CPU "Crash" when AD line goes below GND

[Eugeneo]

Can somebody explain why a 16F819 chip doesn't function correctly when when an ad defined pin voltage is slightly negative with respect to VSS?

I was just expecting the protection diode to kick in and give me a value of 0.

[Ttelmah]

If it is only going fractionally below Vss, what you describe should not happen. 'Fractionally', is down to about -0.3v.
Stop thinking in terms of protections diodes. There are none!. The diode behaviour, is an inherent feature of the structure of the output FETs, not a seperate diode as such. Unfortunately, the presence of both the output FET, and analog input on the same pin, when dealing with -ve excursions, result in the input gate becoming reverse biased _before_ the low side FET starts to conduct. The FET structure protects the chip from destruction, but does not stop an unexpected current flowing into the input multiplexer. Normally if the source has a relatively high impedance, the current is limited to levels that don't stop the chip. There is a 'balancing act' here between the requirements for the source to have a low impedance to give fast acquisition, and avoiding this problem.

Best Wishes

[Eugeneo]

Thanks for the reply. Where did you find out they use a field effect transistor structure in the chip? Do they have a detailed design PDF somewhere?

[asmallri]

What do you call slightly negative? What current is flowing? Good chance you might be inadvertently attempting to destroy the chip?
_________________
Regards, Andrew

http://www.brushelectronics.com/software
Home of Ethernet, SD card and Encrypted Serial Bootloaders for PICs!!

[Ttelmah]

Talk directly to MicroChip. They sent me a 'Spice' model of the input (with a classic "you won't let anyone else have this" requirement), when I needed to work out how a source was going to behave into the analog inputs.

Best Wishes