as mentioned before i have / had one of these in the garage...
for fun i just swapped out one of my qnap 301ws for one of these now that 160mhz is working... btw this extroot is a major pain in the butt if you want to run docker containers on it... i mean, i got it all running but essentially you have to extroot first, then install containderd docker* ipks since the image will bloat to close to 90mb so you wont be able to install it without extroot.
sysupgrades are NOT fun
question : has anyone left theirs with the fan just off, eg: just let her rip? its hitting over 90c... hehe, 30c hotter then the 301w.
Yes, I do. This is one of the reason why i've not make any public compilation. Temp is insane. I don´t know how much temp this device can reached, i made fan works when wifi temp was 95ºC (Room Temp was around 30ºC).
Put something on usb, populate 10gb and 2.5gb port and you will have toaster.
i didnt pull that commit and sysupgrade path is pretty involved for me:
backup my docker containers, flash, re-extroot, re-install the missing docker ipkgs, re-install my missing container images etc etc. ill get around to it
btw, this nbg7815 has a better antenna / radio setup vs the qnap 301w.
this is the identical /etc/config/wireless, this is my 2.4ghz iot ssid:
edit : still running toasty btw... temps hovering around 85c... its in a closed tv stand though... i just tried switching to schedutil and enabling the auto_scale on the nss cores...
I made an observation. When I dismantled router straight after powering off, there are cooling plates on both top and bottom of the router. Top one was warm, but bottom was pretty untouchable (I am guessing that's where wifi chips are sitting). But taking that heat rises, that means bottom plate is slowly cooking the router from inside, with heat being mostly trapped. I know there are venting holes on the bottom (but with rising heat it's probably not helping much) and vents on the edges of the router.
What I have done is put router in a same spot as always, but hang it on the wall. Looking at the holes placement and rising heat principle I thought it may help - and it did I think. At least a bit.
Before that my fan was turning on roughly every hour for around 10 minutes (I am not utilising router heavily). Trigger points were @asvio defaults - 75 for on and 70 for off. With router placed vertically, fans have not come up even once for over 12 hours now, and my cpu temp (according to btop) is steady at 69-70 degrees.
I know that you folks, with much heavier utilisation will still experience quite high temps, but I think it may be worth a try, as it may shave a couple (or a few) degrees from our little space heater
I will also try today to use 12v power supply and see if there is a difference as @anom3 suggested.
ZyXEL's heat dissipation design wasn't particularly good in the nbg6817 already (~10°C higher temperatures than on the r7800), so the trend continues...
Yep, design is not ideal. It's been nearly a couple of days of router being placed on a wall and fan has not kicked in even once... I consider that as a win
Since the introduction of the new drivers defined by @Ansuel there is no predefined color for the LEDs, so if you have not defined them it seems that they do not work when used from the LUCI menu.
If for example in your case you use the command echo "64 0 0" > /sys/class/leds/q1:channel2/multi_intensity
you will define a red color with a LED power of 25%
You can define a different color for each of the 6 existing LED groups.
I have defined them in rc.local so that a specific color is assigned at each start.
Also keep in mind that it is established that if the device has a WAN interface defined, pattern #15 (dark blue with a low brightness) is activated when it starts and pattern #22 (green) when it stops.
Thank you very much for the information and explanation. So how can I learn what colors and values I can identify? Can you share your rc.local script?
I only define some color on led. Nothing especial.
# Put your custom commands here that should be executed once
# the system init finished. By default this file does nothing.
#LEDs setting
echo "80 5 0" > /sys/class/leds/q1:channel0/multi_intensity
echo "64 0 0" > /sys/class/leds/q2:channel0/multi_intensity
echo "0 0 32" > /sys/class/leds/q2:channel1/multi_intensity
echo "0 16 0" > /sys/class/leds/q2:channel2/multi_intensity
#testing latency parameters
sysctl -w net.ipv4.tcp_ecn=1
#nlbwmon fix
sysctl -w net.core.rmem_default=1572864
sysctl -w net.core.wmem_default=1572864
sysctl -w net.core.rmem_max=2097152
sysctl -w net.core.wmem_max=2097152
exit
The 3 numbers on the LEDs mean:
64 0 0
Red Green Blue
Each color can vary between 0 and 255
0 is off
255 is the maximum brightness of that color
the percentage of brightness in this case for red is 64/255 (25%)
In the case of color 80 5 0
The result is orange.
0 0 32 is a weak blue
0 16 0 is a green that despite being only 8% is seen with great strength.
If you use green to mix colors, use it in low values. (no more than 20) but that is already a personal choice.
Hey @asvio , could I check with you, if your recent updates keep thermal thresholds for fan activation same as before? It was 75 on, and 70 off. Is that still the case?
I am asking because with my latest placement of the router vertically my fan hasn't come up even once and I am wondering if this is because of the placement or thresholds were changed?
I am not going to modify it (at least I don't plan to), as those settings works fine for me. I just couldn't believe thermal differences after slapping our toaster on a wall
