My first impressions:
- 3% RMS PRNU is quite high for such a large pixel. Not clear why it's that high.
- 30,000e- full well over 2e- read noise does not make 87db DR, like the table states. May be achieved at slower speeds when the read noise is lower
- -0.4V supply voltage is used for Tx gate bias during integration, I guess.
Thanks to BP for the info!