Hi,
I've been using DSP crossover for my loudspeaker projects as its very flexible and cheap to experiment with prototypes, doesn't mind driver impedances, sensitivities, can do delay, flat response etc. in general mucho flexibility, what ever is the requirement. Alright, please lets not get too deep into whats more expensive or cheap or laborous complicated or sounds better as there are million threads already for this kind of debate.
I've never dug deep into passive crossovers but I can see advantages outside from making some hi/lowpass slopes and notch filters. For example I use L-pad for compression driver to reduce (amplifier) noise along with a capacitor to protect it from mistakes. Quite cheap parts and effective results but doesn't necessarily have nothing to do making actual crossover happen, I've got DSP for it.
There seems to be some more, like adding inductor in series with woofer to reduce impedance variation when BL changes with excursion, or something along the lines, reducing distortion. Joe Rasmussen seems to write about it a lot. I've also seen big series capacitor used series with a woofer, which can change Q of the system and provide some boost allowing using somewhat smaller enclosure (sealed). Anyway, I'm not too deep into this but clearly there seems to be advantages with passive parts, outside of forming filtering as such.
What I'm interested in is get more info about good and bad things using passive crossover parts in active speaker systems, how to use them and what for? Have you used any and why? If you have only passive crossover implemented, what duties your crossover does in addition to just making acoustic slopes and filters happen?
Perhaps all of it is covered in some xo basics readings, didn't come up to one yet
I've been using DSP crossover for my loudspeaker projects as its very flexible and cheap to experiment with prototypes, doesn't mind driver impedances, sensitivities, can do delay, flat response etc. in general mucho flexibility, what ever is the requirement. Alright, please lets not get too deep into whats more expensive or cheap or laborous complicated or sounds better as there are million threads already for this kind of debate.
I've never dug deep into passive crossovers but I can see advantages outside from making some hi/lowpass slopes and notch filters. For example I use L-pad for compression driver to reduce (amplifier) noise along with a capacitor to protect it from mistakes. Quite cheap parts and effective results but doesn't necessarily have nothing to do making actual crossover happen, I've got DSP for it.
There seems to be some more, like adding inductor in series with woofer to reduce impedance variation when BL changes with excursion, or something along the lines, reducing distortion. Joe Rasmussen seems to write about it a lot. I've also seen big series capacitor used series with a woofer, which can change Q of the system and provide some boost allowing using somewhat smaller enclosure (sealed). Anyway, I'm not too deep into this but clearly there seems to be advantages with passive parts, outside of forming filtering as such.
What I'm interested in is get more info about good and bad things using passive crossover parts in active speaker systems, how to use them and what for? Have you used any and why? If you have only passive crossover implemented, what duties your crossover does in addition to just making acoustic slopes and filters happen?
Perhaps all of it is covered in some xo basics readings, didn't come up to one yet
Last edited:
You can shunt/parallel in passive xovers, whereas actives can only attenuate as if in series passive parts were used. Of course, there needs to be a part in series before such circuits as a voltage divider or the results don't work properly.
Since you have the series cap on tweeters, it'll be possible there. Without the coil on a woofer, the uses are reduced. However, this means you can tilt via a CR across a driver. You can notch breakup via an LC(R) across a woofer. You can LC(R) across a tweeter to eliminate the resonances in its stopband, and this is popular among the compression driver xovers as well as home hifi.
Off the top of my head, and I just got up, this is one major factor of difference of the 2 kinds.
Since you have the series cap on tweeters, it'll be possible there. Without the coil on a woofer, the uses are reduced. However, this means you can tilt via a CR across a driver. You can notch breakup via an LC(R) across a woofer. You can LC(R) across a tweeter to eliminate the resonances in its stopband, and this is popular among the compression driver xovers as well as home hifi.
Off the top of my head, and I just got up, this is one major factor of difference of the 2 kinds.
Thanks, it takes some time for me to digest what you wrote
Just in case I was not giving all thoughts on the opening post that I was thinking active system as such that it would work just fine without any passive parts, amplifier per driver. Filters before amplifiers can make the final target frequency response be there passive parts or not. So main interest is in: if using some passive parts in between amplifier and driver on an active multiway speaker would bring additional benefits, such as reduction of noise or distortion or work as protection, affect driver/box Q and alike.
Perhaps should also line out that the passive parts can participate making filtering as part of final crossover slopes but they don't have to, filters before amplifiers can do that part and even compensate for what passive parts do to some extent, like boost back what series resistance took away or make 1st order passive filter a 4th order acoustic filter etc.
I guess the opening post could have also left the DSP stuff out completely and just talk about double duty of passive crossover parts, how they affect multiway loudspeaker system other than just making acoustic slopes. What else we can do with passive parts between amp and driver other than make crossover slopes happen and equalize response? Something that cannot be done with signal manipulation before the amplifier.
Just in case I was not giving all thoughts on the opening post that I was thinking active system as such that it would work just fine without any passive parts, amplifier per driver. Filters before amplifiers can make the final target frequency response be there passive parts or not. So main interest is in: if using some passive parts in between amplifier and driver on an active multiway speaker would bring additional benefits, such as reduction of noise or distortion or work as protection, affect driver/box Q and alike.
Perhaps should also line out that the passive parts can participate making filtering as part of final crossover slopes but they don't have to, filters before amplifiers can do that part and even compensate for what passive parts do to some extent, like boost back what series resistance took away or make 1st order passive filter a 4th order acoustic filter etc.
I guess the opening post could have also left the DSP stuff out completely and just talk about double duty of passive crossover parts, how they affect multiway loudspeaker system other than just making acoustic slopes. What else we can do with passive parts between amp and driver other than make crossover slopes happen and equalize response? Something that cannot be done with signal manipulation before the amplifier.
Last edited:
I think these can be done with DSP filters as well, notch out resonances. Is there additional benefits using the passive parts other than just EQ a peak out? Is there some back-EMF stuff going on or something?
Is this the answer for my question above? If it is, what does it mean in practice?
Here, found something interesting by googling about series and parallel notch filters on cone breakup. It compares various passive notch filter schemes to take out cone break up and shows difference in distortion.
https://purifi-audio.com/wp-content/uploads/2022/03/220211_R05-Notchfilter.pdf
I guess a DSP system would not have such reduction in distortion and distortion would remain higher on system level as well? I'm not sure I understand why the reduction in distortion actually happens here or is it audible or not and what would happen if the peaks were cut with EQ before amplifier.
Alright, gotta compare passive and DSP for a second if there is such an effect for the system from passive parts, actual reduction of distortion or other benefitical "side-effects" than just being a crossover, to make actual measurable and perceptible difference for better, then passive crossovers have advantage over DSP. Reason for the thread. DSP can make the response straight and true but if there is even more that can be done by introducing handful of passive parts, well, perhaps worth it. Or the other way around, not worth to have DSP if response of the drivers is already quite nice and passive crossover doesn't become too expensive. Or third way to look at it, does this mean that if we had very good drivers and unlimited currency units for any number of passive parts there is no way an active (only) system could sound better because it would lack the benefitical side-effects, what ever those are? And my system, not so smooth drivers so DSP is kind of mandatory for smooth response, how can I reap some of the benefits of passive crossovers as a hybrid solution?
https://purifi-audio.com/wp-content/uploads/2022/03/220211_R05-Notchfilter.pdf
I guess a DSP system would not have such reduction in distortion and distortion would remain higher on system level as well? I'm not sure I understand why the reduction in distortion actually happens here or is it audible or not and what would happen if the peaks were cut with EQ before amplifier.
Alright, gotta compare passive and DSP for a second
if there is such an effect for the system from passive parts, actual reduction of distortion or other benefitical "side-effects" than just being a crossover, to make actual measurable and perceptible difference for better, then passive crossovers have advantage over DSP. Reason for the thread. DSP can make the response straight and true but if there is even more that can be done by introducing handful of passive parts, well, perhaps worth it. Or the other way around, not worth to have DSP if response of the drivers is already quite nice and passive crossover doesn't become too expensive. Or third way to look at it, does this mean that if we had very good drivers and unlimited currency units for any number of passive parts there is no way an active (only) system could sound better because it would lack the benefitical side-effects, what ever those are? And my system, not so smooth drivers so DSP is kind of mandatory for smooth response, how can I reap some of the benefits of passive crossovers as a hybrid solution?
Last edited:
The Purifi reading is interresting reading.
Important to understand that it is harmonics at the peaks that makes the distortion.
So it must be done passive or with some very advanced noise cancelling. The fundamental frequencies that generates the harmonics are in the passband.
Important to understand that it is harmonics at the peaks that makes the distortion.
So it must be done passive or with some very advanced noise cancelling. The fundamental frequencies that generates the harmonics are in the passband.
Similar low pass filters or LRC stopband filters can also remove high frequency noise from high output wofers with underdamped highs.
The poweramp makes the noise and the noise are played by the woofer, even if it the signal is filtered out by active filter.
Not much problem with home elements, but pro drivers of let's say 15" can have 110 db sensitivity at high frequencies
The poweramp makes the noise and the noise are played by the woofer, even if it the signal is filtered out by active filter.
Not much problem with home elements, but pro drivers of let's say 15" can have 110 db sensitivity at high frequencies
^I could, especially with some other concept/set of drivers but this one would be quite expensive passive as its 3 way with low crossover point, very mixed sensitivity stuff with passive cardioid mid, lumpy response pa drivers yada yada already have DSP and had the drivers, so going with it. Anyway, got system with DSP and its working nicely and interested on possible further improvements.
already have DSP and had the drivers, so going with it. Anyway, got system with DSP and its working nicely and interested on possible further improvements.Of course you can make a hybrid x/o. Think e.g. of a 3 way system in which you digitally x/o @ 500 Hz, so one leg rus from say 10 Hz to 500 Hz and the other leg runs from 500Hz upwards. You can then passively filter between mid and tweeter at, say, 2500 Hz. You still can equalize and optimize the respective passbands (both Lo and MidHi) digitally, yet use one amp for mid and high and use an (optimized) passive filter for the stopbands.
That was the initial plan but both mid and high have bad response around crossover so didn't happen different set of drivers would work fine on a setup like that, and save on amplifiers and dsp channels.
Alright, but perhaps its not too important what my system is, in a way, looking for stuff thats not doable with DSP I mean, with the knowledge perhaps next systems get better.
different set of drivers would work fine on a setup like that, and save on amplifiers and dsp channels.Alright, but perhaps its not too important what my system is, in a way, looking for stuff thats not doable with DSP
I mean, with the knowledge perhaps next systems get better.
For sure, and its not too far away from it if I try inductor with the mid, few more parts away from making slopes. Mid has obviously some breakup, thats now easily flattened out with DSP but if passive notch has advantages over just equalizing the response then I should try that as well and it would further approach this solution.
The Purifi paper is illustrating the distortion caused by driving variable impedance speaker motors with near zero output impedance voltage drive amplifiers and passive crossovers. (Esa wrote the book on that: https://www.current-drive.info/ ) They are showing how the distortion is reduced by effectively increasing the amplifier output impedance by adding series passive crossover components. The distortion highlighted by Purifi can be avoided from the start by using a transconductance amplifiers ( very high output impedance ). When the variation in driver impedance with cone motion is small with respect to the drive impedance, it won't significantly modulate the current. As cone acceleration is proportional to current, commanding current in the coil produces the best results. Of course the DSP can produce any notch filter with ease and compensated for any low damping of each driver.Here, found something interesting by googling about series and parallel notch filters on cone breakup. It compares various passive notch filter schemes to take out cone break up and shows difference in distortion.
https://purifi-audio.com/wp-content/uploads/2022/03/220211_R05-Notchfilter.pdf
I guess a DSP system would not have such reduction in distortion and distortion would remain higher on system level as well? I'm not sure I understand why the reduction in distortion actually happens here or is it audible or not and what would happen if the peaks were cut with EQ before amplifier.
Alright, gotta compare passive and DSP for a second
I think hybrid designs are useful for adding a fourth driver to a three channel DSP system. If you had woofer, mid woofer, midrange and tweeter. The midrange and tweeter crossover could be passive, as those components are small and cheap. If you are using transconductance amplifiers, that passive crossover would be a series crossover. I don't think a properly designed active system can be matched in performance by a system with passive crossover components.
Last edited:
Dunno if this counts, but I prefer to pad down compression drivers with a bridged balanced T-pad using wire wound resistors rather than decreasing the treble amp amplitude. Does stuff for the (often) complicated impedance curves on these devices that is only achievable through direct interaction with the driver.
^^ Current drive / transconductance amplifiers are interesting but as they are relatively rare I have not come accross too much information about them. There seem to be some projects available on the forum but utilizing class D amplifiers for this would be cool, not sure if its something doable? I remember someone sayin Bruno Putzney tried /used one, perhaps some Hypex module, as current amplifier but thats also something. Hopefully something like that would become more common in future as class-D systems with all kinds of frontend stuff are more and more common, and seem to be quite good quality with low price point. But, I believe its problematic because the speaker /crossover needs to be paired with it, to go with a current amp? Perhaps it stays niche, and would be quite cool for us DIY folks to exlpore and enjoy. DSP based systems would seem good candidate for such amplifier.
Do you have current amp, worth it?
Do you have current amp, worth it?
Last edited:
Whats the benefit of bridged T over simple L-pad? Yes on hifi systems some kind of pad between amp and compression driver seems very logical, I'm using simple L-pad.
Last edited:
Forgot to ask about this, do you know how audible this distortion is? Is it visible with multitone measurement, so that it can be tested / evaluated at home?
I suspect this is something that two way speakers would suffer much more than three way speakers. I guess my broad guestion is if the excursion and related impedance modulation is ever a concern for a three way speaker mid for example? In other words any idea of magnitude of "the problem"?
Well, its interesting enough that I have to order some inductors to play with
On my three way system, 1" CD, 8" mid and 15" woofers, PA drivers, amplifier his/noise is audible from the CD and mid, not much on the woofer in comparison. CD has about 6db pad on it and its still perceptibly louder than whats coming from the mid, perhaps because it has wider polar response past few kilohertz, mid has falling power response due to beaming.
I have quite noisy cheap classD amp with too much gain now so noise is a problem. I have no idea if it would be better having passive xo just for the noise or not, havent tested it. Certainly better low noise amplificatiin will make big difference in case of sensitive speakers.