Go Back   Home > Forums > >

Multi-Way Conventional loudspeakers with crossovers

Have to love the beginner guides out there.
Have to love the beginner guides out there.
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools
Old 1st January 2021, 12:15 AM   #1
nannoo is offline nannoo
diyAudio Member
 
Join Date: Jan 2013
Default Have to love the beginner guides out there.

Hi all, looking around the web for speaker design guides (trying to go back to basics a bit) and found this. So many odd ideas, I don't want to say too much yet until others have had a chance to enjoy;

3 Way Crossover Design Example

Would love to hear your thoughts!
  Reply With Quote
Old 1st January 2021, 12:54 AM   #2
GeoffMillar is offline GeoffMillar
diyAudio Member
 
Join Date: Oct 2016
Location: Melbourne, Australia
Thanks for posting that: it is great to have so much material available for us newbies.

However, that site uses crossover calculators, which are almost guaranteed not to work properly, for a host of reasons. For example, it assumes all drivers are 8 ohms: but that's a nominal 8 ohms and the impedance of the drivers and the finished speakers will vary with frequency, which affects crossover design.

You can prove that by using a program such as Xsim to see what a calculator would give you in parts values for a given speaker, and compare that to what's actually in the measured and tested design.

All that aside, we have access to materials, tools and advice that wouldn't have been dreamed of 20 years ago, and most of them are free. And there are many, many experienced designers and hobbyists who are happy to provide their expertise and time to help us, value!

Geoff
  Reply With Quote
Old 1st January 2021, 08:31 AM   #3
nannoo is offline nannoo
diyAudio Member
 
Join Date: Jan 2013
Hi Geoff,

Thanks for the response I'm guessing it's a little tongue in cheek? I'll edit the OP when a few more people have chipped in
  Reply With Quote
Old 1st January 2021, 10:16 AM   #4
PeteMcK is offline PeteMcK
diyAudio Member
 
Join Date: Feb 2002
Location: Western Sydney
" it assumes all drivers are 8 ohms" - no, the actual calculator asks for impedance values, and is quite usable, provided you have some idea of what's going on.
'So many odd ideas' - what do you find odd? If you list them we may be able to find situations where they're not so odd, assumptions made etc.
If you want to get back to basics, Rod Elliott's articles are quite informative, but written from a theoretical standpoint rather than a practical point of view. I see so many 'you shouldn't do that' type of comments on these forums, simply because it isn't understood that sticking to purely theoretical ideals may not necessarily lead to practical outcomes.
__________________
"A Fenderr Telecasterr has mair penetration than an AK47" - Alex Harvey
  Reply With Quote
Old 1st January 2021, 11:26 AM   #5
GeoffMillar is offline GeoffMillar
diyAudio Member
 
Join Date: Oct 2016
Location: Melbourne, Australia
Quote:
Originally Posted by nannoo View Post
Hi Geoff,

Thanks for the response I'm guessing it's a little tongue in cheek? I'll edit the OP when a few more people have chipped in

Not at all: this is a great hobby, there's a great deal to learn and I never cease to be knocked out by the generosity of people who are willing to share their experience, time and expertise. All my projects are other peoples' .
Geoff
  Reply With Quote
Old 1st January 2021, 01:37 PM   #6
giralfino is offline giralfino  Italy
diyAudio Member
 
giralfino's Avatar
 
Join Date: Feb 2009
Quote:
If you want to get back to basics, Rod Elliott's articles are quite informative, but written from a theoretical standpoint rather than a practical point of view.
My problem with those articles is that they approach crossover design from an ideal point. And this is also the fault at the base of all crossover calculators. I don't have any problem with the math, but the basics on how a crossover works is useful only if you know how to deal with real world drivers, i.e. with non flat FR and impedance. This is where a crossover simulator comes handy, something that is IMHO the great advantage we have today together with affordable measurement gear. The main problem in the DIY world is to perform reliable measurements.

Ralf
  Reply With Quote
Old 1st January 2021, 01:44 PM   #7
Scottmoose is offline Scottmoose  United Kingdom
diyAudio Member
 
Scottmoose's Avatar
 
Join Date: Mar 2005
Location: UK
Ignoring the fact that like all XO calculators it assumes co-incident drivers with a flat on-baffle response and no impedance fluctuation from the nominal Re value other than the tweeter's Fs peak, then the topology is conventional other than the use of a fixed L-section attenuator on the bass driver. Which has been known [very occasionally] but is not generally done. I certainly wouldn't expect this circuit to work well in practice, but it depends on what the on-baffle responses are.
__________________
www.wodendesign.com (commercial site)
www.frugal-horn.com www.frugal-phile.com (community sites)
  Reply With Quote
Old 1st January 2021, 06:16 PM   #8
Dave Bullet is offline Dave Bullet  New Zealand
diyAudio Member
 
Join Date: Jan 2002
Location: Wellington, New Zealand
For anyone reading the link in the OP thinking that is an ok way to design a crossover - it is not. Textbook calculators are absolute crap.

You may as well take all the money spent on the expensive Scanspeak drivers (if you are not going to design a proper crossover) and flush it down the toilet. The noise will be more pleasant.
__________________
"Usual diyaudio train wreck of dubious drivers and just the crossover to sort out. Well, how are you on crossovers and modelling? Pretty green, I reckon" - system7 (Steve)
  Reply With Quote
Old 1st January 2021, 09:23 PM   #9
GeoffMillar is offline GeoffMillar
diyAudio Member
 
Join Date: Oct 2016
Location: Melbourne, Australia
Default Textbook v Real

As a quick illustration of why 'textbook' calculators don't work, I've done a quick Xsim comparison of a textbook (first graph) vs properly measured and designed XO. The second XO takes the raw driver data and applies the correct XO parts, so it's not measured in box: the speaker actually has more BSC than the sim shows.

The red (tweeter) and orange lines (mid woofer) are the driver curves, the blue is the system response. You can see that the first is a real mess in terms of the drivers interfering with each other, plus it has a painful top end. Bleah, as Snoopy would think.

The calculated parts values were 6.8 cap and 0.4mH shunt inductor for the tweeter and 0.4mH and 3.3 shunt cap for the mid. The actual parts values for this speaker - which sounds really nice - are 6.8 series cap, 0.15 shunt inductor and 15 microfarad series cap plus L-Pad for the tweeter, and 1.3mH series inductor and 4.7 microfarad shunt cap for the mid.



I should add the the drivers in question are very smooth and well behaved, without peaks or resonances which need to be tamed. I won't name them as the project is the designers' intellectual property.

As an illustration of what makes this hobby so enjoyable, the project details were kindly provided to me by a well known DIY Audio member.


Geoff
Attached Images
File Type: jpg calculated XO BC25 and 830656.jpg (63.3 KB, 161 views)
File Type: jpg Measured XO 830656 and BC25.jpg (62.4 KB, 160 views)

Last edited by GeoffMillar; 1st January 2021 at 09:35 PM.
  Reply With Quote
Old 13th January 2021, 01:28 PM   #10
nannoo is offline nannoo
diyAudio Member
 
Join Date: Jan 2013
So I deliberately bit my lip because I hadn't taken time to pull this site apart (I hadn't even looked at the crossover calculator). I'm also very aware that I am not a degree qualified audio engineer; there's plenty of scope for me to learn.

All that said however, I'm also aware how this hobby is easy to iceberg.... So many people learn the tip and don't see the rest of what's under the water. Because of that there's a lot of beginners putting advice out there and it seems to be a case of the blind leading the blind.

For what it's worth, if you're an absolute beginner and MUST design using calculators and not measurements, ease do yourself a favour and see the sticky post
"Introduction to designing crossovers without measurement"
The only point I'd add to that is that baffle step compensation should be thought of as mandatory not optional in all speakers, but that's just my experience.

I found the linked site particularly alarming as there were a lot of odd ideas put forward as fact, and some gross errors which put me away from the benefit of the doubt category. All that said, here's a breakdown on the bits that raised an eyebrow for me:

-the author correctly points out that one octave relates to a doubling in frequency, then states that 3Hhz is in the middle of 2kHz and 4kHz.(I know I'm being a pedant on that one)

- "Some people believe that it is best to use a low order crossovers when possible, preferably only 1st order. This does have some benefits. With the greater frequency overlap, voices will not seem to jump from one driver to another as quickly as they would with a steep crossover. It also follows the minimalist approach where the simpler the circuit, the less distortion and modification of the signal is introduced"

In my opinion a true 1st order crossover is 1st order acoustic. That can require a very complex crossover to actually achieve. My experience has been that 1st order electrical is almost always motivated by the fact it's cheap and easy. The idea that voices jump from one driver to another goes against what Zaph said when he described a 2 way 6th order as being like a point source in sound. My understanding is that 1st order appeals for phase alignment reasons, however the also seems to be out on that. Also, I must say that the idea that crossover components introduce audible distortions is such a joke at this point. My experience has shown this isn't true, when something sounds wrong it has always been the crossover design not parts that have caused issue. If each component added more distortion my 22 component 3 way xo couldn't sound better than the 11 component version it replaced.

Also, the signal has to be modified in order for the speaker as a complete system to accurately reproduce it, otherwise no crossover would ever be needed.

"Another belief is that even order (2, 4, 6...) order crossovers should be avoided. Even order crossovers tend to have spikes or dips in the frequency response around the crossover point. These spikes can be as bad as -30db, but can easily be solved by reversing the polarity of only one of the speakers, limiting the spike to about +- 3db."

I read this and feel like the author has misunderstood what an even order crossover is. A 2nd order xo can theoretically have a large null if the polarity of the drivers is not reversed but since when was this a con of 2nd order designs? Further more many well respected companies swear by 4th order LR, hardly in any "should be avoided" category. My understanding is that even order crossovers are the only was to phase align driver output in the listening window (I'm open for input and discussion).I'm aware that Douglas self stated that 4th order LR is really the best crossover we can achieve at the moment, from a technical perspective, just one man's belief I know, but a man with excellent technical ability and understanding.

I would have been far more impressed with the site if it had a calculator that used the driver DCR and inductance and baffle size to make an off the shelf calculator, but when I eventually went to check what calculators they had, they were the generic nightmarish calculators that lead beginners into believing they've done everything right

The last part was the topology with the positive input for the tweeter filter coming from the positive output of the midrange HP. I've never seen anything like that (except in series xo's) and just looks horrid. Rightly or wrongly many people spend a lot of money on fancy caps for tweeter chains, if they were to do this they'd have to ensure the very large, expensive caps in the midrange filter were up to scratch. More importantly, surely this will make a Trainwreck of xo design, as the variable.impedance of the tweeter network would be visible to the midrange output, and the midrange high pass would, to some extent at least, interfere with the tweeter output?

Anyway, just my thoughts...
  Reply With Quote

Reply


Have to love the beginner guides out there.Hide this!Advertise here!
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Beginner Mod Guides? nearlygnarly Class D 7 23rd April 2016 07:39 PM
IC guides/beginner projects AndersJ Parts 8 8th October 2006 07:30 PM
i want to replace my wima caps in my valve amp is there better out there NAP 160 Parts 35 30th November 2002 09:39 PM


New To Site? Need Help?

All times are GMT. The time now is 05:56 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Copyright ©1999-2021 diyAudio
Wiki