Measuring the delay between two rising edges on Discovery 2

[maxwell_]

Hi. I would like to measure the absolute time difference between the rising edges of two digital output pins. The time that the pins stay on are about 100 ms, and I expect the delay to be in the range of 1-10 ms. To be safe, I have to assume that either pin can get HIGH first. 

Additionally, I want to take many measurements (100-1000), it would be great if there is a way to log the data as well. 

 

[attila]

Hi @maxwell_

You can use the following Script with Record and data compression, here you have the workspace: dt.dwf3work
Here I used the Patterns to generated signals with delay.

[maxwell_]

Thanks @attila. What does the 1e-8 correspond to? Because in Patterns 1 of the example workspace you provided, I am not seeing any phase shifts or delays between the two signals. 

Edited March 16, 2022 by maxwell_

[maxwell_]

Never mind, the 1e-8 seems to be the minimum value that can be measured by the timer. By changing the phase shift between the signals, the Script seems to work. Thanks!

[attila]

Hi @maxwell_

Seconds, 1e08 is 10ns

[3D_loves_eng]

On 3/16/2022 at 10:33 AM, attila said:

Hi @maxwell_

You can use the following Script with Record and data compression, here you have the workspace: dt.dwf3work
Here I used the Patterns to generated signals with delay.

Hi Attila,

How to measure the delay of transient slopes (both rising and falling edge) between two channels eg. CH1-CH2 ? I prefer to use a script either in JS or Python.

I currently subtract two channels each other, then use the PosWidth and NegWidth functions to measure delay in Scope. However, I always have to zoom the transient edges to obtain the most accurate results (from 10 to 100ns). So the test procedure is manual and boring in this way.

 

Kind regards,

 

Edited June 20, 2022 by 3D_loves_eng
clarify measurement sources

[attila]

Hi @3D_loves_eng

Like this: dt.dwf3work

[Albertu95]

Hello @attila

On 3/16/2022 at 10:33 AM, attila said:

Hi @maxwell_

You can use the following Script with Record and data compression, here you have the workspace: dt.dwf3work
Here I used the Patterns to generated signals with delay.

Is it possible to do something similar, but with the analog inputs of the scope?
What I want to do is obtain automatically the delay of the rising edge on CH2 compared to the instant of the rising edge on CH1.  
I thought that a possible approach could be using the rising edge trigger of the channels, but I couldn't figure it out how to use the trigger to calculate the delay between the two wave forms.

Can you please help me?
Best regards

[attila]

Hi @Albertu95

You can use something like this to measure the delay with higher approximation/interpolation :

var trig = Scope.Trigger.Level.real;
var c = Scope.Time.Samples.real;
var v2 = 0;
// T0 is aligned to trigger crossing
for(var i = floor(Scope.Time.IndexOf(0))-1; i < c; i++){
    var v = Scope.Channel2.Sample(i);
    if(v<trig){
        v2 = v;
        continue;
    }
    var s = i - (v-trig)/(v-v2);
    var t = Scope.Time.TimeOf(s);
    print(i, t*1e9, "ns");
    break;
}

 

[Albertu95]

Thank you @attila!

[Albertu95]

Hi @attila
I did a simple test code for what you explained to me, using the wave generator of the instrument itself.

 

// This script adjusts the Wavegen offset based on Scope measurement.
clear()
if(!('Wavegen' in this) || !('Scope' in this)) throw "Please open a Scope and a Wavegen instrument";

Wavegen.Channel1.Mode.text = "Simple";
Wavegen.Channel1.Simple.Type.text = "pulse";
Wavegen.Channel1.Simple.Phase.value = 95;
Wavegen.Channel1.Simple.Offset.value = 0.0;
Wavegen.Channel1.Simple.Period.text = "2us";
Wavegen.Channel1.Simple.Symmetry.value= "25"


Wavegen.Channel2.Mode.text = "Simple";
Wavegen.Channel2.Simple.Type.text = "pulse";
Wavegen.Channel2.Simple.Phase.value = 80;
Wavegen.Channel2.Simple.Offset.value = 0.0;
Wavegen.Channel2.Simple.Period.text = "1us";
Wavegen.Channel2.Simple.Symmetry.value= "25"

Scope1.Trigger.Type.text= "edge";
Scope1.Trigger.Condition.text= "rising";
Scope1.Trigger.Level.value=0.5;



Wavegen.run();
Scope.run();

var trig = Scope.Trigger.Level.real;//returns the trigger set value
var c = Scope.Time.Samples.real;//returns the number of samples
var v2 = 0;
print(trig, "trig");
print(c, "samples");// T0 is aligned to trigger crossing
for(var i = floor(Scope.Time.IndexOf(0))-1; i < c; i++){//floor(Scope.Time.IndexOf(0))-1: takes the moment of the trigger -1 as starting point for the for cycle (?)
    t=0;
    var v = Scope.Channel2.Sample(i);// takes i sample from channel 2
    if(v<trig){//if value i of channell 2 is smaller than trigger go further
        v2 = v;
        continue;
    }
    var s = i - (v-trig)/(v-v2);//tries to take the most precise value possible i-((sample bigger than trigger)-trigger)/((sample bigger than trigger) - previous sample) 
    var t = Scope.Time.TimeOf(s);//time raltive to the trigger, and so the delay
    print(i, t*1e9, "ns");
    break;
}
Wavegen.stop();
Scope.stop();

It seems that it works, my only problem now is that if I try to change a parameter, let's say the phase, after the first run of the code it does not work. 
I have to start and stop manually wave generator and scope, and then it works again for one run. Then, if I want to use it again, I have to repeat the operation described before. 

I guess I don't reset properly the instrument in the script, but I cannot figure out what I make wrong.

Can you please help me?

[attila]

Hi @Albertu95

Scope.run();
Scope.wait(); // wait for the capture to be done

[Albertu95]

Thanks! 

[Albertu95]

Here is the complete code if someone needs it:

// This script adjusts the Wavegen offset based on Scope measurement.
clear()
if(!('Wavegen' in this) || !('Scope' in this)) throw "Please open a Scope and a Wavegen instrument";

Wavegen.Channel1.Mode.text = "Simple";
Wavegen.Channel1.Simple.Type.text = "pulse";
Wavegen.Channel1.Simple.Phase.value = 92;
Wavegen.Channel1.Simple.Offset.value = 0.0;
Wavegen.Channel1.Simple.Period.text = "1us";
Wavegen.Channel1.Simple.Symmetry.value= "25"


Wavegen.Channel2.Mode.text = "Simple";
Wavegen.Channel2.Simple.Type.text = "pulse";
Wavegen.Channel2.Simple.Phase.value = 80;
Wavegen.Channel2.Simple.Offset.value = 0.0;
Wavegen.Channel2.Simple.Period.text = "1us";
Wavegen.Channel2.Simple.Symmetry.value= "25"

Scope1.Trigger.Type.text= "edge";
Scope1.Trigger.Condition.text= "rising";
Scope1.Trigger.Level.value=0.5;



Wavegen.run();
Scope.run();
Scope.wait();// wait for the capture to be done

var trig = Scope.Trigger.Level.real;//returns the trigger set value
var c = Scope.Time.Samples.real;//returns the number of samples
var v2 = 0;
print(trig, "trig");
print(c, "samples");// T0 is aligned to trigger crossing
for(var i = floor(Scope.Time.IndexOf(0))-1; i < c; i++){//floor(Scope.Time.IndexOf(0))-1: takes the moment of the trigger -1 as starting point for the for cycle (?)
    t=0;
    var v = Scope.Channel2.Sample(i);// takes i sample from channel 2
    if(v<trig){//if value i of channell 2 is smaller than trigger go further
        v2 = v;
        continue;
    }
    var s = i - (v-trig)/(v-v2);//tries to take the most precise value possible i-((sample bigger than trigger)-trigger)/((sample bigger than trigger) - previous sample) 
    var t = Scope.Time.TimeOf(s);//time raltive to the trigger, and so the delay
    print(i, t*1e9, "ns");
    break;
}
Wavegen.stop();
Scope.stop();

-----------------------------------------------------------------------------------

I was trying to adapt this same code for the falling edge, but the result seems not as precise as the rising edge one.

This is the code I used:

// This script adjusts the Wavegen offset based on Scope measurement.
clear()
if(!('Wavegen' in this) || !('Scope' in this)) throw "Please open a Scope and a Wavegen instrument";

Wavegen.Channel1.Mode.text = "Simple";
Wavegen.Channel1.Simple.Type.text = "pulse";
Wavegen.Channel1.Simple.Phase.value = 10;
Wavegen.Channel1.Simple.Offset.value = 0.0;
Wavegen.Channel1.Simple.Period.text = "1us";
Wavegen.Channel1.Simple.Symmetry.value= "25"


Wavegen.Channel2.Mode.text = "Simple";
Wavegen.Channel2.Simple.Type.text = "pulse";
Wavegen.Channel2.Simple.Phase.value = 0;
Wavegen.Channel2.Simple.Offset.value = 0.0;
Wavegen.Channel2.Simple.Period.text = "1us";
Wavegen.Channel2.Simple.Symmetry.value= "25"

Scope1.Trigger.Type.text= "edge";
Scope1.Trigger.Condition.text= "falling";
Scope1.Trigger.Level.value=0.5;



Wavegen.run();
Scope.run();
Scope.wait();// wait for the capture to be done

var trig = Scope.Trigger.Level.real;//returns the trigger set value
var c = Scope.Time.Samples.real;//returns the number of samples
var v2 = 0;
print(trig, "trig");
print(c, "samples");// T0 is aligned to trigger crossing

for(var i = floor(Scope.Time.IndexOf(0))-1; i < c; i++){//floor(Scope.Time.IndexOf(0))-1: takes the moment of the trigger -1 as starting point for the for cycle (?)
    t=0;
    var v = Scope.Channel2.Sample(i);// takes i sample from channel 2
    if(v>trig){//if value i of channell 2 is BIGGER than trigger go further; TAKE CARE FOR FALLING V>TRIGGER; FOR RISING V<TRIG
        v2 = v;
        continue;
    }
    //var s = i - (v-trig)/(v-v2);//tries to take the most precise value possible i-((sample bigger than trigger)-trigger)/((sample bigger than trigger) - previous sample)
    var s = i - (trig-v)/(v2-v);//tries to take the most precise value possible i-(trigger-(sample SMALLER than trigger))/(previous sample - (sample SMALLER than trigger))
  
    var t = Scope.Time.TimeOf(s);//time raltive to the trigger, and so the delay
    print(i, t*1e9, "ns");
    break;
}
Wavegen.stop();
Scope.stop();

The scope seems not to trigger as precisely as for the rising edge case:

 

as you can see the rising edge trigger seems more precise than the falling edge one.

@attila do you have any sudgestions?

Can I improve this with the code, or it is some kind of hardware limitation?

EDIT: typo

Edited July 12, 2022 by Albertu95

[attila]

Hi @Albertu95

For falling edge set Hysteresis to zero or set level + hyst.
This falling edge level +/- hyst will corrected in the next version.

[Albertu95]

Thank you!