Camera cool-down times (and warm-up times)

[Ags]

Have you ever wondered how long it takes a DSLR to cool down to outside temperatures?

I stuck my camera in the fridge (again ) and set the remote up to take a picture every 15 minutes. These are the temperatures recorded in the CR2 files:

00min 15 C

15min 13 C

30min 11 C

45min 9 C

1hour 7 C

1.25h 6 C

1.50h 6 C

1.75h 5 C

2.00h 5 C

2.25h 4 C

The temps are stable at 2 1/4 hours.

I would say that is a long time to wait - obviously the battery and metal internals are a heat reservoir and the plastic body is a nice thick jumper. (At this point I should add I'm using a 1100D, but this should apply to any plastic body DSLR... in fact the heavier ones should cool down even slower!).

But that is not the whole story - you don't want to cool to ambient temperature, you only want to cool to the temperature the camera will be stable at when it is taking pictures - if it stabilizes at 9 degrees in the imaging run, you don't want half your shots at a different temperature - that will just mess with your darks.

So how fast does the camera heat up? To answer that I shot two minute subs in the fridge with 2 minutes to 'cool down'. Cool down is in quotes because the camera is so bad at losing heat the actual heat loss in the idle periods isn't much. I didn't test it but I think it doesn't matter if the subs were 1 minute with one minute cool down or four minutes with four minute cool downs as the heat loss is so slow and the heat reservoir is so deep the sub length does not have an effect (obviously longer cool downs relative to sub length would have an effect). Here are the temperatures of my simulated imaging run.

100-7014 119 400 4 C

100-7015 119 400 5 C

100-7016 119 400 6 C

100-7017 119 400 7 C

100-7018 119 400 8 C

100-7019 119 400 8 C

100-7020 119 400 9 C

100-7021 119 400 9 C

100-7022 119 400 9 C

100-7023 119 400 9 C

100-7024 119 400 9 C

100-7025 119 400 9 C

100-7026 119 400 9 C

100-7027 119 400 9 C

100-7028 119 400 9 C

100-7029 119 400 9 C

100-7030 119 400 9 C

100-7031 119 400 10 C

100-7032 119 400 10 C

100-7033 119 400 10 C

100-7034 119 400 10 C

100-7035 119 400 10 C

100-7036 119 400 10 C

100-7037 119 400 10 C

100-7038 119 400 10 C

100-7039 119 400 10 C

100-7040 119 400 10 C

100-7041 119 400 10 C

100-7042 119 400 10 C

100-7043 119 400 10 C

So the temperatures stabilized at 10C. For the impatient astronomer this is good news - just over half an hour of cool down got me to 10C, so if I start snapping after that point my camera temperature is stable for the whole imaging run!

THat's all well in theory, but in practice I just want all the data I can get and the thought of 2 minute gaps of getting nothing is appalling. So what temperature does the sensor reach if I just let it shoot continuously? Here is the same run without gaps:

100-7044 119 400 4 C

100-7045 119 400 7 C

100-7046 119 400 8 C

100-7047 119 400 8 C

100-7048 119 400 9 C

100-7049 119 400 10 C

100-7050 119 400 10 C

100-7051 119 400 11 C

100-7052 119 400 11 C

100-7053 119 400 12 C

100-7054 119 400 12 C

100-7055 119 400 12 C

100-7056 119 400 13 C

100-7057 119 400 13 C

100-7058 119 400 13 C

100-7059 119 400 13 C

100-7060 119 400 13 C

100-7061 119 400 14 C

100-7062 119 400 14 C

100-7063 119 400 14 C

100-7064 119 400 14 C

100-7065 119 400 14 C

100-7066 119 400 15 C

100-7067 119 400 14 C

100-7068 119 400 14 C

100-7069 119 400 14 C

100-7070 119 400 15 C

100-7071 119 400 15 C

100-7072 119 400 15 C

100-7073 119 400 15 C

I'm not sure but it looks like the termperature is going up and down between 14 and 15C. So if you just shoot continuously, you can have a stable sensor temperature by starting shooting immediately with no cool down.

This is all assuming you keep your camera in a cold room, like I obviously do :-) If your house is roasting away with central heating to the max, you might want to add another half hour to the cool-down times.

[GlassWalker]

Interesting test. I've always wondered if cooling between subs would give lower noise enough to offset having fewer subs. So here the suggestion is, you can have 5C higher temps for double the exposure time. From memory, I think Gina's earlier testing suggested a factor of 2 change in noise level for a 7C change. If that's correct then running continuously would give a net benefit. It may offer side benefits too, such as having a bigger pool of subs to choose from in case any are spoilt for any reason. The cost may be increased processing times afterwards, but with the weather we get I don't think that's too limiting

[Ags]

Yes, now I can run my noise analysis tool on these subs and make a statement about noise levels between the two strategies - 1x cool subs or 2x hot subs.

[allcart]

Ags, can you run a temp test starting at room temp or maybe 15 C. We don't all keep cams in cold storage. Mine are kept at room temp with negligable cooldown before shooting. It may be interesting to see how high the camera temp will go when started at a higher ambient temp.

[Ags]

Regardless of how hot your house is, if the outside temp is 4C, the camera should stabilise at the temps reached in my tests. That is, if it starts at 23C and shoots continuously in an ambient temperature of 4C, it will cool to 15C, eventually :-)

[Shibby]

Hmm yes very interesting, particularly looking at the difference the 2m pause made...

I've just looked at my subs from Monday night and find that my camera cooled down from room temperature to 6C within about 45minutes, it then stabilised at this temperature (using 4m subs with 15s pauses). Early morning, as temperatures approached 0, exif was reading 4C (I had changed to 10m subs at this point).

[Ags]

Interesting - your camera is much cooler than suggested by my test run. Perhaps the difference lies in the poor air circulation in a fridge, compared to the gentle breeze of the great outdoors. But it is really interesting that your camera was basically shooting continuously and was running at 6C, whereas mine was running at 14-15C. You don't mention the ambient temperature earlier in the evening, but it seems you were at ambient+4 in comparison to ambient+11 in my run.

What camera do you have and does it perhaps have a part-metal body? EDIT: Duh, its a 450D. I read the body is part-stainless steel - that could make a big difference.

[Gina]

Interesting - your camera is much cooler than suggested by my test run. Perhaps the difference lies in the poor air circulation in a fridge, compared to the gentle breeze of the great outdoors. But it is really interesting that your camera was basically shooting continuously and was running at 6C, whereas mine was running at 14-15C. You don't mention the ambient temperature earlier in the evening, but it seems you were at ambient+4 in comparison to ambient+11 in my run.

What camera do you have and does it perhaps have a part-metal body? EDIT: Duh, its a 450D. I read the body is part-stainless steel - that could make a big difference.

I've found the same thing as you Agnes - if fact mine's worse at 14 or 15C above ambient and my cooling to a couple of degrees above ambient produced the expected 4x change in noise level.

[Ags]

Looks like my next camera is going to have a steel/magnesium body! It seems to make a whopping difference.

When you say a 4x change in noise level, was this at any particular ISO?

[Gina]

Looks like my next camera is going to have a steel/magnesium body! It seems to make a whopping difference.

Conducting the heat away from the sensor to the external environment certainly makes a huge difference.

When you say a 4x change in noise level, was this at any particular ISO?

I did tests at 3200 and 6400.

[Ags]

I think those ISOs would have a strong response to temperature. How did you measure the noise?

[Gina]

I think those ISOs would have a strong response to temperature. How did you measure the noise?

I compared the histograms of the different temperatures at the appropriate exposures - the noise produced at the higher temperature was the same as that at the lower temperature with 4 times the exposure. This was covered in my cooling to near ambient thread in the "DIY Astronomer" forum.

[Ags]

I looked at the noise over my widest range, 4 - 15C. This table shows pixels at various levels of illumination (when they should obviously be black). The zero percent column includes a very large number of illuminated pixels that round off to zero. If you include those pixels in the noise count, then noise stays constant more or less across the temperature range.

Here are some graphs showing the same thing. Noise more or less doubles over the temperature range.

Also look at the gap between 10C and 15C - that is the difference between the cooling gaps approach and continuous shooting approach.

[Gina]

This seems very strange - I don't understand it. I've found the noise decreases significantly if I reduce the temperature. But this it the noise level rather than the number of pixels above a threshold. Maybe we're measuring different things.

[Ags]

I think you were looking at very different ISOs and a wider temperature range. I think we need the full data from the set point cooling to reach a better conclusion.

[Gina]

I think you were looking at very different ISOs and a wider temperature range. I think we need the full data from the set point cooling to reach a better conclusion.

Yes, that's true. Ah well, data is accumulating unattended as we "speak"

[Shibby]

Perhaps the difference lies in the poor air circulation in a fridge, compared to the gentle breeze of the great outdoors.

Yeah I bet this does make a difference. It was pretty cold out there when I started the imaging run at ~11pm, probably 2-3C but I can't confirm that for sure... I hadn't realised the body was part steel, it all seems plastic!

[Ags]

Probably plastic with a tiny bit of steel and a lot of marketing hype :-)

But based on Gina similar results for another 1100D, it seems that heat dissipation is a big variable between cameras.

[Gina]

The tripod bush is thermally connected to the main frame in the camera but not to the imaging assembly frame or to the sensor. So I think that with a lot of the metal connected to the tripod bus then if this is cooled by some metal the overall cooling will be better. When the camera is mounter to the scope focuser with the bayonet lens mount, there is less metal to condeuct the heat away. There is some but the main camera part the metal bayonet is screwed to is plastic and doesn't conduct heat. I'll see if I can take some photos of the camera parts and post them.

[Ags]

I was wondering about whether my photo dovetail might act as a heat sink. Back to the fridge...

Another obvious difference is in a real imaging run the shutter is open and cold air can circulate directly into the camera (assuming it is attached to a scope and filters aren't getting in the way. Even so, the metal t adaptor must let out a lot more heat than my lens with lens cap on.