For people who want good sound and strive for really fine points of design, seems pretty odd to be willing to have all kinds of coils, capacitors, and sometimes even resistors between Rice-Kellogg drivers and a very low output impedance amp. Positively quaint, even if cheaper to make and easier to market. Not to mention shortcomings of passive crossovering discussed above.
Seems to me if your voice coil isn't reflecting the activity of the cone (and the sound) then you've got a problem. If it is, then the low output impedance is helping control the voice coil. And more so for Kenwood Sigma Drive and other non-acoustic motional feedback systems. Seems pretty basic and important to me.
Seems to me if your voice coil isn't reflecting the activity of the cone (and the sound) then you've got a problem. If it is, then the low output impedance is helping control the voice coil. And more so for Kenwood Sigma Drive and other non-acoustic motional feedback systems. Seems pretty basic and important to me.
Last edited:
Genelec has had good luck 🙂
Well, I´m talking of slightly different applications where self powered stuff saves time...
Well as long as the design is well thought out all those caps and coils are a part of it. I have a couple of systems that I can run biamped or full passive. The only subjective difference that I hear is in the bass response. I would be curious if I could hear it blind. As I see it you can get stellar results using passive alone, passive and active or completely active. It's all in the implimentation. Done right they all work.
Rob🙂
Can you name ANY design choice in speakers couldn't be substituted for "passive" in your post? Every semi-legitimate option has its proponents and good implementers - look at the original Bose direct-reflecting speakers!
As is happens, you mention the most obvious place where even casual listening hears a difference. Maybe some of us organ music and serious music fans would positively barf at the boomy sound.
What is rare is to hear on such forums is reasoned prioritization of factors that matter. Loss of constraint of voice coils due to series impedance/resistance has to be near the top of my list - and I've experimented with electric motional feedback some. Granted, series impedance/resistance is one of the oldest criteria and so does not engage the hearts-and-minds of many of us.
Last edited:
With LspCAD or SoundEasy and a multichannel sound card, you can draw your active or passive schematic and play the result through the speakers. They will even emulate a DCX if that's what you like. Once you get something with good sound and measurements, you can buy the components and build the crossover.
At least with passives, if the itch strikes to build something new, you can give the old ones away and someone will be able to use them.
Ummmmmm, modeling! Perhaps I am naive, but does your system model the cone motion errors and the consequences of damping or impaired damping on less-than-perfect components? Seems hard to audition the shortcomings of a proposed system by hearing a computed model of the ideal output of that system through a computer sound card of uncertain quality, etc.? Would it make audible (for what "audible" is worth) what Rob reported above about the bass letting go some?
Last edited:
Ben, for your interest in how compromises effect performance, check out these Klippel tests you can take at home. First hand experience is interesting.
Audibility Of Distortion
Sorry these are only on test tones. I can't find the ones they used to have with music. Anyone know anything about that?
Dan
Audibility Of Distortion
Sorry these are only on test tones. I can't find the ones they used to have with music. Anyone know anything about that?
Dan
Your invective is a mighty inappropriate response.
How about addressing my question about how the proposed little sound card is going to "play" the distortion we're talking about so it can be auditioned, as you proposed?
Last edited:
What boomy sound would that be?? You use low dcr coils, problem solved.
Personally I am here for fun not a doctoral disertaion. I agree my post could be more concise and more detailed but frankly I don't have time to writed detailed posts and then get into a discussion that could go for hours face to face. Especially on the internet. To easy to get things wrong.
Are you talking about modelling various DCR values for the same coil to look for an optimum number for the coil/box combo??
Rob🙂
Same way it plays everything else.
Hey, the guy's looking for changes in the sound, the signal.
The sound card's behaviour is a constant. The physical structure of the speakers is a constant. The behaviour of the amp is a constant.
Even a budget card like an M Audio Transit should work fine if the other components are reasonable quality.
I think the point he is getting at would be something like a DCR value of a series inductor for a woofer in a reflex box. The program will give you the appropriate slopes for the drivers but you would not be able to hear the potential audible effects of parasitic components of the passive components used. Such as DCR for an inductor as an example of a "parasitic" parameter.
Rob🙂
Last edited:
It was a joke, sorry if you don't get it. 🙂
Obviously the final step is to build the passive (or active) crossover and listen to it, measure it, etc. You can even do a blind ABX between the computer version and the hardware version to see if those nasty caps and coils have ruined the sound.
As for "little" sound cards you can get a very high quality one for not too much money. It will be much quieter and probably lower distortion than the DCX discussed here and will avoid any extra ADA conversions -- just act as a DAC decoding a digital signal.
Since a kellog-rice transducer is a current device, it can be argued that the best way to do things is to have a driver with proper mechanical damping built-in and drive it with a current source (ie very low damping factor), instead of band-aiding things with a high damping factor amplifier with the inherent side effects.
dave
Gettin' warm. With a negative current source that mirrors the driver.
The driver has mechanical damping and feedback - the spider and surround. It just isn't linear enough and lacks an essential ingredient of feedback: gain. (Did I mention my interest in motional feedback?)
Thank you, Rob. Yes, call it parasitic or error or outside-the-model, but I am asking if there are perturbations that are not in the model, like distortion. If driver distortion were understood, it would be fixed or at least modelled. Until that time, please don't tell me to "listen" to abstract models that, as far as I know, not much, don't have the pieces of interest.
Frank is right that the comparison is the issue, not absolute judgment. But you can't make fine judgments with a poor set-up.
Why not? It would be easy - these are all linear effects we are talking about. It has been shown in several tests that inductors -air or steel - are linear except at very high magnitudes of current. And the nonlinearities are fairly low order - hence not highly audible. I've looked into all this - scientifically there is no case that you can make that active is "better" than passive in function. All the arguments are about implimentation or cost and they can go either way depending on the particular circumstances.
I think you are making assumptions that cause the conclusion you seek which kind of sounds like we are presently living in the best of all possible worlds, not least in how we paste together altogether dissimilar mechanical drivers through big lumps of coils, capacitors, and resistors. I don't really think so.
It would be great if things were all nice and linear or that the simple Klippel distortion factors were the whole story. If not, then I like to keep the driver electrically close to the low output impedance amp - and back to the topic of crossovers.
Hello Earl
What I was talking about was the ability to hear differences in simulated DCR values using SoundEasy. With that software you can "hear" your crossover before you build. The program simulates the curves through the sound card.
Obviously you can model the differences. But I don't see how you could hear them in relation to the damping/changes series resistance can play when you have a woofer in a tuned reflex cabinet. I don't see how the software could possibly model that interaction or how you could preview, listening for audible changes, based on coil dcr in the simulator using the output from the sound card??
Am I missing something??
Rob🙂
Ben not to get sidetracked, but it sounds more like Dr. Geddes is using the world we presently live in--not the best possible world.
The problems I see, as far as perfecting sound reproduction is concerned, is that there is a lot of research that would be nice to have that has never been done. The other is that people refuse to concede to the validity of what has been meaningfully demonstrated and refute it without well-grounded evidence. What a cantankerous conundrum! After every SONAR equipped submersible deployed returns with nothing, some people will still search for Nessie. Hopefully one will at least find some diamonds that cut glass.
Here's a list of DBTs that you may find useful:
ABX Double Blind Comparator Data
It's much easier to complain about what is (or not) than to actually do something about it. 😉 Do you have any ideas you'd like to put forth and/or evidence to present that would strengthen your present thinking? That would be helpful.
Oh, remember that excessive verbosity indubitably dissipates desirable semantic lucidity.
Thanks,
Dan
The problems I see, as far as perfecting sound reproduction is concerned, is that there is a lot of research that would be nice to have that has never been done. The other is that people refuse to concede to the validity of what has been meaningfully demonstrated and refute it without well-grounded evidence. What a cantankerous conundrum! After every SONAR equipped submersible deployed returns with nothing, some people will still search for Nessie. Hopefully one will at least find some diamonds that cut glass.
Here's a list of DBTs that you may find useful:
ABX Double Blind Comparator Data
It's much easier to complain about what is (or not) than to actually do something about it. 😉 Do you have any ideas you'd like to put forth and/or evidence to present that would strengthen your present thinking? That would be helpful.
Oh, remember that excessive verbosity indubitably dissipates desirable semantic lucidity.
Thanks,
Dan
Last edited:
Thank you, Dan. You've made several wise points and done so with a deftness and graciousness that I wish I had.
I think Rob's point of view is that there are subtleties that aren't in models and/or can't be auditioned satisfactorily by "listening to" a model.
Also in practice, you have manufacturing sloppiness too. Somewhere Earl talks about the extraordinary care they routinely take in the acceptance-testing of their passive crossover components and are justifiably proud of their sophisticated crossovers. So what does that mean about other, less quality-conscious manufacturers or DIYer who can't test coils and caps accurately? Slop in impedance values can throw the design way out of whack. And having two speakers (8 "pass" curves on two, 3-driver speakers) with wandering values will harm your sound a lot.
A further point I was trying to make is that it is good practice to have little between the driver's terminals and the low output impedance amp. Maybe that shouldn't be overstated because there always is VC resistance sitting there.
I'm not against progress but then there isn't much in active or passive crossovers that wasn't known and practiced a long time ago. I've been tri-amped (well, 2.5 amped with a mixed-bass sub) since 1967. For sure, once a DIYer goes bi-amped or tri-amped esp. with a multifunction adjustable crossover like the $180 Behringer CX3400, you'll never want to work with inflexible big lumps of coils and capacitors ever again. And if you don't like the sound of the Behringer, once you have found the crossover points and slopes and sub-sonic filtering you like in your music room, you can hardwire your own. For a manufacturer, the issues are different, of course.
The widely accessible new modeling software seems wonderful and as a kind of Rip Van Winkel, I like to learn more about it. The seductive risk with modeling is that you can wrongly suppose it tells you everything.
Thanks for ABX link.
I think Rob's point of view is that there are subtleties that aren't in models and/or can't be auditioned satisfactorily by "listening to" a model.
Also in practice, you have manufacturing sloppiness too. Somewhere Earl talks about the extraordinary care they routinely take in the acceptance-testing of their passive crossover components and are justifiably proud of their sophisticated crossovers. So what does that mean about other, less quality-conscious manufacturers or DIYer who can't test coils and caps accurately? Slop in impedance values can throw the design way out of whack. And having two speakers (8 "pass" curves on two, 3-driver speakers) with wandering values will harm your sound a lot.
A further point I was trying to make is that it is good practice to have little between the driver's terminals and the low output impedance amp. Maybe that shouldn't be overstated because there always is VC resistance sitting there.
I'm not against progress but then there isn't much in active or passive crossovers that wasn't known and practiced a long time ago. I've been tri-amped (well, 2.5 amped with a mixed-bass sub) since 1967. For sure, once a DIYer goes bi-amped or tri-amped esp. with a multifunction adjustable crossover like the $180 Behringer CX3400, you'll never want to work with inflexible big lumps of coils and capacitors ever again. And if you don't like the sound of the Behringer, once you have found the crossover points and slopes and sub-sonic filtering you like in your music room, you can hardwire your own. For a manufacturer, the issues are different, of course.
The widely accessible new modeling software seems wonderful and as a kind of Rip Van Winkel, I like to learn more about it. The seductive risk with modeling is that you can wrongly suppose it tells you everything.
Thanks for ABX link.
Last edited:
- Status
- Not open for further replies.