spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Mon Aug 26, 2019 4:22 pm

Hi


I saw a nice video in youtube, made by Chris ("Explaining Computers" channel),
called "Raspberry Pi 4 Cooling":
https://www.youtube.com/watch?v=AVfvhEJ9XD0

The experiment in the video shows that If you stress the SoC to 100% work, on all 4 cores,
without any type cooling (passive or active),
the SoC will get to 90°C.


It also shows several ways to cool down the SoC,
so you won't get to this too-hot 90°C point.


What the video is not showing, is what are the temperatures the SoC will get to, without any type cooling, when you are doing common operations which are not intensive.

Those not intensive operations/states are quite common, and represent what the majority of users will do with their RPi.
They are for example:

- Immediately after finished booting
- After being Idle for 1 min
- When using a light program like a Text Editor
- When using a Browser (showing text and images, not a video)
- When using a Browser (showing a video)
- When using some IDE



Is there some webpage that someone made, in which he tested the Temperature of RPi's SoC, when in the abive operations/states?
(without any cooling)


The purpose of this question, is to understand what temperatures to expect, when using the RPi in a relatively common and average way,
and not in a CPU intensive way.
It will also help understand, If when using it for such non-intensive work, a cooling solution is needed or not.


BTW,
My question is about the latest RPi (4B),
but if there is a table which shows the temperature measurements not just for the 4B, but for all previous models,
then that would be terrific.


Thank you very much

kflmiami420
Posts: 61
Joined: Thu Mar 21, 2019 6:54 pm

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Mon Aug 26, 2019 4:55 pm

Hi Spaceman5


What are you going to use the pi4 4gig for ?


I ask the question because you do know that is has a built in protection so not having any type of cooling is just going to set off the system to protect it from overheating. and just slow it down until temps come down.

Just get a passive cooling Heatsink like this one https://www.adafruit.com/product/3082 and get a case that has a lot of airflow.
Or run it naked. I have this setp on my pi4 4gig and the original raspberry pi foundation 5v/3amp Power supply and it works .
( no firmware update needed ) I run regular tasks with it. It runs a samba setup , and watch youtube . I have second Pi4 1gig running 5 sensors
bmp388,mcp9808,sht31-d,velm6075,mcp3008 adc, colleting weather data

here is the link info so you can measure your device temp for your application

Code: Select all

Raspberry Pi 4 CPU Temperature Monitor
Author: Darrell Little http://yatb.devcali.co
Date: 07/08/2019
Based on the Raspberry Pi Foundation Project
https://projects.raspberrypi.org/en/projects/temperature-log
Data logged on AdafruitIO
https://io.adafruit.com/dlittle55/dashboards/my-rpi4-dashboard



That way you will save time from reading some results about heat temps at different modes of operation.

I have recorded cpu temps to logs and also have them published to adafruitIo IOT web.
Last edited by kflmiami420 on Tue Aug 27, 2019 5:06 am, edited 1 time in total.

spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Mon Aug 26, 2019 5:47 pm

Hi

Thank you for your reply..

>What are you going to use the pi for?

Simple tasks, nothing CPU intensive..

Mainly experimenting with it, or some simple automation code, and reading sensors, and operating actuators.


So you say even with such non-intensive tasks it might throttle or shut itself down?

kflmiami420
Posts: 61
Joined: Thu Mar 21, 2019 6:54 pm

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Mon Aug 26, 2019 7:37 pm

Mine does not shut it self down. I have it open with out a case to get really good airflow

If it overheats it slows down the mhz rate to decrease the heat but not shutdown.


if you decide to place it in a case with out a fan and ports to suck air in and blow air out then it will over heat. just placing a fan is not enough you have to create positive airflow ..

here read this post , all 17 pages thats all they talk about pi4 and cpu temps

https://www.raspberrypi.org/forums/view ... 8&t=243500

ProDigit
Posts: 374
Joined: Tue Aug 30, 2011 1:24 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Mon Aug 26, 2019 11:54 pm

One of the reasons I didn't get the Pi4 is because of the heat.
I'm sure with the Pi 5, they'll figure out a smaller die solution; or perhaps get a Pi 4B+ with the same specs, but runs cooler.
While the Zero was too slow for many, and the 1, 2, and 3 had an issue of Bluetooth keyboard getting interference with the Wireless signal,
The Pi 4's main issue is heat extraction.

The only pis that can run without cooling fin, are the zero, 1 and 2.
The Pi 3 can run with a cooling fin, as long as it's not in a case.
The 3B+ needs either a large heatsink, or a small active one.
The 4B isn't made to be ran inside a closed case, unless the case itself is the cooling solution (a metal case cooling down CPU and other chips).

If you just want to steer some smaller stuff, why not get a Pi 0W, or Pi 3B+?
They're great alternatives, run much cooler, and still sold today!
I personally prefer my Pi 0 W for most of my work.

User avatar
rpdom
Posts: 15194
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 1:22 am

ProDigit wrote:
Mon Aug 26, 2019 11:54 pm
I'm sure with the Pi 5, they'll figure out a smaller die solution; or perhaps get a Pi 4B+ with the same specs, but runs cooler.
Die shrinkage already happened with the Pi 4, but it has a lot more stuff in it than previous models.
While the Zero was too slow for many, and the 1, 2, and 3 had an issue of Bluetooth keyboard getting interference with the Wireless signal,
The 1 and 2 didn't have Wireless or Bluetooth unless you used USB adaptors.
The only pis that can run without cooling fin, are the zero, 1 and 2.
The Pi 3 can run with a cooling fin, as long as it's not in a case.
All of my Pi 3Bs run just fine in the official case with a small heatsink.
One of my Pi 1Bs has a heatsink on the USB/LAN chip as that gets a bit hot due to a design fault with one of the power rails.
The 3B+ needs either a large heatsink, or a small active one.
How about none? Official case with the lid off works for me.

ProDigit
Posts: 374
Joined: Tue Aug 30, 2011 1:24 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 3:31 am

rpdom wrote:
Tue Aug 27, 2019 1:22 am

The 3B+ needs either a large heatsink, or a small active one.
How about none? Official case with the lid off works for me.
[/quote]

For one, you'll hit thermal throttle.
You want the Pi to run below 60C, to get it's full performance.
That's just not possible with a 3B+.
Even with this relatively 'massive' heatsink, the 3B+ still hits 60C in a closed case:
Image
https://www.amazon.com/dp/B07CTSNZYK/

spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 4:28 am

Thank you all for all the info and tips!

ProDigit wrote:
Tue Aug 27, 2019 3:31 am
Even with this relatively 'massive' heatsink, the 3B+ still hits 60C in a closed case:
Wow..

So a fan is critical..

User avatar
rpdom
Posts: 15194
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 7:46 am

spaceman5 wrote:
Tue Aug 27, 2019 4:28 am
Thank you all for all the info and tips!

ProDigit wrote:
Tue Aug 27, 2019 3:31 am
Even with this relatively 'massive' heatsink, the 3B+ still hits 60C in a closed case:
Wow..

So a fan is critical..
In a fully closed case a fan or heatsink is mostly pointless because the heat has nowhere to go.

In an open case a fan or heatsink is rarely required.

ProDigit
Posts: 374
Joined: Tue Aug 30, 2011 1:24 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 7:54 am

spaceman5 wrote:
Tue Aug 27, 2019 4:28 am
Thank you all for all the info and tips!

ProDigit wrote:
Tue Aug 27, 2019 3:31 am
Even with this relatively 'massive' heatsink, the 3B+ still hits 60C in a closed case:
Wow..

So a fan is critical..
I'd say, a fan is critical if you're running more than 50% constant CPU load.
Passive heat sinks are good if the only CPU load will be program loading, or occasional bursts, but the pi remains mostly passive.
When coding the 3B+ I have, idles around 50c with the large heat sink, which is still ok.
With an open case, around 40c
With an active cooling it should hit close to 30c.

In a closed case, a large passive heat sink works well when the pi runs mostly passive (coding, web browsing, ...).
The heat sink I posted takes 2 full minutes at full load before it hits 60C thermal throttle, or, no throttle if it's set to 70C.
Compare that to a small heat sink, that takes 30-45 seconds to hit 70c thermal throttle.

So a heat sink does matter, even in a closed case. The spread of heat allows more heat to dissipate between the cracks of the connectors, and the LCD hat screen is backheated more uniformly, rather than having a hotspot right above the CPU heatsink.

The 3B+ absolutely needs a heat sink, even in an open case, when you want it to work at peak performance (1.4Ghz), and not throttle down in mere seconds to 1,2Ghz when loading a program.
All Pi's benefit from heat sinks, including the Pi zero, which can reach slightly higher overclocking results with a heat sink (46C) vs without (55C).


spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 8:23 am

rpdom wrote:
Tue Aug 27, 2019 7:46 am
In a fully closed case a fan or heatsink is mostly pointless because the heat has nowhere to go.

In an open case a fan or heatsink is rarely required.
The ideal case is similar to a PC Case:
The case is generally closed, except of openings for air to come in, and come out..

So for RPi, the fan will throw the air out, thru some openings,
and on the other side, there will also be some small openings, to enable air coming in.
thus we get a nice flow, but the case is almost closed, to protect the board.

ProDigit wrote:
Tue Aug 27, 2019 7:54 am
I'd say, a fan is critical if you're running more than 50% constant CPU load.
Passive heat sinks are good if the only CPU load will be program loading, or occasional bursts, but the pi remains mostly passive.
When coding the 3B+ I have, idles around 50c with the large heat sink, which is still ok.
With an open case, around 40c
With an active cooling it should hit close to 30c.
Thank you ProDigit, this is exactly the info I was looking for, in order to understand what to expect.
But you talk about the 3B+,
do you know what happens with the 4B, in similar cases? (non-intensive ork, like program loading, occasional bursts, etc)

Thank you, going to read it

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23663
Joined: Sat Jul 30, 2011 7:41 pm

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 8:36 am

There is a lot of misinformation in this thread. So let set some things straight.
  1. A Pi (any model) without any sort of cooling will not fail through 'overheating'. It will simple throttle itself to cool itself down if it get too hot. This mean that cooling solutions are not 'critical'.
  2. If a Pi goes over a preset limit, it will automatically slow down
  3. If you are running intensive applications all the time, a cooling solution will help with performance.
  4. If you are using an enclosed case, there is nowhere for the heat to go, EVEN with a heatsink. So they will run hotter in an enclosed case.
  5. Heatsinks in an open case work well
  6. Use a fan shim or similar for best cooling, if in a case ensure there are vent holes.
  7. The Pi4 is already on a smaller process than previous models.
  8. The FLIRC case works well but the case itself does get hot. A very good passive cooling solution.
  9. My P4s runs on my desk, no case, rarely overheats, if ever.
  10. Most desktop uses of a Pi rarely use, on average, more than about 25% of CPU. Some video will bump that, as will big compiles etc.
  11. If you are waiting for a Pi5, you are in for a VERY long wait.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 8:54 am

Thank you very much jamesh, interesting info.

kflmiami420
Posts: 61
Joined: Thu Mar 21, 2019 6:54 pm

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 7:03 pm

If you are waiting for a Pi5, you are in for a VERY long wait.
:D :D :lol: :lol: :lol: :lol:

User avatar
rpdom
Posts: 15194
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 7:21 pm

kflmiami420 wrote:
Tue Aug 27, 2019 7:03 pm
If you are waiting for a Pi5, you are in for a VERY long wait.
:D :D :lol: :lol: :lol: :lol:
Yep. In keeping with the Raspberry Pi Foundation's habit of surprising people, they are going to release the Pi 6 before the Pi 5. :lol: :lol: :lol:

ProDigit
Posts: 374
Joined: Tue Aug 30, 2011 1:24 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Tue Aug 27, 2019 10:06 pm

@ spaceman5: My apologies, for some reason I had the impression this all was about the 3B+.
The Pi4 uses slightly more power than the 3B+. I can imagine it has higher idle and peak consumption and thermals.
Though it's power dissipation is spread out over 4 'large' chips, vs 3 chips on the Pi 3B+.

jamesh wrote:
Tue Aug 27, 2019 8:36 am
There is a lot of misinformation in this thread. So let set some things straight.
  1. A Pi (any model) without any sort of cooling will not fail through 'overheating'. It will simple throttle itself to cool itself down if it get too hot. This mean that cooling solutions are not 'critical'.
    They might be critical for a program to work or not. If the Pi doesn't offer the same performance at 70C as it does at 50C, I would say that cooling is integral and absolutely necessary. For most people who are running the 3B+ or 4B in a case, they all will experience CPU throttling; and all will benefit from a heatsink
  2. If a Pi goes over a preset limit, it will automatically slow down
  3. If you are running intensive applications all the time, a cooling solution will help with performance.
  4. If you are using an enclosed case, there is nowhere for the heat to go, EVEN with a heatsink. So they will run hotter in an enclosed case.
    A (large) heat sink will dissipate the heat better. The heat will radiate much more through the plastic. My Pi 3B+ plastic casing feels cold to the touch with a small heat sink, but feels warm with a large heat sink (one that covers nearly the entire board).
    There's more dissipation of heat, so heat gets extracted better. Also, a large heat sink can act as a buffer. if loading programs takes 1 to 2 minutes, a large heat sink can prevent the CPU to reach throttling temperatures, while a small one might start throttling earlier. A large heat sink does need more time to cool down after that, where as a small heat sink cools a lot faster after a CPU load is finished.
  5. Heatsinks in an open case work well
  6. Use a fan shim or similar for best cooling, if in a case ensure there are vent holes.
  7. The Pi4 is already on a smaller process than previous models.
    Though nothing compared to a 'modern' $45-$65 cellphone, using 14 nm lithography. More is possible, and not that expensive in today's terms. Though if in the future the Pi will have a 14nm design, I would hope it'll get something more like a 6-8 core 1,75Ghz CPU. The frequency and responsiveness for a Linux GUI at 1,5-1,6Ghz is quite good. More power can be distributed to more cores. It'll also allow the CPU to remain cooler when running the same load through the same amount of cores as current Pi4.
  8. The FLIRC case works well but the case itself does get hot. A very good passive cooling solution.
  9. My P4s runs on my desk, no case, rarely overheats, if ever.
    There's the case of overheating (going past 85C) and thermal throttling (60-70C on the Pi3B+). While you could go by with a Pi 3B throttling at 1.2 Ghz, some prefer to have maximum performance at all times.
  10. Most desktop uses of a Pi rarely use, on average, more than about 25% of CPU. Some video will bump that, as will big compiles etc.
  11. If you are waiting for a Pi5, you are in for a VERY long wait.
    Yes, I'm expecting a 4B+ first.

wh7qq
Posts: 1326
Joined: Thu Oct 09, 2014 2:50 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Sat Aug 31, 2019 8:00 pm

A closed case is a bad idea for any RPi from the 3B onward. My 3B, even with an open case , would overheat with java and graphics intensive web pages and I would consistently see the little red thermometer in the upper right of the display. That said, I have a B+ running as a pi-hole (headless) in a semi-closed case and no heat sink that runs 24/7 next to my router and never gets above 48 C (ambient is never less than 21 C but often up to 28 C). A headless RPi zero runs 24/7 at ~38 C under the same conditions. The 3B is massive overkill for a pi-hole...it might run well on a zero. I put copper heat sinks on the 3B with the provided black self adhesive and it barely helped but when I replaced the black goo with Arctic alumina epoxy, the temp dropped into the low 60s.

I am particularly impressed with the look of this case:https://www.amazon.com/Metal-Raspberry- ... ay&sr=8-23 and would consider buying it if I didn't already have epoxied on heat sinks on my 3B but there are tons of aluminum cases out there with good ventilation that would work well with the 3B.

ProDigit
Posts: 374
Joined: Tue Aug 30, 2011 1:24 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Sat Aug 31, 2019 10:43 pm

While twice the price, check out how slick this case is:
https://www.amazon.com/gp/aw/d/B077D8K797

What helps lower thermals, is using a thermal pad on the motherboard, connecting the metal bottom to the frame, to dissipate the motherboard's heat better.
It does result in lower CPU temperatures, and costs very little to implement.

spaceman5
Posts: 57
Joined: Sun Aug 25, 2019 5:26 am

Re: Looking for Temperature Measurements of Raspberry Pi in Different States/Operations

Sun Sep 01, 2019 4:05 am

Nice.

wh7qq, what do you do with the RPi Zero?

Return to “Advanced users”