Sensor works much better in open air suspended over water on a boom. We are now testing taking 10 separate reads at each time interval- for trouble shooting information. Although occasional intervals give 300 for all 10 reads, most of the time bad data occurs in only the first 2 or 3 reads at any interval. We are seeing intervals with some bad data values at a frequency of maybe once per day. We will test going up to 25 reads to see if that solves the problem.
The code used for testing is in the attached file. This works as desired (sleep code works) when sonar power is on D18 and signal read is on D19. Alternatively works on D10 (power) and D11 (signal). However, the sleep code does not work if sonar power is provided continuously from the 3.3 V rail. We wanted to test effect of leaving power applied to the sonar all the time to see the effect (might clear off condensation from the face?).