pga450q1

The appropriate Applications Engineer has been assigned to answer your question(s). You can expect a follow-up response within a few days.

its really kind of you :)

if possible please reply asap ... got  some deadlines ... 

thank you :D

My apologies, I wrote out a response for this a couple of days ago, but there must have been an error when trying to post it.

Part of the ultrasonic signal will reflect back and part will pass through the metal when the signal reaches the metal. We've successfully tested detecting objects through a very thin sheet of aluminum placed right next to the transducer.

How thick is the metal you want to pass through and how large is the crack? It might be best to see if this is feasible by running some experiments. Look at the echo data (located in the FIFO RAM) when detecting the metal with and without the crack. If you can see a difference in the echo pattern, then you should be able to device an algorithm to extract that information.

The settings that you will need to adjust are:
PULSE_CNT_A - number of pulses in a burst, more pulses will send out a higher strength signal, but could cause saturation at shorter distances
FIFO_CTRL: sets what portion of the digital datapath is stored in the FIFO RAM. Use the lower 8 bits for more resolution, but if you have saturation issues, you can try the mid 8 bits
DOWNSAMPLE: the ADC samples data every 1us. Use the DOWNSAMPLE setting to control how often a sample is stored in the FIFO_RAM.

This device has OTP; however, you can use the 8k of development RAM to re-write the firmware during development. We can offer sample firmware that shows how to do this. Once you finalize the firmware, just a few changes are needed in the project settings to write the program to OTP instead of the DEVRAM.

Thanks,

-Clancy

Hey Clancy,
We made the module work
We were able to successfully pass sound through metal
This is our project link :
www.youtube.com/watch
TIIC IDC 2015 – Team 1608 - “Ultrasonic Broken Rail Detection”
Thanks a lot for your help :) 

Charu,

That's an impressive project!  I'm curious to know how far ahead in the track you can detect a defect.  Also, I wonder if separate track segments show up differently from cracks.

I have worked with a similar tool in the past called a 'reflectometer' -- but it was designed to find defects in copper telephone wires using electrical pulses.  The range on it was quite impressive (I have seen useful data past 15k feet).

Charu,
That's a cool project! Thanks for sharing. I'm also curious about how far away a defect can be detected.
-Clancy