Wayne from No Life Digital has posted a video to his YouTube channel showing how he got the Canon EOS R5 to record 4 hours of 4KHQ 30p video to an Atomos Ninja V before the camera overheated.
Simply by removing the memory cards from the camera while recording externally lead to a big boost in recording time. This seems like a pretty surprising outcome for something so simple. Is there some kind of software limitation in the camera?
A dummy battery was also used in this test.
Testing methodology:
The Canon R5 does not seem to overheat, externally, if you don't have any cards in the camera. In this video I show (with as much proof as I can), how I'm able to record up to 4 hours in 4KHQ 30fps, 10-bit CLOG, with IBIS and DPAF all on, with no issues like the blackout or screen off issue myself and Gerald were having.
The setup here is as follows – HDMI out of Canon R5, into Atomos, out of Atomos, into ATEM, out of ATEM, into OBS on the PC. This is no different than Geralds live streaming setup, and showcasing the times you can expect as if you were recording on the Atomos. This was to record for this long period of time, without the MASSIVE Atomos file, and still show the camera was in fact running for the entire time.
Although the skipped line 4k is good enough for a lot of work, this is even better. So basically that means all 4k modes except 120p are basically available all the time.
To bad the R6 did not score that great times with the external recorder.
Since R5 is basically the best all-around photo camera today and with this setup I would say it is maybe even the best hybrid camera today it will be a bit large set up for photo/video jobs if you do both at the same time. But If I get two I'm covered.
And if they unlock 120 fps to external recording this is the perfect camera for my needs :D
Thank you Wayne for this info.
Canon R5 Redux: Deep Testing Overheating, Resolution, and Auto Focus.
This is huge news and hopefully we will see more workarounds like this!!
That also indicates that the downsampling doesn’t push the heat up too much, but presumably the combination of h265 and I/o causes it.
Internal writing is compressed to lower the data rate to something cfexpress can cope with. That’s h265 and the I/o which is pushing it over the threshold.
Bottom line, it seems that this might help anyone professional for some modes, but since atomos doesn’t support the 8k modes or 120fps, then not a solution there, and if Canon doesn’t have an hdmi 2.1 then adding that into the body won’t be a quick fix.
Seems if they can lower the h265 overhead and need less I/o we might get some internal recording possible in the higher modes, but how much quality that will impact it is not clear until Canon responds. Plus with a bit of post editing (sharpening) maybe it won’t be noticeable.
There was talk about a new firmware supporting different modes is coming, so maybe this will ease things.
Certainly seems there isn’t one “culprit” to the overheating, rather it is the combination of components based on what you are doing which is causing the system to finally go over the threshold. Maybe Canon can also tweak that threshold but I think they still need to figure out how they can vent the hot air out, maintain the weather sealing (when not venting) and then the recovery time would also be solved.... if it was repeatable venting of the heat in less than 20 mins that also might work for some.
Good news no matter, that others are finding some solutions.
If so, this is not a CF express problem, and it might point towards:
- Perhaps limits are all battery related and we don't know why yet. That should immediately be crossed off the list if running on external power + card in place shows temp warnings (which I presume we would).
- Canon is super conservative with it's firmware limits... to I don't know, limit risk of bricking something? And then somehow they failed to implement limits for this specific use case of no card + external recorder?
- Canon didn't account for cards heating up in general with their mechanical design, found this issue late in the game and just slapped time limits on recording if a card is in there. (I really doubt this.)
- Canon was nerfing some aspect of record to protect another product line... Nah. Seems far fetched for such a mottled spread of time limits.
What else would explain this? Just riffing here.- A
32 minutes here, 47 minutes there, etc. seems like a hyper real data-informed set of limits. Wouldn't a nerf just lock you out right at 15 or 30 minutes on the dot?
The two hours of cooldown seems less of a nerf than a very conservative factor of safety assumption by Canon. That also might imply time limits and not active temp monitoring might be at play, but that seems nuts technically. If a gaming PC far cheaper than an R5 can poll its CPU temp, Canon would be doing the same with it's processor, right?
- A
But do we know the cards are actually getting hot? Hotter than the body adjacent to the Digic?
- A
https://alikgriffin.com/canon-r5-overheating-with-cfexpress-card-are-the-cards-to-blame/
(Note: I don't know how accurate this is. I just saw the link on Fred Miranda forums.)
Plus as mentioned, maybe the hot air can come out through the card slots.
Finally, no h265 encoding over HDMI and that is significant also.
I think it’s a combination of factors, and people are finding ways to tweak the thermal envelope if you will.