Submitted: Monday, 20th April 2015 by Slawomir Lipinski
In 2014 QSI (Quantum Scientific Imaging) announced new 600 series with Sony sensors. All new models have Sony EXview HAD CCD II image sensors; ranging from 2.8 megapixels to 9.2 megapixels. A few months ago QSI added a 12 megapixel model to the 600 family.
Later last year I was in a search for a camera that would provide a larger field of view than my Atik 428 ex, while maintaining its reputable sensitivity and low noise. After considering acquiring a camera with the venerable 8300 chip, I decided that I need a CCD that would work with small 1.25 inch filters; hence I resolved to stick with small but clean and sensitive Sony's CCDs. QSI has good reputation for producing quality cameras and two-stage efficient cooling might prove useful in sub-tropical climate, thus later last year I pulled a trigger and ordered QSI 690i. I already had a filter wheel and did not plan using an OAG, thus I opted for the slim 'i' version.
I have been using QSI 690i camera for a few months and would like to share my experiences with this camera.
Build quality of QSI 690i is first-rate. Power supply and a T-mounting to 2 inch nosepiece adapter are both included. Camera weighs 740 g and feels very solid.
Few Basic Tests
In an attempt to measure the camera, Craig Stark’s method was followed to obtain necessary data (Signal to Noise: Understanding it, Measuring it, and Improving it Part 3 - Measuring your Camera from July 2009).
Since the goal of this investigation, apart from basic curiosity, was to confirm specs provided by QSI, only four data points were taken to estimate system gain.
Data extracted from the above graph, as well as from multiple bias and dark frames, and during ‘normal use’ have been collated in the table below.
In short, specifications given by QSI have been confirmed. Sony does not specify well depth for its sensors; therefore QSI does not provide this information neither.
Cooling and thermal stability
The graph below is a result of measuring mean ADU values for 1-minute dark frames over 25 minutes. It can be noticed that the camera needed only 2 minutes to reach set temperature of -10 C and about 3 minutes to reach thermal equilibrium (ambient temperature was about 30 C, summer in Queensland).
The following image shows a jpg file generated by stacking 160 bias frames in Nebulosity. Histogram values set in Nebulosity: Min 494, Max 745 (out of 64000). These histogram values were based on the min/max ADU numbers in the master bias frame.
There is some gradient evident in the above master bias frame, thus further analysis was performed in Image.
The next graph shows horizontal profile of the same stack of 160 bias frames (saved as a jpg file with histogram values range: minimum 494 and maximum 745). Again, some gradient is apparent; nevertheless, this is a very clean and relatively even bias frame. Maximum Grey Value for jpeg files is 255.
A stack of seventy five (75) 10-minute dark frames at -10 C. Histogram values: Min 494, Max 745 (out of 64000). There is a minor amp glow evident in this master dark frame.
Further analysis of this master dark frame was performed in Image resulting in the image below showing horizontal profile of a stack of 10-minute dark frames.
The jpeg image was generated in Nebulosity, this time with the full range for the histogram values: from 0 to 64000. Again, this is a very even average horizontal distribution of ADU values for pixels across the entire image. Maximum Grey Value for jpeg files is 255.
What does it all mean in ‘real life’?
After some experimentation it has been concluded that subtracting dark frames is completely unnecessary, at least up to 10-minute subs when running the camera at -10 0C. Dark current of 1.09 e- in 10 minutes at -10 C is practically non-existent. There is some minor amp glow that builds up with longer exposures, but I am yet to notice it in my astro-images. Subtracting bias frames has been thus far sufficient and adequate in pre-processing.
Below is a comparison of two images; a crop of a single 10-minute raw frame and the same frame with bias subtracted.
Relatively small pixels of ICX 814 cause bright stars to saturate fairly quickly, nevertheless, clean electronics combined with high sensitivity of ICX 814 chip do not require long exposures, even with narrowband imaging, to capture faint nebulosities.
Below are a few humble examples of images resulting from a range of exposures. All images were taken from a balcony near Brisbane’s CBD, through a 102 mm doublet. Processing was performed in PixInsight.
All in all, high sensitivity of ICX 814 and low noise from QSI 690i combined with good filters allow even budding astro-photographers to capture faint DSOs in relatively high detail.