breakup nodes?

[jliedeka]

I want to get started designing speakers. I've read through Speaker Building 201, several chapters of The Loudspeaker Design Cookbook as well as looking over many published designs on places like Zaph Audio.

My question has to do with choosing drivers. Another thread mentioned you need to take breakup nodes into account when figuring out if you can make a good crossover between two drivers.

Is there some way to know what frequencies those nodes are? Specifically, can I find out in advance before purchasing? Are there any clues in the frequency response charts?

Jim

[Bill Fitzmaurice]

jliedeka wrote:

Are there any clues in the frequency response charts?

Jim

THey appear as bumps towards the upper end of the drivers range. In the vast majority of cases they occur above where the 30 degree off-axis response is down by more than 6dB from on-axis, placing them above where the crossover should occur anyway.

[avaserfi]

Breakup nodes are not intrinsically bad as nearly all drivers will have them at certain frequencies. They just happen to be one indicator of where it is ideal to implement a crossover with specific drivers. As previously stated these resonances will be visible on a frequency chart presuming proper resolution is used along side proper measuring techniques. Sadly, it is not often the case that proper methods are used by the manufacture such that one will know the full story of the driver in question.

My preferred method of finding out true performance of a driver is purchasing a single unit and actually taking various measurements of the unit. A cumulative spectral decay plot (CSD) is a very useful tool if ones goal is to find out resonant behavior of a specific driver and can actually measure the driver appropriately.

[cixelsid]

Frequency response plots or CSD plots don't tell the whole story. It's a good idea to take a close look at the impedance curve. Ripples in the curve are often the first the indicators of cone modes. And on occasion these ripples can be seen in the impedance curve but aren't seen in the frequency response or CSD plots..

[jliedeka]

Thanks, that's some good information.

I should probably stick with woofers tested independently. I don't care for the idea of spending money on something I may not use.

I'm looking to build a set of 5 sealed 2-way monitors. I'm considering the SEAS CA18RLY, a 6.5" woofer with a SEAS 27TDFC with a crossover somewhere between 1500-2000 Hz. My goal is to build some high quality speakers for less that $200 each.

Jim

[Anthony]

Yeah, unfortunately the only way to know about the breakup nodes is to get information from someone who tried, tested, and found them.

In most cases, they show up as a nub, ripple, or bump in the impedance plot. Joe D'appolito stands firm that he can locate every problem in the impedance plot (provided it's accurate enough). Things from frames ringing to breakup nodes to motor problems. So the Z plot is always a good place to start.

Also, breakup nodes are less of an issue with paper or hybrid cones. All metal cones sound like breaking glass when they breakup. Not really, but you get a top layer of shrill distortion. that is pretty noticeable (and it's usually specific -- you'll listen to one passage and the speaker sounds great, then it hits that one note and you want to cringe). But paper cones don't sound this harsh, so something to consider.

Sites like here, DIYaudio, Zaph and others are great resources. Until I have a lot more disposable income to "waste" on drivers I might not use, I'll just stick to pre-approved

[avaserfi]

Anthony wrote:

Also, breakup nodes are less of an issue with paper or hybrid cones. All metal cones sound like breaking glass when they breakup. Not really, but you get a top layer of shrill distortion. that is pretty noticeable (and it's usually specific -- you'll listen to one passage and the speaker sounds great, then it hits that one note and you want to cringe). But paper cones don't sound this harsh, so something to consider.

When used in higher octaves [non-subwoofer] paper cones are more likely than a relatively more rigid cone such a metal one to have these resonances in the intended passband. Thus at times, due to this less rigid structure, implementation of a paper cone will require allowing audible cone resonance to enter the signal path. While using a more rigid material in such an application will typically allow for no breakup nodes in the passband presuming a sufficiently designed crossover and driver. If one is experiencing distortion from cone break up is due to improper driver choice coupled with poor crossover design.

[cixelsid]

jliedeka wrote:

I should probably stick with woofers tested independently. I don't care for the idea of spending money on something I may not use.

I'm looking to build a set of 5 sealed 2-way monitors. I'm considering the SEAS CA18RLY, a 6.5" woofer with a SEAS 27TDFC with a crossover somewhere between 1500-2000 Hz. My goal is to build some high quality speakers for less that $200 each.

Often the mfgr's posted impedance plot is a good starting point.
http://www.madisound.com/catalog/pro...roducts_id=789 Here we see an interesting situation. The FR plot starts getting weird ~500Hz, but the first significant wobble in the impedance plot is ~900Hz.

Another driver you might consider is Peerless Nomex 830875 - 6.5" http://www.madisound.com/catalog/pro...oducts_id=1613 I heard it's little brother the 5-1/4" in a system a couple weeks ago, and it was quite neutral

[Anthony]

My point was, all other things being equal, paper cone breakup is much less noticeable than metal cones. I've heard paper cone speakers with the breakup in the passband and I could not hear the distortion (that I was actively listening for).

You are right though, stiffer materials will have a higher resonance, so the mode itself on metal cones is higher than paper of similar dimensions.

[jliedeka]

cixelsid wrote:

Often the mfgr's posted impedance plot is a good starting point.
(URL redacted because I have <10 posts) Here we see an interesting situation. The FR plot starts getting weird ~500Hz, but the first significant wobble in the impedance plot is ~900Hz.

I see that but I wonder how noticeable that might be. The plot doesn't look terribly detailed.

Another driver you might consider is Peerless Nomex 830875 - 6.5" (URL redacted because I have <10 posts) I heard it's little brother the 5-1/4" in a system a couple weeks ago, and it was quite neutral

This one looks really good from the plots. One thing that concerns me. The EPB is over 100 for this driver. Doesn't that suggest it wouldn't be suitable for a sealed application or isn't that important? It looks like it has plenty of Xmax for sealed though.

Jim

[Bill Fitzmaurice]

jliedeka wrote:

I see that but I wonder how noticeable that might be. The plot doesn't look terribly detailed.

It's detailed enough. In the on-axis plot the breakup is visible starting at 3kHz. The 30 degree off-axis at 3kHz looks fine, so you could cross it there, and with a third order LP most of the breakup would be attenuated, but based on the 60 degree off-axis I'd cross it at 2.5 to 2.8 kHz.

This one looks really good from the plots. One thing that concerns me. The EPB is over 100 for this driver. Doesn't that suggest it wouldn't be suitable for a sealed application or isn't that important? It looks like it has plenty of Xmax for sealed though.

Same crossover for the same reason. As for the EBP, 100 doesn't automatically take it out of consideration for a sealed alignment, that's where modeling comes in.

[cixelsid]

jliedeka wrote:

This one looks really good from the plots. One thing that concerns me. The EPB is over 100 for this driver. Doesn't that suggest it wouldn't be suitable for a sealed application or isn't that important? It looks like it has plenty of Xmax for sealed though.

Since you reference building 5 speakers, are they going to be used with a sub? If so go sealed regardless of the EPB number.

BTW, I see you recently got a Behmor, sounds like fun....

Sid
VBM Super Domobar Levetra...

[mayhem13]

http://zaphaudio.com/6.5test/compare.html-
Don't know if you've seen these tests so hope this helps. Just bought two of these and 2 more on order to go with 27TDFCs in waveguides in sealed 1.5cuft TMMs on top of dual 10" subs actively crossed at 200hz.

[jliedeka]

cixelsid wrote:

Since you reference building 5 speakers, are they going to be used with a sub? If so go sealed regardless of the EPB number.

BTW, I see you recently got a Behmor, sounds like fun....

Sid
VBM Super Domobar Levetra...

Yes, I have a Hsu VTF3-Mk3. I'm replacing a set of low end Klipsch speakers. I'd prefer to build them myself if I can come up with a design that seems to work. I really would like to try to build some really decent sealed monitors. I'm shooting for a Qtc of 0.707 and I'd like to try a lower crossover point to get it out of the 2-4k range. The tweeter I'm considering should be able to handle it.

And I'm still trying to get the hang of the Behmor. I've been doing the heat gun thing for too long. Nice to see someone who shares two of my obsessions.

Jim