I am looking for a text that goes in to more detail over filter design and calculations than a book like Loudspeaker Design Cookbook. I unfortunately sold my Introduction to Electrical Engineering book from college because it was not related to my field at all (All engineering majors were required to take 1 semester of EE)
Most texts I have found are more like a glossary and philosophy of speaker design but no real "meat" or application to them.
I have a better grasp of concepts after working them out by hand. I am even thinking of writing a program that would simulate crossover design over a given range of frequencies at certain intervals (such as 500-900 hz at 25 hz increments) and speaker locations. I do understand that this could take years before I fully understand the topic.
I have found pretty good literature online for frequency response of filters but not so much on phase. Other than you can assume that elements are minimum phase. Which to me means that you can relate it to the instantaneous slope of the frequency, but I could be very wrong! (Hence why I'm looking for good texts)
Anyone have recommendations? Thanks in advance!
Most texts I have found are more like a glossary and philosophy of speaker design but no real "meat" or application to them.
I have a better grasp of concepts after working them out by hand. I am even thinking of writing a program that would simulate crossover design over a given range of frequencies at certain intervals (such as 500-900 hz at 25 hz increments) and speaker locations. I do understand that this could take years before I fully understand the topic.
I have found pretty good literature online for frequency response of filters but not so much on phase. Other than you can assume that elements are minimum phase. Which to me means that you can relate it to the instantaneous slope of the frequency, but I could be very wrong! (Hence why I'm looking for good texts)
Anyone have recommendations? Thanks in advance!
Depends on what sort of "filter design" you have in mind:
- passive filters into loads having complex impedances (e.g. the usual crossover in the loudspeaker you buy at the store)
- passive filters into resitive loads (line level passive crossover)
- analog active filters (passive filters with active buffers or NFB and PFB loops around op-amps)
- Infinite Impulse Response Digital Filter
- Finite Impulse Response Digital Filter
Which is it?
There is already an excellent design tool used by hundreds of DIYers for this:
Loudspeaker Design Software
I suggest you take a look at the PCD User Guide (the last link on that page).
As far as a textbook on the subject, I do not think there is one. It's a bit of a jump from the Loudspeaker Cookbook to using PCD but you have to just jump in there and do it. For assistance, you can post over at PE TechTalk with questions.
Joe D'Appolito "Testing Loudspeakers."
https://www.amazon.com/Testing-Loudspeakers-Joseph-DAppolito/dp/1882580176
Look for publications by Sigfried Linkwitz. I figure if he can co-develop the 4th order filter that bears his name, he might be a good source.
https://www.amazon.com/Testing-Loudspeakers-Joseph-DAppolito/dp/1882580176
Look for publications by Sigfried Linkwitz. I figure if he can co-develop the 4th order filter that bears his name, he might be a good source.
Last edited:
Thanks for the recommendations. I'll check out the Joe D'Appolito text for sure! I was hoping for a Fundamentals of Electrical Engineering recommendation to supplement as well. I might repurchase the old book I owned (Giorgio Rizzoni text) but it is universally panned as a poor text book.
I currently own soundeasy and just downloaded PCD right before this post. I'm pretty much trying to learn exactly what PCD already accomplishes and replicate the results myself. Just for fun... my very own twisted sense of fun.
I currently own soundeasy and just downloaded PCD right before this post. I'm pretty much trying to learn exactly what PCD already accomplishes and replicate the results myself. Just for fun... my very own twisted sense of fun.
Check out this Bell Labs book, "Linear Integrated Networks: Design".
Great stuff. Only $10, but usually it costs much more, even today.
https://www.amazon.com/Linear-Integrated-Networks-Design-Laboratories/dp/0442255829
Last edited:
Worth to test also some other tools than PCD if you're really abandoning SoundEasy. Boxsim, SpeakerSim, VituixCAD, XMechanic, XSim, XSim 3D are a bit more active and popular freeware products at the moment. For example VituixCAD has 500...1000 quite active users. Few "not so small" commercial manufacturers are also changing over to VCAD from obsolete LEAP 5.
I have done this (3 times). For years I could build a speaker that measured well but not one I liked to listen to.
Lots of universities have online courses you, and everyone else, can learn from.
Duckduckgo is your friend.
Your link from writing simulation program to designing well measuring but bad sounding speaker is actually quite funny
Knowledge what measurements correlate with positive listening experience and what the results should be has not been common reality for decades. Using bad acoustic concepts - for example acoustically incompatible or otherwise not so good drivers/radiators but tweaking close to ideal on-axis response still happens, though times are fortunately changing I think and hope. Subjective iterating XO designing method is kinda learning phase or sugar pill for actual illness in case proper objective measurable targets are not yet known.
First, thank you everyone for all the replies!
I wasn't planning on abandoning soundeasy. I just want to know the ins and outs of filters. To me the general loudspeaker design books (I have cookbook and an old Radio shack book) just kind of gloss over the topic. Or they pick a crossover in an example and magically it was a perfect fit! I was hoping that doing the actually work would help me understand how to better select a crossover. I just don't want the program to be a "black box" to where I trust the results (I'm an old cantankerous man about that stuff, ask the people I work with
) So the whole story...
I started with a more difficult build than I should have, but the parts were free. So I've been working with a JBL 2226 and 2446 with a yuichi horn (horn wasn't free, I milled it. I had a jbl 2380 but have a general disdain for that horn from working in a club filled with them) and EV St-350b to work with. Just starting 2-way with the 2226/2446 combo first.
I have also been working with a minidsp and an active setup and the first thing I learned is that phase aligned and time aligned are WAY different. It took a bit to wrap my head around that because I thought speakers would behave closer to an ideal perfect radiator than they do. I did read somewhere online that different frequencies have slightly different acoustic centers which would explain it, but haven't seen it written anywhere else.
So then I was looking at the TAD 2404 crossover as a starting point (6th order LR / 1st order BW an 650 hz I think) and was wondering what tweaks could be made to better suite the JBL drivers if any.
I know how to program in FORTRAN and thought I could compile the results of lots of simulations and then test a few out. Not only looking at different orders LP/HP, asymmetrical crossovers, crossovers that don't share the same xo frequency. I could be a horrible idea, but since I'm starting active it is not like I'll be loosing massive amounts of cash trying all these different permutations.
Plus I find writing programs fun
I have a similar set up, JBL 2226 but 2447/2453 on a home made horn.
I haven't seen much written on crossovers for mid horns to cone woofers, the physical offset is quite different from the usual examples of a baffle mounted small tweeter.
I have started to sim this but in LTSpice, have you considered that option?
I write FORTRAN too but the flexibility of LTspice is a real benefit.
It has pretty decent facilities to plot results, all written for you, for free.
Eventually I plan to use ABEC3 to do full FEM/BEM simulations for the polars.
I understand the desire to write your own code to really learn the subject but I wouldn't reinvent the wheel in this case.
Best wishes
David
Hi Marco, this has reminded me of your "Quasi-optimal" filter thread - on essentially this sort of cross-over.
Are you still interested to discuss the topic?
Best wishes
David
You'll get more information if you don't search on speaker filters but rather on AC component theory and AC networks/filters.
https://www.analog.com/media/en/training-seminars/design-handbooks/Basic-Linear-Design/Chapter8.pdf
I'm assuming you wanted analog. This has passive and active implementations, with the equations, tables, and graphs.
I'm assuming you wanted analog. This has passive and active implementations, with the equations, tables, and graphs.
Hi Don
That is an excellent link but the problem is that typical filter theory papers don't include an issue that is practically unique to speaker crossovers.
This issue is the time/phase offset due to physical speaker position.
Even for flat baffles there is an issue of off-axis response and it's sensitivity to the crossover choices.
Horn speakers like the OP's have an additional aspect - the offset from the depth of the horn can be problematic even on-axis.
On the other hand, it looks possible to exploit the offset to actually improve the speaker performance, I have some ideas for this.
Linkwitz and D'Appolito addressed the flat baffle case.
I don't know a definitive treatment of horn speaker offset, Marco's thread is a nice start - "Quasi-optimal" crossover for high-efficiency loudspeaker system, mentions LeCleac'h.
Vanderkooy & Lipshitz also wrote some informative papers in the JAES, back in the 1980s.
In fact I am just on my way to reread them.
I would love to find more references on the subject, know any?
Best wishes
David
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.