[sdiy] dsPIC CV input protection

[Seb Francis]

Mattias Rickardsson wrote:
> Would it be a bad idea to skip the OP-amp buffer and have a passive
> input attenuation, with two ordinary diodes (connected in reverse to
> ground and dsPICsupply) added as protection?
> (Assuming that the dsPIC survives input voltages reaching one PN
> voltage drop outside the desired input range)
>
>   

The problem with using passive input attenuation is that the dsPIC ADC 
converters need a low impedance source to sample the voltage quickly and 
accurately.

And the problem with using diodes-to-the-rails input protection when you 
have a low impedance source (such as the output of an opamp) is that you 
will get a *lot* of current flowing through the diodes when the voltage 
exceeds the rails.

So really the only sensible solution does seem to be to run the opamp at 
the appropriate supply voltage so it saturates before the voltage goes 
out of range.

> I have no experience with dsPIC, but with the many A/D inputs of
> today's microcontrollers it is nice to have as minimalistic input
> circuitry as possible to reduce outboard component count. :-)
>   

Yep, that's definitely the plan :)
This module I'm designing will have PCB footprints for 12 CV input buffers.

Seb