With the L5 "on the road" as daily driver

I do use my L5 as daily driver here in Germany, but this means that I’m only for short time “on the road” traveling between home and office or going into town (and only a bit usage to make phone calls or check mails with a terminal app mutt).

I was now a week in Cuba and all day outside with a cuban SIM card (to make calls and Internet with low costs). The usage was primary:

  • using Internet with FF to read news in RT DE
  • phone calls
  • telegram messanger
  • read mails with mutt and answer with webmail in FF (because SMTP outbound was blocked in the provider cubacel and I’ve not have time to fix this with another port outbound)

The experience was not good because

  • already after a few minutes the L5 gets hot that the red LED starts to blink sporadic
  • if so, it can’t be charged anymore from a power bank I have, it blinks on/off with 1sec frequense
  • only power-off for a while helps to recharge it again.

I forgot to mention two facts

  • the environment tempreature is 30° C on the day and 20° C in the nights
  • my L5 has a 3D printed case which covers the back and its 4 sides

I think having a case on the L5 isn’t a good idea for this very reason, unless there is some way to design a case with some sort of finned heatsink. As a side note I modified my back cover to attach the flat circular metal disk for the magnetic holder I have and because it kept coming off even with JB weld (the JB weld doesn’t stick well to the plastic of the cover and the adhesive that it comes with eventually degrades because of the heat the phone generates). So what I did was I dilled some holes in the back of the case within the footprint of the metal disk and then cut a circular sized disk of plastic to act as a backing on the inside of the case, then JB welded the whole thing together, so far so good.

Doing this triggered an idea of attaching a larger metallic plate on the inside of the cover and then have a metalic plate on the outside of the back cover would it help dissipate the heat? Now the only caveat would be would it interfere with GPS/Cell/Wifi? I presume that Purism must have spitballed this idea and perhaps tossed it because of this, or maybe didn’t do it because it would mean increase complexity of the back cover and higher manufacturing cost. I should order a bunch of back covers and do some experiments.

1 Like

Would you like to show us some pictures of what you did? And did you do any measurements to see if the temperatural behavior improved?

Watercooled cases as heatsinks!

(Of course it would always have to be mounted to some external source of water input and egress.)

How about a wine skin?

1 Like

Basically I cut a piece of plastic from a container that held a package of screws to act as a backing and then the jb weld both the plastic and metal disks through the back plate which had a bunch of small holes in it so the jb weld would pass through them and connect the metal disk to the backing.


The app Usage shows the temperature, for example in the screen below, of an unused L5. I will do some experiments and measure the temperature increase in the following situations and always the L5 only connected to data mobile, and always to compare the L5 with or without the case for the following situations:

  1. watching a YT video in FF for 30 minutes
  2. reading pages of RT DE for 30 minutes, scrolling up and down and changing the pages
  3. doing a phone call of 30 minutes (have to look for a number without charges)
  4. reading mails with the mutt in the terminal app for 30 minutes
  5. having the L5 in a backpack with display off, but data mobile on for 30 minutes
  6. reading messages/pictures in the app telegram for 30 minutes
  7. something else?

and always before any step let cool down the L5 to the values shown below.

Btw: The Usage app is more or less useless regarding the graph of the temperatures. It lacks any history, a real x/y axis values and different colors for the different points of temperature taken.


Could you please stop product placing this lies** spreading media of a rouge** state in every post?

** The used adjectives are provable with facts.

From your website we could read that you would prefer to live in the GDR instead of modern Germany and regret the fall of the Berlin wall. But like always, instead of living in Belarus (which is as close to GDR as possible) or russia, China, Nord Korea, you instead live in Munich (West Germany, EU, NATO) and use a product made by an US company that you would not be able to buy if you lived in GDR for several reasons:

  1. You would not have the money to afford it.
  2. You would likely not be able to buy an US product from behind the iron curtain.
  3. The totalitarian state might not be allow devices that are not controlled by the state.

You might now starting complaining about censorship ect. but as we know in the West, liars*** like Daniele Ganser, Tucker Carlson ect. are free to spread their lies, while in russia opposition media is under highest restrictions and repressions.

*** Provable with facts.

I would not follow up on this topic to avoid off-topic. I just urge you to restrain from your advertisement. It is sufficient that we know that you use a news website 30 minutes a day.


Glad to live in a state we’re until now we have the right to say what we think. Don’t know, though, how long.


First of all, this is a privately-owned forum, so your “right to say what you think” is entirely irrelevant.
Secondly: stating which news site exactly you were reading during your test is not only entirely off topic, it has also added absolutely nothing to the conversation.
I’m not willing to speculate as to what exactly motivated you to even mention the news site, lest I’d drift even further off topic. Yet, the backlash you’re getting for mentioning the site is not only entirely justified – you also should have expected it.
I’d really appreciate if you kindly kept your politically extremist subtexts to yourself in the future, or took it elsewhere. Thank you.


I totally agree that this is off-topic. It would be ideal if this thread could remain about the L5 as long as there are people wanting to talk about it.

About “justified” and “expected” - I totally expect threads to go off-topic, but that is not always justified by forum rules or etiquette. I should note that the rules don’t say anything about which websites one likes to read, but it do forbid getting personal/political about it:

so I’d appreciate if you kindly kept your unprompted judgements of others’ reading choices to yourselves. Thank you.


Months ago I elaborated a small shell script to read the temperature data directly without the Usage app:

purism@pureos:~$ guru/temp.sh 
cpu-thermal: 42,00°C
gpu-thermal: 42,00°C
vpu-thermal: 42,00°C
bq25890-charger: 35,50°C
max170xx_battery: 39,10°C
purism@pureos:~$ guru/temp.sh | grep thermal | sort
cpu-thermal: 42,00°C
gpu-thermal: 42,00°C
vpu-thermal: 42,00°C

I will write a modified version to collect the data while, for example, reading 30 minutes my preferred news channel. And so I could afterwards draw graphs out of this data.

1 Like

Is there a simple way to log applications used (e.g. in foreground, CPU usage)? Display in/oft can be extracted from journal. Maybe also his, because of power consumption (heat)?

Having temps and CPU (and maybe net) usage in same view/graph would be ideal. None of the others have this. Add warnings and it would be golden!

What is a good way to measure the CPU utilization? Don’t just say top :slight_smile: If you run

top -b -n 1 | grep ^%Cpu

it will give on an otherwise idle L5 always around 70% because it measures the boot up time of the processes top and grep.

You got me there. Could the load be reduced in a meaningful way by defining longer measurement/reporting intervals? I mean, temp doesn’t fluctuate rapidly, while CPU load does, but load averages or even just point measurements (2, 3, 5, 10, 20, 30 or even 60 seconds apart? for data over few hours or days) would give some indication how they correlate. Even though it may be rough, it would be a start. But… this could also be wishful thinking.

I am not an expert but can’t this data not being retrieved from sysfs in a way with lower footprint?

Load averages are in /proc/loadavg and I don’t know if the interval can be changed. CPU usage is in /proc/stat. I don’t know if raw “system load” value is exposed anywhere.

1 Like