Fourth Order Cross over or notch filter?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am building a system with a dual 8" woofer and 1" softdome tweeter (that is already decided and unchanged). It would maybe be more accurate to say I am rebuilding or fixing!

The woofers have a nasty mass-limited breakup centered at 3,750Hz that starts at 3000 and goes to about 4500. The response peaks about 10db at this point.

I want to cross-over the upper woofer and the tweeter at 2000Hz. A second order cross-over does not tame the peak at 3750Hz sufficiently.

I am now faced with either going to a fourth order filter, or putting a notch filter at 3750Hz. Does anyone have experience in a similar situation, and purely from a performance standpoint, which may offer a better "sounding" speaker. The notch filter appears that it will give a smoother amplitude response, but this could be at the detriment of the phase response. I am just learning Calsod and maybe with some playing I will get there, but if you have any suggestions, please put them forward.

If you have any other ideas for how to deal with this, I am all ears!

Well, you can't polish a turd, but if you have to work with that woofer then I'd be inclined to go for a 4th order L-R filter option. The major downside is the number/cost of components but as you pointed out, the phase response is going to suffer more with a notch filter.

Instinct also kind of tells me not to put any more resonant systems into the system (ie. the notch filter) than strictly necessary, especially with passive components.
Joined 2001
Paid Member
alvaius said:
The woofers have a nasty mass-limited breakup centered at 3,750Hz that starts at 3000 and goes to about 4500. The response peaks about 10db at this point.

What kind of woofers? Have you considered damping the resonance at the source. I had good luck smoothing out the resonance on the aluminum MCM 55-1855 using puzzle-coat.

The speakers I have are Zetag Kinima G4. A speaker that is advertised as really good (and not too cheap!), but is in reality seriously deficient. I bought them from the factory for $650 U.S. in a blowout (cash flow issues for Zetag??) so I can't complain about the price. I bought them on a quick listen that had mainly male vocals and guitar.

The cross-overs are seriously deficient - 2nd order LR with no Zobel on the woofer or anything!! Hence once I got them home and expanded my music choice, including female vocals, things did not seem right.

The tweeter in these units is a Vifa 25mm open-weave silk. It is a dual chamber. It was custom made for Zetag. It is a good quality unit. The ex president of Zetag sells pieces and he claims the FO is 1400Hz, but in my measurements, It was near 900Hz. In fact, these units have the smallest resonance impedance peak of any tweeter I have ever seen.

The woofers are also custom made for Zetag to their specifications. If I had to go on a limb, I would say they are Seas 21cm Excel units. Seas makes 21 cm woofers (Excel series) with Magnesium cones that look darn similar including the curves that I have measured.

To that end, I looked at some of the kits on the Seas web-site and they generally use a 1st order cross-over or a second order LR with a Zobel Network to maintain impedance, and a parallel LC to tame the units output at about 4KHz, just like mine. if you want to look at them. Expensive if you are thinking you are getting top notch speakers. I guess as a kit, they are not too bad. The cabinetry is top notch.

For 21 Seas Excel Magnesium driver you'll get nice result by using 6dB/oct. with LC resonance trap. There is just one thing: the x-over point should be at or just bellow 2kHz and I'm affraid your tweeter wouldn't match. You can try it anyway- inductor should be around 2mH and LC values are 0,15mH and 4,7-6,8uF (try it).
good suggestion.

there is no rule that the XO freq. of the woofer and tweeter must be teh same. in fact if there is a suck out in the 2-4k region it might tame nasty horns and shrill voices.

remember we are hear to listen to music hence we are looking at muscial speakers. accurate speakers while they measure well and sound great will well recorded music reveal every flaw of poorly recorded stuff. this is another thread (see "square wave").

given the inverse law "a recording is as bad as the performance is good" tha applys we know there is very few good recordings that are also good music.

the demo version will do for this purpose. there is a lite version too. remembe you are trying to something relative all you need to do is a make a filter that has a faster roll off than the present XO.

first build present XO in LSPCAD and then change XO for faster roll off LSP allows you to see changes before you actually implement them.
I found out who made the woofers. They are made by Visaton in Germany. They appear to be similar to their high end 8" aluminum cone units though the cone color is gold and the peak at 3,750 is more pronounced. They are a custom unit, but I am hoping that a pleasant letter may get me the specifications.

I was able to determine that they are Visaton units as their is a small Visiton logo on the front on the edge.

Does anyone have any experience with Visiton? Comments?

I have the same problem

Alvaius, what did you end up doing about your cross-over problem? I have the same speaker and my cross-over doesn't make sense to me but then again it was made by a guy who spent 40 years making speakers and I just built them, they were my first pair. But it might be a bad cross-over. I measured the speakers and they had a perfectly flat frequency response, but I don't thing I did it right. They sound pretty good though, I think.

But the cross-over is for the tweeter is(again, I think):

1.1 uF cap and 1 mH inductor

for top woofer:

1.1 uF cap and 2 - 1 mH inductors wired in parellel on pos wire

for bottom woofer:

2.2 uF cap and 2 -1 mH inductors and one about .5 mH inductor; all inductors are wired on pos wire in parellel

And there's a resistor wired in the tweeter and one wired into the woofers, I think.

What do you think is going on? Does this cross-over make sense or is this zetag speaker guy the mad scientist of loudspeakers?
No body wants to help me?

I was wondering from the post I posted above, if that x-over made any sense? The speakers don't sound that bad, but I supect that they have an frequency spike like Alvaius says at 3750 and you can hear it with female vocals. I tested the speakers using the MLS near field method (1 cm away from each driver) with the program Loudspeaker Lab 2.0. All the frequency responses I got were almost perfectly flat from 10 Hz to 100 Hz, but I suspect that the microphone I used wasn't good quality (it came bundled with my soundblaster sound card).

Can anyone tell me if there is a possiblity this cross-over makes any sense and/or if I did the frequency response is accurately?

If this is not a good sound or cross-over I will build a 4th order LR cross-over at 2 KHz or 2.3 KHz like Alvaius said. Or a notch filter, but I don't think the peak at 3,750 Hz will be such a bid deal. I figure at 2 KHz with a 24 db/octive slope at 3,750 Hz the spl should be down at least 20 Hz.

Please help me. Look at that sad little yellow circle face thing, that is me.
This is the response of the woofer

This was done outside with a flat response microphone, a Philips WM-60A electret.

This was done with the speaker in the enclosure without the cross-over. Straight wire into the amplifier.

As you can tell, that peak is nasty!



  • g4 - visaton bass.jpg
    g4 - visaton bass.jpg
    74.2 KB · Views: 403
Well I am not sure who exactly designed the speakers. I know Graham at Zetag seemed to be heavily involved, but his son Phil did the cross-over and he certainly does not have 40 years of cross-over experience.

Just because someone has speaker design experience does not mean they have experience in cross-overs to tame the peaks in a metal cone speaker which seem pretty severe with this speaker.

I think Zetag had a someone "minimalist" approach. That just does not work with this driver. Frankly, while the capacitors are of good quality, the inductors are not.

I have asked Zetag for a response curve. They have not provided one. I am not sure they have the equipment to do one. If they do, then I put the question out here, please show me a response curve for a G4.

As to your cross-over, to increase inductance, you put them in series. In parallel reduces the total inductance. Flat from 10Hz to 100Hz, what did you really mean? The odds are the microphone that came with you soundblaster is not even remotely flat over about 1-2 KHz with huge peaks of its own in the 3-7KHz range (though not a spike like this woofer response), so I am not sure how you got a flat response. Can you post what you measured?

Can you tell me what cross-over components you used and maybe where you bought them? I'm going to build the same x-over as you (if you don't mind telling me what components you used) cause I'm obviously not a compittent speller or speaker builder/designer. I love how on this site and other DIY speaker sites some people take pictures of every construction step so that other people can see what they did. I should have done that so people can laugh and see what not to do. Like giving up on screws half way through building the speakers and let me tell ya, those routers man they can really get way from you. Although I was glad I only cut myself with the router once and also glad it wasn't not spinning at the time.

I agree Zetag has a minimalist approach. My x-overs were built by Graham or atleast that's what he wrote in an e-mail "I'm sweeting like a bull" with "no AC" after building my cross-overs. But of course now I know that doesn't mean they're good.

I was also supprised that the inductors were 18 gauge when I found out. I don't know if that's what you were referring to when you said they were not good quality.

Zetag has equipment to do frequency response curves. I saw a freq. response curve for one of their speakers. I can ask for the freq. response curve of the G4, while I ask this him on his take on his cross-over. He is obviously not ignorant about how a standard LR cross-over should be built cause even I know that a 1.1 cap doesn't go with a 1 mH inductor as described at every website I've ever visited pretaining to cross-overs.

My microphone is not a good one, I think. Even the frequency response for my cheap computer speakers are pretty flat. I did not know that putting inductors in series increases inductance and in parrellel decreases the inductance, but here's what I know about wiring inductors I know, it's pretty limited to say the least.
I don't have a good microphone and this person has the same speaker as me. Thank you for your willingness to help, but I don't think it will be necessary if Alvaius at least tells me the details of his cross-over. Thanks for the link too. I'm always interested in loudspeaker programs. I use LoudSpeaker Lab 2.0 demo now and I can't save anything cause it's a demo.

I think eventually I will get a microphone or build one like some one suggested at this website I'll see how the program you just told me about is. And if it's not confusing to me, which I doubt, I will get a good microphone and do some measurements.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.