Non-linearity analysis of time to voltage converter

[dirac16]

How can I measure the linearity of a time to voltage converter? For ADC we can measure its DNL, INL, gain error, ENOB etc. to assess its linearity performance. But as for a TVC both time and voltage vary continuously it does not make sense to define DNL/INL and so on. So what are the alternative methods to derive the linearity of the TVC?

I want to design a two-step time to digital converter using a time to voltage converter followed by an ADC. So the non-linearity analysis of the TVC is important to select the right number of bits for the AD conversion.

 

[danadakk]

TVC PWM implementation with a LPF post output ? If so its absolutely monotonic
so clock uncertainty would be primary factor.


Regards, Dana.

 

[Dominik Przyborowski]

INL is an dispersion of actual value from ideal one, and is applicable to every linear system, fully continuous as well. Simply, you will not normalised it to LSB but to range and express in % or ppm.

 

[dirac16]

INL is an dispersion of actual value from ideal one, and is applicable to every linear system, fully continuous as well. Simply, you will not normalised it to LSB but to range and express in % or ppm.

Right, if I am only able to draw INL in percentage then I don't know how that helps in knowing the maximum applicable resolution of the forefront ADC. For example, I don't know what the maximum resolution of the ADC can be based on the linearity of the TVC. I would need to analyze the TVC transfer curve and give an estimate of its linearity which imposes limitations on the number of bits in the ADC.

 

[FvM]

But as for a TVC both time and voltage vary continuously it does not make sense to define DNL/INL and so on.

Makes no sense. It's well possible to measure time-to-voltage transfer characteristic and determine DNL/INL. As already stated, because the curve is monotic by design, you can focus on INL.