sensor for detection wire length

[c_mitra]

In principle it is simple: you send the pulse as a digital write bit on a given pin. That output is connected to the wire whose length is to be determined. You also start the counter (clock at the same time).

Connect the same end to another pin and configure that as a digital read: as soon as you read 1 you stop the counter.

In reality, you may be needing some schmitt trigger or a monostable to remove noise and select a threshhold. Sometimes the reflected signal may get inverted (the reasons for this we need not discuss).

Finally you need to correct for various delays in the software.

 

[FvM]

The velocity of the pulse is same as the velocity of light in vacuum but it is about 10% in metals. Light travels 30cm/nS in vacuum and as a rough estimate, the electrical pulse will travel about 3 cm/nS.

How did you arrive at this result? Propagation in transmission lines with dielectricum is slowed down by a factor of squareroot(Er), about 1.6 - 1.8 for usual cables, gives about 20 cm/ns. TDR resolution is respectively about 0.1m per ns, as previously derived in this thread.

The intended time resolution can be hardly achieved with a microprocessor timer, except for the dedicated high resolution timers offered by some modern processors like STM32F334x.

Microchip has an application note AN1375 that describes how to achieve 1 ns TDR resolution with PIC 24 built-in time to digital converter (CTMU). It's also described in an ECN article **broken link removed**

 

[c_mitra]

How did you arrive at this result? Propagation in transmission lines with dielectricum is slowed down by a factor of squareroot(Er), about 1.6 - 1.8 for usual cables, gives about 20 cm/ns. TDR resolution is respectively about 0.1m per ns, as previously derived in this thread.

Sorry for mistake. My apologies.

 

[asa1365]

How did you arrive at this result? Propagation in transmission lines with dielectricum is slowed down by a factor of squareroot(Er), about 1.6 - 1.8 for usual cables, gives about 20 cm/ns. TDR resolution is respectively about 0.1m per ns, as previously derived in this thread.

The intended time resolution can be hardly achieved with a microprocessor timer, except for the dedicated high resolution timers offered by some modern processors like STM32F334x.

Microchip has an application note AN1375 that describes how to achieve 1 ns TDR resolution with PIC 24 built-in time to digital converter (CTMU). It's also described in an ECN article **broken link removed**

Thank you , your opinion I can produce a strong pulse with stm32f429 with 168MHz
http://stm32f4-discovery.com/2015/01/library-48-measure-pwm-input-signal-stm32f4/
, (i have it now ) or do I buy stm32f334 ? i have not programmer for Microcontrollers pic

 

[FvM]

As far as I see, STM32F429 has no high resolution timers. Not suitable for short range TDR.

 

[c_mitra]

A super simple experiment is described at: https://sharepoint.umich.edu/lsa/ph...y of an Electrical Pulse in a Wire (New).aspx

It is not dealing with reflections, though.