For evaluating the precision of a DAS setup with a AD574 and 89c51, I am looking for a solid and steady source.

In the current field location, there is no access to a miilivolt source and I was wondering if the following can be used as quick and dirty solution :

Use two numbers 9 VDC batteries in series as the supply. Then use a AD581 bandgap refernce. Record this 10V input as the input to the ADC ( with suitable scaling )over a period of time and then evaluate the results. Since the temperature coefficient of the AD581 is detailed in the data sheet, the ambient is measured near the case of the AD581 and then use this correction in the noted readings.

( It must be noted that the ADC+ MCU setup is being evaluated. I am only looking at the way to make a good voltage source from available components for field use without resorting to Fluke / Tektronix and the like )

Any tips / pitfalls to avoid in the above proposal ?

Thanks

Raghu