Compensating measurement timing errors

[frankqt]

I have a system where I sample some data periodically (every 10usec). The shape of the data is triangle, in other words it linearly increases and decreases in time. (Both theory and practice are in agreement, I plot the ADC output and I clearly see linear increase and decrease although there is some non-linearity but let's ignore that for the sake of discussion) However the frequency of the signal is not fixed. It may increase very fast or very slow. (narrow or wide triangle)

I am trying to locate the true peak point of the data with 500nsec accuracy, however my ADC samples every 10usec. One item on my favor is that the data values increase and decrease in sequence. I get worst case 50 samples, best case 1000 samples. (in worst case, I will have 25 samples in increasing order and 25 in decreasing order)

How can I approach this issue using DSP techniques? One idea I had was take the derivative of the final sample set and find out the side angles (rising and falling) and calculate peak point from those angles. (I basically extend the left and right side of the triangle and look at where those lines cross each other). I cannot think of any other way. I welcome suggestions.

Thx

 

[FvM]

One idea I had was take the derivative of the final sample set and find out the side angles (rising and falling) and calculate peak point from those angles.

Or something like that. Derivative smells like amplifying accidental measurement errors. For mimimal error, you would want to calculate lines of best fit for both edges and find the intercept point.

The mathematics for line of best fit is called linear regression and can be found in a high school or university mathematics respectively statistics handbook.

 

[frankqt]

Or something like that. Derivative smells like amplifying accidental measurement errors. For mimimal error, you would want to calculate lines of best fit for both edges and find the intercept point.

Ok. You better explain that I did. The only question I have is that since I am sampling at a lower frequency, wouldn't everything be shifted the same amount and I end up getting the mid point that is also shifted.

 

[FvM]

The only question I have is that since I am sampling at a lower frequency, wouldn't everything be shifted the same amount and I end up getting the mid point that is also shifted.

You need to determine the true sampling time and compensate possible linear distortions of the measurement channel (frequency dependent group delay).

 

[frankqt]

You need to determine the true sampling time and compensate possible linear distortions of the measurement channel (frequency dependent group delay).

I get what you mean by group delay but I don't fully understand what you mean by true sampling time. Can you elaborate?

 

[FvM]

I don't fully understand what you mean by true sampling time.

I'm referring to your original post.

I am trying to locate the true peak point of the data with 500nsec accuracy.

To determine the true "peak point", you have to know the true sample points before. The question presumes that there's a time reference for the peak points, so you will be able to assign a time stamp to each sample, respectively for the starting point of a samples sequence

 

[frankqt]

Now I got it. Yes, I have my internal sampling reference time and I can estimate the peak point based on my reference and correlate with real time.