R5 recording limits temp/sensor correlations (continued)

[lightnchade]

I just threw the R5 with version 1.1 in my fridge. It just recorded 45 minutes of 8K without an overheat warning. At room temperature, I got 24:53 before full shutdown. Still plenty of experimenting to do, but this is quite interesting.

Perhaps add one of these to the bottom of your camera bag for your next shoot:

Maybe someone industrious will make one that fits as a base grip.

 

[JakeJY]

So much for the conspiracy theories. You don't want to cook your electronics. Its how you do accelerated life testing...

Yes, when I looked at this chart late last night, and no subsequent posters had yet replied, I thought it quite odd that the chart didn't generate an immediate response that a supposed "workaround" like this would likely fry the R5's electronics in a fairly quick manner. These internal temperatures are well above ones that would ordinarily prompt any manufacturer to set limits on operation, whether through heat sensors or timers. You may argue that the design should have allowed better cooling and faster temperature cool downs, but that would have required any or all of a lot of less than ideal changes: more space for heat sinks and fans in a bigger body, possible noise and vibration problems and poorer weather sealing, or limitations on bit depth, and ommision of specifications which are uniquely high in the case of the R5. But the main take-away here should be: given the reality of what the R5 actually is, rather than what some people had wished it to be, that Canon is trying to protect buyers, not trick them.

If the temperatures reached were dangerous enough to fry the R5's electronics, then no workaround should have been possible. The camera should shut down based on the temperature sensors. However, by bypassing the timer based part of the algorithm (which relies on variables stored in non-volatile memory), it was possible to record for much longer.

Also prior to this update, the ambient temperature (even fridge temps) appeared to play zero role and the only thing that had an effect is using the freezer, which is not practical for most usage.

 

[Peak freak]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

<Record scratch>

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

Yes. From user reviews which happens all of the time. The reset time and a few other things needed to be looked at. Conspiracy theorists that blew this way out of proportion wanted a professional cinema camera and they will never get it out of the R5. For some reason no matter how many times people would say (and said months before) you will never be able to run 8K continuously with a small body with no ventilation it didn't matter.

I followed the build up to the release of the R5 pretty closely, partly because I simply couldn't figure out how Canon was going to do it. 'It' being release a camera with all those specs without some kind of 'secret' limitation or revelation.

You're correct in that early on, many people said it couldn't be done because of the heat generated, but after some surprisingly direct marketing by Canon, these concerns were pretty much allayed.

Can you point me to someone who, just prior to launch was still telling everyone that Canon couldn't do what they were suggesting because of heat management?

In other words, was there anyone out there who specifically predicted the time limitations with some accuracy? I think Canon had for the most part convinced 'us' that heat wouldn't be a problem. At least, not the problem it has been.

[FWIW I thought it would be stupid expensive, rather than have the limits. I just knew it couldn't be as good as some people hoped].

 

[Hoka Hey]

So much for the conspiracy theories. You don't want to cook your electronics. Its how you do accelerated life testing...

Yes, when I looked at this chart late last night, and no subsequent posters had yet replied, I thought it quite odd that the chart didn't generate an immediate response that a supposed "workaround" like this would likely fry the R5's electronics in a fairly quick manner. These internal temperatures are well above ones that would ordinarily prompt any manufacturer to set limits on operation, whether through heat sensors or timers. You may argue that the design should have allowed better cooling and faster temperature cool downs, but that would have required any or all of a lot of less than ideal changes: more space for heat sinks and fans in a bigger body, possible noise and vibration problems and poorer weather sealing, or limitations on bit depth, and ommision of specifications which are uniquely high in the case of the R5. But the main take-away here should be: given the reality of what the R5 actually is, rather than what some people had wished it to be, that Canon is trying to protect buyers, not trick them.

If the temperatures reached were dangerous enough to fry the R5's electronics, then no workaround should have been possible. The camera should shut down based on the temperature sensors. However, by bypassing the timer based part of the algorithm (which relies on variables stored in non-volatile memory), it was possible to record for much longer.

Also prior to this update, the ambient temperature (even fridge temps) appeared to play zero role and the only thing that had an effect is using the freezer, which is not practical for most usage.

Not true. My experiment of pre-chilling my R5 with FW 1.0 as reported by Horshack in part one of this thread resulted in 69 minutes of 8K done in 3 minute segments.

--
Joe

 

[gavin]

Yes. It does what it does not what you wish... I think it is fine for me but we are in a photo tech forum.

 

[Peak freak]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

<Record scratch>

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

I think the 'truth' is somewhere in the middle.

Canon has to mange heat and in turn the stability and longevity of the product.

But I am also convinced that this significant FW update, being so soon after release, is due to 'feedback' shall we say. Remember, Canon didn't just release the R5 without any testing or strategy - there had been months, even years of planning.

I can't believe for one second that Canon don't know exactly what the R5 can handle in terms of heat. The time extensions are not something they 'just figured out'!

I think people who continually say 'nothing to see here' should have to pay for the update.

 

[JakeJY]

So much for the conspiracy theories. You don't want to cook your electronics. Its how you do accelerated life testing...

Yes, when I looked at this chart late last night, and no subsequent posters had yet replied, I thought it quite odd that the chart didn't generate an immediate response that a supposed "workaround" like this would likely fry the R5's electronics in a fairly quick manner. These internal temperatures are well above ones that would ordinarily prompt any manufacturer to set limits on operation, whether through heat sensors or timers. You may argue that the design should have allowed better cooling and faster temperature cool downs, but that would have required any or all of a lot of less than ideal changes: more space for heat sinks and fans in a bigger body, possible noise and vibration problems and poorer weather sealing, or limitations on bit depth, and ommision of specifications which are uniquely high in the case of the R5. But the main take-away here should be: given the reality of what the R5 actually is, rather than what some people had wished it to be, that Canon is trying to protect buyers, not trick them.

If the temperatures reached were dangerous enough to fry the R5's electronics, then no workaround should have been possible. The camera should shut down based on the temperature sensors. However, by bypassing the timer based part of the algorithm (which relies on variables stored in non-volatile memory), it was possible to record for much longer.

Also prior to this update, the ambient temperature (even fridge temps) appeared to play zero role and the only thing that had an effect is using the freezer, which is not practical for most usage.

Not true. My experiment of pre-chilling my R5 with FW 1.0 as reported by Horshack in part one of this thread resulted in 69 minutes of 8K done in 3 minute segments.

Your pre-chilling was done with a freezer not a fridge however right? Or I would stand corrected if it was just at fridge temps.

Edit: quick search found freezer

https://www.dpreview.com/forums/post/64286798

My impression is that fridge temps were not enough to get the Canon algorithm to budge, only the freezer temps were, which is far less practical. If it was fridge temps, some colder places can be at those temps or there may be other ways to get those temps (like AC or a cooler) without needing access to a freezer.

 

[Jotoco]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

I think the 'truth' is somewhere in the middle.

Canon has to mange heat and in turn the stability and longevity of the product.

But I am also convinced that this significant FW update, being so soon after release, is due to 'feedback' shall we say. Remember, Canon didn't just release the R5 without any testing or strategy - there had been months, even years of planning.

I can't believe for one second that Canon don't know exactly what the R5 can handle in terms of heat. The time extensions are not something they 'just figured out'!

I think people who continually say 'nothing to see here' should have to pay for the update.

I believe Peter McKinnon got his R5 back in March if I'm not mistaken.

If influencers got theirs in March, before lockdown, canon must have had engineering samples from back in January or December 2019.

Edit: probably those engineering samples and even Peter 's weren't final hardware and software versions.

Maybe that is why Peter managed to get a lot of 4k120 video? Maybe that is why he didn't mention much overheating at first? Maybe even some of these early units died? A lot of conjecture, but might explain the over conservatism of Canon.

 

[David Hull]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

I think the 'truth' is somewhere in the middle.

Canon has to mange heat and in turn the stability and longevity of the product.

But I am also convinced that this significant FW update, being so soon after release, is due to 'feedback' shall we say. Remember, Canon didn't just release the R5 without any testing or strategy - there had been months, even years of planning.

I can't believe for one second that Canon don't know exactly what the R5 can handle in terms of heat. The time extensions are not something they 'just figured out'!

I think people who continually say 'nothing to see here' should have to pay for the update.

I believe Peter McKinnon got his R5 back in March if I'm not mistaken.

If influencers got theirs in March, before lockdown, canon must have had engineering samples from back in January or December 2019.

Edit: probably those engineering samples and even Peter 's weren't final hardware and software versions.

Maybe that is why Peter managed to get a lot of 4k120 video? Maybe that is why he didn't mention much overheating at first? Maybe even some of these early units died? A lot of conjecture, but might explain the over conservatism of Canon.

Some of these guys probably discovered the issues that Canon later fixed in FW.

 

[Hoka Hey]

So much for the conspiracy theories. You don't want to cook your electronics. Its how you do accelerated life testing...

Yes, when I looked at this chart late last night, and no subsequent posters had yet replied, I thought it quite odd that the chart didn't generate an immediate response that a supposed "workaround" like this would likely fry the R5's electronics in a fairly quick manner. These internal temperatures are well above ones that would ordinarily prompt any manufacturer to set limits on operation, whether through heat sensors or timers. You may argue that the design should have allowed better cooling and faster temperature cool downs, but that would have required any or all of a lot of less than ideal changes: more space for heat sinks and fans in a bigger body, possible noise and vibration problems and poorer weather sealing, or limitations on bit depth, and ommision of specifications which are uniquely high in the case of the R5. But the main take-away here should be: given the reality of what the R5 actually is, rather than what some people had wished it to be, that Canon is trying to protect buyers, not trick them.

If the temperatures reached were dangerous enough to fry the R5's electronics, then no workaround should have been possible. The camera should shut down based on the temperature sensors. However, by bypassing the timer based part of the algorithm (which relies on variables stored in non-volatile memory), it was possible to record for much longer.

Also prior to this update, the ambient temperature (even fridge temps) appeared to play zero role and the only thing that had an effect is using the freezer, which is not practical for most usage.

Not true. My experiment of pre-chilling my R5 with FW 1.0 as reported by Horshack in part one of this thread resulted in 69 minutes of 8K done in 3 minute segments.

Your pre-chilling was done with a freezer not a fridge however right? Or I would stand corrected if it was just at fridge temps.

Edit: quick search found freezer

https://www.dpreview.com/forums/post/64286798

My impression is that fridge temps were not enough to get the Canon algorithm to budge, only the freezer temps were, which is far less practical. If it was fridge temps, some colder places can be at those temps or there may be other ways to get those temps (like AC or a cooler) without needing access to a freezer.

This is not correct. My tests were done in the meat drawer of my refrigerator. Just above freezing. There was some confusion at the time which I corrected in a post in that thread.

--
Joe

 

[Noah M]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

I think the 'truth' is somewhere in the middle.

Canon has to mange heat and in turn the stability and longevity of the product.

But I am also convinced that this significant FW update, being so soon after release, is due to 'feedback' shall we say. Remember, Canon didn't just release the R5 without any testing or strategy - there had been months, even years of planning.

I can't believe for one second that Canon don't know exactly what the R5 can handle in terms of heat. The time extensions are not something they 'just figured out'!

I think people who continually say 'nothing to see here' should have to pay for the update.

I believe Peter McKinnon got his R5 back in March if I'm not mistaken.

If influencers got theirs in March, before lockdown, canon must have had engineering samples from back in January or December 2019.

Edit: probably those engineering samples and even Peter 's weren't final hardware and software versions.

Maybe that is why Peter managed to get a lot of 4k120 video? Maybe that is why he didn't mention much overheating at first? Maybe even some of these early units died? A lot of conjecture, but might explain the over conservatism of Canon.

Some of these guys probably discovered the issues that Canon later fixed in FW.

The EXIF data shows one or a combination of internal temperatures. It could be Canon saw they were getting too high and adjusted the logic last minute. Who knows.

 

[rrc1967]

On balance the new thermal details provided by Canon and the description of how that logic changed in today's firmware sounds a bit contrived to me. Thermals must have been one of the biggest challenges during the development and testing phases of the project. How could it be that they overlooked some elemental relationships between the three sensors and also not have sampled them frequently enough? It doesn't make much sense to me. IMO it sounds like misdirection to simply obscure the fact they may have simply loosened the reins a bit on the temperatures allowed, either peak or more likely total heat (temp x time). The fact the cooldown periods were most affected by the new firmware (shortened) suggest they're simply allowing the camera to operate warmer for longer. The fact the temps usually plateau near the end of the recording time allowed also supports the notion it's a total heat consideration rather than a peak or "overheating" consideration.

yes, but all of these could have been caught by Canon in their final pre-production testing but too late to affect the first wave of cameras out the door.

Faulty assumptions do occur in software development.

You do realize they released a camera like no other during a global pandemic that shut the world down. I can just imagine the pressure everyone was under. No one seems to be taking this under consideration.

I agree, which is why I think there could have just been problems with 1.0 and even 1.1.0 isn't what they want it to be. it takes time to release firmware.

 

[rrc1967]

So much for the conspiracy theories. You don't want to cook your electronics. Its how you do accelerated life testing...

Yes, when I looked at this chart late last night, and no subsequent posters had yet replied, I thought it quite odd that the chart didn't generate an immediate response that a supposed "workaround" like this would likely fry the R5's electronics in a fairly quick manner. These internal temperatures are well above ones that would ordinarily prompt any manufacturer to set limits on operation, whether through heat sensors or timers. You may argue that the design should have allowed better cooling and faster temperature cool downs, but that would have required any or all of a lot of less than ideal changes: more space for heat sinks and fans in a bigger body, possible noise and vibration problems and poorer weather sealing, or limitations on bit depth, and ommision of specifications which are uniquely high in the case of the R5. But the main take-away here should be: given the reality of what the R5 actually is, rather than what some people had wished it to be, that Canon is trying to protect buyers, not trick them.

If the temperatures reached were dangerous enough to fry the R5's electronics, then no workaround should have been possible. The camera should shut down based on the temperature sensors. However, by bypassing the timer based part of the algorithm (which relies on variables stored in non-volatile memory), it was possible to record for much longer.

it's not a question of frying the electronics. it's a case of many different elements.

a) cmos sensor can't get to warm.

b) outer shell cannot exceed 48C

c) operationally, long term internal temperature shouldn't exceed around 65C

d) when you quickly turn off video and go back to stills there has to be a "envelope" that still allows you to shoot stills.

Also prior to this update, the ambient temperature (even fridge temps) appeared to play zero role and the only thing that had an effect is using the freezer, which is not practical for most usage.

this isn't true at all.

 

[Peak freak]

USA firmware is up i just upgraded my camera! Based on the review he is able to record many more shorter clips of 4k 120 than before because the recovery time has improved due to better sensor reading.

I really enjoyed all the technical discussions here and the firmware proved there is really no conspiracy theory that all the trolls are alluding to. With what is in the r5 and the form factor unlimited 8k raw is impossible.

Had it not of been for all the upset and conspiracy theories in the first place would Canon of made any changes?

I think the 'truth' is somewhere in the middle.

Canon has to mange heat and in turn the stability and longevity of the product.

But I am also convinced that this significant FW update, being so soon after release, is due to 'feedback' shall we say. Remember, Canon didn't just release the R5 without any testing or strategy - there had been months, even years of planning.

I can't believe for one second that Canon don't know exactly what the R5 can handle in terms of heat. The time extensions are not something they 'just figured out'!

I think people who continually say 'nothing to see here' should have to pay for the update.

I believe Peter McKinnon got his R5 back in March if I'm not mistaken.

If influencers got theirs in March, before lockdown, canon must have had engineering samples from back in January or December 2019.

Edit: probably those engineering samples and even Peter 's weren't final hardware and software versions.

Maybe that is why Peter managed to get a lot of 4k120 video? Maybe that is why he didn't mention much overheating at first? Maybe even some of these early units died? A lot of conjecture, but might explain the over conservatism of Canon.

I find the PMK example interesting.

Peter is a great YouTuber who presents very watchable content. It makes sense for Canon to have him promoting their product, except...

I seem to remember Pete mumbling something about it getting hot after some 4K120, like he didn't fully understand the implications of what was happening - did he even know about the time limits (back then)?

You may remember, this was touted as the holy 'grail camera'. I can be pretty sure being locked out of 4K120 after a bit of use wasn't part of the plan.

I can sympathize with PMK. He was there to promote the camera, not delve into its engineering, but I would love to know if he always knew there would be limitations, or it came as a bit of a surprise to him as well.

[Matti Haapoja's take is pretty telling: He was all set for an R5, but is now looking closely at the A7S3. It is important to note that doesn't mean the R5 is a bad camera, just not as suitable for someone like Matti as the A7S3. However, the Canon hype machine sure sucked a few people in pre launch].

 

[mechanical_EYE]

Not sure what Canon witchery is going on now, but if you want to keep your battery pull timer reset abilities Hold off on updating firmware. It seemed to work when I tried a few times earlier, but after using the R5 throughout the day trying different modes... don't update, and if you did... MY BAD.

 

[Wing2]

How could it be that they overlooked some elemental relationships between the three sensors and also not have sampled them frequently enough? It doesn't make much sense to me.

This statement only make sense if one of these sensor's data was never read or taken into consideration in firmware version 1.0 - sampling frequency = 0 Hz.

Many tests using 1.0 firmware indicated ambient temperature has little effect on shooting time remaining. I would not be too surprised that data from sensor measuring external temperature was not used in version 1.0 firmware.

If data from this sensor was sampled say once every 10 seconds = 0.1Hz in firmware 1.1 , Canon is logically correct(but a bit misleading) to say 0.1Hz is more frequently enough than 0Hz.

 

[koenkooi]

How could it be that they overlooked some elemental relationships between the three sensors and also not have sampled them frequently enough? It doesn't make much sense to me.

This statement only make sense if one of these sensor's data was never read or taken into consideration in firmware version 1.0 - sampling frequency = 0 Hz.

Many tests using 1.0 firmware indicated ambient temperature has little effect on shooting time remaining. I would not be too surprised that data from sensor measuring external temperature was not used in version 1.0 firmware.

If data from this sensor was sampled say once every 10 seconds = 0.1Hz in firmware 1.1 , Canon is logically correct(but a bit misleading) to say 0.1Hz is more frequently enough than 0Hz.

The pre-cooling experiment with 1.0 seems to point to that the sensor(s) are read on startup, but only after a timer has lapsed.

 

[Handsome90]

On balance the new thermal details provided by Canon and the description of how that logic changed in today's firmware sounds a bit contrived to me. Thermals must have been one of the biggest challenges during the development and testing phases of the project. How could it be that they overlooked some elemental relationships between the three sensors and also not have sampled them frequently enough? It doesn't make much sense to me. IMO it sounds like misdirection to simply obscure the fact they may have simply loosened the reins a bit on the temperatures allowed, either peak or more likely total heat (temp x time). The fact the cooldown periods were most affected by the new firmware (shortened) suggest they're simply allowing the camera to operate warmer for longer. The fact the temps usually plateau near the end of the recording time allowed also supports the notion it's a total heat consideration rather than a peak or "overheating" consideration.

yes, but all of these could have been caught by Canon in their final pre-production testing but too late to affect the first wave of cameras out the door.

Faulty assumptions do occur in software development.

You do realize they released a camera like no other during a global pandemic that shut the world down. I can just imagine the pressure everyone was under. No one seems to be taking this under consideration.

Sony a7s3, black magic 12k, red komodo....

The R5 was designed for a different target audience than either of the cameras you mention. Since the A7sIII was announced, I have tried to imagine who that thing is for (I haven't had a 12 MP camera since the original 5D in 2005). I then remembered that Sony inked a contract with the AP to be their photographic equipment supplier and realized that the A7sIII is probably the perfect PJ camera. A similar comment can be made for the Red.

I think that the internal design goal for the R5 was to be mirrorless equivalent to the 5D series DSLR's. It seems to me that they met that goal.

The other guy was saying that it was a feat producing a camera during covid, I was merely mentioning that many companies did the same.

Canon could have waited a few more months, make the video modes more usable by tweaking it further and also release it with clog3. They didn’t have to release it now.

 

[janez]

On balance the new thermal details provided by Canon and the description of how that logic changed in today's firmware sounds a bit contrived to me. Thermals must have been one of the biggest challenges during the development and testing phases of the project. How could it be that they overlooked some elemental relationships between the three sensors and also not have sampled them frequently enough? It doesn't make much sense to me. IMO it sounds like misdirection to simply obscure the fact they may have simply loosened the reins a bit on the temperatures allowed, either peak or more likely total heat (temp x time). The fact the cooldown periods were most affected by the new firmware (shortened) suggest they're simply allowing the camera to operate warmer for longer. The fact the temps usually plateau near the end of the recording time allowed also supports the notion it's a total heat consideration rather than a peak or "overheating" consideration.

yes, but all of these could have been caught by Canon in their final pre-production testing but too late to affect the first wave of cameras out the door.

Faulty assumptions do occur in software development.

You do realize they released a camera like no other during a global pandemic that shut the world down. I can just imagine the pressure everyone was under. No one seems to be taking this under consideration.

Sony a7s3, black magic 12k, red komodo....

The R5 was designed for a different target audience than either of the cameras you mention. Since the A7sIII was announced, I have tried to imagine who that thing is for (I haven't had a 12 MP camera since the original 5D in 2005). I then remembered that Sony inked a contract with the AP to be their photographic equipment supplier and realized that the A7sIII is probably the perfect PJ camera. A similar comment can be made for the Red.

I think that the internal design goal for the R5 was to be mirrorless equivalent to the 5D series DSLR's. It seems to me that they met that goal.

The other guy was saying that it was a feat producing a camera during covid, I was merely mentioning that many companies did the same.

Canon could have waited a few more months, make the video modes more usable by tweaking it further and also release it with clog3. They didn’t have to release it now.

They should have solved overheating before they released the camera. They could make also some hardware changes. Now they are limited to software fixes only.

 

[Hoka Hey]

Canon could have waited a few more months, make the video modes more usable by tweaking it further and also release it with clog3. They didn’t have to release it now.

Glad they didn’t wait. I’ve had a month of fun with the R5! It’s the best stills camera that I’ve ever had the pleasure of shooting.

Since I rarely shoot video, I can’t say whether I would have preferred to wait a few months or have it now, but from the quality I’ve seen from it, I’m guessing that a lot of videographers are loving it.

--
Joe