Stevenson Screen 87mm Diameter with solid base for window ledge

Update:
I made a new TPU version (https://www.thingiverse.com/thing:4906214) of this
Running the two with identical sensors in parallel shows: PLA wins. More below.

The problem with sensor housings is that they always add to the failure margin of the sensor's measurements. This one does, too. Just a little less.

I keep my Luftdaten - dust sensor in the proposed airRohr design which is ugly but holds strong against UV and rain. However, the temperature rises to high peaks in that dark PVC tube and even in shadow the temp is too high by 2.5 to3°C, according to control measurements done with a DS18B20 with no housing. Similar measuring saw an temperature increase of ~0.4°C with this housing.

So I've tried to find an alternative for the temp/humidity sensor and found the source for my remix which comes as an openSCAD file. You can enable/disable features in the python source code and I used it to create a more compact size. the design is modular, you print a top and several middle parts and you snap them together to form a kind of pagoda. The connection is not very firm and I ended glueing the parts at the snaps.

I added a Y-shaped thin piece that has a hole to hold a 8x Dupont base to plug in the sensors. (It is a very tight fit, you may need to carve out the corners to get it in.

Another bit I added is the base which is meant to stand on a window ledge and keep the pagoda from flying away. As my window ledge has an 10° inclination so has the base.

I printed the parts in white PLA and the base in TPU. Printing the rings takes ~45min each while the bottom printed 2,5h. The .stl are named x30 cause 30mm width was my input, the result has a diameter of 87 mm.

Freecad sources are included so you can adapt it, if, however, you want to change the general size of it I'd recommened to use the source's scad file.

Compairing PLA and TPU:
I have 2 almost identical sensor cases, both just a window apart on the same side of the house, one in PLA, the other in TPU. I plotted the temperatures on a graph, during night * hours you see the sensor in the PLA case reports about 0,6°C higher temps than the other, the sensor in the TPU case and a DS18B20 sensor (as a control in the PLA case) report identical.
As soon as the *sun appears the picture changes. The sensor in the TPU case goes up and reports higher than those under PLA, and the more sun there is, the higher the difference.
After the sun weakrns, 18:00 at this time of the year, things reverse once again.
So this looks like a clear proof that PLA is better suited to isolate against heat from solar radiation than TPU. Didn't expect that, the tests on durability and rain resilience are still running.

Update:
The thing on top is a lingam & yoni (https://www.thingiverse.com/thing:4915081) The lower semi-sphere is double walled so it helps shielding the thermo sensor from the radiation and heat of the sun and I hope it will guide some rain away from the structure and make it drop down at the side.

-