Canon EOS R5: a new teardown video calls the nature of its recording-time limitations into question

[Redaktion]

The EOS R5 is a recently-released, near-US$4000 Canon full-frame mirrorless camera that can shoot 8K RAW - but for only about 20 minutes straight. This, according to the OEM, is due to the need to protect its processor from overheating. However, a teardown documented on Baidu has found that this same silicon may have been underserved in the way of thermal protection.

https://www.notebookcheck.net/Canon-EOS-R5-a-new-teardown-video-calls-the-nature-of-its-recording-time-limitations-into-question.486505.0.html

[rhyme]

I suspect that the operating time is being limited to prevent thermal noise.  Thermal paste and copper will just conduct heat better and make it happen faster. Those are to be used with active cooling to help get rid of heat faster.  I think it can be mitigated in firmware, but I do not think it will be as easy as people think.  It is kind of strange to me that people who shoot photos and videos know more about computers than cameras these days.

[S.Yu]

Throw it in the freezer and see if there's a difference I suppose.

[Gerald Undumb]

So you are saying Sony crippled that camera even more because the Canon can do 8K Raw for 20 minutes and the A7SIII can do it for 0 minutes?

"It can record in up to 8K RAW at 30fps - but is only rated to do so for 20 minutes at a time, whereas its rival the Sony a7s III is not. "

LOL, exactly. The Sony A7S III is limted to ZERO MINUTES at 8K and has NO internal Raw.... and it's a crippled 12 MP sensor.

The Sony equivalent to the R5 is the A7RIV (or the A7RIII) and those are both limited to basic 4K 30 maximum... and still can overheat in some cases.
They can't even do 4K 60 or 120 let alone 8K.... Is Sony purposely crippling their hybrids with super low resolution photos, or only 4K 30?

And why is Sony not putting in cinema colours in any of their mirrorless cameras when Canon is putting in the same colour in their cinema line all the way down to their rebel line? Hmmm? Even the A7SIII fails miserably, as it still has green skin problems (Parker Wallbeck confirmed this).

The time limits exist for a reason, and it's not because of "crippling". Moderate temps can affect sensor quality, and the R5 is a photo camera, and they've worked hard to give it the best IQ of any full frame camera to date (currently has the best dynamic range of any full frame colour according to Photons to Photos, as just one example).

The people writing this conspiracy stuff need to go back to hiding from vaccines and flat earth santa, and continue to hoard their toilet paper.

[_MT_]

You really shouldn't comment on things that you don't understand. I don't know what they are trying to do with those two pads. I doubt they're there to provide cooling for the processor. It's passively cooled. All you are trying to do is to reduce thermal resistance between package and surface structure (you're trying to provide an efficient path for the heat). And PCBs actually play major role in spreading heat. They're full of copper and often used as heat sinks. Solder is also a good thermal conductor. It's entirely feasible to pull heat through a PCB and drain it away from the other side. It would have been even better (thermally) to have the processor mounted on the other side (assuming a suitable package). But that pad there is probably thermal. It's in the right place and connects to the metal chassis which is exactly what you want. The metal chassis is going to act as a heat sink. A pad directly on the processor would make no sense if there is no good sink to connect to. In this case, it appears there is a second board covering the processor area (plus a shield) and then rear panel assembly. Which is not going to be better than going through the board the chip is soldered to.

Not to mention that to evaluate this, you would have to measure temperature internally. It's been a very long time since I have seen any physics, but the surface temperature shouldn't really change as the difference between surface and ambient dictates how much heat is being dissipated. And that should remain the same. What changes is the difference between package temperature and surface temperature. In other words, how hot the chip needs to get to achieve the necessary surface temperature to move the heat it's producing (= achieve equilibrium). Of course, real life is always more complicated. Surface temperature won't be uniform. It might get cooler in one spot and hotter in another.

Thermal design is not easy and even plenty of graduated electronics engineers (which I'm not one of) struggle with it (I would go as far as saying most). Like RF, it's one of the things considered "dark arts" which require "wizards".

Whether an improvement in cooling would make any difference depends on how is the protection implemented. Some solutions are dumber than others yet still effective. They usually don't design to allow easy hacking (if it can be easily hacked, then it's typically a happy coincidence, not intention - they aim to make their life easy, not hackers'). It's also entirely possible that the marketing department is miscommunicating the reason. It's not uncommon. Communication between engineers and people with little to no technical knowledge is challenging.