means everything is "normal".

 

I jumped into a quick sim with this driver, and I can see how you would be tempted to go that way with this driver.   A few comments:

 

3ohm // and the series inductor together mean you have an impedance of 2ohms (as you can see in your sim).   This is quite low, but no problem if your amplifier can handle it stably (many can't).    Will it mean more "distortion"?   In general, no.  Only if your amplifier is clipping, or if you believe that resistors have audible distortion.   It does make the load on the amplifier a lot "easier" (less reactive), as you can see by the flatter impedance phase when you add the parallel R (the light grey line).

 

Are you using the datasheet rfd (frequency response) file? .... This problem doesn't include (enough) cabinet diffraction (aka. "baffle step") and so your 3dB rise below 1khz might not be enough, and the speaker might sound a bit "thin".

 

Ideally you should be simulating the speaker based on acoustic measurements of your drivers in your cabinet....  but what you're doing is obviously a good/gentle start.   :thumb:?

 

Yes, I would take out the parallel R, and start again in the simulator....

 

Link to comment
Share on other sites

1 hour ago, Lynx4 said:

Awesome thank you. I'll sketch that out unless you already have a copy of the circuit diagram? 

Built it in vcad and post a pic.   Then people can make their suggestions.

 

If you start with the circuit VA described, the you'll find you should be able to "simplify" it quite a bit, while keeping close enough to the same outcome.

Link to comment
Share on other sites

So I have come up with this one let me know what you guys think or where I should make changes. Just so you know the resistor before the inductor is the dc resistance of the inductor and is not accurate until I search for parts. Same with everything else I have cross referenced yet with a parts list so some values might change slightly 

20191110_210547.jpg

20191110_210536.jpg

Link to comment
Share on other sites



Again the crossover cct is weird. I would suggest that if you are relying on simulation that you use a conventional network such as the Butterworth  second order with the component values  outlined in a  previous post.  Then  add in the impedance correction network for the Dayton woofer  followed by the attenuator pad for the tweeter .  If you run  an  impedance  graph as well as the frequency response , this will  help you  understand  what  happens although  it is not always  the same as measuring the  actual speaker system.

 

  • Like 1
Link to comment
Share on other sites

1 hour ago, VanArn said:

Again the crossover cct is weird. I would suggest that if you are relying on simulation that you use a conventional network such as the Butterworth  second order with the component values  outlined in a  previous post.  Then  add in the impedance correction network for the Dayton woofer  followed by the attenuator pad for the tweeter .  If you run  an  impedance  graph as well as the frequency response , this will  help you  understand  what  happens although  it is not always  the same as measuring the  actual speaker system.

 

Jeez I suck at this ok. I followed the other one on the simulator with the spl graph and impedance graph. Let me know if this is right. There is a large spike at about 1700hz plus the top end is still a little sensitive compared to the lower mids. 

Sorry I'm not very smart I guess lol. Also red is woofer green is tweeter for reference 

20191111_120618.jpg

20191111_120722.jpg

20191111_120626.jpg

Link to comment
Share on other sites

The  impedance ( plus phase  )  correction circuit, consisting  of the 4.5 ohm resistor  in series with the 36 mFd capacitor needs to be connected  across the bass speaker terminals. Put another way, the speaker positive terminal should go to the junction of the 0.3mH inductor  with the  9.4mFd capacitor    (and the L.H. point of the 4.5 ohm resistor).

The attenuator pad  needs to be shifted to a position after the  capacitor inductor junction. To clarify this, the 1.3 ohm resistor connects from that junction to the ' top' terminal of the tweeter and the 10 ohm resistor is connected across the neg. and positive terminals of the tweeter. It should be noted that it is normal for second order crossovers  ( two-way)   to have  the tweeter connected out of phase with the bass section.  I believe your crossover cct. does this , presuming 1 is the positive input and 2 is the negative return terminal 

Link to comment
Share on other sites

1 hour ago, Lynx4 said:

Sorry I'm not very smart I guess lol.

You'll get there.   A simulator makes it easy to tinker, but doing a good original design takes a lot of experience.

 

Here's how the impedance correction bit of the woofer needs to be (approx values).    ie.  Both parts across the driver (in series with each other)

 

See how it flattens the impedance out at higher frequencies.

This means that if you add a "textbook" crossover filter (which is designed to work on constant impedance) that it will work as expected.

 

As mentioned earlier, it can often be simplified by combining elements of the impedance correct and XO.... but don't worry about that for now.

 

image.thumb.png.98bb529155ce2feca0ee9565727c5ce7.png

 

 

 

Probably the best advice I can give, is don't fall into the trap of adding parts tinkering to get the response you want, without also looking at the impedance.   You want to keep the impedance generally either high, or flat .... if it both low, and with peaks and dips, then it's probably bad.

Link to comment
Share on other sites

10 hours ago, davewantsmoore said:

You'll get there.   A simulator makes it easy to tinker, but doing a good original design takes a lot of experience.

 

Here's how the impedance correction bit of the woofer needs to be (approx values).    ie.  Both parts across the driver (in series with each other)

 

See how it flattens the impedance out at higher frequencies.

This means that if you add a "textbook" crossover filter (which is designed to work on constant impedance) that it will work as expected.

 

As mentioned earlier, it can often be simplified by combining elements of the impedance correct and XO.... but don't worry about that for now.

 

image.thumb.png.98bb529155ce2feca0ee9565727c5ce7.png

 

 

 

Probably the best advice I can give, is don't fall into the trap of adding parts tinkering to get the response you want, without also looking at the impedance.   You want to keep the impedance generally either high, or flat .... if it both low, and with peaks and dips, then it's probably bad.

Ok ok I have done this now I think only issue is my SIM won't allow me to place resistor on the other side of the capacitor for the correction. This is what I'm looking at now. Still have a spike in impedance at the top end of frequency is this the crossover network? To me it's looking good and seems to work? 

20191112_000603.jpg

20191112_000610.jpg

20191112_000617.jpg

Link to comment
Share on other sites



It does not matter which way, the impedance correction network  is connected to the bass speaker terminals. However you have an error in the placement of the tweeter attenuation pad.

The inductor should be shifted to connect between the 9.4 mFd capacitor and the return line ( tweeter negative ). The attenuator 'L' pad is for the tweeter to drop  its' sensitivity to  match that of the bass speaker while  keeping the impedance at  4 ohm throughout the crossover region. It can  only  be connected at the tweeter  for it to work as intended.

Link to comment
Share on other sites

3 hours ago, VanArn said:

It does not matter which way, the impedance correction network  is connected to the bass speaker terminals. However you have an error in the placement of the tweeter attenuation pad.

The inductor should be shifted to connect between the 9.4 mFd capacitor and the return line ( tweeter negative ). The attenuator 'L' pad is for the tweeter to drop  its' sensitivity to  match that of the bass speaker while  keeping the impedance at  4 ohm throughout the crossover region. It can  only  be connected at the tweeter  for it to work as intended.

So should the L pad go before or after the crossover (capacitor and Inductor) 

Link to comment
Share on other sites

  Bravo, Lynx4!

   The impedance curve is what I would expect of a crossover designed to suit specific speaker units. In this case the manufacturer's data has been relied on and  this information can be inaccurate  as there can be  production changes and material variability can  also effect results.  It is a good starting point for your speaker build project .

  • Like 2
Link to comment
Share on other sites



3 minutes ago, VanArn said:

  Bravo, Lynx4!

   The impedance curve is what I would expect of a crossover designed to suit specific speaker units. In this case the manufacturer's data has been relied on and  this information can be inaccurate  as there can be  production changes and material variability can  also effect results.  It is a good starting point for your speaker build project .

Now the only trouble is finding parts to suit i have tried to match everything up best I could but it has slightly changed the impedance will this be much of an effect? Or would it be alright?

20191112_122159.jpg

20191112_122203.jpg

Link to comment
Share on other sites

For prototyping a speaker crossover, especially when you have access to many components, the values can be tightly controlled by selection with the aid of an RLC meter . The better the instrument, the tighter  the  tolerance limits for the crossover network . If you are prepared to buy a few extra resistors, it is possible to make up or select from a batch, an exact match (at least to within 5% or better) of the calculated  design values.  For example, two 2.7 ohm resistors in parallel can be used to make up the 1.3 ohm value and you can use the lower cost 5 watt wire wound types , which should be cheaper than a 10 watt rated resistor.  The 4.5 ohm resistor can be made up of  10 and 8.2 ohm resistors in parallel and again the 5 watt wire wound types can be used.The 9.4 mFd capacitors can be made up of two 4.7 mFd caps. connected in parallel (Polypropylene capacitors of at least  a 250 volt DC rating).  I would prefer the inductors to have a lower DC resistance ,but this would mean a heavier gauge wire and of course, a higher cost.

What components do you have already ? It is  a simple task to adjust  to other values by the addition of other parts as shown in the  forgoing discussion.

 In any case, using the values that are shown in your schematic are  very close and the results may not be distinguishable from a more  rigorous assembly.

Edited by VanArn
additional material
  • Like 1
Link to comment
Share on other sites

4 hours ago, VanArn said:

  Bravo, Lynx4!

   The impedance curve is what I would expect of a crossover designed to suit specific speaker units. In this case the manufacturer's data has been relied on and  this information can be inaccurate  as there can be  production changes and material variability can  also effect results.  It is a good starting point for your speaker build project .

... and the biggest issue, is that it doesn't take into account the response on the intended cabinet shape.... everything > 1 khz is going to look a lot different when mounted on a small box, vs the (assumed) IEC baffle or similar that the datasheet measurements were done on.

Link to comment
Share on other sites

That's alright I'm not after an audiophile grade set of bookshelfs this was just a project to introduce myself into the world and already I have learnt a tone thanks to all you guys helping out I can not thank enough! ?

  • Like 1
Link to comment
Share on other sites

2 hours ago, VanArn said:

For prototyping a speaker crossover, especially when you have access to many components, the values can be tightly controlled by selection with the aid of an RLC meter . The better the instrument, the tighter  the  tolerance limits for the crossover network . If you are prepared to buy a few extra resistors, it is possible to make up or select from a batch, an exact match (at least to within 5% or better) of the calculated  design values.  For example, two 2.7 ohm resistors in parallel can be bused to make up the 1.3 ohm value and you can use the lower cost 5 watt wire wound types , which should be cheaper than a 10 watt rated resistor.  The 4.5 ohm resistor can be made up of  10 and 8.2 ohm resistors in parallel and again the 5 watt wire wound types can be used.The 9.4 mFd capacitors can be made up of two 4.7 mFd caps. connected in parallel (Polypropylene capacitors of at least  a 250 volt DC rating).  I would prefer the inductors to have a lower DC resistance ,but this would mean a heavier gauge wire and of course, a higher cost.

What components do you have already ? It is  a simple task to adjust  to other values by the addition of other parts as shown in the  forgoing discussion.

 In any case, using the values that are shown in your schematic are  very close and the results may not be distinguishable from a more  rigorous assembly.

That was the other way I was thinking but space as well as cost is an issue already with the componates listed in the above crossover I'm totalling about 100bucks and I'I'nmot chasing something that is mind blowing just a good sound in a great looking cabinet. Once I have more experience and knowledge behind me then I'll probably attempt to make great sounding set of speakers but most likely a few years away as I do have alot to learn still! Haha. 

I think I'll go ahead and purchase what I have now at least then I should have some alright sounding bookshelfs by the weekend. 

20191112_175137.jpg

Link to comment
Share on other sites



  • Recently Browsing   0 members

    • No registered users viewing this page.




×
×
  • Create New...
To Top