Help with first order series crossover?

majerjack

Member
2010-02-20 11:57 am
I've been considering using a first-order series crossover on a pending speaker project. Having never used one before, I have a question about implementation.

I would like to keep the crossover as simple as possible, and to that end I wish to use a parallel resistor on the tweeter to flatten the impedance with which the crossover reacts. I am having difficulty predicting what, if any, effect the extra resistor in the series crossover will have on the response.

I know the resistor will dissipate some amplifier power and that the power rating of the resistor must be high enough to do so safely. I know that I must cross the tweeter high enough above resonance to keep the power applied to a low level and to come as close as possible to an acoustical first-order response. My question is: will a parallel resistor on the tweeter alter the characteristics of a first-order series crossover?

If anyone has specific knowledge on the subject, I would appreciate your reply. If you have experience with series crossovers, I would be interested in hearing about that also. Thanks to all.
 
Hi,

A parallel resistor with a tweeter achieves next to nothing.
You need series or series+parallel to adjust tweeter level.

For a typical 2 way 1st order series will not work well.
BSC is one difficult issue, driver issues also come
into play regarding the actual acoustic slopes.

rgds, sreten.
 
There are pretty good reasons you don't see many series crossovers. You also don't see too many " simple as possible" that are successful. Simplified, yes. First order crossovers will require exceptional drivers with far wider bandwidth than are easy to find.

So, what is your motivation for going series, and what drivers do you have in mind?
 
Hi,

A parallel resistor with a tweeter achieves next to nothing.
You need series or series+parallel to adjust tweeter level.

I'm sorry, but this is wrong. As long as there is a component in series to create the voltage divider, and not necessarily a resistor, the shunt resistor will attenuate the tweeter.

In the network of a SXO, this is an example of one:
XenumRev3-Invert.jpg


In this example, the 3.3 ohm is the series L-pad, and the 30 ohm damps the Fs mainly for a flatter tweeter-system impedance. Yes- this is a second-order electrical xover, BW3 acoustic.

Later,
Wolf
 

majerjack

Member
2010-02-20 11:57 am
Thanks to Odougbo, sreten, and tvrgeek for your replies.

I hope to receive an answer specific to my question. If anyone can offer one, I would appreciate it. In addition, if you have built a system with a first-order series crossover, I would be interested in hearing about the experience and the results.

Edit: Thanks also to wolf_teeth. Your post went up before I had finished mine.
 
Last edited:
Why do you think that it is necessary to place a resistor in parallel circuit with the tweeter? Is this based on knowledge of the impedance vs frequency characteristic of the tweeter, above the resonant frequency? Normally the impedance of a tweeter at frequencies an octave or more above the resonant frequency is flat enough to not require correction.

-Pete
 
cT- the reason for the 30 ohm resistor is not so much for level, because that will remove less than a dB from the general output level above the xover. Being that the XT25 series of drivers do not have ferrofluid, and have a rather high peaking Fs magnitude, they tend to 'buzz' in systems where it's not compensated if you don't xover high enough. The 30 ohm in parallel with 4 ohms isn't much different, but in the range where the impedance hits 30 ohms, this is reduced to 15 ohms max- a great difference. Above Fs, it's not really about Z-comp. (Another note is that even though the tweeter could be flat impedance, it may have a rising response, and a CR shunt will remedy that. This in no way purposed to flatten the Le.)

Otherwise, in some circuits where you have either an increasing Le (more than typical for a tweeter, and I've encountered a few), a higher impedance of 8 ohms, or a high enough xover to allow usage of a parallel resistor without a series resistor in place (or minimal series before xover and greatly reducing parallel resistor), you can weigh the balance of cost in terms of xover parts. The lower impedance makes caps larger, and coils smaller. Since caps can be inexpensive and copper is not always such, you can reduce cost to build at times with this method.

Other reasons include that a resistor in series before the xover or one in parallel with the driver are more likely to not affect spectral tilt. Sometimes you get all 4...

Maybe using a 20 ohm resistor across an 8 ohm tweeter nets a 5.x nominal impedance, reduces the tweeter-system's Fs magnitude, sets the correct sensitivity, reduces parts cost, all while retaining the correct spectral tilt- while only using a single parallel resistor.

Does that answer your question?
Wolf
 
maj.
Yes, I built several series crossovers about 35 years ago when I started this hobby. I was not at all happy with them, which is why I have not used them since. Granted, I did not have the kind of test and simulation tools we have now and we can have much better drivers than back then. I just don't see anything they do that is better than parallel.

What drivers are you considering and what crossover point do you think is appropriate? Maybe we can be more specific is helping why several of us consider this not to be a good idea. If you were using a Fountek FE89 and an SB29, maybe you could pull it off at 4K. If you were using a Dayton RS150 and an XT25, you are in for a serious letdown. All very good drivers, if used correctly.

Wolf,
I doubt the tweeters Fs was the issue. Welcome to the club of us who found this is to an easy driver to use well; but if so, very good indeed. The trick is keeping the the XT25 above 3K. It just plain won't stop flopping in the breeze below that. Look at it's CSD plot. Second order electrical at 4K makes it much happier.

I want to try it first order at 5K, but can't settle on a mid. If a shunt helps, that would be a great discovery as the lower the crossover, the more mids can keep away from their breakup. Off to test that now.
 
Majerjack, I think Wolf_teeth's remarks on a resistor paralled with the tweeter are right on the money. Sometimes it is the most economical solution.

One word of caution on the use of of a parallel 1st order passive xover: without drivers mounted, it will create a dead short at the xover frequency. It can kill your amp.

As to 1st order in general, as has been remarked before, it only works with carefully selected drivers with flawless operation up to at least 2 octaves outside the xover point. And even if you would find such drivers, it will be very difficult to find a solution to compensate for the increasing beaming of the low driver, this in combination with the profound lobes you create with low order filters. This will necessitate you to work with small drivers, closely spaced with a low xover frequency, 2KHz or so. This also shows the impractability of a 1st order setup; you would need a tweeter that still does its work well at 500 Hz (2 octaves below 2KHz)
 
Tvrgeek, that looks like an interesting project, but if I may make some suggestions: before you make a passive filter, first do some measurements and listening tests with an active setup using something like a MiniDSP. This will show you which filter slopes and xover frequencies will work best. Don't just measure on axis, but also look at what happens off the horizontal and vertical axis. 5K from mid to tweeter is in my experience way too high, as you will note from such tests.

A very nice little speaker that would match up good with your other components is the Fountek FR88EX they have at Madisound. It is advertised as fullrange, which it is. It does have some breakup above 10KHz and ultrasonic it goes nuts, but it is a very good midrange.
 
Here's a pretty good thread link. However, the Acoustic Reality XO site is gone. They had some great instructions on how to empirically select the xover components.

Ooops! sorry for forgetting to paste the link prior. Here it is...

http://www.diyaudio.com/forums/multi-way/27969-series-crossover.html

The link below is to a much larger thread that basically debates the merits of series xo's. Some heavy hitters participated in this one. You should pay attention to what they wrote. Particularly JohnK....

http://www.diyaudio.com/forums/multi-way/206843-sreten-speakerman-go-series-xos.html
 
Squabbling as usual...:rolleyes:

Back on topic. The Acoustic Reality XO from my archive, since it getting hard to find:

[URL="[IMGDEAD]http://img202.imageshack.us/img202/1988/acousticrealitycrossove.jpg[/IMGDEAD]

I expect it would work best with a good flat polycone.

The gurus tend to do something more complicated. This is a Troels Gravesen design IIRC:

[IMGDEAD]http://imageshack.us/a/img851/5273/troelsgravesencrossover.jpg[/IMGDEAD]

Tony Gee did this one at Humble Homemade Hifi:

[IMGDEAD]http://imageshack.us/a/img37/8500/tonygeeclassiccrossover.jpg[/IMGDEAD]

Tricky things. :)