> While conducting our internal engineering load tests of the Raspberry Pi 3 Model B, it became evident that the Raspberry Pi 3 REQUIRES heatsinks. Without them, the board will overheat and experience thermal shutdown, corrupt HDMI signal, and video corruption. SoC diode temperatures are also incorrectly reported lower than actual temperatures. A test of 4 threaded sysbench resulted in the imagery below.
> the board will overheat and experience thermal shutdown, corrupt HDMI signal, and video corruption
Wow.
The test mentioned in the Reddit post involved CPU load only.
I dont know why somone wouldnt toss a heatsync on "the hot chips" right from the start, it is after all an experimenters board kind of thing. If a person was going to turn it into a desktop computer.
Snapdragons (and most other phone SoCs) don't normally do this. They'll throttle quite heavily, but it's extremely unusual for them to shut down hard due to overheating.
> The temperature sensor on the BCM2837 becomes inaccurate after 80C. When the SoC has hit 100C, it is still reporting 83C. This prevents the 85C firmware throttling from kicking in.
(emphasis original)
Broadcom: "Oops."
Any idea what the TDP of the BCM2837 is? There seems to be some serious thermal issues with the Raspberry Pi 3.
That this kind of overheating is possible under normal operating conditions (case attached, high ambient temperatures) is really surprising to me.
@Psycogeek It's designed for education first and foremost.
I'll have to get back to work soon. See ya in a bit.
BCM2837 reporting wrong Tjunction is a serious fault.
@allquixotic @Bob For Ashes of the Singularity, Stardock folks during their Twitch TV chat were recommending that you use the 362 driver for best performance. (source)
The latest NVIDIA graphics drivers as of this writing are broken.
There have been quite a few reports of 364 series drivers causing all sorts of graphical issues. Version 364.72, in particular, has been widely reported to cause serious graphical defects and even hardware failure. Your best bet i...
Again, appears to be problems with specific hardware configurations. You probably got lucky. I'd recommend that you roll back while you still have a chance.
It's also possible that the driver is bugged in such a fashion that it's commanding the card to drive unsafe voltages through the GPU or is erroneously overclocking them to unsafe clock frequencies.
Couldn't really find much on this particular issue.
@allquixotic: Remove the 364.72 drivers, then install 362.00. Don't just roll back.
> To reinstall a previous or older NVIDIA graphics driver, do not use the Windows rollback feature. This method will not reliably restore all the previous driver files.
Instead, use the Windows Add and Remove programs to remove the current driver, and then install the older driver using setup.exe.
had a power hiccup lasting long enough to cause air circulation fans (not CPU fans) to go off; microwave forgot what time it is; and my bluetooth Imperial BART 1 connected to a powered USB hub got reset
@Bob when I first looked into UPSes, I ordered one without explicit PFC support. I then refused shipment after they sent the gigantic box out to my house (kinda felt bad), then a few months later I ordered one with PFC (still Cyberpower, just higher-end model). It's line-interactive.
And it works perfectly
I'm still not sure how long the battery would last with the considerable combined load of my peripherals, monitor and beefy computer, but probably on the order of 5 minutes -- just long enough to shut down Windows and say "Shut down anyway" to all the programs that try to keep from being terminated
anyway, for momentary hiccups, battery life is almost irrelevant, and it did just fine for that
and I primarily bought it to address hiccups, since we rarely get actual outages
(but hiccups are common enough during some seasons as to occur multiple times per week)
I think that this would be perfect for reading(like E-ink with backlight), and would lower eye strain in situations where high refresh rate is useless. It would be even better if the refresh rate could change automatically depending on current movement on the screen.
Those were intentionally going hand in hand, because device manufacturers were lazy and slow, so by taking direct control Google can also push updates out directly
I feel like buying another 500GB SSD. Hmm. But I shouldn't.
@kerbalspacebecky True, but by taking the apps out of the realm of open source, now manufacturers that do want to keep things up to date can't anymore.
Apparently the IP68 rating on the S7 is inferior to the IP67 rating on the S5. S7 uses microscreens and not actual sealed speaker unit, and tends to die if water is thinner than normal
@Bob There's a dozen of them, I think one of the earlier teardowns pointed them out
> I read in an article recently that IP68 certification has nothing to do with steam or mist which can still penetrate areas like speaker grills that are otherwise capable of withstanding liquid water.
Urgh why does Samsung have to make this shit so broken and complicated.
My S4 I just rooted on the way home from the shop on the bus. None of this Knox crap, or wiping encryption, or Samsung Pay being disabled, or locked bootloaders, or stinking SELinux or all that crap
@Bob look at more reviews for that stuff on amazon (even if the first 20+ reviews on amazon will be people getting freebies) amazon.com/Ringke-Snug-Fit-Tailored-Ultra-Thin-Coverage/… the tpu style has the flexability, but people are saying it slips off easily, would depend on how it is used. Some cases did have enough Edge on an edge to keep the face from being hit first . . .
And to reduce accidental clicks.
some of the reviews discuss the issue of the port holes not being large enough, I guess those people do not yet own a razor blade :-)
I like how they are saying (often) this Branded item is always good, as if these are being made by them, and not by 2-3 total china factories, making them. Like toasters and tennis shoes , find me one Any one that has different materials anymore.
it is weird how they are saying silicone en.wikipedia.org/wiki/Silicone and TPU en.wikipedia.org/wiki/Thermoplastic_polyurethane in the same sentance, when as far as I know they are 2 different species, that could have much the same properties but simply are not made of the same chemicals. Generally the TPU leans more towards stiff , but can be flxable, and silicone leans more towards flexable but can be made more rigid.
https://en.wikipedia.org/wiki/Water_Resistant_mark if we go back to the older method of water resistance , which was used for watches, it is interesting to note that a "Water Resistant 10 atm or 100m Suitable for recreational surfing, swimming, snorkeling, sailing and water sports. Not suitable for diving"
100meters with it just having been dropped there, but that doesnt provide an ability to swim with it under water.
forum.xda-developers.com/galaxy-s7/how-to/… but back at this link, wouldnt you say that the person had overly high expectations, for a phone that is tested by "placing it" in a few feet of water?
I think that this would be perfect for reading(like E-ink with backlight), and would lower eye strain in situations where high refresh rate is useless. It would be even better if the refresh rate could change automatically depending on current movement on the screen.
