Jim, move the sensor back 6" slowly while you're still looking at the par numbers and watch it jump up and then back down. It's hard to take readings in a tank and post numbers with it aquascaped, as the rock work is naturally pushing your sensor to the front of the tank, rather than underneath the sensor. Not saying it's good or bad, just stating that one par test of a fixture could have been done directly under the fixture and someone posted the numbers on their tank, then the next par test could have been taken with the sensor not directly under the fixture, and more out towards the front.
You can have a sensor at 6" under the surface of the water, and move it 3" to the front and see a 300% drop in PAR, and then back 5", and see a 450% jump. I think the most accurate way to test this would be on a tank with no aquascape to ensure where the sensor is in relation to the fixture, and then test a new fixture (same model), in the exact same manner. Or maybe even better yet, to ensure accuracy of the comparison test between this fixture and a new one, to just test it in air, not even over a tank, then you can chart where the sensor is in relation to the fixture to be consistent.
This isn't anything that's just related to LED's either. Throughout the years that I've done testing on metal halide bulbs and different reflectors, and I can make PAR readings jump by as much as 30% by just moving a sensor an inch left or an inch right, or front to back a very small amount, etc. It's just due to the sensitive nature of the testing meter.
Let me know if I can do anything to help. I will test my 48" Illumina that I have here to see what it comes out to when I can. HTH!
