Why uncooled thermal cameras need shutter?

[videvug]

I understand a shutter is needed for NUC, because sensor pixels behave differently in different temperatures, and may produce fixed pattern noise. But many (old) imagers have both sensor temperature stablizer and shutter at the same time, why? If the sensor can have a stable working temperature, then it should have a known behavior for every pixels, and can work without a shutter, like cooled ones, right?
Also, newer uncooled cameras seems to get rid of the temperature stablizer, and recently some shutterless TICs appeared, how do these happen?

[bap2703]

The lens temperature isn't controlled for example

[Bill W]

There are two considerations, temperature accuracy as noted by bap, but also fixed pattern noise on the image.

The individual column amplifiers are not temperature stabilised enough to prevent the drift in those that causing the fixed pattern noise as they change in offset.
Temperature stabilisation is mainly helpful from the point of keeping pixel responsivity constant.

Once running at a steady temperature you could reduce the shutter operation down to once in 5 or 10 minutes on a stabilised camera (even ASi) but would not go over 2 minutes on a fully unstabilised one.

Bill

[videvug]

Thank you everyone, but cooled ones also don't cool lens, however they don't need shutter at all, why?

[Spirit532]

Because lenses don't have rapidly changing and very dramatic nonuniformity.

[David Hess]

I have an old cooled one and am amazed that uncooled sensors work at all.

[wadizuwaro]

They work in different ways:
Cooled MWIR TICs detect photons hit on the sensor, similar to visible-light cameras. The reference point is zero.
Uncooled LWIR TICs also focus the energy on the sensor like cooled ones, but uncooled ones detect the temperature difference caused by the energy, comparing with the original sensor temperture as a reference.

[Ben321]

There are two considerations, temperature accuracy as noted by bap, but also fixed pattern noise on the image.

The individual column amplifiers are not temperature stabilised enough to prevent the drift in those that causing the fixed pattern noise as they change in offset.
Temperature stabilisation is mainly helpful from the point of keeping pixel responsivity constant.

Once running at a steady temperature you could reduce the shutter operation down to once in 5 or 10 minutes on a stabilised camera (even ASi) but would not go over 2 minutes on a fully unstabilised one.

Bill

What do you do then if you want to film a continuous video (like with a camcorder) without that annoying shutter forcibly pausing the video stream (making it obvious when things seem to jump in their movement) once a minute? Like creating an LWIR video for posting on Youtube and want it to look like it's taken with decent equiepment? The answer is first let it warm up for like 10 minutes, and then also be willing to accept less accurate temperature measurements, and disable the shutter. Then record your video.

[Bill W]

What do you do then if you want to film a continuous video (like with a camcorder) without that annoying shutter forcibly pausing the video stream (making it obvious when things seem to jump in their movement) once a minute? Like creating an LWIR video for posting on Youtube and want it to look like it's taken with decent equiepment? The answer is first let it warm up for like 10 minutes, and then also be willing to accept less accurate temperature measurements, and disable the shutter. Then record your video.

Even if temperature reading is irrelevant to 'nice pictures' as it mainly appears as an overall image shift and most cameras will deal with that very well unless set for a fixed scene temperature span black-white. 

However you will still get column fixed patterns emerging fairly quickly on a fully unstabilised sensor unless you have it in a very well stabilised emvironment - such as running open frame in an aircon office.


Bill