Hi Eelco
This statement is basically right.
The main problem is that there is not a lot of literature on how to implement active filter circuits to achieve target acoustic slopes.
For those who are skilled in circuit design it is not that hard to achieve but if one has to rely on textbooks it is not that easy.
There was a good paper by John Kreskovsky where he showed how to derive the crossover transfer function for a given driver and a given target response, but it does of course not show how to get from that crossover function to a crossover circuit, simply because this part alone could fill books. One has to be aware that the circuit topology for achieving a desired target response can be quite different for different drivers in the worst case - not only the dimensioning of said circuit.
I once opened a thread showing how acoustic target slopes could be approximated for a transient-improved FAST system as an example:
http://www.diyaudio.com/forums/full-range/104958-transient-perfect-2nd-1st-order-crossover.html
Regards
Charles
+1 to Charles
Books on crossovers discuss how to build the functions, so once you know what you need that tells you the ways of doing it. Books on how to build active speakers that go into any depth are very few and far between. it's a x-functional discipline. And I agree that most of the literature takes the std studio 3-way design where you are crossing over well within band and then just glomming an LR4 on.
Or to put it another way a good passive crossover designer uses components to coax the drivers into harmony. A textbook active forces them to your will.
I think miniDSP and its brethren have really opened up the ability for people to experiment and certainly in the DIY arena the knowledge on how to get new levels of optimisation with the rapid prototyping abilities of DSP is expanding and that is good. Whether at the end of the day someone then builds passive or active they can get to a better level faster this way.
It's all good!
Books on crossovers discuss how to build the functions, so once you know what you need that tells you the ways of doing it. Books on how to build active speakers that go into any depth are very few and far between. it's a x-functional discipline. And I agree that most of the literature takes the std studio 3-way design where you are crossing over well within band and then just glomming an LR4 on.
Or to put it another way a good passive crossover designer uses components to coax the drivers into harmony. A textbook active forces them to your will.
I think miniDSP and its brethren have really opened up the ability for people to experiment and certainly in the DIY arena the knowledge on how to get new levels of optimisation with the rapid prototyping abilities of DSP is expanding and that is good. Whether at the end of the day someone then builds passive or active they can get to a better level faster this way.
It's all good!
Hi Charles and billshurv,
Those were exactly my points. Even if you pick the right topology, then the accuracy of your filter-functioning determining R's and C's becomes your your next issue.
Actually a fully optimized (=re-dimenioned) Bainter a la Hardman with BSC on the first opamp will get you well under way. But the again, many here have issues with 8th order acoustic.
In Speaker Builder, around 1990, a certain Lamy(?) described a state of the art analogue active 4 way filter (plus Motional Feedback for the woofers) including a host of gyrators to equalize driver imperfections. Opamp count in the dozens..
Regards,
Eelco
Those were exactly my points. Even if you pick the right topology, then the accuracy of your filter-functioning determining R's and C's becomes your your next issue.
Actually a fully optimized (=re-dimenioned) Bainter a la Hardman with BSC on the first opamp will get you well under way. But the again, many here have issues with 8th order acoustic.
In Speaker Builder, around 1990, a certain Lamy(?) described a state of the art analogue active 4 way filter (plus Motional Feedback for the woofers) including a host of gyrators to equalize driver imperfections. Opamp count in the dozens..
Regards,
Eelco
Well drive units are rarely matched to 1% so you can over worry.
Where it gets fun for me is where you have a situation that you use a hybrid. Take for example this Zaph speaker Zaph|Audio You can get LR2 response with minimal components but could do all sorts of optimisation upstream.
Where it gets fun for me is where you have a situation that you use a hybrid. Take for example this Zaph speaker Zaph|Audio You can get LR2 response with minimal components but could do all sorts of optimisation upstream.
This reminds me of the 90's when I told everyone on rec.audio.pro to sell their tape machines now and buy a PC or Mac because within a few years nobody would even make tape. They had hissies. Analog crossovers are on the way out. Soon more than obsolete as well as external amps. You can spit more n an hour than a fine amplifier weighs and driver will always need more correction than a reasonable passive can do. The Audio Infidel thinks passive is already as good as gone. Put wireless interconnect in the mix and it is a done deal. We are having an antique conversation now.
Ok, so I have a typewriter. A manual one. I wouldn't build a typewriter (umm 🙄..) and should have sold it while ribbons were plentiful. Tape decks are currently nice to look at too.
But an analogue crossover is always going to be a unique but useful masterpiece of DIY art, and that's the way I like it 😀
But an analogue crossover is always going to be a unique but useful masterpiece of DIY art, and that's the way I like it 😀
I've tried to address the problem of making an analogue active filter that gives a target acoustic slope by designing an active filter that mimics a passive one. Using Gyrators and FDNR's.... the FDNR's are the problem, they are finicky little things and difficult to get stable.
I just used speaker workshop to design my filter and then implemented that The coils and caps calculated for the passive just convert to resistors in the active. Hopefully over the next few weeks I will actually get it implemented (prototype has been done).
I'm not 100% happy with the objective performance of the FDNR, and will probably continue to search for a better implementation, but I think it will do me for now 😉
Tony.
I just used speaker workshop to design my filter and then implemented that The coils and caps calculated for the passive just convert to resistors in the active. Hopefully over the next few weeks I will actually get it implemented (prototype has been done).
I'm not 100% happy with the objective performance of the FDNR, and will probably continue to search for a better implementation, but I think it will do me for now 😉
Tony.
So like valves and vinyl? This is a DIY site not tech projections for mass market carp.
I see that many comments underestimate the power of dsp crossover systems. The genius lies in flexibilyty and versatility. The user must know what she/he is aiming for. multi-way loudspeakers are acoustic complexes, and the sound radiating from them follows acoustic laws, despite of crossover system controling them.
DSP-controlled xo, eq, delay and spl management must be done with sane acoustic evaluation/measurements. Actually then we don't care about drivers' efficiency, Fs, impedance, response curve etc at all. Distortion behaviour is most important "mechanical" factor that we cannot handle/overcome with dsp.
So, the challege with dsp-control is to master acoustic measurements and the rest is just trial and error! (If cabinet construction and driver choices were wise.)
Topic of this thread is "what quality of dsp wins over passive". I believe that every applicable dsp system has adequate quality. The operator is in charge and perhaps softwares are easy or difficult.
DSP-controlled xo, eq, delay and spl management must be done with sane acoustic evaluation/measurements. Actually then we don't care about drivers' efficiency, Fs, impedance, response curve etc at all. Distortion behaviour is most important "mechanical" factor that we cannot handle/overcome with dsp.
So, the challege with dsp-control is to master acoustic measurements and the rest is just trial and error! (If cabinet construction and driver choices were wise.)
Topic of this thread is "what quality of dsp wins over passive". I believe that every applicable dsp system has adequate quality. The operator is in charge and perhaps softwares are easy or difficult.
Not sure about 'mass market carp'.
When i was looking for a high-quality, analogue crossover with balanced in/outs I noticed that there were none left in production. The usual suspects Klark Teknik and BSS had switched to DSP years earlier while others just disappeared and were replaced by digital ones from new manufacturers (XTA et al).
May be they'll be making a nostalgia-based return like valves and vinyl did but I doubt it.
That's because the original poster asked about his active analog (SX45) and wanted to know about the components (opamps, caps) to surpass passive.
That would be an interesting topic.
I agree I like antiques. However I do come here for bleeding edge speaker tech fix. I have learned enough here get a degree!
My original questions was about active analog xover to surpass passive speaker network. But the thread evolved and now there are two more contenders: DSP and passive line level xover.
I will not be able to determine which is better, especially to find answer to my original question, without knowing flat response out of my speaker system, so I would try miniDSP first (which I ordered already).
I will not be able to determine which is better, especially to find answer to my original question, without knowing flat response out of my speaker system, so I would try miniDSP first (which I ordered already).
But, that would not be an analogue. I could recommend a cost effective unit, but if my memory serves correct, you want a unit that has at least one x-over frequency for a sub woofer (?)
Give "balanced" it's full title and it's implementation becomes more obvious.
Balanced impedance connection.
As Bills says it's the IMPEDANCE that must be "balanced".
Signal level has nothing to do with it.
The signal could be zero and the balanced impedance minimises the interferences that leak into the processor.
Note in that Gword link that the connections are "balanced", but internally the circuit converts to unbalanced to allow the volume control to happen accurately and then there is a passive unbalanced to balanced conversion, to make up the balanced impedance output connection.
That passive unbal to bal is in the an003 application note from Jensen.
https://www.google.co.uk/url?sa=t&r...vSIqHvaWA&sig2=_lUFLCDhcgKBz6uBDzXuMA&cad=rja
section 2.4 a simple alternative.
If it's good enough for Jensen and Putzeys, it should be good enough for all of us.
Ah !!!!!
Jensen have given in to our protestations and allowed access to their papers
http://www.jensen-transformers.com/...n-of-balanced-and-unbalanced-equipment/an003/
Last edited:
And yet, my petetrolly, on another thread you claim a creative unit with lots of DSP is 'perfection'. http://www.diyaudio.com/forums/loun...-presents-return-hifi-masses.html#post4593104 make your mind up!
In volume production maybe not, although I am sure Manley labs will still do tubed stuff! However for the DIY'er I think analog crossovers are only just starting. I tend to include line level passives in there, as they have an active device at each end.
It's also a valid stance to say 'Linkwitz etc have gone digital therefore analog is dead' but I actually think that the availability of miniDSP will over time increase the use of analog crossovers in DIY. But we shall see.
https://passlabs.com/products/preamplifiers/speciality/xvr-1
Bryston Limited: 10B-STD Electronic Crossover Can be ordered w/balanced I/O
electronic crossover, PLLXO, passive crossover, active crossover, custom amplifier Offer balanced option
Even Ashly offers balanced
http://ashly.com/wp-content/uploads/2015/08/xr-1001-2001-4001-crossovers-r01.pdf
- Status
- Not open for further replies.