Hi everybody!
After years learning from www.diyaudio.com it's time to give something back!
First I want to thanks Rod Elliott and Ti Kan, who inspired me to do this project, the third of my state variable crossovers.
Here is the Xover44, my last project, to be shared to everybody.
A 2/3/4/n ways MCU controlled state variable active crossover.
The project is still a work in progress. I hope somebody will be interested to this. sorry for my English that is not the best.
Features for every crossover point:
- 2nd order (12db/oct) or 4th order (24db/oct) variable configuration.
- 4 different selectable resonances: Linkwitz-Riley, Butterworth, Bessel, etc.
- 32 selectable crossover frequencies.
The whole circuit is analog, based on dual op-amps like OPA2134, NE5532, or similar for characteristics and pin-out.
Everything is controlled trough high quality latched relays, driven from an I2C digital controller.
The interface use a 7'' VGA LCD monitor controlled straight from an Arduino DUE, multiplexed by some MCP23017 each one driving two ULN2804A.
I'm still building the prototype but I want to start to share the project with you guys.
Here is the free site I'm slowly building for the project:
Xover44
I have no clue if the site works but we will figure that out together!
Here are three pictures of the prototype:
After years learning from www.diyaudio.com it's time to give something back!
First I want to thanks Rod Elliott and Ti Kan, who inspired me to do this project, the third of my state variable crossovers.
Here is the Xover44, my last project, to be shared to everybody.
A 2/3/4/n ways MCU controlled state variable active crossover.
The project is still a work in progress. I hope somebody will be interested to this. sorry for my English that is not the best.
Features for every crossover point:
- 2nd order (12db/oct) or 4th order (24db/oct) variable configuration.
- 4 different selectable resonances: Linkwitz-Riley, Butterworth, Bessel, etc.
- 32 selectable crossover frequencies.
The whole circuit is analog, based on dual op-amps like OPA2134, NE5532, or similar for characteristics and pin-out.
Everything is controlled trough high quality latched relays, driven from an I2C digital controller.
The interface use a 7'' VGA LCD monitor controlled straight from an Arduino DUE, multiplexed by some MCP23017 each one driving two ULN2804A.
I'm still building the prototype but I want to start to share the project with you guys.
Here is the free site I'm slowly building for the project:
Xover44
I have no clue if the site works but we will figure that out together!
Here are three pictures of the prototype:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Lt spice simulation
An externally hosted image should be here but it was not working when we last tested it.
an overview of the 2/3/4way configuration
Every filter board (6 in this case) is a mono crossover point.
Left and right channels have different ground-planes with independent loop breakers (giving more flexibility to the project).
An externally hosted image should be here but it was not working when we last tested it.
Every filter board (6 in this case) is a mono crossover point.
Left and right channels have different ground-planes with independent loop breakers (giving more flexibility to the project).
Last edited:
> Click the image to open in full size.
All your image links seem to be 403:
(Even using "open in new tab")
"Your client does not have permission to get URL /dH_jMB_ZMbNe2BoRJcSPFST3rxH7PUqYjCrA5IPzKrzW2hNjoff7nnE2wXHLe5YJEIG34 mHOi6KqMY5GN8cEN0hRXNlVPihCEswWwjTW_oZY42FYOH2n=w1175 from this server."
My guess is that you need to set these links Public in your Google User Content control panel.
All your image links seem to be 403:
(Even using "open in new tab")
"Your client does not have permission to get URL /dH_jMB_ZMbNe2BoRJcSPFST3rxH7PUqYjCrA5IPzKrzW2hNjoff7nnE2wXHLe5YJEIG34 mHOi6KqMY5GN8cEN0hRXNlVPihCEswWwjTW_oZY42FYOH2n=w1175 from this server."
My guess is that you need to set these links Public in your Google User Content control panel.
The Modular and Stackable Filter Board
Here are the schematics and a couple of pictures of the filter board:
Filter Board Schematics
Here are the schematics and a couple of pictures of the filter board:
Filter Board Schematics
Is that even still working? Moving on from Picasa
Why not use the Attach function in this forum's Advanced post window? That way they stay with the post as long as DIYAudio exists.
FWIW: Google Classic sites can give public-sharable images. The URLs are ugly but it works, I think?
An externally hosted image should be here but it was not working when we last tested it.
Here is the tower of filter boards in test!
I tested board by board not-stacked first and then stacked just to be sure.
Till now everything seems to work properly and the sound quality is good to be using just OPA2134s!!!
The mistake I've done may be is to tune the last board too low...
(I made confusion about cutting frequency and crossover frequencies when I chose the resistors of the bass boards...).
Fortunately an easy to fix mistake: changing the fixed resistors will help me a lot without changing any capacitor.
with my actual configuration i have a 3 bit choice (8 possibilities) starting from 80Hz to 420Hz. that would be good as cutting frequency but is actually a crossover frequency.
That is not terrible in my case because I will use 15'' or 18'' inches woofers in my configuration.
But I can easy fix that adding a 33k resistor in parallel with every 19k1 fixed resistors (FA, FB.... to FF) to obtain 12k ohm resistors.
that will turn my filter, using 100 nF capacitors, to behave from 131Hz to 469Hz approximately, as crossover points (and not as frequency cut. Those are the kind of stupid problems I have when I can't do a project all at once.)
The result, with the 33k resistors added in parallel to the 19k1 ones, will shift everything 50Hz higher more or less.
there are also series resistors in case you want to change the "curvature".
If the series resistors are just jumpers we have linearity in our frequency options.
Adding them we will have an anti-log behavior.
My question is how to obtain a good logarithmic scale to place our frequency options?
In theory would be enough to have a negative resistor in place of the series resistors...
Is there any simple way to have the frequencies options placed in a "log" way instead of a linear way? may be just with passive components would be awesome. (right now all the frequencies options for every cut are distant almost the same amount of Hz among each other).
would be more convenient to have it in the "log" way!
Cheers!
Testing the "Thing!":
Thanks to everybody to get interested on the topic!
I tested board by board not-stacked first and then stacked just to be sure.
Till now everything seems to work properly and the sound quality is good to be using just OPA2134s!!!
The mistake I've done may be is to tune the last board too low...
(I made confusion about cutting frequency and crossover frequencies when I chose the resistors of the bass boards...).
Fortunately an easy to fix mistake: changing the fixed resistors will help me a lot without changing any capacitor.
with my actual configuration i have a 3 bit choice (8 possibilities) starting from 80Hz to 420Hz. that would be good as cutting frequency but is actually a crossover frequency.
That is not terrible in my case because I will use 15'' or 18'' inches woofers in my configuration.
But I can easy fix that adding a 33k resistor in parallel with every 19k1 fixed resistors (FA, FB.... to FF) to obtain 12k ohm resistors.
that will turn my filter, using 100 nF capacitors, to behave from 131Hz to 469Hz approximately, as crossover points (and not as frequency cut. Those are the kind of stupid problems I have when I can't do a project all at once.)
The result, with the 33k resistors added in parallel to the 19k1 ones, will shift everything 50Hz higher more or less.
there are also series resistors in case you want to change the "curvature".
If the series resistors are just jumpers we have linearity in our frequency options.
Adding them we will have an anti-log behavior.
My question is how to obtain a good logarithmic scale to place our frequency options?
In theory would be enough to have a negative resistor in place of the series resistors...
Is there any simple way to have the frequencies options placed in a "log" way instead of a linear way? may be just with passive components would be awesome. (right now all the frequencies options for every cut are distant almost the same amount of Hz among each other).
would be more convenient to have it in the "log" way!
Cheers!
Testing the "Thing!":
Thanks to everybody to get interested on the topic!
P.S.
in the SCHEMATICS
you can see the series resistors FA, FB.... to FF (now jumpers in the prototype) as:
R15, R16, R20,R21,R28,R29.
They are on the right side, just coming from the top of every state variable filter, feeding the variable resistors banks.
in the SCHEMATICS
you can see the series resistors FA, FB.... to FF (now jumpers in the prototype) as:
R15, R16, R20,R21,R28,R29.
They are on the right side, just coming from the top of every state variable filter, feeding the variable resistors banks.
Here is how I calculated my resistors using the resistor calculator of the site:
Bass board:
Serie Resistor = 0 Ohm
Filter Capacitors = 100 nF
Highest resistors = 33200 Ohm
Fixed Resistors = 12000 Ohm
Mid board:
Serie Resistor = 0 Ohm
Filter Capacitors = 33 nF
Highest resistors = 47500 Ohm
Fixed Resistors = 47500 Ohm
High board:
Serie Resistor = 0 Ohm
Filter Capacitors = 10 nF
Highest resistors = 105000 Ohm
Fixed Resistors = 182000 Ohm
Bass board:
Serie Resistor = 0 Ohm
Filter Capacitors = 100 nF
Highest resistors = 33200 Ohm
Fixed Resistors = 12000 Ohm
Mid board:
Serie Resistor = 0 Ohm
Filter Capacitors = 33 nF
Highest resistors = 47500 Ohm
Fixed Resistors = 47500 Ohm
High board:
Serie Resistor = 0 Ohm
Filter Capacitors = 10 nF
Highest resistors = 105000 Ohm
Fixed Resistors = 182000 Ohm
Last edited:
It took me more then 30 minutes o translate to English.... I'm getting bad!
Here is how I calculated my resistors using the resistor calculator of the site:
Bass board:
Series Resistor = 0 Ohm
Filter Capacitors = 100 nF
Highest resistors = 33200 Ohm
Fixed Resistors = 12000 Ohm
Mid board:
Series Resistor = 0 Ohm
Filter Capacitors = 33 nF
Highest resistors = 47500 Ohm
Fixed Resistors = 47500 Ohm
High board:
Series Resistor = 0 Ohm
Filter Capacitors = 10 nF
Highest resistors = 105000 Ohm
Fixed Resistors = 182000 Ohm
the resistors values at the end of the spreadsheet will be approximated to Vishay/Dale CMF (industrial) or RN (MILITARY) series typical values.
I recommend to use the best resistors having the following properties: a very low noise figure and with a very small tolerance, still having an acceptable impedance.
Translated: Almost the best resistors you can get, not just for price! 0.5%, tolerance would be the best 1% is acceptable.
Notice that sometimes 0,5% is cheaper than 1% in shops!!!
As filter capacitors in the state variable filters should be better a modern technology Polypropylene MKP with minimum 1% tolerance. Matching can improve but is not necessary.
Other option that could result in better quality is a polystyrene set with 1% tolerance, but they are pretty retro, hard to find, expensive, and less durable.
On the Op-Amps power supplies and on the digital side of the filter boards, I recommend X7R ceramic caps in order to have a higher resistance, making with their 100nF capacitance a decent filter.
If anybody want to build this Dream-Machine contact me on the site! (I Know some purists don't like the fact we use Op-Amps. But they are the good way to achieve what we need, and with modern op-amps designs it's hard to beat them!!!)
Just for the people who doesn't believe on CHIPS and Op-Amp: look outside! It is plenty of discrete Op-Amps out there you could try! Just let us know what is your experience with them in a unity gain configuration!
I would like a more precise op-amp then the opa2134 from Burr-Brown. I'm sure there is one for this kind of applications. But till now it is the best I could try in that price range.
I installed them straight on the board without a socket, because there are already to many not-soldered contacts for my standars. This is why I recommend golden contacts on every molex-kk-connectors, and on the headers when possible. Especially the 2 position one giving signal from a board to another. That could be maid of a twisted pair or coaxial cable too.
Here is how I calculated my resistors using the resistor calculator of the site:
Bass board:
Series Resistor = 0 Ohm
Filter Capacitors = 100 nF
Highest resistors = 33200 Ohm
Fixed Resistors = 12000 Ohm
Mid board:
Series Resistor = 0 Ohm
Filter Capacitors = 33 nF
Highest resistors = 47500 Ohm
Fixed Resistors = 47500 Ohm
High board:
Series Resistor = 0 Ohm
Filter Capacitors = 10 nF
Highest resistors = 105000 Ohm
Fixed Resistors = 182000 Ohm
the resistors values at the end of the spreadsheet will be approximated to Vishay/Dale CMF (industrial) or RN (MILITARY) series typical values.
I recommend to use the best resistors having the following properties: a very low noise figure and with a very small tolerance, still having an acceptable impedance.
Translated: Almost the best resistors you can get, not just for price! 0.5%, tolerance would be the best 1% is acceptable.
Notice that sometimes 0,5% is cheaper than 1% in shops!!!
As filter capacitors in the state variable filters should be better a modern technology Polypropylene MKP with minimum 1% tolerance. Matching can improve but is not necessary.
Other option that could result in better quality is a polystyrene set with 1% tolerance, but they are pretty retro, hard to find, expensive, and less durable.
On the Op-Amps power supplies and on the digital side of the filter boards, I recommend X7R ceramic caps in order to have a higher resistance, making with their 100nF capacitance a decent filter.
If anybody want to build this Dream-Machine contact me on the site! (I Know some purists don't like the fact we use Op-Amps. But they are the good way to achieve what we need, and with modern op-amps designs it's hard to beat them!!!)
Just for the people who doesn't believe on CHIPS and Op-Amp: look outside! It is plenty of discrete Op-Amps out there you could try! Just let us know what is your experience with them in a unity gain configuration!
I would like a more precise op-amp then the opa2134 from Burr-Brown. I'm sure there is one for this kind of applications. But till now it is the best I could try in that price range.
I installed them straight on the board without a socket, because there are already to many not-soldered contacts for my standars. This is why I recommend golden contacts on every molex-kk-connectors, and on the headers when possible. Especially the 2 position one giving signal from a board to another. That could be maid of a twisted pair or coaxial cable too.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.