Hi,
Its completely meaningless irrelevance, as real drivers are not pistons.
But it shows the case assuming drivers are used in pistonic regions,
for a 3 way probably only true for the bass driver.
The real off axis response of drivers is very different to the above.
If you think you can learn how to build a proper 3 way in a thread you
are sadly mistaken, you will get more misinformation than information.
undefinition (see FAQs)
Zaph|Audio
FRD Consortium tools guide
RJB Audio Projects
Speaker Design Works
HTGuide Forum - A Guide to HTguide.com Completed Speaker Designs.
Humble Homemade Hifi
DIY Loudspeaker Projects Troels Gravesen
Quarter Wavelength Loudspeaker Design
The Frugal-Horns Site -- High Performance, Low Cost DIY Horn Designs
Linkwitz Lab - Loudspeaker Design
Music and Design
rgds, sreten.
Last edited:
The answers you got on this are in the right direction but are incomplete. Some things which immediately spring to mind as left out are
1) passive component nonlinear distortions --- largely avoidable through capacitor and indictor selection
2) impedance shifts due to heating in the crossover passives and voice coils --- moves the crossover and passive EQ off target
3) removal of backwave crosstalk --- if the cabinet has poor return loss (most do) the backwave tries to move the cone and the power amp acts to suppress the resulting back EMF which is fine except 1) the crossover leaks some of that back EMF into the other drivers and 2) the crossover allows some of the power amp's correction into the other drivers
3) is not high SPL specific but since driver nonlinearity increases with excursion the convolution will widen somewhat and you'll see SFDR decrease.
There are more effects, such as microphony or magnetostriction, but those are usually negligible given good design.
No Sreten, I intend to learn to build a proper 3 way by building one (or a dozen).
I do, however, enjoy the discussion around it.Thanks again for the links - they are not wasted!
Last edited:
Re Twest820:
1) Are you speaking of using not high enough rated components? By 'magnetostriction' does this fall into this catgegory with non air-core inductors?
Speaking of which, just how much of a problem is that? I've read various opinions, and I have a background in electronics, but I just don't have any experience with it - and obviously it is topical for my 'Classic 3 way'.
I can certainly see how a non air-core inductor can solve the insertion loss problem, but can also see how it might create a saturation problem.
How big of a problem is this in poeple's experience?
2) Not sure what I can do about this.
3) Isn't this mitigatable by utilizing proper cabinet damping?
1) Are you speaking of using not high enough rated components? By 'magnetostriction' does this fall into this catgegory with non air-core inductors?
Speaking of which, just how much of a problem is that? I've read various opinions, and I have a background in electronics, but I just don't have any experience with it - and obviously it is topical for my 'Classic 3 way'.
I can certainly see how a non air-core inductor can solve the insertion loss problem, but can also see how it might create a saturation problem.
How big of a problem is this in poeple's experience?
2) Not sure what I can do about this.
3) Isn't this mitigatable by utilizing proper cabinet damping?
Saturating the core effectively removes it because it can no longer flux, and this dynamically varies the inductance.
I have noticed an apparent reverberation effect due to what I suspect was interaction between iron cored inductors.
Inductor resistance tends not to be critical in series resonant sections in parallel with a driver. I prefer air core here simply to reduce interaction.
With regards to voice coil heating, use large high efficiency drivers.
and, I find that damping material works well enough for me. I like fibreglass and rockwool.
I have noticed an apparent reverberation effect due to what I suspect was interaction between iron cored inductors.
Inductor resistance tends not to be critical in series resonant sections in parallel with a driver. I prefer air core here simply to reduce interaction.
With regards to voice coil heating, use large high efficiency drivers.
and, I find that damping material works well enough for me. I like fibreglass and rockwool.
1) There are plenty of cap measurements and inductor design discussions around that can introduce you the behavior of different dielectrics in caps and mutual inductance management. Magnetostriction's more likely in self wound air core inductors as the winding has more degrees of mechanical freedom but, as I mentioned, is typically negligible. (Lafe, I guess it's a matter of how low low resistance is; there's not exactly a shortage of zip cord inductors.)
2) I would think an active crossover would be an obvious solution. Granted, you're at a power point where the additional amp channels will cost more than the passives, but a lot of folks end up voicing speakers with DSP for the flexibility. Once you've done that, well, there's no engineering reason to actually build a physical crossover. Cost reasons, yes. Or I want my project this way, yes.
3) Not wholly, since achieving good return loss in the low bass requires enclosures which are usually impractically large. Look at the Linkwitz Pluto and B&W Nautilus for closed box design with decent return loss. Open baffle is often the easiest solution, though it's specifically excluded in this case.
2) I would think an active crossover would be an obvious solution. Granted, you're at a power point where the additional amp channels will cost more than the passives, but a lot of folks end up voicing speakers with DSP for the flexibility. Once you've done that, well, there's no engineering reason to actually build a physical crossover. Cost reasons, yes. Or I want my project this way, yes.
3) Not wholly, since achieving good return loss in the low bass requires enclosures which are usually impractically large. Look at the Linkwitz Pluto and B&W Nautilus for closed box design with decent return loss. Open baffle is often the easiest solution, though it's specifically excluded in this case.
its the reverse, and the question should be; what is the advantage of solid core over stranded
I don't really know the answer
other than simple cheap solid cores are always reliably good
and good enough to build the best possible speaker
I wouldn't waste my money on foil inductors etc
for the real big ones, I use wire wound toroid core
no, I'm saying that cheap solid core wire wound air core does the job
even if you want to build world class speakers design is 99%, and component quality/price is of very little importance
Ahhh.
My particular problem will be the series pass inductor for the woofer. I haven't calculated it yet, but it stands to reason that it will be pretty large.
My dilemma, which I understand is pretty common, is that the larger an air core inductor, the higher the resistance, which equates to insertion loss, which can have a pretty big effect on bass extension and overall speaker efficiency when the rest of the drivers are matched to the output of the woofer.
The common remedy to this is to use an iron or ferrite core inductor which decreases the resistance, but incurs the dangers of inductor saturation, right when and where you need it least!
Like I said, I don't know much about audio but I do have a background in electronics - I think it would be a neat trick if I could figure out a large low resistance inductor.
Just seems like it would be the best solution - and a solution to a common problem at that.
My particular problem will be the series pass inductor for the woofer. I haven't calculated it yet, but it stands to reason that it will be pretty large.
My dilemma, which I understand is pretty common, is that the larger an air core inductor, the higher the resistance, which equates to insertion loss, which can have a pretty big effect on bass extension and overall speaker efficiency when the rest of the drivers are matched to the output of the woofer.
The common remedy to this is to use an iron or ferrite core inductor which decreases the resistance, but incurs the dangers of inductor saturation, right when and where you need it least!
Like I said, I don't know much about audio but I do have a background in electronics - I think it would be a neat trick if I could figure out a large low resistance inductor.
Just seems like it would be the best solution - and a solution to a common problem at that.
One advantage of solid core is resistance verses cross sectional area, where the closer you can wind the coils, the greater the inductance and the less turns needed.
You'll often find that an inductor of double the inductance but otherwise similar has less than double the resistance. You can put two such inductors in parallel. Regardless, insertion loss can be figured in.
You'll often find that an inductor of double the inductance but otherwise similar has less than double the resistance. You can put two such inductors in parallel. Regardless, insertion loss can be figured in.
This will halve the resistance, but unfortunately diminishes the inductance as the reciprocal of the added reciprocals which halves the inductance as well! To get around that you would have to use inductors 4 times as large, which would give the original inductance but also twice the original resistance!
-edited-
-edited-
Last edited:
Ah...but when you produce an inductor of twice the inductance, it doesn't necessarily take twice the copper as the extra windings over the same area are more efficient.
My local sells 0.8mm air cores in common values for between 3 and 15 dollars each. 3mH has a DCR of 1.36 ohms, but 0m1H has a DCR of 0.28 ohms.
Anyway, I think DCR is a little overrated when at least specified sensibly.
My local sells 0.8mm air cores in common values for between 3 and 15 dollars each. 3mH has a DCR of 1.36 ohms, but 0m1H has a DCR of 0.28 ohms.
Anyway, I think DCR is a little overrated when at least specified sensibly.
Great thread, my main reason for posting is so that I can follow it..
There are a lot of options in terms of inductors for your X-O and frankly the DCR of the coil can be quite useful in managing the Q of the filter being implemented. In any event the DCR of the driver is likely to be 5 - 10X (or more) that of a reasonably chosen inductor.
Madisound has all sorts of very nice air core, foil, iron cored, etc., inductors, and often at very reasonable prices.
I've found film caps to be worth the extra money in a lot of instances (stable over time, and no inherent wear out mechanism if made correctly). I particularly like the Clarity SA series which are very reasonably priced and have reasonably unimpeachable technical performance. (And have fewer audible colorations than a certain equally inexpensive film cap from France for about the same money.)
There are a lot of options in terms of inductors for your X-O and frankly the DCR of the coil can be quite useful in managing the Q of the filter being implemented. In any event the DCR of the driver is likely to be 5 - 10X (or more) that of a reasonably chosen inductor.
Madisound has all sorts of very nice air core, foil, iron cored, etc., inductors, and often at very reasonable prices.
I've found film caps to be worth the extra money in a lot of instances (stable over time, and no inherent wear out mechanism if made correctly). I particularly like the Clarity SA series which are very reasonably priced and have reasonably unimpeachable technical performance. (And have fewer audible colorations than a certain equally inexpensive film cap from France for about the same money.)
Well Allen I think in my 'Classic 3 way' it is a definite consideration.
One calculation sheet I looked at recommended an 8 mH inductor. The resistance of an 8 mH inductor on partsexpress is 1.75 ohms. Lets say the impedance of my woofer at the crossover frequency is 8 ohms (which I think is fair) - that's nearly 22% of the amp output voltage being dropped at the crossover frequency. It's even worse from Fs to the crossover frequency, as the woofer impedance dips nearly to Re before raising again towards the crossover frequency!
For my application, obviously every care must be taken to eliminate insertion loss - which primarily is in that one single inductor.
I'll tell you a handy way to cheat I see in the math...
It's true if you add inductances in parallel both the resistance and the inductance halve - or is one third with three, one fourth if four, etc... but the trick here is to remember the rule of thumb about keeping inductors separated to keep them from interacting. What happens if you put them together?? Well, they interact, of course!
So, if you take your parallel inductors and put them closely together - zip tying and shellacking them right side by side would be best - the resistance would be half but the inductance now depends not just on their individual inductance but also on their magnetic coupling - which in this case would probably result in close to 99% of the original inductance value.
Additional inductors added to the pile would just keep decreasing the inductance, but as long as the mutual coupling was kept very close, the inductance would still stay close to the target value. With careful design and construction this could be easily accounted for, resulting in air core inductors easily one fourth or more of the original resistance.
The down side to this, of course, is that it can take a LOT of copper - but for someone who wants to build something very good, this will likely be no impediment.
I like this idea - in fact, I think I'll call it the 'Lafe Eric Inductor'.
One calculation sheet I looked at recommended an 8 mH inductor. The resistance of an 8 mH inductor on partsexpress is 1.75 ohms. Lets say the impedance of my woofer at the crossover frequency is 8 ohms (which I think is fair) - that's nearly 22% of the amp output voltage being dropped at the crossover frequency. It's even worse from Fs to the crossover frequency, as the woofer impedance dips nearly to Re before raising again towards the crossover frequency!
For my application, obviously every care must be taken to eliminate insertion loss - which primarily is in that one single inductor.
I'll tell you a handy way to cheat I see in the math...
It's true if you add inductances in parallel both the resistance and the inductance halve - or is one third with three, one fourth if four, etc... but the trick here is to remember the rule of thumb about keeping inductors separated to keep them from interacting. What happens if you put them together?? Well, they interact, of course!
So, if you take your parallel inductors and put them closely together - zip tying and shellacking them right side by side would be best - the resistance would be half but the inductance now depends not just on their individual inductance but also on their magnetic coupling - which in this case would probably result in close to 99% of the original inductance value.
Additional inductors added to the pile would just keep decreasing the inductance, but as long as the mutual coupling was kept very close, the inductance would still stay close to the target value. With careful design and construction this could be easily accounted for, resulting in air core inductors easily one fourth or more of the original resistance.
The down side to this, of course, is that it can take a LOT of copper - but for someone who wants to build something very good, this will likely be no impediment.
I like this idea - in fact, I think I'll call it the 'Lafe Eric Inductor'.
You can still make your inductor, but why don’t you try a massive laminated iron core inductor such as Erse Super Q? I have an Erse Super Q 13 mH inductor that has around or less than 0.5 ohms resistance.
Additionally I have a 10 mH inductor with 2.2 kg (~5 pounds) transformer core made from laminated amorphous iron. Still for avoiding any chance of saturation it has 1.5mm spacers between the two parts of E's and I's. It has resistance of 0.33 ohms with 120 windings of 1.2 mm wire (~0.05 in). Without the spacers it is a 30 mH inductor again with 0,33 ohms of resistance, or it could be a 10 mH one with ~0.1 ohms resistance (40 instead of 120 windings with fully closed core without spacers)...
Here it is: http://www.diyaudio.com/forums/full-range/180491-l-here-l-there-l-everywhere-2.html Never mind that is has a second winding, without it is a simple inductor.
You could try with iron core and adjust the crossover while you construct your inductor, but i doubt you would want an air core once you have tried a good big iron core.
The engineers who constructed it for me said that this thing had such low distortion, that it could be disregarded. Distortion was almost if not completely equal to a comparable air core.
Oh, by the way, what are you considering for dealing with the baffle step and different driver efficiency? A good method for dealing with baffle step is choosing a midrange that is ~6 db less efficient than the woofer, and crossing around the beginning of the baffle step - that way all added complications are avoided and for the untrained eye, the design stays simple as if the issue is not addressed.
By the way, although I am much younger, my understanding for a proper speaker is the same as yours... even though we both know that the listening at 120 db will be the first few times and for "scaring" of friends but there is a general rule (with many exceptions though) that who is capable of more is capable of better too...
Wish you luck!
Additionally I have a 10 mH inductor with 2.2 kg (~5 pounds) transformer core made from laminated amorphous iron. Still for avoiding any chance of saturation it has 1.5mm spacers between the two parts of E's and I's. It has resistance of 0.33 ohms with 120 windings of 1.2 mm wire (~0.05 in). Without the spacers it is a 30 mH inductor again with 0,33 ohms of resistance, or it could be a 10 mH one with ~0.1 ohms resistance (40 instead of 120 windings with fully closed core without spacers)...
Here it is: http://www.diyaudio.com/forums/full-range/180491-l-here-l-there-l-everywhere-2.html Never mind that is has a second winding, without it is a simple inductor.
You could try with iron core and adjust the crossover while you construct your inductor, but i doubt you would want an air core once you have tried a good big iron core.
The engineers who constructed it for me said that this thing had such low distortion, that it could be disregarded. Distortion was almost if not completely equal to a comparable air core.
Oh, by the way, what are you considering for dealing with the baffle step and different driver efficiency? A good method for dealing with baffle step is choosing a midrange that is ~6 db less efficient than the woofer, and crossing around the beginning of the baffle step - that way all added complications are avoided and for the untrained eye, the design stays simple as if the issue is not addressed.
By the way, although I am much younger, my understanding for a proper speaker is the same as yours... even though we both know that the listening at 120 db will be the first few times and for "scaring" of friends
but there is a general rule (with many exceptions though) that who is capable of more is capable of better too...Wish you luck!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.