Further Crossover Help for Modified Golden Boys

[xluben]

I still need some guidance on what values to use for the tweeter's high pass crossover in my modified Golden Boys. Here's a link to my other post:

<A HREF="http://www.pesupport.com/cgi-bin/config.pl?read=270383">http://www.pesupport.com/cgi-bin/config.pl?read=270383</A>

I've decided to go with a 2nd order high pass at around 6000Hz. This is what I've drawn up:

<A HREF="http://img80.imageshack.us/img80/135...smedium2cq.jpg">http://img80.imageshack.us/img80/135...smedium2cq.jpg</A>

The schematic for the stock tweeter had the polarity backwards for some reason. Should I also do this for the dome tweeter I am going to be using?

Right now my dilemna is finding the correct values for C4 and L4. These two tables seem to be quite off from each other:

<A HREF="http://www.partsexpress.com/resources/xover.html">http://www.partsexpress.com/resources/xover.html</A>

<A HREF="http://www.partsexpress.com/tech/260-122.html">http://www.partsexpress.com/tech/260-122.html</A>

I'd just like a little help figuring out what values to use to get good integration with the midrange. Part numbers would be great, also.

Thanks

[Chris Roemer]

> I still need some guidance on what values to
> use for the tweeter's high pass crossover in
> my modified Golden Boys. Here's a link to my
> other post:

>
> <A HREF="http://www.pesupport.com/cgi-bin/config.pl?read=270383">http://www.pesupport.com/cgi-bin/config.pl?read=270383</A>
> I've decided to go with a 2nd order high
> pass at around 6000Hz. This is what I've
> drawn up:

>
> <A HREF="http://img80.imageshack.us/img80/135...smedium2cq.jpg">http://img80.imageshack.us/img80/135...smedium2cq.jpg</A>
> The schematic for the stock tweeter had the
> polarity backwards for some reason. Should I
> also do this for the dome tweeter I am going
> to be using?

No one can really say for sure w/o phase plots.
Wire both tweeters the same. If you think something's not right in the 4k - 8k band, try reversing both of them.

> Right now my dilemna is finding the correct
> values for C4 and L4. These two tables seem
> to be quite off from each other:

>
> <A HREF="http://www.partsexpress.com/resources/xover.html">http://www.partsexpress.com/resources/xover.html</A>
>
> <A HREF="http://www.partsexpress.com/tech/260-122.html">http://www.partsexpress.com/tech/260-122.html</A>
> I'd just like a little help figuring out
> what values to use to get good integration
> with the midrange. Part numbers would be
> great, also.

> Thanks

Use a .15mH coil, #20 will be fine.
Use a 2.2uF cap.
I think you'll need SOME attenuation for your tweeter. I'd put a 1 ohm resistor right in front of your cap for starters. Work out the phase thing (tweeter polarity), then, if you think the top end is too "sizzley", up that resistor to 2 or 3 ohms, however you like it.

[xluben]

> No one can really say for sure w/o phase
> plots.
> Wire both tweeters the same. If you think
> something's not right in the 4k - 8k band,
> try reversing both of them.

> Use a .15mH coil, #20 will be fine.
> Use a 2.2uF cap.
> I think you'll need SOME attenuation for
> your tweeter. I'd put a 1 ohm resistor right
> in front of your cap for starters. Work out
> the phase thing (tweeter polarity), then, if
> you think the top end is too
> "sizzley", up that resistor to 2
> or 3 ohms, however you like it.

How did you choose those values? Accroding to this chart:

<A HREF="http://www.partsexpress.com/resources/xover.html">http://www.partsexpress.com/resources/xover.html</A>

The 2.2 uF cap seems right, but from those values it seems the inductor should be around 3.0 mH

Also, what tells you that I need to attenuate the highs? How much will each extra ohm take off? How much should I take off?

[Chris Roemer]

> How did you choose those values? Accroding
> to this chart:

>
> <A HREF="http://www.partsexpress.com/resources/xover.html">http://www.partsexpress.com/resources/xover.html</A>
> The 2.2 uF cap seems right, but from those
> values it seems the inductor should be
> around 3.0 mH

> Also, what tells you that I need to
> attenuate the highs? How much will each
> extra ohm take off? How much should I take
> off?

Well, (as you probably know) that cap, SR (series resistor), and PR (parallel resistor) on the original tweeter is NOT very standard, unless the tweeter's a piezo.

Your tweeter is about 1dB louder than that midrange. A midrange usually picks up a dB or 2 due to what's known as "bandpass gain". I THINK the reason it's got a 4 ohm resistor in-line with it, is mostly to cut back that BP gain so it's not louder than the woofer.

The freq. plot that Darren has looks like the G Boys run roughly 85-86 dB. Your tweeter is hotter than that, so . . . I say start with a 1 ohmer in-line with it, and if it sounds too hot, up it to a 2 or 3 ohmer. Those resistors are almost free they're so cheap.

I don't do crossovers off of charts or formulas anymore. I use "XOverSim" from the FRD Consortium site along with frequency and impedance data. I have none for all your Goldwood drivers, so I took an educated guess. If you have (or know where I can get) .frd and .zma files, OR EVEN JUST FREQ AND IMPEDANCE PLOTS (graphs), let me know, and I'll run the sim for you.

Chris

[xluben]

> Well, (as you probably know) that cap, SR
> (series resistor), and PR (parallel
> resistor) on the original tweeter is NOT
> very standard, unless the tweeter's a piezo.

> Your tweeter is about 1dB louder than that
> midrange. A midrange usually picks up a dB
> or 2 due to what's known as "bandpass
> gain". I THINK the reason it's got a 4
> ohm resistor in-line with it, is mostly to
> cut back that BP gain so it's not louder
> than the woofer.

> The freq. plot that Darren has looks like
> the G Boys run roughly 85-86 dB. Your
> tweeter is hotter than that, so . . . I say
> start with a 1 ohmer in-line with it, and if
> it sounds too hot, up it to a 2 or 3 ohmer.
> Those resistors are almost free they're so
> cheap.

> I don't do crossovers off of charts or
> formulas anymore. I use "XOverSim"
> from the FRD Consortium site along with
> frequency and impedance data. I have none
> for all your Goldwood drivers, so I took an
> educated guess. If you have (or know where I
> can get) .frd and .zma files, OR EVEN JUST
> FREQ AND IMPEDANCE PLOTS (graphs), let me
> know, and I'll run the sim for you.

> Chris

Thanks, for the help.

This is what is in the project plans about the 4 ohm resistor in series with the midrange:

"A 4 ohm padding resistor reduces the output by several dB, while propping up the impedance throughout the midrange."

I have no idea where to get data on these drivers. I'md guessing with a company like Goldwood, and such cheap drivers (that aren't popular) no one would have taken the time to measure them.

So, I think whenever I get around to doing this project I'll just go ahead with the values you suggested for the tweeter, and or 1, 2, 3 ohm resistors to try out.

[wg_ski]

> Thanks, for the help.

> This is what is in the project plans about
> the 4 ohm resistor in series with the
> midrange:

> "A 4 ohm padding resistor reduces the
> output by several dB, while propping up the
> impedance throughout the midrange."

Because the "bandpass gain" effect is really a misnomer - the extra output you get isn't free. The bandpass network just presents a lower impedance to the amp. All even-order bandpass networks do this.

> I have no idea where to get data on these
> drivers. I'md guessing with a company like
> Goldwood, and such cheap drivers (that
> aren't popular) no one would have taken the
> time to measure them.

I wouldn't say they're not popular - PE has expanded the line in a huge way over the past 5 years. I'm sure they sell a lot of them. At their performance point, a crossover designed from T/S parameters and Re/Le data alone and not stretching the useable frequency limits can be quite listenable. It's the higher performance (and more expensive, more revealing) stuff that requires heroic measures in the crossover and can't be adequately designed without acoustic phase data.

[xluben]

> Because the "bandpass gain" effect
> is really a misnomer - the extra output you
> get isn't free. The bandpass network just
> presents a lower impedance to the amp. All
> even-order bandpass networks do this.

> I wouldn't say they're not popular - PE has
> expanded the line in a huge way over the
> past 5 years. I'm sure they sell a lot of
> them. At their performance point, a
> crossover designed from T/S parameters and
> Re/Le data alone and not stretching the
> useable frequency limits can be quite
> listenable. It's the higher performance (and
> more expensive, more revealing) stuff that
> requires heroic measures in the crossover
> and can't be adequately designed without
> acoustic phase data.

So do you think a high pass at 6kHz will work for the tweeter/mid combo I am going to use?

And do you think the C and L values Chirs suggested will work well?

Also, how do I make sure the polarity of my tweeter is correct?

[wg_ski]

> So do you think a high pass at 6kHz will
> work for the tweeter/mid combo I am going to
> use?

Really anything 3-6k will "work" with a well-behaved paper cone 5" mid and a budget tweeter.

> And do you think the C and L values Chirs
> suggested will work well?

I haven't checked the math, but probably.

> Also, how do I make sure the polarity of my
> tweeter is correct?

The $64000 question. Answer - try it both ways and whichever sounds better, use it. If it's backwards, it will be off by *at least* 90 degrees and one way will sound noticeably better - even with cheap drivers. It may take more than one listening session and several types of material to determine which way to put it, but you'll eventually get it. If the crossover was designed without compensating for excess phase, it could be off by +/-45 degrees even when you get it "right". And contrary to popular opinion, it's *not the end of the world*.