Astrophotography cameras quantum efficiency

When someone asks you “what qualities should a camera have?”, probably there will be many different answers. Some like it small, others would like to have it smart. Another valuable quality may be the long battery life, WiFi connection, or smartphone control application. But when someone asks you “what qualities should an astrophotography camera have?”, then sensitivity (quantum efficiency) should be at the top of the list. Because at the end the quantum efficiency determines how many photons that landed on the sensor will be transformed into the effective signal. Assuming the dark noise is not a big issue in modern cameras, and readout noise is usually low and can be escaped by longer subframe exposure, the quantum efficiency should not be underestimated.

And the latest generation of CMOS sensor is very well equipped with quantum efficiency 🙂

There are a lot of discussions on the web about CMOS cameras, and usually, the main topics and concerns are readout noise and image sampling rate. That makes sense in most cases because the sensitivity of the modern cameras is often quite high and at the same level. However, if we take a look little at the past and compare them to premium astrophotography cameras that were developed 10 or more years ago, we can see how much have changed.

Not so long time ago cameras with 50% quantum efficiency just started to appear in the market. Well-known Kodak chips KAF-11000 and KAF-8300 barely reached 50%, popular ICX285 (Atik314) peaked at 65% and had excellent noise properties, but it was physically small and had modest resolution – even then.  Extremely popular SBIG-2000 was also able to reach 50% around the green spectrum range. The common pain of all these sensors was a quite deep sensitivity drop at the red spectrum area. There were also sensors considered as more “scientific” with QE up to 80% but came at a price – no anti-blooming gates. Missing ABG was a no-go for general astrophotography because those kinds of artifacts were really hard to remove.

A little bit later new CCD sensors from Sony were released – the famous ICX694 family. QE at the level of 75% was a big advantage, but these chips also had low noise, and the sensitivity in the Ha band was at an impressive level of 65%. The only problem that remained was the sensor size, which was still just modest. So the amateurs had a dilemma – to catch more light in a small field of view, or to get a bigger picture while spending more time on data capturing.

At that time CMOS sensors were already present in the market, but at small sizes – mainly for planetary imaging and guiding. It quickly turned out, that the low readout noise of CMOS sensors can be a big advantage. Plus all the consumer market cameras already were using CMOS technology, and the availability and quality of CMOS sensors started to increase very fast. One of the market breakers was the MN34230 sensor from Panasonic that is used in the extremely popular CMOS cameras like ASI1600, QHY163, or later Atik Horizon. It is a medium format sensor in the size of KAF-8300, and just a little bit better QE, but the new camera pricing was quite tempting.

And then CMOS sensors avalanche started to roll. In the beginning, it was hard to get monochromatic sensors in a larger format, but that situation has fortunately changed and currently, we have a decent selection.

The table below contains quantum efficiency values in percent for selected cameras and their sensors at different spectrum bands. L channel is averaged QE over the visual range 400-700nm. R, G, and B channels are averaged QE over these color bands (about 150nm wide). H alpha and Oiii QE are values read for that wavelengths – 650 and 500nm respectively.

The table is ordered so the higher QE sensors are also higher in the table.

What does it mean in real life? Well, if you take a look at Atik 383 or ASI1600 and compare it to recent QHY268 or QHY294, then you can see, that QE is about 60% higher in every band. So using the same telescope you can capture the same amount of data with QHY294 in 2 hours, as with ASI1600 in 3 hours. Or – from a different point of view – your telescope just gained 25% of aperture, but kept the same size and weight 🙂

Another example – if you are still using famous full-frame Atik 11000 and decide to switch to full-frame CMOS QHY600 or ASI2600, then your QE will increase by 90% on average, but over 140% in the red part of the spectrum, where the H alpha band is captured! That is almost 40% aperture “increase” on average, and over 50% aperture “increase” for red color. Or just getting the same result in half of the time.

In the plot above there is QE for selected sensors presented as the graph. The source of the data presented in the plot is listed at the end of the article. It is clearly visible how the sensor sensitivity has increased, and especially in the red part (H alpha band around 650nm), the QE is now at very reasonable levels around 70-80%.

So if you plan to upgrade your camera, take a look at the QE parameter as well. If you select the camera with 80% QE, then it may stay with you for a while because there is already not much to gain in this area – you just cannot break the 100%, and probably it will be quite hard to get significantly closer to this value.

CMOS cameras of course have some weaknesses. Some of them come from the small pixel size – the pixel capacity is lower, and also for bigger telescopes it is easy to fall into the oversampling area. But since the readout noise is low, the moderate oversampling is not a big issue for CMOSes. And the same low readout noise allows to capture shorter subframes, and then lower pixel capacity is not a real problem anymore. Another feature of small pixels is that they reveal optical system flaws, that were not visible before. But this is not the CMOS sensor that should be blamed for optic problems, right? Well, many amateurs disagree 🙂 Some CMOS sensors have also other problems like amp glow (can be removed with proper dark calibration) or random/walking noise (can be removed with dithering and stacking with pixel rejection algorithms). One way or another it does not matter if you have many or a few clear nights, the large QE camera will always help. And proper calibration and dithering – that is not a big price for having better images.

You may check other entries about CMOS sensors:

To CMOS or not to CMOS
CMOS calibration challenge
Pixel scale and resolution

Sources:

• http://blog.astrofotky.cz/pavelpech/?p=864https://astronomy-imaging-camera.com/product/asi178mm-mono
• https://www.cloudynights.com/topic/754121-starting-mono-asi1600mm-pro-vs-asi294mm-pro/
• https://www.qhyccd.com/measuring-the-absolute-qe-of-the-qhy268m/
• https://astronomy-imaging-camera.com/product/asi294mm-pro
• https://www.baader-planetarium.com/en/blog/comparing-the-imx455-industry-grade-and-kai-11002-35mm-format-monochrome-sensors/