Why the first radiation shield read three degrees too warm
14 June 2026
The original shield was a stack of seven plates, 90 mm across, printed in white PLA. On still, sunny afternoons it reported up to 3.1 °C above the reference thermometer hung in full shade two metres away. On overcast days the two agreed within 0.2 °C, which is what made the cause obvious in hindsight: the shield was heating, not the sensor.
What actually fixed it
Three changes, applied together because the summer was ending and we wanted the season:
- Plate diameter increased from 90 mm to 150 mm, spacing from 12 mm to 20 mm.
- PLA replaced with ASA. The PLA plates had also begun to sag noticeably by August.
- The sensor moved to the middle of the stack instead of the second plate from the bottom.
The remaining daytime error is about 0.4 °C at noon in July, which is roughly what a passive shield of this size is expected to do. Aspirating it would close most of that gap, and would also cost the station its ability to run for six days on the battery. We kept the error.
The measurement we should have made first
A second temperature sensor taped to the inside face of the top plate. It reads up to 11 °C above air temperature in direct sun, and it would have pointed at the shield on day one instead of week five.