[sdiy] OpAmp Overvoltage / Power Up (was: Re: Pots read by microcontroller and sourcing question)

[rsdio at audiobanshee.com]

Pots that will be read by a uC should be powered by the same rails as the uC, NOT from the 12 V or 15 V analog rails.

Of course, if you need these pots to drive CV directly as well as being read by the uC, then you'll need scaling.

If you're worried about over-voltage, then provide a series resistance that limits the current to the maximum that the input can handle. This might not work for every op-amp or uC (check the data sheets), but you can feed a 5 V op-amp from a +/-15 V op-amp if the correct series resistor is used (this also assumes that you won't typically send more than 5 V, but just want to protect from accidental over-voltage). A Zener might add extra protection if the input doesn't already have it, but it might be over-kill.

Brian


On Feb 5, 2015, at 12:12 AM, Roman Sowa <modular at go2.pl> wrote:
> All those concerns are unimportant when you use rail-to-rail in-out opamp supplied from the same voltage as the micro, which is what I always do when there's external opamp involved in uC design.
> This way whatever happens, opamp output will never exceed uC rails.
> 
> Roman
> 
> W dniu 2015-02-05 o 00:09, Michael Zacherl pisze:
>> 
>> On 3.Feb 2015, at 10:11 , Oakley Sound <oakleylist at btinternet.com> wrote:
>> 
>>> One problem with an op-amp per pot is that you need some sort of over voltage protection on the ADC or the Mux. An op-amp on power up, or partial power up, can do some horrid things.
>> 
>> I tested OpAmp circuits with one rail disconnected and in most cases the output went up/down straight to the other rail.
>> That indeed can be way too much for a 3.3 or 5V µC in a 12 or 15V system.
>> For a good reason every now and then I can read warnings about making sure both voltages are present in a system otherwise parts can/would be damaged.
>> I wonder how's that at power-up time?
>> Due to tolerances etc. both supply voltages never would come up exactly at the same
>> time, but that's probably a different situation with capacitors charging, rails poperly connected etc?
>> 
>> Michael.
>