Previous versions of TopPixel's message #15486895 « A Technical Look at CF Express Power Consumption and Why the Sony Alpha 1 May Be a Power Miracle »
TopPixel Offline [X]
A Technical Look at CF Express Power Consumption and Why the Sony Alpha 1 May Be a Power Miracle
Jesse Evans wrote:
As with all other costly lunches, you can see that the performance at write time scales roughly linearly, and the CFExpress Type B cards offer higher performance per transfer speed over SD. This makes sense since CFExpress uses NVMe rather than SATA, and NVMe is a more efficient protocol as it was developed specifically for flash storage.
1000MB/s is 4.13x the speed achieved by the UHS-II card, and in the worst case you pay a 3.16x power penalty.
When comparing the CFExpress Type A and Type B cards, you can expect that the power usage will be dictated by the transfer rate, and not the form factor.
Yes, it will use less power, because you are transferring data more slowly.
If one were to software limit the transfer rates for a Type B card, you would likely see nearly the same power usage.
Actually, I wrote in the original post that even the standby and sleep power of CF Express B is orders of magnitude higher than a UHS-ii sd card in standby/sleep, so it is not just about transfer rate.
And from what people report about excessively hot CF Express B cards coming out of the Canon R5 not even writing to it and from users with CF Express B card readers just having the card inserted report very hot cards. This means the CF Express cards are not even going into the P3 and P4 standby and sleep states that would be measured in mW instead of multiple watts. The CF Express B cards on the market are consuming multiple watts of power not even doing anything.
In theory, you *COULD* use less power in the CF Express format compared to UHS-ii for the same amount of data written since CF Express writes much faster so can go to standby/sleep much faster. But from what we see out in the real world with very hot CF Express cards, this is not happening consistently or maybe even at all, resulting in constant high power usage. The infamous very hot CF Express B cards out of the Canon R5 doing nothing from the 8K testing debacle shows this. It may not even be possible to put CF Express B cards reliably into and out of the standby P3 and P4 states in time for burst shots and might be why Canon has the CF Express running full power all the time.
Also, I see no mention of reduced transfer rate modes in the CF Express card specifications I linked to before. The Delkin one does show that is has thermal throttling safeguards but those only appear to be triggered when temperature limits are reached.
And if it's a pair of ProGrade CF Express B cards, they would be consuming over 10 watts in operation and bursting up to 15 watts combined. I don't think a Sony Alpha 1 body can handle this kind of heat without something melting or heating up the image sensor and cause thermal noise. Also, the FZ100 battery only has a 16.4 Wh capacity, slightly bigger than the Canon R5, so it's not like we have 95 Wh batteries from laptops where we can waste energy.
And that thermal noise, knowing now how much power CF Express B cards consume, I actually think it is not a good idea to use them in their curent state for a camera since our image sensors are so sensitive to noise generated by those conditions.
Feb 01, 2021 at 06:43 PM
TopPixel Offline [X]
A Technical Look at CF Express Power Consumption and Why the Sony Alpha 1 May Be a Power Miracle
Jesse Evans wrote:
As with all other costly lunches, you can see that the performance at write time scales roughly linearly, and the CFExpress Type B cards offer higher performance per transfer speed over SD. This makes sense since CFExpress uses NVMe rather than SATA, and NVMe is a more efficient protocol as it was developed specifically for flash storage.
1000MB/s is 4.13x the speed achieved by the UHS-II card, and in the worst case you pay a 3.16x power penalty.
When comparing the CFExpress Type A and Type B cards, you can expect that the power usage will be dictated by the transfer rate, and not the form factor.
Yes, it will use less power, because you are transferring data more slowly.
If one were to software limit the transfer rates for a Type B card, you would likely see nearly the same power usage.
Actually, I wrote in the original post that even the standby and sleep power of CF Express B is orders of magnitude higher than a UHS-ii sd card in standby/sleep, so it is not just about transfer rate.
And from what people report about excessively hot CF Express B cards coming out of the Canon R5 not even writing to it and from users with CF Express B card readers just having the card inserted report very hot cards. This means the CF Express cards are not even going into the P3 and P4 standby and sleep states that would be measured in mW instead of multiple watts. The CF Express B cards on the market are consuming multiple watts of power not even doing anything.
In theory, you *COULD* use less power in the CF Express format compared to UHS-ii for the same amount of data written since CF Express writes much faster so can go to standby/sleep much faster. But from what we see out in the real world with very hot CF Express cards, this is not happening consistently or maybe even at all, resulting in constant high power usage. The infamous very hot CF Express B cards out of the Canon R5 doing nothing from the 8K testing debacle shows this. It may not even be possible to put CF Express B cards reliably into and out of the standby P3 and P4 states in time for burst shots and might be why Canon has the CF Express running full power all the time.
Also, I see no mention of reduced transfer rate modes in the CF Express card specifications I linked to before. The Delkin one does show that is has thermal throttling safeguards but those only appear to be triggered when temperature limits are reached.
And if it's a pair of ProGrade CF Express B cards, they would be consuming over 10 watts in operation and bursting up to 15 watts combined. I don't think a Sony Alpha 1 body can handle this kind of heat without something melting or heating up the image sensor and cause thermal noise. Also, the FZ100 battery only has a 16.4 Wh battery, slightly bigger than the Canon R5, so it's not like we have 95 Wh batteries from laptops.
And that thermal noise, knowing now how much power CF Express B cards consume, I actually think it is not a good idea to use them in their curent state for a camera since our image sensors are so sensitive to noise generated by those conditions.
Feb 01, 2021 at 06:41 PM
TopPixel Offline [X]
A Technical Look at CF Express Power Consumption and Why the Sony Alpha 1 May Be a Power Miracle
Jesse Evans wrote:
As with all other costly lunches, you can see that the performance at write time scales roughly linearly, and the CFExpress Type B cards offer higher performance per transfer speed over SD. This makes sense since CFExpress uses NVMe rather than SATA, and NVMe is a more efficient protocol as it was developed specifically for flash storage.
1000MB/s is 4.13x the speed achieved by the UHS-II card, and in the worst case you pay a 3.16x power penalty.
When comparing the CFExpress Type A and Type B cards, you can expect that the power usage will be dictated by the transfer rate, and not the form factor.
Yes, it will use less power, because you are transferring data more slowly.
If one were to software limit the transfer rates for a Type B card, you would likely see nearly the same power usage.
Actually, I wrote in the original post that even the standby and sleep power of CF Express B is orders of magnitude higher than a UHS-ii sd card in standby/sleep, so it is not just about transfer rate.
And from what people report about excessively hot CF Express B cards coming out of the Canon R5 not even writing to it and from users with CF Express B card readers just having the card inserted report very hot cards. This means the CF Express cards are not even going into the P3 and P4 standby and sleep states that would be measured in mW instead of multiple watts. The CF Express B cards on the market are consuming multiple watts of power not even doing anything.
In theory, you *COULD* use less power in the CF Express format compared to UHS-ii for the same amount of data written since CF Express writes much faster so can go to standby/sleep much faster. But from what we see out in the real world with very hot CF Express cards, this is not happening consistently or maybe even at all, resulting in constant high power usage. The infamous very hot CF Express B cards out of the Canon R5 doing nothing from the 8K testing debacle shows this. It may not even be possible to put CF Express B cards reliably into and out of the standby P3 and P4 states in time for burst shots and might be why Canon has the CF Express running full power all the time.
Also, I see no mention of reduced transfer rate modes in the CF Express card specifications I linked to before. The Delkin one does show that is has thermal throttling safeguards but those only appear to be triggered when temperature limits are reached.
Feb 01, 2021 at 06:24 PM
Previous versions of TopPixel's message #15486895 « A Technical Look at CF Express Power Consumption and Why the Sony Alpha 1 May Be a Power Miracle »
