snapsy wrote: Gary Irwin wrote: snapsy wrote: sturgis wrote: snapsy wrote:
Jay and I collaborated before on an R5 test (typical video run-time tests) and I pinged him today with the idea of defeating the battery door. He was game and tried it today and by initial measures it appears to work.
He documented the full experiment here:
Notes:
In the video he demonstrates how the camera flashes the overheat warning and is close to shutting down. He pulls the battery with the door-sensor taped down and when the camera comes back it shows 5:00 available. He doesn't demonstrate it in the video but he did that three times in a row back-to-back, each time recording about 6-7 minutes and then pulling the battery before the camera would shtudown - without the hack he only gets 1:00 after restarting the camera (he demonstrates this later in the video)...and it would take an hour to get 5:00 of available indicated time back
Also not in the video but he then let the camera cool for an hour and without the hack it only showed 5:00 but after recording for close to 5:00 and doing an orderly restart it came up with the warning immediately and only showed 1:00 and that's about what it lets him record, repeating what he's seen previously in actual use. So the hacked behavior is definitely giving him a lot more time.
I'm guessing the camera is showing 5:00 instead of the full cooled-down time due to the actual temperature of the camera triggering the throttle in real-time vs the usual thermal management algorithm heuristics limiting the time prematurely. He's planning some more experiments, including doing more than three back-to-back 5:00 records w/hack to see if the camera lets him do that repeatedly.
Another possibility for the 5:00 is that the camera may be periodically updating NVRAM even during the session with the thermal stats...so pulling the battery at the end of the video recording session may only be inhibiting the most recent update, and since that update would be at 5:00 -> 0:00 available that's why the camera returns to 5:00 after restarting.
Interesting thanks. Another option is a $30 dummy battery.
Is the hacked behaviour indicating that we are tricking the algorithm which manages the heat but at the risk of damaging something inside? Maybe there aren’t as many internal temp sensors as we might hope, and as you suggest the orderly shutdown writes timings to nvram whereas the trick prevents this. Which may indicate as others have suggested that Canon controls it more by firmware and timings and less by temp sensors, and thus tricking may impact the longevity of some components inside the camera?
Yep, a dummy battery should work as well, for the same reasons.
I don't think it's a lack of thermal sensors but do believe it's Canon trying to protect the electronics and media cards, and not just from peak high temperatures but also total heat x total time. Roger C. has commented in the past how Lens Rentals would get the most issues on DSLR bodies that were used for video and many of those issues were PCB-releated. This is why I asked Jay to put the warning at the start of the video.
More likely Canon is trying to protect their cine market. Anyway, as a compromise I expect Canon will release a firmware update that will extend recording times somewhat, but there's no way Canon's going to take the video choke chain completely off the R5 IMO.
If it was Canon trying to protect their cine market they wouldn't have given the R5 nearly-unlimited recording time over HDMI. Most professional videographers use external HDMI monitors+recorders.
Perhaps, but I'm thinking the R5 is such a ground breaking camera for Canon maybe they were a little unsure as to how to best "fit it" within the broad imaging market. From what I can see Canon, more than some other manufacturers, has a tendency to "engineer" cameras to fit a specific market slot (i.e. by making use of the infamous "cripple hammer"). By initially restricting video performance through firmware gives them time to figure things out. It also leaves Canon plenty of room to release future firmware updates with "dramatic" video recording improvements thus extending the market life of the R5. Regardless, it seems there's a lot more going on here than just thermal issues.
Aug 22, 2020 at 08:29 AM
Previous versions of Gary Irwin's message #15324413 « R5 artificial time limit »
