fundamental limits of detection.

[Anonymous]

Hi all
I have a problem that I need a little help with. If I have series of parallel copper tracks 2mm wide with 2mm gap between them, what is the furthest distance I would be able to resolve each track? One can suppose that the surroundings are non metallic. What about alu? Or tracks 10mm wide with 10mm gaps?
Many thanks (brilliant forum by the way)
GaD

 

[Anonymous]

I presume your talking about using an eddy-current based metal detection device ("metal detector"), and not for instance something that works in the microwave region.
The answer depends on whether you know in advance the dimensions of the tracks and gaps; also depends on what kind of sensing mechanism; also depends on what kind of algorithm is used to resolve the separation.
In other words, there isn't an absolute limit; however, as a practical matter, I doubt you'd be able to resolve the individual tracks at a distance greater than about twice the distance between the track centerlines, without using some pretty fancy software analysis.
Problems that are mathematically similar to this are common in image processing. Perhaps someone with experience in that field can provide a better answer.
--Dave J.

 

[Anonymous]

Actually the more track there are the easier the problem becomes. Reason is that they can be detected with an optical grating. The grating will interact w/ the light from the parallel Cu tracks and form Moire pattern (interference fringes) w/ are macro compared to the Cu tracks. If there are only a couple of tracks or in general you need to know what kind of optics you have, their quality, the type of detection system and its resolution capability. With signal processing you can pull lots of images out of the noise. I worked with images where we got good data with a 1/1000 signal to noise ration (SNR). Basically you have asked a very complex question.
Dan

 

[Anonymous]

Hi
There has got to be a limit, surely? with 2mm tracks and gaps, I'm not gonna find them at 3 metres. I thought that the field would be essentially circular? As I pass a coil across the tracks at a fixed distance, I need the field to cut one track and not the two either side. More importantly I need to detect the gap in between. The field must not cut the tracks otherwise it will start to look like a track. Does this make sense?