ABX Box

Hello all,

I'm making a Arduino controlled ABX box that will allow blind tests by yourself. I just have one one question; I'm planning on using a reed relay to switch between two unbalanced connections. I am just curious do I need to switch the ground along with L+ and R+ or can I simply connect the grounds from each source together?

 

You need to switch the grounds from each source.

 

So this is my proposed circuit at the moment, the idea is that it will randomly swap the sources in response to a user request. This schematic is only for the left channel. It will be duplicated for the right channel. LA and LB will be connected to a 4 pole rotary switch along with RA and RB.

Does anyone have any thoughts? I'm thinking I may need protection diodes on the muting relays.

My Arduino program will go as follow.

1) User requests next random trial
2) Set pin 3 low to mute both sources
3) Wait 20ms to allow mute to occur
4) Set pin 2 high or low randomly to swap sources
5) Wait 20ms to allow swap to occur
6) Set pin 3 high to un-mute both sources

 

I am completely unable to read these schematics, I do applaud the project though. I would guess there is some log of the succession of A and B during testing? Otherwise how would you know how well you've done afterwards? How is that log accessed? I'm assuming this is a headless device?

 

Yes, the arduino will send the A or B configuration to the computer where a program will be running for the user to record their guess. I'll probably make it completely computer based in the second revision using a simple USB gpio board rather than an arduino.

I'll make the program in Java as that's what I do for work. My main concern at the moment is preventing any damage to the source amps when the relays switch.

 

Neat project. Keep in mind that you can hear a relay click meaning that a user will know if nothing has switched. So between selections, you should deselect all or something along those lines.

Sources and amps should be fine (relay switches are like any other switches) -- but you do need protection diodes across the coils. Also, avr pins do not like to source or sink that much current, so it is better to use them to control a transistor and not the relay directly.

 

So that was part of the idea behind the initial muting relays, in addition to preventing the amps from interacting at all. I'm actually going to change the DPST muting relays to DPDT make before break relays with one set of poles going to a 300ohm resister to simulate load. I've read that tube amps can be damaged if unloaded so MBB into a load should prevent unloading.

 

I don't mean that you'd hear it through the speaker, but rather than the relays themselves click audibly.

If an amp can be damaged by being unloaded, the designer will place a high default load on it to keep that from happening. But I don't really think this is actually a thing.

 

So I've updated the design to use MBB relays. Regarding clicking noises, with my current design there will always be two clicks regardless of if the sources have been swapped.

Psuedo code:

Code:

boolean oneActive = 0;

function nextTrial(){
    oneActive = random()
    switchOn(oneActive);
    record(oneActive);
}

function record(oneActive){
    if(oneActive){
        sendToComputer(1)
    }
    else{
        sendToComputer(2);
    }
}

function switchOn(boolean oneActive){
    allOff();
    wait(20);
    if(oneActive){
        Pin.D2 = 1;
        Pin.D3 = 1;
    }else{
        Pin.D4 = 1;
        Pin.D5 = 1;
    }
    wait(20);
   
}

function allOff(){
    Pin.D2 = 0;
    Pin.D3 = 0;
    Pin.D4 = 0;
    Pin.D5 = 0;
}

function swapAB(){
    oneActive = !oneActive;
    switchOn(oneActive);
}

Schematic:

 

Once again, have a new design in the works using 4 3pst relays. I'm wondering though if someone can give me some pointers on how to get a PCB design done. I have zero experience with PCB routing, ideally I'd love a tool that takes my schematic and parts footprints and lays them out auto magically on the PCB. Anyone know of anything like that?

 

I've done my own and used ExpressPCB. They also supply free schematic and layout SW.
http://www.expresspcb.com/

 

Great, I'll check it out! I was using Upverter and 123d but they were incredibly buggy. Then I tried eagle but it was way too deep.

 

If you wanted to get fancy, you could have an ADC sense the source signal and not switch until both left and right signal were near or at zero crossing.

 

I've evolved the design a bit from what was posted above. I wanted to add a rotary switch into the mix so people could change between A and B at will. So in my new design I was planning on basically turning off all four relays, waiting 50 ms and then opening up two of them. Basically, each input would be connected to two relays, those two relays in turn would be connected to opposite outputs. I'll post a schematic tonight.

I'm using 4 x 3pst reed relay, and one as yet to be determined 3p3t rotary switch (with 1 throw being off)

 

Did you get this completed? I'm interested in making one myself.

 

I've done stuff like this, but not ARDUINO. And my opinion is.

If you want to test the acoustics as accurately as possible, you do not need any devices. You need to invite a few women and give them the opportunity to listen to your acoustics.
Women's hearing is more specifically the male and female emotions or rather any devices. If a woman liked your acoustics, then you did a great job and your calculations are correct.

Upon determining that sad for me the fact I no longer develop devices for the blind auditions.

 

Do you marry the one that liked your acoustics or do you have her do some ABX tests first to make sure she's legit?

 

I've been married twice and I'm smart enough not to tell my wife I bring women.