Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Please, support PV!
It allows to keep PV going, with more focus towards AI, but keeping be one of the few truly independent places.
Canon EOS R5 - Overheating video failure
  • 250 Replies sorted by
  • ha ha ha this is awesome

  • Level of journalism reached new low

    What's interesting is that the exterior of the cameras don't get especially hot when shooting for extended periods. We're only speculating, but this could indicate that Canon has tried to isolate the camera's internals from external temperature fluctuations, with the down-side that they can't then dissipate internally produced heat.

    https://www.dpreview.com/articles/8653213690/canon-r5-r6-overheat-claims-tested-stills-shooting-setup-quickly-cut-into-promised-capture-times

    What exact idiot wrote this?

    LSI shutdown temperature is around 95-105 degrees celsius, and they wanted to isolate this from 35 degrees (even 45-47 degrees if you leave it ont he direct sun)?

    Doing proper cooling require extremely good heat transfer to the outside air (as otherwise each and every camera will shut off like this R5 id doing).

    My source in big reseller told me that we now observe intentionally defect construction intended to remove part of top management (that is part of big management wars that also include some investors). Note that severe hack and encryption of computers happened rigth after this in most important branch - Canon USA.

  • Up to 30% of all comments in R5 related reviews can be written by Canon owned bot crowds. Some of it are advanced computer bots (many registered years ago and behaving like normal people). Some are hired people using bots accounts.

    Last year I talked with one ex marketing guy and he told funny example - under one article it had been no independent human comments at all and mroe than hundred of comments by advanced bots with one half belonging to one company and another to different one. Virtual fight. Only around 20% of comments had been made by human operators helping bots.

    Last year at IMDB it had been aroound 95% of fake paid for reviews (from one week before release to around 2 weeks into release) for each major Hollywood piece. Parts of text is now made by specially tuned neural networks.

  • Anyway, the Android app connected successfully and we began to experiment. A further 15 mins later and the record time in 4K HQ was down to 10 mins with not a frame recorded. Soon after that the camera overheating warning was triggered and the record time started dropping rapidly from the 5 minute mark. 4,3,2,1 then zero. About 1-2 minutes after the record time went down to zero, the camera shut off automatically. Not a single minute of video was recorded and not a single still frame shot.

    https://www.eoshd.com/news/i-bought-a-canon-eos-r5-and-it-overheated-in-the-wifi-menu-also-a-look-at-potential-solutions/

    My souce in retail told me that Canon talks about main LSI defects (as I referenced before) where parts of the chip are not adaptively turning off, and instead ot os consuming lot of power and generating big amount of heat.

    It is probable that they could find some software workaround, but info leaked tell that it is severe chip design defect. And hence fix can come only with new chip release, or even never as Canon does not have free funds now due to horrible sales numbers. Heatspreader and thermal paids are also horrible, btw.

    In video most strange is how thermal pads are applied. It requires 3 thermal pads of different thinkness - two for RAM and one for LSI, it is obvious.

    https://www.bilibili.com/video/BV1Uv411q7CA

    image

    sa14046.jpg
    800 x 474 - 61K
  • seems that the display has a bigger role in overheating than anything else

  • Results of cuts on engineers

    image

    PCB board sits on top of LSI. Never in cameras history we had such bad thermal design.

    image

    sa14054.jpg
    579 x 423 - 57K
    sa14055.jpg
    650 x 529 - 43K
  • After switching on from a cold start and being left for a couple of minutes, at 46C, I recorded 8K for 15 mins and the camera temperature only increased to 62C which is very comfortable for pretty much any kind of electronics.

    An internal temperature of 46C isn’t too hot for a CPU or memory card, or anything else, to operate normally.

    62C isn’t too hot to carry on recording in 8K. Had the camera reached 95C, I could have accepted what Canon is telling us.

    First, 95 C can be top temperature only if you have sensor directly onthe LSI die.

    Second, reported temperature can be anything - including sensor located on sensor or in some place on board.

    Third, modern computers and devices frequently have up to 15 temperature sensors.

    This literally means that if your isolated sensor cooled down and is showling 30 degrees, it does not matter that chip itself and area around it cooled already.

  • Canon PR sent to some sites info about future firmware

    • Addition of Cinema RAW light option to 8K recording
    • Addition of CLog 3
    • Increased record time limits (how much is the question)
    • Various bug fixes

    Can be small bluff to divert attention.

  • Andrew now thinks that limits are software

    • After 5 JPEGs (5 minutes): 34 °C
    • After 10 JPEGs (10 minutes): 34 °C
    • After 15 JPEGs (15 minutes): 34 °C
    • After 20 JPEGs (20 minutes): 34 °C
    • After 25 JPEGs (25 minutes): 34 °C
    • After 30 JPEGs (30 minutes): 34 °C
    • (AT THIS POINT 8K IS CAPPED AT 5 MINUTES)
    • After 35 JPEGs (35 minutes): 34 °C
    • After 40 JPEGs (40 minutes): 34 °C
    • After 45 JPEGs (45 minutes): 34 °C
    • After 50 JPEGs (50 minutes): 34 °C
    • After 55 JPEGs (55 minutes): 34 °C
    • After 60 JPEGs (60 minutes): 33 °C
    • (AT THIS POINT 8K IS DISABLED – OVERHEATED! SHUTTING DOWN)

    https://www.eoshd.com/news/canon-eos-r5-overheated-in-my-fridge-after-just-60-jpegs-4-c-ambient/

    Issue actually can be that temperature written in EXIS is using one thermal sensor located on motherboard somewhere and CPU temperature is not related to it.

  • Even if the temp sensor is on the PCB, you'd think it would be influenced by a nearby CPU running 80+

    Instead it didn't increase from 34C and actually went down 1 degrees.

    Also the fridge achieved a 12C cooling effect on the internal PCB compared to running the camera at room temp.

    But still the camera claims a thermal death is nigh after 0 mins of video and 60 JPEGs

    Yet the Tilta fan claims to achieve longer record times also with external cooling - how?

  • Even if the temp sensor is on the PCB, you'd think it would be influenced by a nearby CPU running 80+

    Not necessary, can tell you as guy who repaired notebooks and motherboards (and build around 300 of them).

    If it is on other side and other end of the board it can be almost zero influence, especially in the fringe.

    But still the camera claims a thermal death is nigh after 0 mins of video and 60 JPEGs

    You need to have measurements of multiple on die sensors (for modern LSI it is usually such).

    Issue can be even with some area of LSI.

    You already saw horrible cooling design, worse I ever saw.

  • Yes the internal layout is a mess. Is it purposeful, or just incompetence? Or both?!

  • @AndrewReid_EOSHD

    I highly doubt idea of crippling things here. As Canon badly needed any sales and cinema cameras are too niche things that now sell extremely bad due to most serious production still stopped and almost no big weddings and such.

    Issue with heat happened as almost all ambassadors know about coming huge cuts in PR, marketing and... ambassadors department, so most of them wanted to lick the ass as deep as they can. Considering that business of most of them is destroyed now. So, multiple cameras and nice lenses can be actually ALL, 100% income (besides goverment help) they got in few months time.

    As for design - we can have design made by new, young engineers, it is exactly change of generations coming at all main Japanese manufacturers, can see it on all interviews during last 2-3 years.

    If you sit in facebook or 4 hours a day and spent 7-8 hours a day playing games during shcool and university - it is not much of surprise. I know lot of student personally , it is much fewer people who love their work and not emulate it.

  • Canon did a silent recall.

  • @hoodlum

    I have big issue with this guy. As in Olympus case he (as around 50 other video bloggers) retold PV post, here without referencing to real source.

    Looking at other videos I can 99.99% tell that it is total clickbait bullshit.

  • It makes 100% sense and all retailers I spoke to including some of the biggest have said batches were not arriving at all, or had been much smaller than expected.

    It makes sense for Canon to try and do it silently without whipping people up into even more of a frenzy by admitting there is a major problem with the LSI or hardware design, which they'd have to do if they announced a recall.

    They will try and fix it silently and delay shipments for a few weeks, blaming corona.

    Or they may do nothing.

    I don't see where else that PCB on top of the CPU is going to be positioned... unless they insert a folded copper heat sink around it to transport heat off the CPU and onto the back casing.

    It's still a major re-tooling of the internal PCB even just to attach it and to reconfigure the assembly robots.

    It might be best at this point if Canon just removed the 8K and 4K HQ modes altogether, or capped each at 2 mins maximum with shorter recovery times between clips.

    I still suspect firmware limitations and intentional segmentation, because Canon has 10 years of form on this. Why change now.

    Almost all tests show there is more to the overheating behaviour than just the hardware.

  • @AndrewReid_EOSHD

    You are jumping to conclusion as it proves some of your points, but channel of the guy really proves opposite. His words about Olympus are clear 100% lies, same as his words about Canon source. He is jumping to hot topics to get views and income. It is lot of such people now. Many don't have any other income, but his channel in last months. So, they will do such stunts.

    I still suspect firmware limitations and intentional segmentation, because Canon has 10 years of form on this. Why change now. Almost all tests show there is more to the overheating behaviour than just the hardware.

    For now you do not know exactly, as tests are not scientific.

    As far as my sources go - it is no any recall going on now, it is shortages - yes. I also never hear about Canon informing anyone about redesign.

    If you ask me - they just wait for wave to subside and it can work.

    They will happyly sell it to rich people who use camera without big load and such.

  • Let's not forget our friend's in China and their findings - thermal paste on the CPU, measuring hot spots with an infrared thermometer... their conclusions about firmware are worth taking into account! No, we don't know for sure what goes on in the blackbox of Canon's firmware (or top management) - but it is clear even with the DIY modification, firmware timers exist and Canon had no willingness to mitigate this with proper thermal design and better layout.

    I have a Xiaomi Mi 10 Pro with vapour chamber. Shoots 8K in a much more confined chassis than the EOS R5. Meanwhile Sony's new smartphone overheats just doing 4K H.265 after 3 minutes.

    I do think Vitaliy you are right when you say the younger generation is coming through at the major companies, but that too many of them are chinless wonders. There by family, education, background, class but not talent.

  • I have a Xiaomi Mi 10 Pro with vapour chamber. Shoots 8K in a much more confined chassis than the EOS R5.

    First, Snapdragon in your phone is 7nm chip using most advanced tech available today (Samsung 7nm EUV). Compare it to 28-35nm of Canon LSI (28nm is tops that is available for camera companies due to development and masks cost), actually, nothing changed since Samsung NX days in this regard.

    Second, vapour chamber is directly attached to main LSI and is spreading heat to large area, properly designed smartphone is much more efficient in spreading heat (long term) compared to any large camera similar to Canon one.

    Third, Sony smartphones are going way of the dodo. :-) Amount of money Japanese government and investors poured into this department is staggering, even moved former Sony digital cameras boss into managing smartphones. Did not help much.

    Final part. R5 thermal design (and overall design) is rushed, made without good engineers and with lack of resources. Yet it had been done in amazingly good days for Canon (development started in 2018). It is very hard to think that will be result of 2020 cameras development.