INTL 3:770/1 3:770/3
REPLYADDR tnp@invalid.invalid
REPLYTO 3:770/3.0 UUCP
MSGID: 311c7e3c
REPLY: 2e614da0
PID: SoupGate-Win32 v1.05
On 09/04/2024 21:40, druck wrote:
> On 09/04/2024 13:52, Adrian wrote:
>> Apart from the Pi4, to upgrade, I know that I will need a bigger power
>> supply and a HDMI adapter lead, but after, things are a bit vague.
>> Can I get away with the Pi4 without any additional cooling, or do I
>> need a heat sink or fan, and if so, how do they get on with the hat ?
>> The space it is in has had a temperature range of -1 to 40 degrees C,
>> the mean across that time is about 15.5C.
>
> 40C is quite a high ambient temperature.
Not where I have worked...
Many years ago a small piece of equipment I had designed, housed in a
totally enclosed steel rack mount case failed after it had been on for
over a day.
It came back to me. It was working perfectly.
So I left it on for a day in the lab with a thermometer probe inside,
It got to well over 90°C inside.
Beyond the upper limits of the commercial chips it was built from. They
were well over 100°C case when opened it up.
We changed the case design and punched some air holes in it, and the
problem never reoccurred.
There are three relevant points here.
The first is that silicon chips themselves survive extremely high
internal temperatures. I have had a power Schottky diode unsolder itself
when connected across a battery. The chips will generally suffer
permanent damage at somewhere around 180°C . Which is a good thing since
they are soldered to a board using hot air at a lot more than that.
The second thing us that they will start to exhibit seriously out of
spec. behaviour at their rated *ambient* temperatures - usually +75°C
for ICs. Inside the chip may well be a lot warmer than that.
The final thing is to understand the difference between a closed box,
thermal convection cooling and forced air cooling. the is roughly a
25:5:1 ratio in cooling between the three cases.
How many Pis are in *closed* boxes *without* fans?
It's less an issue of 'do I need a fan?' or 'its 40°C ambient' than 'is
any hot air going to escape?'
With case temperatures up to well over 100°C possible for power devices,
the difference to an ambient of 40°C is not the much worse than say,
25°C. at a device case temperature of 100°C it is the difference between
a 75°C delta and 60°C delta - not that great really
And a Pi draws somewhere in the 3-5W region. And has lots of board
surface area to dissipate that.
From the Pi forum
"I have put the Pi4 in an airtight enclosure (with a big passive
heatsink), locked the frequency to 1500Mhz, overvoltage=4 and have
cpuburn running for about two hours. Ambient temp inside the box is 54
degrees celsius now.
vcgencmd measure_temp and vcgencmd measure_temp pmic show:
Pi4 CPU temp=124.0'C
Pi4 PMIC temp=129.7'C
And it's still running fine, have a chromium open and can surf the web.
I'm quite amazed."
https://forums.raspberrypi.com/viewtopic.php?t=264786
The whole thread is well worth reading.
And complements what I am saying, that only in a sealed enclosure where
no convection is possible can you get to super high temeperatures, and
in normal use the CPU will in any case throttle back
If you are going to run the Pi
> 4B hard with the likes of matplotlib, passive cooling my not be enough
> to stop it throttling.
That is a very hand wavey statement.
Passive cooling in a car in the sun with the windows closed yes, passive
cooling in a well ventilated case in a room at 40°C, no way.
> One of mine has a large heat sink case with no
> fan and never goes over 60C, but that's at a 20C ambient. For 40C
> ambient, I would recommend a fan case, go for a 40mm fan if you want it
> to be quiet, 30mm ones can be noisy.
>
Personally year of designing high power audio amplifiers and dealing
with computer hardware says 'bullshit'. You simply don't need fans for a
board doing 5W. You can get that out of a single small transistor with
a heatsink clipped on it. You can get that out of a TO3 packaged power
transistor with no heat sink *at all*.
And a PI4B board is a *lot* more surface area than that.
Currently this rooms is at 24.2°C according to a thermometer equipped
pico PI and the resident 4B that is in free air is showing this:
me@Coriolanus:~$ vcgencmd measure_temp
temp=53.5'C
me@Coriolanus:~$ vcgencmd measure_temp pmic
temp=45.8'C
So at idle -ish the chip internals are 29.3°C - call it 30°C over ambient.
at 40°C ambient the chip internals will be up around 70°C.
That is nothing.
"we throttle the CPU when it get to 85 to keep the temperatures down.
I've never seen anything over 90 in normal usage"
Is what the Raspberry Pi engineer says the chips do, in that thread I
referred to above.
In short *unless* you put the pi in an airtight case and wrap it in
insulation or leave it in the midday sun, it doesn't need a fan.
But people are just used to old school SPARC and INTEL chip that kicked
out upwards of 50W at idle, needing a fan. In hard use some SPARCs
would draw up to 300W!!!
Of all the customer returns on SPARC workstations I can remember they
were all dead fans or dead cpus (with dead fans as well).
--
“There are two ways to be fooled. One is to believe what isn’t true; the
other is to refuse to believe what is true.”
—Soren Kierkegaard
--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)
SEEN-BY: 10/0 1 15/0 90/1 103/705 105/81 106/201 128/260 129/305 135/220
SEEN-BY: 135/225 153/757 7715 218/0 1 601 700 840 870 930 220/70 221/1
SEEN-BY: 221/6 360 226/17 30 100 227/114 229/110 111 112 113 200 206
SEEN-BY: 229/307 317 400 426 428 470 550 616 664 700 240/1120 266/512
SEEN-BY: 267/800 282/1038 291/111 292/854 301/1 113 812 310/31 320/219
SEEN-BY: 322/757 335/364 341/66 342/200 396/45 460/58 633/280 712/848
SEEN-BY: 770/1 3 100 330 340 772/210 220 230 5020/400 1042 5058/104
SEEN-BY: 5075/35
PATH: 770/3 1 218/840 221/6 301/1 218/700 229/426
|
