Very noisy fans

[magostinelli]

Hello,

my nas is too noisy, do you know if it's possibile to change the fans with others that is silent?

 

Regards.

[Werner]

Not entirely sure but I seem to remember that I catched something about fan speed adjustment is not support (yet).

[ShadowDance]

I replaced the stock fans with Noctua NF-A8-PWM fans. These have slightly lower CFM (airflow) at 32.5 vs 35 (stock) however they have a high static pressure rating at 2.37 mm H₂O. For me they produce better cooling than the stock fans. Their noise level is also rated at 17.1 dB(A) so you will barely hear them even at full speed.

 

I also had a pair of NF-R8-PWM (discontinued model) from the old days which I tried. They are very close in CFM at 31.4 but have worse static pressure at 1.41 mm H₂O. These fans produced worse cooling than the stock fans.

 

One additional change I did was place the metal fan grills (finger protectors) on the outside because then the fan is allowed to produce a seal against the case. I thought it was a small design miss to leave a gap between the case and the fan because it might allow air to move back inside. Aesthetically it's nicer to have the grill on the inside (IMO), so as an alternative fix one could design a 3D printed piece to fill the gaps.

 

It's also possible to adjust speed by modifying the `/etc/fancontrol` configuration and restarting the service (`systemctl restart fancontrol`), but I would not recommend this unless you're using better than stock fans. If the CPU is working full throttle you will want the fans running at full speed to sufficiently extract the heat from the CPU area.

Edited November 14, 2020 by ShadowDance

[SIGSEGV]

@ShadowDance

Thanks for the input regarding the fans.

I'll give it a try if the noise goes up on my system once the ambient temperature starts to rise here in couple of months, for now I'm concerned about the dust.

[gprovost]

For now fan speed control is based on SoC temperature which is to be honest not the smartest approach since SoC temp is not representative of the ambient temp of the enclosure.  But it was the wisest approach to make the baseline Armbian image clean without too many tweaks.

 

What matters the most in a NAS is the HDD temperature since it's what impacts the most the MTBF of a storage array.

 

We will provide soon the instruction on how to control the fan speed based on HDDs temperature. This will result in a more constant airflow (less variation in speed) and therefore a more quiet operation.

 

It's actually ok to let the SoC goes a bit high in temperature but no the HDD.

[SIGSEGV]

Now it makes sense why the fans ramp up when login into the system, hahaha.

An algorithm that takes into account both the CPU and HDD temperatures would be a nice solution, prioritizing HDD temps since these need to lower than the temp for the SoC.

That being said - is there a way to monitor if the HDDs are idle/sleep and the heads parked to save energy? There are periods of inactivity where it might make sense to stop the HDDs from spinning, I'm sure that this will add a bit of startup/lag when they are accessed after being idle, so far I haven't been able to tell if they are have stopped spinning (they only become moderately loud when starting up the NAS).

[barnumbirr]

11 hours ago, gprovost said:

For now fan speed control is based on SoC temperature which is to be honest not the smartest approach since SoC temp is not representative of the ambient temp of the enclosure.  But it was the wisest approach to make the baseline Armbian image clean without too many tweaks.

 

What matters the most in a NAS is the HDD temperature since it's what impacts the most the MTBF of a storage array.

 

We will provide soon the instruction on how to control the fan speed based on HDDs temperature. This will result in a more constant airflow (less variation in speed) and therefore a more quiet operation.

 

It's actually ok to let the SoC goes a bit high in temperature but no the HDD.

Also running Noctua NF-A8 PWM for about a month now with "custom" fancontrol config. @gprovostis there a way for /etc/fancontrol not to be overwritten at every kernel update until the instructions to control fan speed based on HDD temperatures are live?

[SIGSEGV]

@barnumbirr

Could you share your custom config for /etc/fancontrol?

[retrack]

Also running a pair of Noctua NF-A8 PWM (80mm) and it is more silent.

 

I also had a look at the design and I am considering printing a duct to one fan from the SOC heatsink. In the current setup the fans are pretty inefficient at getting airflow through the heatsink fins so a duck could improve that. As the 2 fans can be separately controlled, one could have the fan with the duck react to CPU temperature and the second one to HDD/SDD temperature. From my experience these are running cool (1 SDD 8TB + 4 x 16 TB enterprise HDD).

[gprovost]

I don't really see the point to bother with duct air flow specifically to the SoC.

 

The maximum rated junction temperature of the RK3399 is 125C.

 

Currently on a system loaded at 100%, with PWM value of 100 on both fan which is relatively quiet (ambient room temp 26C), SoC temperature is around 73C and all HDDs are 29-30C.

 

 

 

It's completely OK to have the SoC running at 72C during maximum load till it doesn't increase heat in the enclosure and impact HDD temp. In idle the SoC will immediately go back to 40-43C.

 

[gprovost]

Looking into the fancontrol configuration, I think we were way too conservative on the settings making the fan going too early in higher RPM.

Plus we are not taking into consideration that the heatsink is already doing a pretty good job in passive mode.

 

I would recommend the following fancontrol settings until we provide a way to control the fan based on HDD temp instead :

 

# Helios64 PWM Fan Control Configuration
# Temp source : /dev/thermal-cpu
INTERVAL=10
FCTEMPS=/dev/fan-p6/pwm1=/dev/thermal-cpu/temp1_input /dev/fan-p7/pwm1=/dev/thermal-cpu/temp1_input
MINTEMP=/dev/fan-p6/pwm1=40 /dev/fan-p7/pwm1=40
MAXTEMP=/dev/fan-p6/pwm1=110 /dev/fan-p7/pwm1=110
MINSTART=/dev/fan-p6/pwm1=60 /dev/fan-p7/pwm1=60
MINSTOP=/dev/fan-p6/pwm1=40 /dev/fan-p7/pwm1=40
MINPWM=40

 

[ShadowDance]

@gprovost thanks, that fancontrol config looks good. I thought I had read somewhere on this forum that the SoC was rated for operation at 85C which is why I previously commented that the fans need to run at full speed during heavy load. Glad to know this isn't the case.

 

@retrack if you want to increase airflow to SoC, an alternative approach would be to create the funnel at the front. The gap is the size of a HDD and could be more focused so that the air flows better through the SoC heatsink. As an added bonus, this might also improve HDD cooling because less (but more focused) air is flowing to the SoC.

 

Or a second option, a 3D printed mount for a secondary (small) fan that sits on the SoC heatsink. You could use one of the Noctua splitters to power it.

[barnumbirr]

21 hours ago, SIGSEGV said:

@barnumbirr

Could you share your custom config for /etc/fancontrol?

Sure thing! Here it goes:

Spoiler

# Helios64 PWM Fan Control Configuration
# Temp source : /dev/thermal-cpu
INTERVAL=10
FCTEMPS=/dev/fan-p6/pwm1=/dev/thermal-cpu/temp1_input /dev/fan-p7/pwm1=/dev/thermal-cpu/temp1_input
MINTEMP=/dev/fan-p6/pwm1=40 /dev/fan-p7/pwm1=40
MAXTEMP=/dev/fan-p6/pwm1=75 /dev/fan-p7/pwm1=75
MINSTART=/dev/fan-p6/pwm1=125 /dev/fan-p7/pwm1=125
MINSTOP=/dev/fan-p6/pwm1=100 /dev/fan-p7/pwm1=100
MINPWM=125

 

With the Noctua fans, this config makes the Helios64 silent most of the time while keeping SoC temperatures below 75C and HDD temperatures below 35C.

[gprovost]

15 hours ago, ShadowDance said:

I thought I had read somewhere on this forum that the SoC was rated for operation at 85C which is why I previously commented that the fans need to run at full speed during heavy load. Glad to know this isn't the case.

 

RK3399 commercial grade is rated as 0 - 80C ambient operating temperature. To not be confused with max Temperature junction (Tj) which is 125C ;-)

[retrack]

On 11/18/2020 at 4:08 AM, gprovost said:

 

RK3399 commercial grade is rated as 0 - 80C ambient operating temperature. To not be confused with max Temperature junction (Tj) which is 125C ;-)

Then yes the default fancontrol was super conservative. I have been running your suggestion a few days and I am now graphing temperature of SOC and disks to get better insights.

[allen--smithee]

I'm digging up this topic

 

I have installed new fans but I can't stop them ! 

Intro
 

when I remove the power supply the fans run without humming (it's inaudible) during idle (pwd <70) or when the unit is turned off.
See the problem here H64 off (same as the unit is turned on idle) the power supply affects the fan, they get annoying.

damné

 

How can I adjust it? 

@aprayogaIs there a solution or do I need to replace my new fans (Coolink SWiF2-80P 80mm PWM)?

 

@gprovostThe originals Fan were very good in idle but I couldn't find the right compromise for the acoustics during the charges with Fancontrol. I try to find the best compromise because helios is in a very quiet living room.

 

 

[allen--smithee]

On 11/14/2020 at 11:40 PM, ShadowDance said:

I replaced the stock fans with Noctua NF-A8-PWM fans. These have slightly lower CFM (airflow) at 32.5 vs 35 (stock) however they have a high static pressure rating at 2.37 mm H₂O. For me they produce better cooling than the stock fans. Their noise level is also rated at 17.1 dB(A) so you will barely hear them even at full speed.

 

I also had a pair of NF-R8-PWM (discontinued model) from the old days which I tried. They are very close in CFM at 31.4 but have worse static pressure at 1.41 mm H₂O. These fans produced worse cooling than the stock fans.

 

Hi, @ShadowDance

Yours Noctua NF-A8-PWM fans stops completely the same as the original ventilators when PWM tends towards 0 ?

 

[gprovost]

@allen--smithee Based on what you described and your video, the fans type you are using are Type-A while the stock ones are  Type-C.

Refer to this page to understand the difference : https://wiki.kobol.io/helios64/pwm/#pwm-fan-implementation

 

With Type-A, the fans will always spin, they can't be stopped even at PWM 0.

[allen--smithee]

Hi @gprovost

I know the different types by heart since 1 year but I misunderstood because I always thought that the power curves were managed only by the unit and not also by the ventilators.
These fans (Type A) are very good in idle they emit no sound unless the unit is connected to the power supply in which case there is a very annoying humming in a quiet room (unit power off or on), as I can't use the unit on battery permanently I'll change them.

Thank you for this quick precision.

[ShadowDance]

On 1/6/2021 at 10:19 PM, allen--smithee said:

 

Hi, @ShadowDance

Yours Noctua NF-A8-PWM fans stops completely the same as the original ventilators when PWM tends towards 0 ?

 

You probably have the answer you're looking for already, but yes, they do stop :).

[allen--smithee]

@ShadowDance
Finally, I installed the original fans which were already very good,
but this time I pushed further by adding a MAXPWM line limiting the maximum rotation of the fans.
This changed my life

I feel terribly stupid !!

 

# Helios64 PWM Fan Control Configuration
# /etc/fancontrol

# /usr/sbin/fancontrol
# /var/run/fancontrol.pid

# Command
# ls -l /sys/class/hwmon | grep fan
# cat /sys/class/hwmon/hwmon4/pwm1 && cat /sys/class/hwmon/hwmon5/pwm1
# systemctl stop fancontrol.service
# systemctl start fancontrol.service
# echo 255 > /sys/class/hwmon/hwmon4/pwm1 && echo 255 > /sys/class/hwmon/hwmon5/pwm1
# hdparm -H /dev/sdd | grep celsius
# hddtemp /dev/sd[a-e]

INTERVAL=10
# Temp source : /dev/thermal-cpu
FCTEMPS=/dev/fan-p6/pwm1=/dev/thermal-cpu/temp1_input /dev/fan-p7/pwm1=/dev/thermal-cpu/temp1_input

# temp1_critique = 85° < big downclock 1.8Ghz > 1.62Ghz
# P6 => Fan down (2HDD+CPU) | P7 => Fan Up (3HDD)
#Temp Limit PWM increase
MINTEMP=/dev/fan-p6/pwm1=55 /dev/fan-p7/pwm1=65
#Temp Limit PWM = MAXPWM
MAXTEMP=/dev/fan-p6/pwm1=85 /dev/fan-p7/pwm1=85

#Fan Type "C" 350-3600 PWM = 25 = 350 RPM| PWM = 255 = 3600RPM
#Fan Start PWM = 25
MINSTART=/dev/fan-p6/pwm1=30 /dev/fan-p7/pwm1=30
#Fan Stop PWM = 15
MINSTOP=/dev/fan-p6/pwm1=35 /dev/fan-p7/pwm1=35

#PWM Minimal < MINSTOP (SSD P7 = 0 / HDD P7 = 35)
MINPWM=/dev/fan-p6/pwm1=35 /dev/fan-p7/pwm1=0
#PWM Maximal (78°< Tcpu100% <85°)
#T = 20° PWM = 120 | T = 25° PWM = ?? | T = 30° PWM = ??
MAXPWM=/dev/fan-p6/pwm1=120 /dev/fan-p7/pwm1=65

 

[bunducafe]

Hi folks,

 

I am fairly new here, reading, trying to understand but do have some knowledge with Unix and some SBC. Now the Helios64 is on arrival and I would like to know if I should change the fans immediately in order to get a quieter case. 
 

Is the NF-A8-PWM significantly less noisy than the stock ones? Or is there maybe a different route to go?! 

[allen--smithee]

hi @bunducafe

in order of price, time and efficiency. 

a) you can play with the parameters of fancontrol to find a passive / active fan compromise, and limit the rotational speed and therefore the noise of the fans. (/etc/fancontrol). 

b) you can also lower the frequency of the processor to reduce its TDP, (armbian-config)

c) i do not know the specifications of the thermal bridge (W.mK and thickness) between the CPU and the radiator, maybe this can be replaced to improve thermal conductivity between them. (<10euro)

d) New Fans (>30euros)

 

 

[bunducafe]

Thank you for the insight. So I might opt in for the Noctua directly and play with the config afterwards anyway.

 

Helios case is arriving tomorrow and shipment was way faster than expected, so I will be able to set it up later next week.

[allen--smithee]

@gprovost

Is it possible to know the characteristics of the thermal bridge?

(W.mK et thickness)

 

@bunducafe

You can use your helios64 without fan with a CPU limitation following :

# 816Mhz and 1Ghz (eco)

# 1.2Ghz CPU 100% Temp <82 ° (safe)

# 1.41Ghz CPU 100% Temp <90 ° (Limit)

 

not viable :

# 1.6Ghz CPU100% Temp > 95 ° reboot and reset fancontrol config /etc and there I regret that it is not a K.

Nevertheless, it is reassuring to see that the protection method works.

 

From my own personal experience, it is not necessary to change the fans immediately

and you will quickly agree thereafter that the noise of the fans is erased by the noise of a single rotating hard disk.

But if you use it also as a desk with SSD like me, you would definitely want to use it without limitation.

In that event you must familiarize yourself with Fancontrol otherwise you will also be disappointed with your Noctua or any fan.

[gprovost]

16 hours ago, allen--smithee said:

Is it possible to know the characteristics of the thermal bridge?

(W.mK et thickness)

 

Around 3W/m-K. Dimension and thickness below , the small 21x21x1mm square is the one above the RK3399 SoC :

 

 

16 hours ago, allen--smithee said:

You can use your helios64 without fan with a CPU limitation following :

 

Let's clarify, you can use Helios64 board in passive cooling mode on its own, but if the board is installed in the provided enclosure along HDDs device then you definitely need active cooling.

[bunducafe]

Am 4.2.2021 um 12:59 schrieb allen--smithee:

 

 

@bunducafe

You can use your helios64 without fan with a CPU limitation following :

# 816Mhz and 1Ghz (eco)

# 1.2Ghz CPU 100% Temp <82 ° (safe)

# 1.41Ghz CPU 100% Temp <90 ° (Limit)

 

not viable :

# 1.6Ghz CPU100% Temp > 95 ° reboot and reset fancontrol config /etc and there I regret that it is not a K.

Nevertheless, it is reassuring to see that the protection method works.

 

From my own personal experience, it is not necessary to change the fans immediately

and you will quickly agree thereafter that the noise of the fans is erased by the noise of a single rotating hard disk.

But if you use it also as a desk with SSD like me, you would definitely want to use it without limitation.

In that event you must familiarize yourself with Fancontrol otherwise you will also be disappointed with your Noctua or any fan.

Thanks for the info. I’d rather use the case with fans. I won’t use it as DAS and it will be in the hallway under the router and stuff. I am now tempted to use the stock fans and see how it goes instead of change everything in the first place... if the fans are too noisy in the end then I will change them for some Noctuas 

[allen--smithee]

After one hour of disassembly, cleaning, cutting, installation and reassembly.

My helios64 has gelid thermal pad GP-Ultimate solutions whose manufacturer boasts a small 15W.mK.

 

I think I just solved my ventilation problem for this generation and the following ones.

I will do the tests within a week but for the moment Helios64 remains very cold even under full load.

[SIGSEGV]

14 minutes ago, allen--smithee said:

After one hour of disassembly, cleaning, cutting, installation and reassembly.

My helios64 has gelid thermal pad GP-Ultimate solutions whose manufacturer boasts a small 15W.mK.

I think I just solved my ventilation problem for this generation and the following ones.

I will do the tests within a week but for the moment Helios64 remains very cold even under full load.

 

I'm interested on your results @allen--smithee

[allen--smithee]

@SIGSEGV

The temperature rise of the cpu is now slow and gentle.
I haven't heard my fans all day with these parameters.

 

Divulgacher

 

Thermal Pad Quick Test

Small test UP: Cpu on-demand 400-1800 & fanless & stress -c 6	Small test Down: Cpu on-demand 400-1800 & fanless &	The average test: Cpu performance 400-1800 & Fanless & mp4 & stress -c 6 & Firefox 6 tabs
T-0 = 44° T0 = 50° T1 = 52° T3 = 54° T4 = 57° T5 = 59° T6 = 61° T7 = 62° T8 = 63° T9 = 64° T10 = 65° T15 = 68° T25 = 72° T30 = 73° T35 = 74° T50 = 75° T60 = 75°	T-0 = 75° T0 = 65° T1 = 63° T2 = 61° T3 = 60° T4 = 58° T5 = 57° T10 = 53° T15 = 50°	T-0 = 38° T0 = 48° T1 = 51° T2 = 53° T3 = 55° T4 = 57° T5 = 60° T6 = 63° T7 = 64° T8 = 65° T9 = 66° T10 = 67° T15 = 71° T20 = 73° T25 = 75° T45 = 77° T60 = 80°