Crossover Confirmation for Accuracy

[jerimiah]

Why is it that every time I go to a different website or Crossover Chart everybody has different values to get certain crossover points. For example, for a 2-way crossover at 2,500 Hz; every site I go to has a different setting. Is this stuff not universal that on an 8 ohm speaker crossover at 500Hz 6db a 2.55mh is needed. How come it varies depending on what chart you are looking at. I am try to look at how to put a 3-way crossover together and because almost every site has something different I am not sure if I am using the right parts. So my question is; what chart is mor accurate for me to use for a 3-way and 2-way???

[dennismurphy]

Are you saying that you've visted several sites that puport to list "text book" values for crossover values for a given slope and fc, and they're all different, or are you saying that you've checked out different speaker designs that use different crossover values? You would certainly expect to see different values for the latter, since the drivers and baffles, and a whole lot else, are different. But if a site is telling you what inductor value, in theory, would give you a 6 dB Butterworth with an fc of 2500 Hz assuming constant 8 ohm impedance and perfectly flat driver response, then they should be the same. The figures wouldn't be of much use in the real world, but they should be the same.

[jerimiah]

So where can I find a the text book values for crossover points? Also another question; if I take a two 1.8mH 14 GA. PERFECT LAYER INDUCTOR and put them in parrell will that be the 500 hz crossover point for a 12db and since one of those handles 800watts would that then be able to handle 1,600 watts. I am trying to use this for some dual 18" subs running at 8 ohms rated at 800rms on each speaker.

> Are you saying that you've visted several
> sites that puport to list "text
> book" values for crossover values for a
> given slope and fc, and they're all
> different, or are you saying that you've
> checked out different speaker designs that
> use different crossover values? You would
> certainly expect to see different values for
> the latter, since the drivers and baffles,
> and a whole lot else, are different. But if
> a site is telling you what inductor value,
> in theory, would give you a 6 dB Butterworth
> with an fc of 2500 Hz assuming constant 8
> ohm impedance and perfectly flat driver
> response, then they should be the same. The
> figures wouldn't be of much use in the real
> world, but they should be the same.

[billfitzmaurice]

Two problems. First, one typically uses 18" subs to no higher than 150Hz, and more often than not only 100Hz. Second, one typically bi-amps subs and mains using an electronic crossover. With compression horn drivers in a very high power situation a tri-amped system is the norm.

[wg_ski]

> So where can I find a the text book values
> for crossover points? Also another question;
> if I take a two 1.8mH 14 GA. PERFECT LAYER
> INDUCTOR and put them in parrell will that
> be the 500 hz crossover point for a 12db and
> since one of those handles 800watts would
> that then be able to handle 1,600 watts. I
> am trying to use this for some dual 18"
> subs running at 8 ohms rated at 800rms on
> each speaker.

The text book values depend on what kind of fiter you're trying to implement. They call them "text book filters" because they used to have filter tables in old engineering texts. They give theoretical low-pass values for a 1 Hz cut off and a 1 ohm load, and you scale to get the frequency and impedance you want. You look up from whichever table based on slope and Q. For high pass or band pass there are standard conversion algorithms that you need a few basic EE classes to understand. Not as simple as it sounds anymore, is it? So you get different values for differnet on-line calculators? That's because they make different assumptions as to the desired slope and Q. They can't read your mind and they don't know what will sound best with any given drivers.

[jerimiah]

If I take a two 1.8mH 14 GA. PERFECT LAYER
INDUCTOR and put them in parrell will that
be the 500 hz crossover point for a 12db and
since one of those handles 800watts would
that then be able to handle 1,600 watts.

[Chris Roemer]

> Two problems. First, one typically uses
> 18" subs to no higher than 150Hz, and
> more often than not only 100Hz. Second, one
> typically bi-amps subs and mains using an
> electronic crossover. With compression horn
> drivers in a very high power situation a
> tri-amped system is the norm.

But the main point you're missing here is that a "textbook" crossover almost NEVER works the way you'd expect it to.

Looking at someone's 18"er, the impedance isn't 4 ohms, or 6 or 8 ohms, its 12 ohms at 20 Hz, peaks at 24 ohms at 26 Hz, drops to a minimum of 4 ohms from about 100 Hz to 200 Hz, then rises to 6 ohms at 500 Hz (where you'd like to cross yours). At 20,000 Hz this driver has an impedance of 45 ohms.

Textbook crossover calculators don't take this into account. They ASSume that the driver's impedance profile is FLAT at its nominal rated impedance (like 8 ohms), but they NEVER are (except for some special types of drivers, which you aren't using). They also assume that the driver has a flat response level throughout its pass band. Never seen it.

The 18" I'm lookin' at has an output of 83dB at 20 Hz, rises to a "hump" of 92dB at 150 Hz, bottoms back out around 88dB from about 400 Hz to 700 Hz, then has a bad "spike" up to ONE HUNDRED dB at about 2200 Hz before dropping like a stone down to 70dB at 6000 Hz (and beyond).

Get it?

Three ways have 3^3/2^2 or 27/4 or 7 times the chance of not working compared to a two way, and that's where you say you're going. Do you know anything about bandpass gain on your midrange? Are you going to reverse the polarity on that driver? How did you size your cabinet for your woofers? Are you going sealed with the mid?

Why not tell us what drivers you're using so we can give you some good advice? Otherwise continue wasting your time, our time, and your money by going down a path you're not quite ready for.

Just trying to help.

[wg_ski]

> If I take a two 1.8mH 14 GA. PERFECT LAYER
> INDUCTOR and put them in parrell will that
> be the 500 hz crossover point for a 12db and
> since one of those handles 800watts would
> that then be able to handle 1,600 watts.

For a 500 Hz 2nd order with an 18" PV lowrider you use 3mH in series, with a series combination of 2 ohms and 65 uf in shunt with the woofer. For dual drivers you use half the inductance (1.5mH), double the cap to 130 uF, and half the resistor to 1 ohm. Just the coil (and at 0.9 mH for the parallel combination) won't even be close. Values were arrived at by circuit simulator - Load at 6.21 ohms + 0.87 mH, set db(vout) to 0dB up to 250Hz, 6dB down at 500 Hz, 24dB down at 2k, and hit "optimize". You can't and won't do better than this without actually taking measurements.

Trying to operate a heavy cone 18" up to 500 Hz isn't going to sound very good, however.

[Paul O]

I can confirm you are determined to build this your way, experts be damned. A 6db filter at 500hz is all but useless for sub duty, when you hear this you WILL be suprised just how much midrange sound will be coming out of your "subs". That notwitstanding..

A 6db/oct low pass filter at 500hz into 4ohms will require a 1.27mH inductor. To get more power handling you can put 2 inductors in parallel, but you have to use double the value to get back to your desired target inductance, so in this case you would use two 2.54mH inductors in parallel.

[brianwalter]

> If I take a two 1.8mH 14 GA. PERFECT LAYER
> INDUCTOR and put them in parrell will that
> be the 500 hz crossover point for a 12db and
> since one of those handles 800watts would
> that then be able to handle 1,600 watts.

Assuming 8 ohms, 500 hz, and a Butterworth filter, you would need 3.6mh inductor, so you would need to wire the coils in series, not parallel. The combination should handle the 1600 watts, but this is not a good idea. I don't know what the series resistance of the 1.8 mh 14 ga. inductor is, but assume it's 0.5 ohms, if you wire the two coils in series you will have 1.0 ohms of resistance in series with your subwoofer and this will totally change the Thiel Small parameters of your sub, and it will have a big affect on the performance. You need to take this series resistance into account with you box design.

You would probably be better off going with a steel laminate core inductor to keep the series resistance low.

I suggest that you pick up a copy of a good speaker design book, such as the "Loudspeaker Design Cookbook" or another book which I don't remember the name of (someone help me here). Otherwise, if you tell us what drivers you want to use and approximately what your goals are, there are several people here that could get you much closer to a decent crossover than what you'll get using text book values.

Sincerely,

Brian Walter

[johnastockman]

> Assuming 8 ohms, 500 hz, and a Butterworth
> filter, you would need 3.6mh inductor, so
> you would need to wire the coils in series,
> not parallel. The combination should handle
> the 1600 watts, but this is not a good idea.
> I don't know what the series resistance of
> the 1.8 mh 14 ga. inductor is, but assume
> it's 0.5 ohms, if you wire the two coils in
> series you will have 1.0 ohms of resistance
> in series with your subwoofer and this will
> totally change the Thiel Small parameters of
> your sub, and it will have a big affect on
> the performance. You need to take this
> series resistance into account with you box
> design.

> You would probably be better off going with
> a steel laminate core inductor to keep the
> series resistance low.

> I suggest that you pick up a copy of a good
> speaker design book, such as the
> "Loudspeaker Design Cookbook" or
> another book which I don't remember the name
> of (someone help me here). Otherwise, if you
> tell us what drivers you want to use and
> approximately what your goals are, there are
> several people here that could get you much
> closer to a decent crossover than what
> you'll get using text book values.

> Sincerely,

> Brian Walter

...which is probably the best choice for someone just starting on this path. Very beginner-friendly, it can give you a better understanding of the values/components. But I do agree that because of the varying resistance of any driver across the freq. range, textbook on-line calculators won't give you the proper info. It might be OK for a starting point for a PA system, but I'm far from an expert in this area...but to double the value of inductors, you have to wire in series, not parallel.

John A.