[volt-nuts] Offset voltage temperature drift in rail-to-rail opamps

[Charles Steinmetz]

Attila wrote:

> I am currently looking at the LT1677/1678/1679 family of opamps, which
> have an increadibly low tempco for the input offset voltage (0.4µV/°C typ).
>      *  *  *
> But, when the input voltage comes close within the upper or lower rail,
> the input offset voltage specs get much worse.    *  *  *
> is my assumption correct, that with worse offset spec, close to the
> rails, also the offset voltage drift specs get worse?

That is almost certainly true, but it is not specified by the 
manufacturer.  You would have to test for  yourself to know for sure.

Note that, in addition to the offset voltage changes, the input current 
and input offset currents also change with the common-mode voltage. 
Both are due to the rail-to-rail *input* architecture, and are common to 
RRI opamps.

Compare this to the AD8675, which is RRO but not RRI.  Its input offset 
tempco (typical) is even better than the LT1677 (0.2uV/C), as well as 
its input noise (2.8nV/sqrtHz with a very low 5Hz 1/f corner frequency), 
but without the input common mode effects of the 1677.

If RRI is important in your application, you may be able to do without 
an actual RRI opamp by running the opamp on higher supply voltages than 
the rest of the circuit.

Best regards,

Charles