In November 2014 I have been invited by Nikon Japan to evaluate 2 pre-series models of the D810A. First tests have been acomplished at Panther-Observatory with my 16" at f/3 in my home village, in January 2015 I travelled to Chile and could do extensive testing of the new models including a set of Nikon lenses (14-24mm, 58mm, 200mm)and with the Tak FSQ-85 at our remote observatory at CTIO known as Chart32

The camera is featuring 36.3 MP in full frame format and placed at the high end DSLR segment.
The following features have been changed in comparison to the D810 model:
* removal of low-pass filter in front of the CMOS chip
* new improved IR filter with high transmission for H-alpha at 656 nm, see filter curve here
* higher review magnification of up to 10x for analyzing images via the internal display;
* new M* mode allows automatic internal exposures up to 15 min;
* unlimited internal interval timer for creating time-lapse videos or multiple exposures
* electronic first curtain to remove vibrations.

The Nikon D810A weighting 980 g uses the same 36.3 megapixel CMOS sensor (7360x4912 pixel) in FX format (36x24mm) as the D810.

Recording takes place in JPG, TIF or NEF format

The metal and very solid SLR type body enables optimum coupling to any telescope or non Nikon lenses via a FX/T2 adapter ring.

Highly recommended accessory: remote power supply for continuous use or several spare batteries,
one battery lasts for about 3 h of continuous use.
The remote control software is included in Nikon supply.

Interesting links for the D810A:

Official Nikon information: here
Review by Ken Rockwell: here
Review by Göran Strand: here
Review by Adam Woodworth: here

What I am still missing:
* moveable back monitor
* subframe center readout (ROI) for planetary imaging in video mode

Dark # in series

First

Last

Average of 10

Nikon D810A (NEF)

26.3 / 6.3

26.7 / 6.9

26.3 / 5.4

Nikon D810A (jpg)

18.6 / 14.3

19.7 / 15.4

17.2 / 7.4

Canon 6D (raw)

8.6 /17.6

9.7 / 21.0

9.1 / 16.5

Canon 6D (jpg)

7.5 / 10.5

14.7 / 22.0

10.6 / 15.1

Mean value and standard deviation measured in Photoshop CS5
Nikon D810A darks simultaneously recorded in NEF and jpg-fine. NEF conversion to 16 bit tiff in ViewNX2
Canon 6D darks simultaneously recorded in raw and jpg-fine. Canon Raw to 16 bit tiff in Photoshop via Adobe DNG converter.
Min-max excluded average combine in Imagesplus

Conclusion:
Internal heating during multiple exposures minimal for D810A (7%), significant for Canon 6D (20-100%)
Creating Masterdarks is most valuable for D810A, the D810A CMOS camera behaves similar as a CCD astrocamera.
For the Canon 6D in JPG mode a preprocessing and blurring within the camera is evident, 95% of the pixels count to zero, so no information can be used during averaging the darks, the blurring seems to reduce the standard deviation but detail information is lost. Surprisingly the Canon 6D masterdark in JPG mode shows the same noise level as the single dark in the middle of the series.
For the Canon 6D in RAW mode there is no clipping in the dark area and the heating effect during the series counts less. Creating RAW Masterdarks is helpful but to limited extent (25% improvement).

The noise level in the D810A images is up to 40% higher for JPG compared to NEF mode in the averaged image, 100% difference when single darks are used, another benefit is the higher dynamic range for NEF mode so it should be used always except for very bright objects. However the calibration and processing of NEF images converted to 16 bit tiff is very demanding for the computer hardware, a single 16 bit D810A tiff file is 207 MB in size.
Also for the Canon 6D the raw mode setting is highly recommeded when using in low light environment.

Overall the the new Nikon D810A clearly is outperforming the Canon 6D.

Find the processed videos and images with the D810 here.