You're welcome to be as pedantic as you want. The site said the Sun sets on the equinox and shows a photo of it on that day.
Incorrect. Had you bothered to click on the image you posted here from the site, it would have enlarged, showing that the image was not snapped on the equinox day.
http://www.arctic.noaa.gov/npole/2002/images/noaa-2002-0830-0140.jpg Now let's agree that clouds, irregular ice surfaces, and other minutia can alter the scenario. You're still challenged to show on your model that the Sun can be seen from both poles on the same day to match the evidence, even if there's a day or two difference.
The evidence which has been provided for that assertion has been a website which states that predicting the position of the sun under RET is a "guessing game," an observation of the sun from the SP by a NOAA employe, a site which states that at the NP the sun sets before the Equinox occurs, and a NP webcam picture which was taken almost a month before the day of Equinox.
I'm afraid the ball is still in your court on this.
Why can't you produce a illustration of the Sun's location and area of illumination of September 20-30 and demonstrate that required flexibility in your model's predictions? I just love the idea that your model requires the Sun to act as in such contorted ways as to defy all reasonable expectations. The Sun in your model as midnight UT is "above" the NP in June, drawing a line down across the NP to London, where it's dark, is closer than the Sun is to Hawaii, where it is light. Quite embarrassing for the bi-polar model.
When the sun is over the mid Pacific, it is not light in London.
So I repeat the challenge: I challenge you to show, and then explain, the area illuminated by the Sun at 23.00 UT on March 21, 2014. People in Sydney, Australia saw the Sun to the east. Yet by your explanation the Sun was on the Equator on the western part of the FE.
Who saw what, where? Were you in Australia on that date?