Pulse detection using 8051 microprocessor

Hi,

I am currently using the philips 8051 to detect a magnet passing an AMR sensor. Does anyone have any ideas or recommended reading to get near 100% detection and near 0% false activation from noise.

In the perfect situation the digitized signal is a single sin-wave pulse with polarity depending on the magnet orientation. In reality this get distorted by magnet speed/offset/angle/etc

The 8051 is 22Mhz with 1kB RAM so I can't really do any complex spectral analysis.

Any suggestion would be wonderful.

Gabriel

Hi, I'm not sure what you mean by "digitised signal" in this application.If you have allready used analouge to digital conversion then all you need to do is trigger the 8051 from 1 of the higher data bits. If you are using the analouge signal directly then you can use a schmit trigger to remove the noise.

There is no magic circuitry involved.  It's strictly a matter of signal to noise ratio (SNR).  You don't stipulate what near 100% means, but a 99% probability of detection with a 1% false alarm rate requires around 7:1 SNR.  

This can only come from a stronger signal, which means a stronger or closer magnet or a lower noise detector, which means possibly a more sensitive detector with integrated low-noise preamp located director at the detector.  

TTFN

You can reduce the noise by
1.) shielding 2.) filtering 3.) amplitude and duration
windowing

<nbucska@pcperipherals.com>