I tried to use simple geometry to figure out how the PAR reading should vary with the length of the can, at a fixed distance. I used my last data point, with the lightly sanded can as a starting point. From this I figured that a 12 inch long can should give me a maximum PAR in the lighted circle of 20-25. So, I made a 12 inch long can from the rest of the dryer vent pipe I have, and set it up at 36 inches and measured the PAR again:
You can see that I did get the 20-24 PAR that I should get. Of course now the circle of full PAR light is only about 12 inches in diameter, a bit small for an 18 inch front to back 65 gallon tank, but workable I think.
The chart of PAR vs can length, with the red data points the two can lengths I used, and the red line is what I calculate should be the variation with can length, is:
Now, I need to decide whether it is worth it to finish the project. This seems to be a good lighting arrangement for a 40B tank, only 17 inches high, but very questionable for a 65 gallon tank, 24 inches high.