Yes I think the loop reading the feedback from the ADC and converting it for display appears to have an issue.

Though I have decided to correct the ADC readout as below, I would like to know if any better method exists :

ADC in +/-10V bipolar mode.

Thus when the reading is below 800H, I subtract reading from 800H and save it as ADCVAL. ( 16 bit variable)

WHen reading is 800H I set ADCVAL to zero

When reading is more than 800H, I subtract 800H from reading and save it as ADCVAL.

Then I multiply ADCVAL by 488 ( 1bit = 4.88mV and to increase precision I use a 100 multiplier) and divide the answer by 1000 before converting it to BCD and displaying it. Thus a count of 2048 decimal will yield ( 2048 x 488 /1000 = 999 ) which is displayed as 9.99V. Polarity is deduced based on the earlier branchout.

Is this method OK ?

Also I am currently checking the interrupt priorities - there is one master ISR using T0 ( 5ms )to set the ADC capture event trigger once every 100ms and another using the T1 for the UART comm at 9600 Baud. Not sure if there is a clash anywhere . Checking.

Thanks for your inputs.

Raghu