Re: Do I need to worry about voltage swing?
Shortly after making the original post I did the calculations and discounted battery internal resistance as the cause of the problem.
Dived in to debug the code by using an if statement to gradually invoke the code that was processed in the main loop. This involved about 10 iterations. Every iteration except the last revealed a voltage swing of just .05V. The last iteration brought me back to the 0.25V swing which was causing my problems with HLVD.
The problem piece of code that's causing the voltage to drop by 0.2V is
The voltage dip here could well be more than the 0.25V I measured with the multimeter.
Seems bizarre, won't have time to another look at the problem until later this week. Closing USART was supposed to reduce power consumption before going into sleep, but what it appears to be doing is introducing a large dip in voltage that triggers an interrupt from HLVD. I've yet to measure the impact on power consumption of not closing USART before entering sleep mode.
The circuit is still on a breadboard. I have checked that all the wires are properly seated.
With the evidence I've seen so far and once I've understood what's going on with Close1USART() I can move to a more permanent solution for the breadboard.