Hi Dino,
I have done my measures with a Spyder 3 Elite (don't tell my wife it costed me extra 200€).
I must admit the 'Dino test' is less expensive. But I like technology, I can't resist. 🙂
6600 Kelvin average on top (rather uniform) and 6200 average bottom *but not uniform* (6350 left and 6100 right).
The panel passes the Dino test, I write on the Post Message form full page : no sign of yellowish.
But the gray bar test shows a slight difference Top-Bottom and a warmer bottom-right corner.
Try to borrow a colorimeter -maybe from Apple store 🙂 - and test your unit.
Here are a few evidences we have confirmed so far about the defect :
= it varies widely from one unit to another,
= early units are better (week 40-41 2009),
= it's a color shift to yellow (meaning not a LCD problem but a thermal issue with LEDs - my opinion),
= the shift is stronger in the bottom-right corner (on most units).
The color shift to yellow (color tint) is meaning a lot (starting from here my theory).
There are no white LEDs, but different ways to make white light using LEDs. The most common (and economical) way to produce white backlight is to use blue Leds (InGaN) with a yellow phosphor (YAG) on the top of it.
The combination of blue light emitted from the InGaN and yellow light from the YAG phosphor produces a light that appears white.
Among other causes, there is a color shift associated with the LED temperature; the dominant blue wavelength of the InGaN is increasing (as the intensity of the YAG ) resulting in the shift to yellow.
All this seems IMO to point on a thermal management issue, with the iMac panel.
See this article in The “Led Journal” and particularly Fig 2
http://www.ledjournal.com/images/PDFs/Online%20Issues/Articles/LEDJA08Microsemi.pdf
Cheers!
Jacques.