Active Crossover schematic for SDA 2.3TL attached here
daifanshi
Posts: 46
After a few nights of work I was able to implement the passive crossover of the 2.3TL as an active network. The main 2nd order filters of the tweeter, dimensional, and stereo arrays are modeled using Sallen-Key type active filters. Consider this the first revision of the design. Minor changes are sure to happen. Values need to be standardized and there is room for parts reduction. A schematic is attached.
The design is custom and off-the-shelf active crossover units will require SIGNIFICANT modification if you choose to use them.
Each speaker "box" needs three amplifiers. One for the tweeter array, one for the stereo array, and one for the dimensional array. All amplifiers should have the same gain and polarity and be of reasonably good quality. I would say 60 watts per channel is enough. More than 100 watts is probably excessive. These amplifiers should be stable down to 4ohms. But they do not need to be Krells that work at 1 ohm. The amplifier loads are VERY friendly and there is an improvement in the speaker system efficiency with the active approach.
The entire dimensional and stereo crossovers are bypassed and are active, but we are keeping a few passives for the tweeter array for reasons that are discussed below. The SDA "umbilical" is no longer required since it exists in the active crossover.
The schematic is "bare-bones" and does not include some additional features that I would like to add such as adjustable channel gain and phase adjustments for the tweeter and dimensional arrays. Also the power supply is not included. But this should give you an idea of what the complexity is. I have left some placeholders for level adjustment.
All caps and resistors should be close to 1%. Resistors can all be 1/4 watt leaded or surface mount. Caps should be of a stable plastic film type and voltage ratings do not need to exceed 20 volts. In many cases they can be rated as low at 6.3V. There are NO lumped inductors in the design. There are many op-amp options available and I chose the
high performance OPA2134 for their low cost and wide availability (and I have a stick of them in my garage.
). You'll probably want at least +10V/-10V rails on the op-amps. These supplies should be linearly regulated and clean.
A few notes and descriptions follow:
1. The tweeter array is driven through a 2nd order highpass filter. Through simulation of the original passive crossover, I have reproduced the amplitude response, Q and susbsequent phase relationships using op amps. For the 2nd order woofer and tweeter response to sum correctly, the tweeter array has reverse polarity as it was in the original passive design.
To limit the required amplifier channels to six instead of ten(!), the tweeter amplifier will also drive the connected series elements of the three tweeters for the "disappearing
point source" response. The 2nd-order highpass in the tweeter crossover is active and in the schematic. In my experience midrange and woofer drivers tend to gain the most benefit from active crossovers. We can always eliminate all the passives in the future if we wish.
2. The midbass stereo array is driven directly by an amplifier without any passives in between. There is a 2nd order low-pass filter in the active crossover board. Again, the electrical behavior of the original passive network was reproduced.
3. The dimensional array filter is by far the most complex. The passive crossover design demonstrates some non-obvious behavior. In addition to opposite speaker IAC signal, the local stereo signal is also intentionally driven into the local dimensional array. The response is a shelving lowpass where the local bass signal is sent to the dimensional drivers to help the bass. The midrange from the local stereo signal is suppressed around 6-8dB after which there is an additional 2nd order low pass. This makes sense. No point in throwing away the bass potential of the dimensional drivers.
Important also is that the actual dimensional drive signal from the opposite speaker is NOT a simple inverted low-pass of the opposite side stereo signal. There is a shunt inductor in the original passive crossover that produces a first order high pass function to suppress
the bass signal (<100Hz) from the far speaker. The reason for this is intuitively obvious. Since bass is typically monaural and not stereophonically localized, Polk did NOT want the dimensional drivers to cancel the bass from the speakers. It robs amplifier power and probably produces some weird bass effects. I must apologize if any of this is obvious and has been stated/observed/surmised before in a paper or this forum. Undesired bass cancellation is something we can probably improve at a later time.
The circuit replicates both these "features".
Both the processed local stereo and opposite channel dimensional drive signals are summed before going to the dedicated dimensional amplifier. There are no passives between the amplifier and drivers. (unless you count wire and terminals as passives...)
If there is enough interest in what I am doing I can make the schematic more "complete" by adding everything that I think should be included. I can also produce a block diagram for those less technical. I have simulation results of just about everything in the schematic.
Send me a PM if you're interested and know what to do with a pSpice netlist. The design methodology is straight forward and we can discuss that privately if you wish. Also any of the other SDA speakers can be done with active crossovers.
Please be aware that this is NOT a plug and play modification. There are some electrical modifications needed on the passive crossover in the speaker (bypassing and disconnecting).
For those who are sensitive to such things, please understand that there is no sawing, drilling or other physical changes to the speakers. But I do suggest a new terminal cup be installed that includes a third set of terminals for the dimensional drivers. The construction of the active crossover module will require the ability to solder, test, and troubleshoot.
The advantages (and a few disadvantages) to using active crossovers are very well documented in many places and technical papers. Please read up and research before posting negative knee-jerk responses. I found this on the web:
http://sound.westhost.com/biamp-vs-passive.htm
I have no affiliation and it's a good overview. He missed a few advantages, but it's mostly complete. There are other similar documents available elsewhere.
If you're genuinely interested in contributing to this project, please PM me. There are a lot of practical circuit issues that need to be addressed before anything is built. The more people that are involved the better.
And please ignore this thread if you don't like this post or the premise or anything related to active crossovers.
At some point it may make sense to start posting followups in the DIY forum. Thanks!
The design is custom and off-the-shelf active crossover units will require SIGNIFICANT modification if you choose to use them.
Each speaker "box" needs three amplifiers. One for the tweeter array, one for the stereo array, and one for the dimensional array. All amplifiers should have the same gain and polarity and be of reasonably good quality. I would say 60 watts per channel is enough. More than 100 watts is probably excessive. These amplifiers should be stable down to 4ohms. But they do not need to be Krells that work at 1 ohm. The amplifier loads are VERY friendly and there is an improvement in the speaker system efficiency with the active approach.
The entire dimensional and stereo crossovers are bypassed and are active, but we are keeping a few passives for the tweeter array for reasons that are discussed below. The SDA "umbilical" is no longer required since it exists in the active crossover.
The schematic is "bare-bones" and does not include some additional features that I would like to add such as adjustable channel gain and phase adjustments for the tweeter and dimensional arrays. Also the power supply is not included. But this should give you an idea of what the complexity is. I have left some placeholders for level adjustment.
All caps and resistors should be close to 1%. Resistors can all be 1/4 watt leaded or surface mount. Caps should be of a stable plastic film type and voltage ratings do not need to exceed 20 volts. In many cases they can be rated as low at 6.3V. There are NO lumped inductors in the design. There are many op-amp options available and I chose the
high performance OPA2134 for their low cost and wide availability (and I have a stick of them in my garage.
). You'll probably want at least +10V/-10V rails on the op-amps. These supplies should be linearly regulated and clean.A few notes and descriptions follow:
1. The tweeter array is driven through a 2nd order highpass filter. Through simulation of the original passive crossover, I have reproduced the amplitude response, Q and susbsequent phase relationships using op amps. For the 2nd order woofer and tweeter response to sum correctly, the tweeter array has reverse polarity as it was in the original passive design.
To limit the required amplifier channels to six instead of ten(!), the tweeter amplifier will also drive the connected series elements of the three tweeters for the "disappearing
point source" response. The 2nd-order highpass in the tweeter crossover is active and in the schematic. In my experience midrange and woofer drivers tend to gain the most benefit from active crossovers. We can always eliminate all the passives in the future if we wish.
2. The midbass stereo array is driven directly by an amplifier without any passives in between. There is a 2nd order low-pass filter in the active crossover board. Again, the electrical behavior of the original passive network was reproduced.
3. The dimensional array filter is by far the most complex. The passive crossover design demonstrates some non-obvious behavior. In addition to opposite speaker IAC signal, the local stereo signal is also intentionally driven into the local dimensional array. The response is a shelving lowpass where the local bass signal is sent to the dimensional drivers to help the bass. The midrange from the local stereo signal is suppressed around 6-8dB after which there is an additional 2nd order low pass. This makes sense. No point in throwing away the bass potential of the dimensional drivers.
Important also is that the actual dimensional drive signal from the opposite speaker is NOT a simple inverted low-pass of the opposite side stereo signal. There is a shunt inductor in the original passive crossover that produces a first order high pass function to suppress
the bass signal (<100Hz) from the far speaker. The reason for this is intuitively obvious. Since bass is typically monaural and not stereophonically localized, Polk did NOT want the dimensional drivers to cancel the bass from the speakers. It robs amplifier power and probably produces some weird bass effects. I must apologize if any of this is obvious and has been stated/observed/surmised before in a paper or this forum. Undesired bass cancellation is something we can probably improve at a later time.
The circuit replicates both these "features".
Both the processed local stereo and opposite channel dimensional drive signals are summed before going to the dedicated dimensional amplifier. There are no passives between the amplifier and drivers. (unless you count wire and terminals as passives...)
If there is enough interest in what I am doing I can make the schematic more "complete" by adding everything that I think should be included. I can also produce a block diagram for those less technical. I have simulation results of just about everything in the schematic.
Send me a PM if you're interested and know what to do with a pSpice netlist. The design methodology is straight forward and we can discuss that privately if you wish. Also any of the other SDA speakers can be done with active crossovers.
Please be aware that this is NOT a plug and play modification. There are some electrical modifications needed on the passive crossover in the speaker (bypassing and disconnecting).
For those who are sensitive to such things, please understand that there is no sawing, drilling or other physical changes to the speakers. But I do suggest a new terminal cup be installed that includes a third set of terminals for the dimensional drivers. The construction of the active crossover module will require the ability to solder, test, and troubleshoot.
The advantages (and a few disadvantages) to using active crossovers are very well documented in many places and technical papers. Please read up and research before posting negative knee-jerk responses. I found this on the web:
http://sound.westhost.com/biamp-vs-passive.htm
I have no affiliation and it's a good overview. He missed a few advantages, but it's mostly complete. There are other similar documents available elsewhere.
If you're genuinely interested in contributing to this project, please PM me. There are a lot of practical circuit issues that need to be addressed before anything is built. The more people that are involved the better.
And please ignore this thread if you don't like this post or the premise or anything related to active crossovers.
At some point it may make sense to start posting followups in the DIY forum. Thanks!
Post edited by daifanshi on
Comments
-
Very interesting. I don't know enough about the technical aspects of your project to comment directly. But, I applaud your interest in SDA's and your interest in experimenting with them in this crazy audio hobby we have.
Do you plan to build the active crossover and give it a try?Carl -
Too bad you're in CA, I'd love to compare your finished project with mine. What amps are you considering? Do you plan on adding additional circuits to run those 6 amps? What do you plan on using for interconnects and speaker cables? What is the estimated cost for the crossovers, power supplies, additional amps, additional circuits and all necessary cables?
The above are direct questions, so how about some direct answers this time.Political Correctness'.........defined
"A doctrine fostered by a delusional, illogical minority and rabidly promoted by an unscrupulous mainstream media, which holds forth the proposition that it is entirely possible to pick up a t-u-r-d by the clean end."
President of Club Polk -
Too bad you're in CA, I'd love to compare your finished project with mine. What amps are you considering? Do you plan on adding additional circuits to run those 6 amps? What do you plan on using for interconnects and speaker cables? What is the estimated cost for the crossovers, power supplies, additional amps, additional circuits and all necessary cables?
The above are direct questions, so how about some direct answers this time.
Expensive high current amps are not necessary for this application since the loads seen by the amplifiers will be the same as the raw drivers in whatever series/parallel combination they will be in. (8 or 4 ohms with no dips.)
I have gotten along fine using an old 1990's Kenwood 6 channel home theatre amp that puts out about 100 w/ch. And I plan to reuse this with the design. But in all-in-one speakers I have used inexpensive "chip-amps" with excellent results. Build more if you need to.
There are a lot of old 4-channel HT amps around in the used market. Assuming you already have a suitable two-channel amplifier, an additional 4-channels should not cost you more than 200 dollars.
There will be no additional circuits required for running all six channels. The schematic drive signals are connected directly to the amplifier inputs with standard cables. A preamp feeds the input. That's all.
The actual cost of the crossover is obviously dependent on what you have lying around. But if you have nothing, 10 dual OPA2134 op-amps will run you about $25. The resistors from my local surplus place are 10 cents each so lets say $8 dollars for those. The caps are about 30cents from my local shop a piece so lets say $15 bucks. A used 10 watt 24VAC wall wart will cost you 5 bucks maybe. The linear regulators will cost you $2 bucks for both. The RCA phono jacks are about $10 bucks total where I am. Cables shouldn't cost anything. Inexpensive 14gauge speaker wire is MORE than sufficient as are phono cables that came with your VCR. Reuse those...
If you use a custom PCB it will add 200 dollars more. Use your proto-board or the dead-bug approach and it is cheap. The circuit is compact, you should be able to find a 10 dollar piece of copper clad board that will fit the circuit.
So I would say conservatively the crossover/cable cost is under $100. Add $200 bucks for four more amp channels.
But I'm sure others could find a way to spend more...
cheers. -
Very interesting. I don't know enough about the technical aspects of your project to comment directly. But, I applaud your interest in SDA's and your interest in experimenting with them in this crazy audio hobby we have.
Do you plan to build the active crossover and give it a try?
Yes, but it may be a few weeks before I can begin. I plan to start by modifying one speaker in steps and then do comparisons with the unmodified one for voicing purposes. Then I'll finish the second. The basic character of the speaker should not change with the modifications. The purpose isn't to create a whole different sound. -
Thanks for your answers, about what I expected. Have fun, I'm outta here.Political Correctness'.........defined
"A doctrine fostered by a delusional, illogical minority and rabidly promoted by an unscrupulous mainstream media, which holds forth the proposition that it is entirely possible to pick up a t-u-r-d by the clean end."
President of Club Polk -
When it's all done, how about you bring your stuff by and we can try it on a real 6 channel amp.
"He who fights with monsters should look to it that he himself does not become a monster. And when you gaze long into an abyss the abyss also gazes into you." Friedrich Nietzsche -
The passive crossover design demonstrates some non-obvious behavior. In addition to opposite speaker IAC signal, the local stereo signal is also intentionally driven into the local dimensional array. The response is a shelving lowpass where the local bass signal is sent to the dimensional drivers to help the bass. The midrange from the local stereo signal is suppressed around 6-8dB after which there is an additional 2nd order low pass. This makes sense. No point in throwing away the bass potential of the dimensional drivers.
Important also is that the actual dimensional drive signal from the opposite speaker is NOT a simple inverted low-pass of the opposite side stereo signal. There is a shunt inductor in the original passive crossover that produces a first order high pass function to suppress
the bass signal (<100Hz) from the far speaker. The reason for this is intuitively obvious. Since bass is typically monaural and not stereophonically localized, Polk did NOT want the dimensional drivers to cancel the bass from the speakers. It robs amplifier power and probably produces some weird bass effects. I must apologize if any of this is obvious and has been stated/observed/surmised before in a paper or this forum. Undesired bass cancellation is something we can probably improve at a later time.
Now this is something I've been very curious about. Being less than a year into this loudspeaker hobby (on a somwehat serious basis), I don't fully possess the knowledge to completely get this yet. But, I'm almost having a eureka moment here. I do have a couple questions; maybe in another couple months the answers will gel in my head.
By the way, I'm glad you stuck around. No negative comments will come from me! That's not why I joined this forum. I look forward to seeing your progress. For my 2.3tl's, I was going to rebuild the crossovers externally on prototyping boards, using high-quality passive components in an aesthetically beautiful layout. But, at my experience level, an 'active' approach is something that hadn't entered my mind. So, I am quite intrigued. Polk SDA SRS 2
Polk RTA 15tl
Polk Monitor 7C
Polk Lsi9
Infinity RS-II (modded)
Infinity RS-IIIa (modded)
Infinity RS 2.5 x 2
Magnepan 1.6QR (modded)
System: http://forum.audiogon.com/cgi-bin/fr.pl?vevol&1290711373 -
Very interesting. Will you be holding auditions?HT: SRT's.....Pioneer SC95, Oppo 105D; ps4; Mitsubishi 1080p 82" DLP; DirecTV
2 ch: SRS 1.2tl's.....Odyssey Stratos monos, Dynaco PAS pre w/Mods, Arcam DV-135, Music Hall MMF 5, AQ wires
Bedroom: SRS 3.1tl's, RTA 8T (phantom center), Pioneer sc-lx701, Mitsubishi 1080p 82" DLP, DirecTV, ps3
Car: SR 6500's.....Planet Audio HVT752 tube hybrid amps (2), Alpine dvd-a double din HU, two 12" Kicker's -
dacoupedeville wrote: »Very interesting. Will you be holding auditions?
They likely won't be fully ready for several months. But I'll let others listen to them who are in the area (Silicon Valley).
If somebody can help me with the logistics and has good connections with a good audio store locally, I could place them there for a couple of weeks or so.
