Is there a way to cool the camera without buying a new hardware? How do you determine that the camera is NOT cool ATM? What temperature can be considered hot for the camera sensor? Does cooling really reduce the noise in the pictures?
In general, the camera's manual will tell you the operating temperatures for the camera.
Like all pieces of electronics, excessive heat isn't good for it. That said, I've only seen anecdotal evidence of heat adversely affecting image quality. It seems more common for batteries to stop working.
As for noise, you're probably going to have this issue in broad daylight, maybe in a desert. Your problem will be to get less light to the sensor, not worry about low light levels where noise is usually a problem.
I'd get a decent camera bag, in a lighter colour, and keep my gear in the shade whenever possible. Avoid using icepacks or similar or you will probably get problems with condensation.
Cooling is used in astrophotography, but those sensors are mostly CCD's, and the exposure times and low light levels are an order of magnitude longer and lower, respectively. I don't know whether CMOS sensors have the exact same issues as CCD's with heat when used for long exposures.
Assume you are taking video or very many exposures very quickly and the camera is magnesium and you don't want to deal with taking dark frames for astronomy or wait between shots much.
Heat does add noise, but slightly more at high ISO.(which essentially is gain via on-sensor amplification)
If the camera is of magnesium alloy 50-100 W/(m·K) construction the thermal conductivity is higher than plastic, almost that of Aluminum's <150 W/(m·K)- which seems at first to be a boon but in reality is not because the surface is painted, and also smooth (low surface to air area) which doesn't work well with small portable fans ~8cm.
It generally will absorb heat readily from your hand and (if it is hotter) the environment but not dissipate it nearly as well even with a fan because I noted in another question though that the shape of a camera doesn't lend itself to much exposed area. The lcd, handgrip stickers, buttons, knobs ect cover almost all of the camera and the only exposed area are corners and edges.
If you are shooting in an even/cooler temperature and are trying to combat sensor emission then you are out of luck. The sensor is mounted on a pcb whose average conductivity is only 50 W/(m·K), which is then mounted on steel brackets to the magnesium housing.
All in all the sensor is not in a really happy place but the only variable I am not sure of is the sensor's temp ratings on spec sheet.
Cooling the housing with a fan doesn't mean the sensor will get any of it very effectively. The contact point between the sensor and pcb is large, so say 50 W/(m·K) effective conductivity, then it goes down to 17 for the steel mount to the housing, and from the photos of the cutout I'm not seeing a whole lot of surface area in contact between the pcb to steel, and steel to magnesium. What you gain though, is that you are no longer holding the camera with your hot hand if theres a fan there.
A fan of 14cm or more mounted on a separate cheap tripod will do the trick if you don't move anything around, but for general photography it's not going to work out too well.
This post at a telescope forum shows a 5d2 with a temp probe on the back of the sensor These folks are serious about their cooling.
The graphs show temp rises with 5 minute exposures and some cooling time between shots. Kind of simulates shooting HD video.
This is in a fairly cool chamber so if you extrapolate rise above ambiance temps of 14C to a real world warm environment it could go to 45C or something which doesn't look too promising. Again unless you are shooting HD like there's no tomorrow in 30C room at max camera ISO without a fan then the graph shows that the sensor does cool off quite a bit from a long exposure in less than a minute.
If you have an old unused/broken computer you can recycle a fan. With a bit of DIY you can attach it to the back of the camera (more or less in front of the LCD screen). Two things to note:
Since it is not unusual to use a computer in astrophotography (where cooling is often used) these can be more or less minor issues, but obviously it depends on what you are doing.
As an avid astrophotographer, I was happy to find that my cameras already record what they believe is the CCD temp in the EXIF of each picture. You can use software like PhotoMe or EXIFtool to view/extract that data. This is invaluable for keeping a library of dark files for temperature matching with dark frame subtraction.
For my camera (Pentax K10D) I see the CCD getting to 23-27C dependent on ambient temp, ISO setting (higher = more heat) and the exposure length. Although, if the ambient temperature is low and the ISO setting is less than 400, it will stabilize at about 23C and not get hotter.
To actively cool a camera, I've done some pretty basic things, including keeping the LCD off at all times and even strapping an icepack onto the body.
Here are two links for some projects that people have done to make cooler boxes with peltier devices: http://ghonis2.ho8.com/rebelmod450d16c.html and http://www.cloudynights.com/ubbthreads/showflat.php/Cat/1,2,3,4,5,8,9,10/Number/3005776/Main/2993595