DIY Water-Cooled Canon R5 Set-up Results in Unlimited 8K Recording

Imagine being able to film unlimited 8K video, or capturing images of deep-sky objects all night with almost zero noise. While there may be better ways to manufacture a camera without heat issues, there is a market for a camera that could be cooled beyond the average current camera housing. Water-cooled cameras are not anything new and for that matter creating a custom water cooling system for a DSLR has been done many times before too. Dedicated cameras in the astrophotography market often come water-cooled in order to reduce the heat that leads to increased noise during long exposures. To that end, several astrophotographers have set out to create their own version with a popular DSLR for astrophotography. As 8K has become more mainstream in smaller form factor cameras so have heat issues. Even when those issues don't cause the camera to shut down completely they do lead to other issues like higher noise and hotter devices. 

The projects over at DIY Perks are often experiments in the possibility rather than the best solution for a problem and this new video is no different. It's really interesting to see different cooling mods and the results in recording times that can be achieved. It is unlikely everyone is going to go out and build water-cooling setups for their cameras but the results speak for themselves. So maybe there is a market for a 3rd party company to produce a new camera back with built-in water cooling. Or at the very least the video shows a vast improvement from just improving the heat transfer plate that would be pretty simple for a company to produce and consumers to install. 

While I don't think the vast majority of consumers have a need for any changes to most existing cameras on the market there are niche uses like astrophotography that would benefit from cooling modifications and maybe this could lead to a new product market similar to the one built up around PCs. 

If Canon wants to send me a new Ra I'd be happy to water-cool it and do some testing to see what kind of improvements can be achieved. 

Michael DeStefano's picture

Michael DeStefano is a commercial/editorial photographer focusing on Outdoor Lifestyle and Adventure. Based in Boston, MA he combines his passion for outdoor sports like climbing and surfing into his work. When not traveling or outdoors he is often found geeking out over new tech gadgets.

Log in or register to post comments

I had to sign in just to give this guy a thumbs up on his video. Amazing work.

Yep amazing. He has a lot of really great videos and one particularly on converting old lcd tvs into daylight lights which could be great for studio work.

So many people still claiming the overheat issue is not faked.
"“This proves that the camera with the initial firmware actually did not measure the temperature to decide when to shut down. It was programmed to shut down after a certain time and implement a long recovery time.”"
Do you really need another proof?

Without going to the full water cooling system; a simple heatsink upgrade gives you almost twice recording time.
That's embarrassing for Canon, at least.

"Canon chose a timer solution because they'll have done tests to work out how long it takes to overheat and then add a safety margin."

Thats interesting. I always try to check the facts. Can you please give me the source of this that I can check it.

Or is this just a guess?

Did you watch the video?

Of course, but I thin Perks is not a Canon official also.

Of course, let's get someone from Canon marketing to chime in.


"But looking at Canon's decisions and what happens seem to suggest the bleeding' obvious. Sometimes the obvious is right in front of us if we don't try and see around it."

I am not a Canon shooter, but I observed the discussion from both sides and I see that for lot of people a "cripple hammer" is as obvious as your arguments. And for them it is also obvious that top engineers never will come up which such a bad heatsink solution like on the R5.

Another issue is that the camera, as it currently sized, probably would not be able to dissipate the heat in an efficient manner without the whole case also getting too hot to handle. There were lots of tradeoffs that Canon had to make. The timer implementation probably came at the late stage of development when they truly found out there was a heat issue and scrambled for a way not to damage the camera.

As far as the "cripple hammer" comments, those are wannabe engineers with no design experience in electronics or thermodynamics. EOSHD.COM said it was a total hoax, same as some really stupid YouTubers (Tony Northrup, Matt Granger, Andrew Reid...)

Camera is overheating in all HQ mode, even in 24p.
And, please, can you spot me any other 3C device that use a timer to deal overheating protection.
Can you?
There is none. Not one single device do so cause it's simply impossible to simulate heating in a realistic way.
Canon have thermal sensor but added these fake limitation.
Go check magic lantern if you like more details.

Fact is that Canon first released a camera with an extra counter to stop people from recording 4K HQ even inside a fridge. (But no limitation in 8K with external recorder, what a joke)

They changed later with firmware 1.1.1

Are you complaining about the actual camera overheating or the fact they implemented a mickey mouse procedure to try to save the camera? Overheating is an issue with the camera. That is a fact. They implemented a poor method to stop they camera from overheating which they corrected with some firmware updates. A better heatsink may not be the answer. You do realize that a more efficient thermal design means less delta from the hottest component to the case temperature? Would you be happy if the case temperature rises to 120 degrees? There were several tradeoffs that Canon had to do.

Funny how you keep talking about 8K
The camera oveheat in 24p HQ.
Yes, this is something I'm using quite a lot these day.
I really don't care about 8K at all.
This is used as bait for troll to exlain how good Canon is.


An no, not one single 3C product is using time as oveheat protection.

Even Canon now have kinda lift this infamous "recovery timer" of 40mn.
Yes, 40mn recovery time!!! Can you explain me?
My oven that is designed to keep heat inside is colder than a camera recording 4K24p after 40mn.

Canon marketed the camera with a big 8K logo. But people are stupid asking for this feature?

Taking picture at very low temperature only for the span of one hour gives you 0 minute of 8K video.
How is this fair?

Why don't you answer about the 40mn recovery time?
How this is possible? Except fabricated limitations with no relationship with actual temperatures?

On my YouTube channel, i recorded 8k video for 30 minutes until the camera stopped due to the video recording time limit. It then showed me i had another 30 minutes of 8k record time in the r5. Not kidding! My channel is . I published it a couple of days ago. No hack involved!

Would't it be nice if a startup company would produce a "new heatsink pack" for Canon R5?

don think so...

"there are niche uses like astrophotography that would benefit from cooling modifications and maybe this could lead to a new product market similar to the one built up around PCs. "
There is camera like this already. And it's different to this mod.
Astronomy camera need the cooling at the SENSOR to lower the electronic noise of device (which corresponding to temperatures). There is company moves the sensor and mount out of camera to put them in a cooling chamber (Peltier).
With the EOS R, there is a space inside the body, under the sensor (according to Lensrental teardown) so it's possible for a thin heatpipe over there.