Hi All,
I've completed 2 stereo amp boards and power supply based on a group buy from years back. I'd like to do an active crossover before the 2 stereo channels to allow for bi-amping speakers. I've looked around and although there are a lot of schematics, I don't see many PCBs available for purchase so I was thinking of starting a group buy. I can organize it and probably complete the schematic and PCB layout if no one else is able to. Wondering if people are interested in a group buy such as this. I was thinking a single board with line in and two lines out filtered at the crossover frequency. If you want to do two speakers bi-amped then you would get two boards. I was also thinking of making it adjustable for the crossover frequency and have not looked into the order or type of crossover model yet. Please post with any interest. It would probably be limited to USA and Canada due to shipping.
Edit>> I can setup a web site for this too if it gets off the ground.
Edit>> So after researching and looking at my particular application and possibly others needs, I am thinking of including an AC to DC circuit for a low current transformer as well as a voltage regulator circuit. This would allow for powering the active crossover directly from your main transformer without any additional components. I'm thinking this would fit onto one small board with a header connector that would allow you to us the AC to DC circuit and voltage regulator for other purposes as well. Not much interest so far but I'll see if anything picks up.
I've completed 2 stereo amp boards and power supply based on a group buy from years back. I'd like to do an active crossover before the 2 stereo channels to allow for bi-amping speakers. I've looked around and although there are a lot of schematics, I don't see many PCBs available for purchase so I was thinking of starting a group buy. I can organize it and probably complete the schematic and PCB layout if no one else is able to. Wondering if people are interested in a group buy such as this. I was thinking a single board with line in and two lines out filtered at the crossover frequency. If you want to do two speakers bi-amped then you would get two boards. I was also thinking of making it adjustable for the crossover frequency and have not looked into the order or type of crossover model yet. Please post with any interest. It would probably be limited to USA and Canada due to shipping.
Edit>> I can setup a web site for this too if it gets off the ground.
Edit>> So after researching and looking at my particular application and possibly others needs, I am thinking of including an AC to DC circuit for a low current transformer as well as a voltage regulator circuit. This would allow for powering the active crossover directly from your main transformer without any additional components. I'm thinking this would fit onto one small board with a header connector that would allow you to us the AC to DC circuit and voltage regulator for other purposes as well. Not much interest so far but I'll see if anything picks up.
Last edited:
Hi Fred,
Thank you for your post. I'll upload the schematic when I have it complete. As far as design criteria, I'm adjusting that so that there will be two channels of active crossover and one AC-DC power supply as well as a voltage regulator on one board whose preliminary size is 2 inches x 4.5 inches.
Thank you for your post. I'll upload the schematic when I have it complete. As far as design criteria, I'm adjusting that so that there will be two channels of active crossover and one AC-DC power supply as well as a voltage regulator on one board whose preliminary size is 2 inches x 4.5 inches.
A couple of things to consider-
Unless the drivers are flat almost two octaves past the crossover frequency, you will need EQ, and possibly different slopes on high pass and low pass filters. You can accomplish EQ a number of ways, adjustable Q, shelving filters, notch and peaking filters and Linkwitz transform to name a few.
Phase alignment is a must (at least according to Sigfried Linkwitz) - two ways to achieve this are with all pass filters and split crossover points.
Adjustable level - use a multiturn trim pot so that you can more easily set the exact attenuation you need.
Sallen Key topology sounds better than state variable to me. Some like multi feedback, I haven't tried it.
In a Sallen Key filter it is difficult to get a potentiometer to track well enough to allow even a single channel 2nd order filter adjustment. I suggest that you allow for swappable 1% resistors. Either a pluggable module or simply snap header sockets at the appropriate locations.
Allow plenty of room for quality capacitors. MKP caps sound better , but tend to be a bit bigger than MKT.
It sounds like you plan to use amplifier rails or at least the main AC for power. Check the maximum voltage of regulator chips. Most I know of are limited to ~35V. You can use a simple resistor or resistor and zener shunt pre-regulator to drop the voltage to something suitable for your regulator. Be aware of the dissipation and size your parts accordingly. Check your opamp data sheets for current draw.
Even with opamp crossovers, the quality of the power supply makes a difference. There your proposed board doesn't seem to have room for a discrete regulator, but check the threads for suggestions.
Watch cost - your competition is miniDSP. At around $110 enclosed it's awfully convenient. If your design ends up costing anywhere near that it had better offer a heck of a lot and sound an awful lot better.
Unless the drivers are flat almost two octaves past the crossover frequency, you will need EQ, and possibly different slopes on high pass and low pass filters. You can accomplish EQ a number of ways, adjustable Q, shelving filters, notch and peaking filters and Linkwitz transform to name a few.
Phase alignment is a must (at least according to Sigfried Linkwitz) - two ways to achieve this are with all pass filters and split crossover points.
Adjustable level - use a multiturn trim pot so that you can more easily set the exact attenuation you need.
Sallen Key topology sounds better than state variable to me. Some like multi feedback, I haven't tried it.
In a Sallen Key filter it is difficult to get a potentiometer to track well enough to allow even a single channel 2nd order filter adjustment. I suggest that you allow for swappable 1% resistors. Either a pluggable module or simply snap header sockets at the appropriate locations.
Allow plenty of room for quality capacitors. MKP caps sound better , but tend to be a bit bigger than MKT.
It sounds like you plan to use amplifier rails or at least the main AC for power. Check the maximum voltage of regulator chips. Most I know of are limited to ~35V. You can use a simple resistor or resistor and zener shunt pre-regulator to drop the voltage to something suitable for your regulator. Be aware of the dissipation and size your parts accordingly. Check your opamp data sheets for current draw.
Even with opamp crossovers, the quality of the power supply makes a difference. There your proposed board doesn't seem to have room for a discrete regulator, but check the threads for suggestions.
Watch cost - your competition is miniDSP. At around $110 enclosed it's awfully convenient. If your design ends up costing anywhere near that it had better offer a heck of a lot and sound an awful lot better.
Last edited:
The board Jens Rasmussen and I collaborated on is here: http://www.diyaudio.com/forums/group-buys/74420-active-filter-board-gb.html If that meets your needs, I can order the number of boards you coordinate. PSU not included.
Thank you
Hi Bob,
Thanks for your post and helpful info. I'll check out the link you posted. I'll continue with the schematics but if there is not enough interest than I will have to shelve the project.
Edit >> Hi Bob, I checked out the link to the group buy post you posted but am having difficulty locating the files associated with the project, namely the schematic. Feel free to post here or provide a link. It seems as though I might be trying to reinvent the wheel, so to speak, so maybe we can just go with another group buy on the board you guys did. I can certainly make a website and organize.
Hi Bob,
Thanks for your post and helpful info. I'll check out the link you posted. I'll continue with the schematics but if there is not enough interest than I will have to shelve the project.
Edit >> Hi Bob, I checked out the link to the group buy post you posted but am having difficulty locating the files associated with the project, namely the schematic. Feel free to post here or provide a link. It seems as though I might be trying to reinvent the wheel, so to speak, so maybe we can just go with another group buy on the board you guys did. I can certainly make a website and organize.
Last edited:
Since you'll be competing with miniDSP, one of the things people have an issue with miniDSP is that when there is a lot of eq (like dipoles.) the average level is limited. An analog output stage with a bit of (adjustable) gain would be a good idea. Something that can swing at least +/- 5V ought to do it. Of course that means you need more than a single 5V supply for the output stage.
Another issue that would be solved by using a bipolar supply is the low end response. I don't own one but I have hear the extreme low end is weak. A DC coupled output would fix that.
The thing that makes miniDSP so popular is ease of use. Many people lack programming skills, so to attract the most people you need an easy to use interface like miniDSP. If you're really good at programming perhaps you could make your system work with one of the popular measurement packages so you could use measured response to help drive your crossover design.
Good luck.
Another issue that would be solved by using a bipolar supply is the low end response. I don't own one but I have hear the extreme low end is weak. A DC coupled output would fix that.
The thing that makes miniDSP so popular is ease of use. Many people lack programming skills, so to attract the most people you need an easy to use interface like miniDSP. If you're really good at programming perhaps you could make your system work with one of the popular measurement packages so you could use measured response to help drive your crossover design.
Good luck.
Last edited:
I would be in depending on the circuit design.
You are probably already aware of this but here goes anyway.
Douglas Self made a very compelling talk on the theory and practical design considerations for active crossovers at a recent Burning Amp Festival. This is the link:
https://www.youtube.com/watch?v=f_zkE_42BV4
This might be a state of the art circuit topography to consider.
You are probably already aware of this but here goes anyway.
Douglas Self made a very compelling talk on the theory and practical design considerations for active crossovers at a recent Burning Amp Festival. This is the link:
https://www.youtube.com/watch?v=f_zkE_42BV4
This might be a state of the art circuit topography to consider.
Hi Bob,
Thanks again for your input. I do VB.net programming so a custom app might be feasible but I have not found much info on that as of yet. I'm just starting to dig into the Analog site. The Sigma DSP software is pretty nice and might be all that is needed to start with. I downloaded it and laid out a crossover schematic pretty easily.
For the boards, Sunstone PCBs couples with Screaming Circuits and produces PCBs with any number of components placed and soldered. I was thinking of producing boards with just the components that would not be able to be hand soldered pre-assembled. So far the prices look reasonable but when all is said and done, it's probably going to approach the miniDSP price level so again, it might not make sense to re-invent the wheel. But correct me if I'm wrong, the miniDSP offers only a crossover component??
Thanks again for your input. I do VB.net programming so a custom app might be feasible but I have not found much info on that as of yet. I'm just starting to dig into the Analog site. The Sigma DSP software is pretty nice and might be all that is needed to start with. I downloaded it and laid out a crossover schematic pretty easily.
For the boards, Sunstone PCBs couples with Screaming Circuits and produces PCBs with any number of components placed and soldered. I was thinking of producing boards with just the components that would not be able to be hand soldered pre-assembled. So far the prices look reasonable but when all is said and done, it's probably going to approach the miniDSP price level so again, it might not make sense to re-invent the wheel. But correct me if I'm wrong, the miniDSP offers only a crossover component??
AFAIK miniDSP does crossover, EQ, delay, shelving filters, etc. with the right software module. (their website seems to be down at the moment.) The signal level issue can easily be dealt with by adding a buffer/gain stage. The extreme low end can be as simple as bypassing the caps on the miniDSP board.
Not sure what else to add other than the analog stage and/or a better DAC - like the SABRE32. It's a nice 8 channel chip that Oppo uses in the BDP-105. The Twisted Pear guys parallel four channels x 2 to make it an even lower noise stereo DAC.
Since you have to go ADC, there is probably a better quality ADC available as well. I don't have any chips in mind, but I'd bet ESS has something at the same level as their DAC chips.
The ability to accept and select digital inputs and control the volume would be a good thing. Probably way too much to ask, but accept DSD over HDMI for those of us who enjoy SACDs.
All of this puts you above the price point of miniDSP, but to me an audiophile equivalent seems the best option for something to have much interest. Anyone else care to chip in?
Not sure what else to add other than the analog stage and/or a better DAC - like the SABRE32. It's a nice 8 channel chip that Oppo uses in the BDP-105. The Twisted Pear guys parallel four channels x 2 to make it an even lower noise stereo DAC.
Since you have to go ADC, there is probably a better quality ADC available as well. I don't have any chips in mind, but I'd bet ESS has something at the same level as their DAC chips.
The ability to accept and select digital inputs and control the volume would be a good thing. Probably way too much to ask, but accept DSD over HDMI for those of us who enjoy SACDs.
All of this puts you above the price point of miniDSP, but to me an audiophile equivalent seems the best option for something to have much interest. Anyone else care to chip in?
Analog Devices DSP Chip
http://www.analog.com/static/imported-files/data_sheets/ADAU1701.pdf
Looking at this chip currently. I'll fill you guys in on my thoughts.
SigmaStudio? Graphical Development Tool | Analog Devices
This is the software that is available.
http://www.analog.com/static/imported-files/data_sheets/ADAU1701.pdf
Looking at this chip currently. I'll fill you guys in on my thoughts.
SigmaStudio? Graphical Development Tool | Analog Devices
This is the software that is available.
OT post removed.I might be interested in a set of boards to try for the active cross over.
This project with boards raised here is interesting as well along with DSP project
Mini dsp is very interesting modeling tool indeed although in comparison to the hi end dacs tue dac and adc conversion suffers. I have tried mini DSP 10x10 connected to aya dac and degradation is obvious. Not surprising the dac costs as whole mini dsp itself.
It might be interesting to get a dsp solution with i2s in and several i2s out options allowing to use your own dacs on the output similar to what Najda does
Sent from my ONE A2003 using Tapatalk
We might need to raise a separate thread to discuss the DSP on its own, let me introuce some concepts here:
High level
There would be a software component allowing to configure filters and export them to the hardware DSP to run
Software component should be platform agnostic and might potentially support different hardware implementations via different interfaces
Hardware component can be based on different chips, but it should be modular/
It can support
Digital input (I2S) and digital Output (I2S) lets say 6-8 channels to allow 3-4 way systems.
This component might either be able to get one stream via USB/Etherenet or I2S and decode it into 3-4 stereo pairs. Say if the cross over is done by Jriver or another software program
Or support a stereo feed and split it into multi channel by running DSP code
So the concept looks like this
For digital cross over
I2S Input(Multiple streams via 1 I2S channel) -> Multiple I2S Outputs (3-4 chanels of stereo I2S Output each)
Or it can do DSP in between I2S Input -> DSP -> Multiple channels Output
On top of it it can support the third party dacs or dac modules offered together with DSP
ADC Input module looks like an interesting option as well to allow analog input
All in All Inputs and outputs might be
Inputs
I2S
SPDIF
AES
TOSLINK
ANALOG
Outputs actually the same through different modules
High level
There would be a software component allowing to configure filters and export them to the hardware DSP to run
Software component should be platform agnostic and might potentially support different hardware implementations via different interfaces
Hardware component can be based on different chips, but it should be modular/
It can support
Digital input (I2S) and digital Output (I2S) lets say 6-8 channels to allow 3-4 way systems.
This component might either be able to get one stream via USB/Etherenet or I2S and decode it into 3-4 stereo pairs. Say if the cross over is done by Jriver or another software program
Or support a stereo feed and split it into multi channel by running DSP code
So the concept looks like this
For digital cross over
I2S Input(Multiple streams via 1 I2S channel) -> Multiple I2S Outputs (3-4 chanels of stereo I2S Output each)
Or it can do DSP in between I2S Input -> DSP -> Multiple channels Output
On top of it it can support the third party dacs or dac modules offered together with DSP
ADC Input module looks like an interesting option as well to allow analog input
All in All Inputs and outputs might be
Inputs
I2S
SPDIF
AES
TOSLINK
ANALOG
Outputs actually the same through different modules
The software part might be split into
GUI allowing to input different filter types and exporting them to the hardware language used in DSP
Need to consider platform independent languages like Java or Python, which can allow to implement nice looking GUI.
Database of the filters
Export optons to the formats supported by hardware DSP
An interesting hardware solution is XCore 200 series
They have a language suited for parallel execution and it includes DSP functions already
It can be used as a starting point for the filter database if this solution is taken
Evaluation boards are around 200 usd for this chip and diyinhk has multi channel usb based on it already
GUI allowing to input different filter types and exporting them to the hardware language used in DSP
Need to consider platform independent languages like Java or Python, which can allow to implement nice looking GUI.
Database of the filters
Export optons to the formats supported by hardware DSP
An interesting hardware solution is XCore 200 series
They have a language suited for parallel execution and it includes DSP functions already
It can be used as a starting point for the filter database if this solution is taken
Evaluation boards are around 200 usd for this chip and diyinhk has multi channel usb based on it already
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.