bela cape - analog input, crossfeed between channels

giuliomoro

D300 I´m having a potentiometer of 100k, connected to 3.3V and GND to alter the voltage input.

Any analog inputs that are left unconnected will be influenced by the voltage at the inputs next to them and you should not be concerned by that. In fact, you shouldn't be looking at the readings from any of the inputs that you are not feeding a signal into.
In your description and video you do not specify two key points:

what input channel is the potentiometer connected to?
what input channels are being sent to the scope and in what order? Do you want to show the code that you are using to do that?

Keep in mind that scope channels start counting from 1, while ADCs start from 0 (or from 3 in Pd) so there is room for confusion.

D300 Can i use any ground (GND) on the board for the analogReads?

Yes, though the preferred one is the one that is on the analog input socket itself.

D300

giuliomoro what input channel is the potentiometer connected to?

in the previous post, when i state "on channel 1)" the poti is connected to channel 1

giuliomoro what input channels are being sent to the scope and in what order? Do you want to show the code that you are using to do that?

i scope like this:

scope.log(ch1, ch2, ch3, ch4, ch5, ch6);

in the project i have 6 sensor values, 3 of them i need to read out permanently.
the other 3 are part of a resistor network, consisting of three voltage divider circuits. circuit is attached.
the resistor network contains 2 switches, swichting in pingpong, enabling to measure the resistances in the circuit.

maybe lets ignore the crossfeed then, there is still a big difference in ranges comparing channel 1 & 2 and channel 4 for instance. how can this be?

R12,R23,R13 are ch. 4 to ch6. the resistors in the inner triangle are variables. the outer ones are the reference resistors.
alt text

giuliomoro

D300 so far i would say ch3 is damaged.

that would be an uncommon failure mode.

D300 is there any routine of the manufacturer with which i could test if ch3 is broke? eg. a feedback loop.

Sure, check examples/Extras/cape-test. Not sure it is fine-tuned enough to discover the partial failure of one channel, though. The best option is probably to scope all channels and ground all of them. Then, one at a time, connect a channel to the 3.3V line and make a note of the level that channel reaches in the scope. Verify that all channels reach the same level on the scope when connected to 3.3V.

D300 / this scopes on channel 3 and 4, oscilloscope colours 3 & 4(green orange)

No, it doesn't. Any call to scope.log() will send its arguments to scope channels starting from 1. And you should have only one such call per each frame. When you initialise the scope with scope.setup() you are setting the number of channels. Successive calls to scope.log() should include that exact number of arguments. So if you have 4 channels in your scope because you called scope.setup(4, ...), then you should call scope.log(a, b, c, d). If you call scope.log(a, b), then anything you see on channels 3 and 4 of the scope will be garbage and should be disregarded.

D300 giuliomoro what input channels are being sent to the scope and in what order? Do you want to show the code that you are using to do that?

i scope like this:
scope.log(ch1, ch2, ch3, ch4, ch5, ch6);

Can you send the full code please?

Your drawing refers to Arduino, it seems, can you show an equivalent with labels for Bela? Note that "R12" and "R23" are in the title but not in the diagram, where you have generic "Resistor", so you could also clarify that in the next iteration of the diagram you are going to send.

Also if appropriate you could send us a picture of your setup, although that may be complicated to parse, or at least indicate in the diagram exactly what analog input channels you are connecting to.

D300

giuliomoro One relay board seems to have power while the other one doesn't.
the relay boards are powered all the time via DC+ and DC-. do you mean the state of the switches CH1 and CH2 ?

giuliomoro What is this trying to achieve?

as I described in the beginning of this converstation:

D300 the resistor network contains 2 switches, swichting in pingpong, enabling to measure the resistances in the circuit.

probably it was a bit unclear because the diagram did not differ between reference and variable resistors.
attached another diagram, in which i rectangled the variable resistor i want to measure in blue.

alt text

the outer resistors, at the the 5V line, are the "known resistors" 10k in the voltage dividers.
the main thing is to measure the resistance between the points P1,P2 and P3 when connecting them, eg. by hands.
here are some pictures of the signals:

base values, note: green and red are at the top, blue at the bottom, why that? .. in arduino all signals are at the bottom at 0V
alt text

touching P23 - green signal
alt text

touching P12 - red signal, this signal range is roughly a third of the green one
alt text

touching P13 blue signal, range of blue signal is similar to green signal range
alt text

D300

i did some further testing:

// test the scope!
A12 = analogRead(context, n/gAudioFramesPerAnalogFrame, 3 ); // 3
A13 = analogRead(context, n/gAudioFramesPerAnalogFrame, 4 ); // 4
			
scope.log(A12,A13); // this scopes on channel 3 and 4, oscilloscope colours 3 & 4(green orange)

the green signal on ch3, has again as in the prev. post, a minimum of the range as the orange signal although both potis get the same voltage gnd. beyond that the ch3-signal sinks back do zero within 3-5 seconds.
if i pull up the poti on max. the green signal goes high as possible but then sinks, no matter what the value of the poti is. I have then to "catch" the value to bring it up again. ( video in google drive --> "bela_100k_3.3V_ch3_sinking_signal_tiny_range.mov").

oscilloscope scale argument per channel is all set to 1 of course.

so far i would say ch3 is damaged. is there any routine of the manufacturer with which i could test if ch3 is broke?
eg. a feedback loop.

D300

giuliomoro what the maximum number of letters allowed? always hitting error code 418

D300

what tool do you use to mockup a circuit with bela digitally ?

giuliomoro

we have used fritzing in the past (https://learn.bela.io/community/resources/#fritzing-files)

D300

giuliomoro thanks I´m working on it. is there any difference between the analogInputs of the arduino and the one of the BELA/BBB ?

D300

giuliomoro i have Version 0.9.10 of Fritzing and cant find BELA or a Beaglebone... should it be in it?

giuliomoro

D300 Thanks I´m working on it. is there any difference between the analogInputs of the arduino and the one of the BELA/BBB ?

Sure, sampling rate (22.05kHz on Bela, variable on Arduino), bit depth(16-bit on Bela, 10-bit on Arduino Uno) , sampling time(1us approx on Bela IIRC, don't know on Arduino). What else?

D300 i have Version 0.9.10 of Fritzing and cant find BELA or a Beaglebone... should it be in it?

The page I linked to linked to this repo where you'll find them https://github.com/BelaPlatform/fritzing

D300 what the maximum number of letters allowed? always hitting error code 418

it's not a number of characters, it's some weird pattern matching mechanism on the server that I haven't been able to figure out yet.

D300 This code looks kind of OK, although it's weird that you only update A12 and A23 when status is 0 and only update A13 when status is 1. This leads to discontinuities in the scope, but I see no crosstalk with proper wiring. If you get crosstalk, then there's something fishing with the wiring.

D300 here is a foto of the setup.

Now for the wiring...

I see something suspicious going on with the power of the relay board. I found a similar board online and it seems that it is supposed to be powered from 5V, but signal inputs can be as low as 3.3V. Is this the same board you have? I cannot see clearly in your wiring, but it seems that you are only taking 3.3V from the Bela headers. If you are using that 3.3V line to power the relay board (and the relay board you are using is the same I linked above), you shouldn't. You should instead use a proper 5V line, which is P9.07. CAREFUL when operating with 5V and Bela board. You should not connect 5V to any of the pins on the P8 and P9 headers apart from P9.07 or you will kill the board. Analog inputs on J7 are 5V-tolerant, but the full-scale range is 4.096V, so if you use a 5V reference for the signals going into them, you may get clipped values. The underpower of the relay board may mean it is not actually switching loads as you expect it to.
One relay board seems to have power while the other one doesn't.
where's the Arduino code? What is this trying to achieve?

D300

#include <Bela.h>
#include <libraries/Scope/Scope.h>
#include <cmath>


// instantiate the scope
Scope scope;

int gAudioFramesPerAnalogFrame = 0;
int gCount = 0;

/* RESISTOR NETWORK */
// 4 digital pins setting the solid state relais to set the resistor network

// arduino pins 5,4,3,2 become 3,2,1,0 on BELA/BBB
// SSR P3, 
int gR_P3_V_ch_1 = 0; // digital pin 0 - check the pin diagram in the IDE
int gR_P3_GND_ch_2 = 1;

// SSR P1, P2
int gR_P1_p12_ch1 = 2;
int gR_P1_p13_ch2 = 3;

float SSR_pinpong_Interval = 0.02; // seconds
float delay_stabilize = 0.001; // seconds
int gStatus = 0;

// analogReads Pins
int A_P1 = 0;
int A_P3 = 1;

// 2 analog Reads as A0 and A1 of the resistor network
float A12 = 0.f;
float A23 = 0.f;
float A13 = 0.f;


bool setup(BelaContext *context, void *userData)
{
	// tell the scope how many channels and the sample rate
	scope.setup(3, context->audioSampleRate);
	
	// Check if analog channels are enabled
	if(context->analogFrames == 0 || context->analogFrames > context->audioFrames) {
		rt_printf("Error: this example needs analog enabled, with 4 or 8 channels\n");
		return false;
	}

	// pin setup
	pinMode(context, 0, gR_P3_V_ch_1, OUTPUT); 		// dig pin 0
	pinMode(context, 0, gR_P3_GND_ch_2, OUTPUT);	// dig pin 1

	pinMode(context, 0, gR_P1_p12_ch1, OUTPUT);		// dig pin 2
	pinMode(context, 0, gR_P1_p13_ch2, OUTPUT);		// dig pin 3


	// Useful calculations
	gAudioFramesPerAnalogFrame = context->audioFrames / context->analogFrames;

	rt_printf("SOG Setup Done\n");

	return true;
}

void render(BelaContext *context, void *userData)
{
	// if ( SSR_pinpong_Interval < delay_stabilize )
	// 	rt_printf("stop - stabilization time frame needs to me smaller then SSR ping pong interval");

	for(unsigned int n = 0; n < context->audioFrames; n++) 
	{	
		// // resistor network
		if(gCount >= (int)(context->digitalSampleRate * SSR_pinpong_Interval)) //if enough samples have elapsed
		{ 
			gCount = 0; //reset the counter
			rt_printf("reset cntr\n ");

		// 	//toggle the status
			if(gStatus == 0)
			{
				gStatus = 1;
				rt_printf("gStatus PING: %i\n", gStatus);
				
				// ==> LOW/0 means CONNECTED !
				// phase 1
				digitalWrite(context, n, gR_P1_p12_ch1 , 0);  // A12
    			digitalWrite(context, n, gR_P1_p13_ch2 , 1);   // not A13

    			digitalWrite(context, n, gR_P3_V_ch_1 , 0);   //
    			digitalWrite(context, n, gR_P3_GND_ch_2, 1);

				/*
				stabilize
				measure
				scope / control sound
				delay for stabilization
				*/

			}
			else if (gStatus == 1)
			{
				gStatus = 0;
				rt_printf("gStatus PONG: %i\n", gStatus);
				
				// phase 2
				// set circuit paths
				digitalWrite(context, n, gR_P1_p12_ch1 , 1);  // not A12
				digitalWrite(context, n, gR_P1_p13_ch2 , 0);   // A13
				digitalWrite(context, n, gR_P3_V_ch_1 , 1);   //
				digitalWrite(context, n, gR_P3_GND_ch_2, 0); // to ground..

				/*
				stabilize
				measure
				scope / control sound
				delay for stabilization
				*/
			}
		}
		gCount++;
		// rt_printf(" gCount: %i\n", gCount);
		
		
		// READ SENSORS IN ANALOG FRAMES NOT AUDIO FRAMES!
		if(gAudioFramesPerAnalogFrame && !(n % gAudioFramesPerAnalogFrame)) 
		{
			// Depending on the sampling rate of the analog inputs, this will
			// happen every audio frame (if it is 44100)
			// or every two audio frames (if it is 22050)

			// read resistance values
			if (gStatus == 0) // phase 1 , add stabilize delay here as condition
			{
				// A12 = analogRead(context, n/gAudioFramesPerAnalogFrame, A_P1 )* (3.3 / 1023.0);
				A12 = analogRead(context, n/gAudioFramesPerAnalogFrame, A_P1 ) * 1.0f; // 4
				// A23 = analogRead(context, n/gAudioFramesPerAnalogFrame, A_P3 ) * (3.3 / 1023.0);
				A23 = analogRead(context, n/gAudioFramesPerAnalogFrame, A_P3 ); // 5
			}
			else if (gStatus == 1)
			{
				// A13 = analogRead(context, n/gAudioFramesPerAnalogFrame, A_P3 ) * (3.3 / 1023.0);
				A13 = (analogRead(context, n/gAudioFramesPerAnalogFrame, A_P3 ) * 1.0f); // 5 // -A12 ?
			}

			scope.log(A12, A23, A13);

			// test the scope!
			// chan3 not working?
			// A12 = analogRead(context, n/gAudioFramesPerAnalogFrame, 0 ); // 3
			// A13 = analogRead(context, n/gAudioFramesPerAnalogFrame, 3 ); // 4
			
			// scope.log(A12,A13); // this scopes on channel 3 and 4, oscilloscope colours 3 & 4(green orange)
		}
	}
}

void cleanup(BelaContext *context, void *userData)
{

}

D300

here is a foto of the setup.
alt text

I´m using the same breadboard cicuit with arduino and BBB and have different behaviour of the analogInput readings, considering that the scope output is correct. I dont understand why.

this is how the values should look like:
alt text

current they look like this:
(google drive link will follow)

D300

giuliomoro hey, appreaciating your help so much!!

here is the arduino code:

// INTERACTIVE RESISTOR NETWORK
int A_P1 = A0;
int A_P3 = A1;

// look at circuit.pdf to understand naming and logic

int R_P1_p12 = 4;
int R_P1_p13 = 5;

int R_P3_V = 2; // P3 V
int R_P3_GND = 3; // P3 GND

 
// define the value of the known resistor
int knownResistor = 10000; // 10K ohm resistor

float R12;
float R23;
float R13;

void setup() 
{
  Serial.begin(9600);

  pinMode(R_P1_p12 , OUTPUT);
  pinMode(R_P1_p13 , OUTPUT);

  pinMode(R_P3_V, OUTPUT);
  pinMode(R_P3_GND, OUTPUT);

  delay(1000);

  Serial.print("SETUP END\n\n\n");
}

void loop() {

  // INTERACTIVE RESISTOR NETWORK

    // PHASE 1 ( writePin, stabilize, measure, print )
    // ==> LOW means CONNECTED !

    // set circuit paths
    digitalWrite(R_P1_p12 , LOW);  // R12
    digitalWrite(R_P1_p13 , HIGH);   // not R13

    digitalWrite(R_P3_V , LOW);   //
    digitalWrite(R_P3_GND, HIGH);

    // stabelize
    delay(10);

    // measure
    R12 = measure_resistance(A_P1);
    R23 = measure_resistance(A_P3);


    // print
    Serial.print("R12 ");
    Serial.print(R12);
    Serial.println("\r"); //  Carriage Return
  
    Serial.print("R23 "); 
    Serial.println(R23);
    Serial.println( '\r' ); //  Carriage Return

    delay(10);

    // PHASE 2

    // set circuit paths
    digitalWrite(R_P1_p12 , HIGH);  // not R12
    digitalWrite(R_P1_p13 , LOW);   // R13

    digitalWrite(R_P3_V , HIGH);   //
    digitalWrite(R_P3_GND, LOW); // to ground..

    // stabelize
    delay(10);

     // measure
    R13 = measure_resistance(A_P1);

    // print
    Serial.print("R13 ");
    Serial.print(R13);
    Serial.println("\r"); //  Carriage Return

    delay(10);

    // 40 ms per loop
}

float measure_resistance(int analogPin)
{
    // read the voltage at the analog pin
  int analogValue = analogRead(analogPin);

  // convert the analog value to voltage
  float voltage = analogValue * (5.0 / 1023.0);

  // calculate the resistance of the unknown resistor using the voltage divider formula
  float unknownResistor = ((5.0 - voltage) * knownResistor) / voltage;

  return unknownResistor;
}

you are right with the 5v powerig issue i guess.. will test

giuliomoro

A diagram with annotated resistor names
and values instead of "resistor" could help.
Also your relays are not SPDT as those switches in the diagram. Are they wired as one SPDT? Note that in your picture one of the relays has several unconnected terminals. Until you sort out the relay power thing you should probably avoid using hem in your tests. You should put together a little test to be able to verify the relays are working, e.g.: reading the resistance across terminals with a voltmeter or place 3.3V with a 220ohm resistor on one terminal and an LED to ground on the other terminal: does the led turn on when switching the relay?

D300 base values,

What does this mean?
What's the wiring for "base values"?

D300 touching P23 -

D300 touching P13 blue

What does "touching" mean? Is it equivalent to toggling some of the switches in the diagram, i.e. changing the wiring? Or is it about touching the copper plates with your hand? What are the relays doing in the meantime?

D300

i just connected two potentiometers on channel 0 and 1

// analogReads Pins
int A_P1 = 0;
int A_P3 = 1;

and here i did not have the "inverted range" as the blue signal has in the picture before.

thus, the problem has maybe to do with the difference how arduino and bela read out the signal. or with grounding?

giuliomoro

D300 thus, the problem has maybe to do with the difference how arduino and bela read out the signal. or with grounding?

D300 what about the behaviour described with the pictures of the scope? does this make sense to you that the blue signal "starts from bottom"?

If the signal reads 0V when "not touched", it would mean that when you are not touching, the signal is pulled down to ground by a resistor (however large). Sounds like a wiring issue to me.

D300

giuliomoro Note that in your picture one of the relays has several unconnected terminals.

what photo?
in the beginning i had mechanical relais, then mosfets, now SSRs because with SSRs i had the best result because there was a very low internal resistance.

giuliomoro What's the wiring for "base values"?

with base values i´m meaning the analog read values i measure when there is "no touch". "touch" means connecting the copper surfaces P1, P2 and P3 with hands, and consequently closing the circuit. the human body is the resistor in this case. "no touch" means not touching the copper foils.

giuliomoro What are the relays doing in the meantime?

the relais are switch in ping pong.
refering to this picture there are two phases - ping phase and pong phase.
in ping phase i measure R12 and R23. in pong phase R13. R12 means the resistance between the points P1 and P3
alt text

i did test the relais, but will do again in respect to the pingpong interval which is around 20 ms right now...

giuliomoro

D300 what photo?
in the beginning i had mechanical relais, then mosfets, now SSRs because with SSRs i had the best result because there was a very low internal resistance.

the only photo above with the solid-state relays. The rightmost SSR unit seems to have several free positions and possibly two wires going into the same terminal ?

D300 i did test the relais,

did you sort out the relay power issue? are you still powering them from 3.3V?

D300 "touch" means connecting the copper surfaces P1, P2 and P3 with hands, and consequently closing the circuit. the human body is the resistor in this case. "no touch" means not touching the copper foils.

You could start by measuring actual resistors of known values to validate the behaviour of the relays and the code before moving to measuring the resistance of the body.

D300

giuliomoro i measured with the oscilloscope while the circuit is connected via arduino and BBB.
the signals measured at P1 and P3 concerning the attached diagram are the same, except that the amplitude of the signal is slithly higher with arduino, because the sensor lines are powered with 5v on arduino and 3.3v on BBB.

100% sure that wiring is correct. correct in the sense that it is the wiring concerning to this image:

D300

giuliomoro So that would mean that the "Arduino" is wrong?

i started this project and its circuit design with the arduino and for me this is the "working" signal, ready to map on sound. now I would like to understand why there is this difference in signal interpretation of the arduino uno.

D300

giuliomoro the only photo above with the solid-state relays. The rightmost SSR unit seems to have several free positions and possibly two wires going into the same terminal ?

the perspective of the photo is too bad to see every wire - terminal - connection. but its connected right. this circuit works as expected with arduino.

giuliomoro did you sort out the relay power issue? are you still powering them from 3.3V?

i powering the relays now with 5V as you suggested. now the ranges of the signals are better. weirdly it also worked with just 3.3v.

giuliomoro You could start by measuring actual resistors of known values to validate the behaviour of the relays and the code before moving to measuring the resistance of the body.

thats a good idea.
what about the behaviour described with the pictures of the scope? does this make sense to you that the blue signal "starts from bottom"?

D300

giuliomoro Sounds like a wiring issue to me.

i´m sure its about the difference how the analog inputs work under the hood.
I measured with an oscilloscope. what the bela scope shows is the same as the oscilloscope shows.

I´m wondering.. i guess the question to ask is:
what does the arduino analogRead() under the hood?
how to "simulate" this behaviour to get the values as with arduino?