Onwards and upwards. There's a topic in the forum today which runs:
What's a "How to Build to the Limits of Possibility" Steen Duelund loudspeaker design all about?
It turns out to be VERY INTERESTING indeed. You use 4th. order filters in a time aligned setup. The two way Linkwitz-Riley is relatively straightforward to grasp and has 4 components in the bass and 4 in the treble as you might expect. You need good drivers with overlap and impedance equalisation to get it to work. Steen gives the formulas to calculate values, BTW. He also tells you how to pad components with resistors to allow the real drivers' properties to modify it.
Where Steen really carries the ball forward is with three way systems. He really thinks about the demands on the drivers in practise. You should read the article, but the bottom line is that the more you ask the midrange to do, the nearer a fullrange unit it has to be. Just when I was thinking a 3-way is easier, I realise the problem is entirely more subtle.
Just my opinion, but I hate stuff like that. It is selfishly posed, and uninformative to everyone else who is learning here. I hope he/she gets what they deserve.
What's a "How to Build to the Limits of Possibility" Steen Duelund loudspeaker design all about?
It turns out to be VERY INTERESTING indeed. You use 4th. order filters in a time aligned setup. The two way Linkwitz-Riley is relatively straightforward to grasp and has 4 components in the bass and 4 in the treble as you might expect. You need good drivers with overlap and impedance equalisation to get it to work. Steen gives the formulas to calculate values, BTW. He also tells you how to pad components with resistors to allow the real drivers' properties to modify it.
An externally hosted image should be here but it was not working when we last tested it.
Where Steen really carries the ball forward is with three way systems. He really thinks about the demands on the drivers in practise. You should read the article, but the bottom line is that the more you ask the midrange to do, the nearer a fullrange unit it has to be. Just when I was thinking a 3-way is easier, I realise the problem is entirely more subtle.
An externally hosted image should be here but it was not working when we last tested it.
Guess you never met Lynn Olson.
(He's much taller in person)
LOL, only by Email. He is taller than I expected.
How far out does it need to be to support 4th order? Is that 4th electrical because I always took it to mean that you would use its natural premature roll off to be part of the 24db fall. A 2nd order electrical combined with natural roll off would result in a 4th order acoustic.
I dont get this whole complexity thing with crossovers...(bear in mind I am clueless)...but heres my take:
If you have drivers with clearly defined properties...like freq. resp., and say your doing a 3 way (ar 12 way for that matter)...why would you not just design a crossover where there is a slight over or under lap with each drivers freq. resp.?
You're covering the whole range right? Each driver is doing its thing in its correct range right? Assuming you get the box vol right etc., whats the big deal?....(hehe I like my ending on this one).
Ok, I know this is over simplification..but someone please explain.
If you have drivers with clearly defined properties...like freq. resp., and say your doing a 3 way (ar 12 way for that matter)...why would you not just design a crossover where there is a slight over or under lap with each drivers freq. resp.?
You're covering the whole range right? Each driver is doing its thing in its correct range right? Assuming you get the box vol right etc., whats the big deal?....(hehe I like my ending on this one).
Ok, I know this is over simplification..but someone please explain.
Why are there almost no high-end repeatable designs on the Internet ?
There are literally thousands of designs that compete with the Lo-Fi end of the commercial sector, ie sub. £2K, but nothing seems to come out that competes with the leviathans.
It seems that we need a Nelson Pass of the Loudspeaker World.
I've just sold my B-I-L's B&W 683s, I thought they were good but someone described them as Low-Fi.
The B&W 802Ds are MUCH better (OK I didnt pay £18K for them, I got them for a steal), but someone must have a DIY design than can match these BRILLIANT speakers.
There are literally thousands of designs that compete with the Lo-Fi end of the commercial sector, ie sub. £2K, but nothing seems to come out that competes with the leviathans.
It seems that we need a Nelson Pass of the Loudspeaker World.
I've just sold my B-I-L's B&W 683s, I thought they were good but someone described them as Low-Fi.
The B&W 802Ds are MUCH better (OK I didnt pay £18K for them, I got them for a steal), but someone must have a DIY design than can match these BRILLIANT speakers.
Last edited:
There's plenty of such designs that would wipe the floor with those if you're willing to accept that you don't need thousand dollar drivers to do it.
FWIW, the 683s may be lo-fi but that does not mean the 2K price point is lo fi. The best commercial stuff at that price points is literally "this" close to world class, normally only limited by the bass extension and max output. I can name a handful of < $2000 speakers that would have better fidelity over 802Ds
Could you please provide a few examples.
DALEKs - Aesthetics are only important to women.
I HATE BOSE - Some guys like them, mostly disco club owners.
Last edited:
Again - no definite designs.
Just corny comments.
What I am after is a REPEATABLE - NO EXPERIMENTING - HI END - DIY DESIGN.
Any experimenting means that the design is not yet complete.
OK so the designer may feel that he wants money for his PERFECT design - No longer a DIY design in my eyes (or ears).
I don't mind paying £10K for a BRILLIANT speaker, but I do not want to waste £2K on a poorer DIY design that needs £100s to get it working.
Just corny comments.
What I am after is a REPEATABLE - NO EXPERIMENTING - HI END - DIY DESIGN.
Any experimenting means that the design is not yet complete.
OK so the designer may feel that he wants money for his PERFECT design - No longer a DIY design in my eyes (or ears).
I don't mind paying £10K for a BRILLIANT speaker, but I do not want to waste £2K on a poorer DIY design that needs £100s to get it working.
Last edited:
We really shouldn't just opinionate without backing it up with solid reasons why a design is good. Let's look at the crossover problem first. This is BEFORE considering drivers that have the right properties and natural slopes.
Here's an idealised second order butterworth frequency response 2 way crossover I have been modelling:
Crossover 2.8kHz and 4.5dB bafflestep correction and in practise a 3dB peak at crossover. What's wrong with it? Group delay, the green curve.
The bass arrives later than the rest due to inductor bafflestep, and the peak at 2kHz is due to the Q=0.707 of the filter. The first dip and the eventual rollloff is inevitable just by it being a second order filter.
You can improve this with various strategies, bafflestep delay being rather easy to tame. But strangely, speakers with a good time response sound rather dull and the room characteristics become more important.
Oh, an example. A duelund implemented:
http://htguide.com/forum/showthread.php4?t=27218
Very nice.
Here's an idealised second order butterworth frequency response 2 way crossover I have been modelling:
An externally hosted image should be here but it was not working when we last tested it.
Crossover 2.8kHz and 4.5dB bafflestep correction and in practise a 3dB peak at crossover. What's wrong with it? Group delay, the green curve.
An externally hosted image should be here but it was not working when we last tested it.
The bass arrives later than the rest due to inductor bafflestep, and the peak at 2kHz is due to the Q=0.707 of the filter. The first dip and the eventual rollloff is inevitable just by it being a second order filter.
You can improve this with various strategies, bafflestep delay being rather easy to tame. But strangely, speakers with a good time response sound rather dull and the room characteristics become more important.
Oh, an example. A duelund implemented:
http://htguide.com/forum/showthread.php4?t=27218
Very nice.
Not looking for complex problems.
ONE). Make cabinet with proven deign.
TWO). Buy EXCELLENT transducers that actually are EXCELLENT.
THREE). Build CRossover from proven design.
As I keep saying - A REPEATABLY EXCELLENT - WORLD CLASS design.
I think that I can speak for most of the DIY fraternity. If this was possible then it would be a winner.
ONE). Make cabinet with proven deign.
TWO). Buy EXCELLENT transducers that actually are EXCELLENT.
THREE). Build CRossover from proven design.
As I keep saying - A REPEATABLY EXCELLENT - WORLD CLASS design.
I think that I can speak for most of the DIY fraternity. If this was possible then it would be a winner.
I read this same thing in Vance Dickason. Yes - can someone please explain?
For active crossovers will time delay on higher freq drivers fix that so the bass arrives when it should?
Before even getting into the speaker design you need to address the other variable factors that need to be universally standardized. Room design/size/shape/construction needs to be standardized between all DIYers. Then you need to have an agreed upon standard for amplification (tube/solid state/topology, etc.). If you can get the members of the forum to agree to these two things, then you've probably missed your calling as an international diplomat. Now try to get everyone to agree on what "good" sounds like, what "affordable" is, and what is aesthetically pleasing.
System7 your references to Duelund crossovers looks very interesting.
I just searched this forum for threads about Duelund and came up with a half a dozen or so examples all of which are very short.
Would you consider starting and contributing to a thread that breaks down Duelund crossovers for math and speaker idiots like me?
I would very much appreciate it, and I bet others would too!
I just searched this forum for threads about Duelund and came up with a half a dozen or so examples all of which are very short.
Would you consider starting and contributing to a thread that breaks down Duelund crossovers for math and speaker idiots like me?
I would very much appreciate it, and I bet others would too!
I'm working SO HARD at this, but you're not really paying attention. You gotta read the references.
Get this. A crossover is a filter. An enclosure is a filter. A room is a filter. A drive unit is a filter. Steen Duelund is saying that a Q of 0.5 is about as good as it gets. You want smooth slopes, not abrupt steps at crossover for good time delay characteristics.
Here's a Dynaudio D54 AF dome midrange, not a very good driver really, but it makes the point and looks a bit like a gaussian shape, let's guess it's 6dB per octave rolloff at either end:
So when you add the third order 18dB filter to the driver in an ideal world, you get the overall acoustic rolloff response of 24dB per octave.
Get this. A crossover is a filter. An enclosure is a filter. A room is a filter. A drive unit is a filter. Steen Duelund is saying that a Q of 0.5 is about as good as it gets. You want smooth slopes, not abrupt steps at crossover for good time delay characteristics.
Here's a Dynaudio D54 AF dome midrange, not a very good driver really, but it makes the point and looks a bit like a gaussian shape, let's guess it's 6dB per octave rolloff at either end:
An externally hosted image should be here but it was not working when we last tested it.
So when you add the third order 18dB filter to the driver in an ideal world, you get the overall acoustic rolloff response of 24dB per octave.
if you use the same drivers since many hi end design use scanpeak and seas , built good boxes and use decent crossover parts , why pay 18k for speakers ?
if you must buy i like JBL speakers , good price and sound quality
i´m happy with a decent 3 way and sub(s) , no need to sell the house car and family to buy speakers
if you must buy i like JBL speakers , good price and sound quality
i´m happy with a decent 3 way and sub(s) , no need to sell the house car and family to buy speakers
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.