The fan in my RN214 is too lound once it gets going, so I took a chance at replacing it with a Noctua NF-A9.
It usually works fine, the resported RPM in the dashboard is typically a bit lower than the stock fan, but the noise level is substantially quieter. The HDD temps increased slightly, but still typically <45C.
However, the dashboard over the past month has reported twice that the fan has stopped. Changing the setting to "cool" (I assume this would push the max voltage) does not get the fan going again, a shutdown/restart did. Obviously not ideal.
I assume the voltage to the fan is dropping below the threshold for the Noctua fan, which requires a higher min voltage than the stock fan (Delta Electronics AFB0912HH).
I searched the forum and found a thread on the old NV+ units for setting minimum fan speed in a config file, but could not find a corresponding conf file on the RN214 using find via SSH.
Any thoughts, or even better, options for finer control over the fan speed (minimum voltage setting)?
The fan in my RN214 is too lound once it gets going,
Fans are always to loud as noise is involved. The fan system is designed to move a certain mass of air over time, to cover all possible (supported) conditions (environment temperature, installed storage blocks, maximum storage activity, ...).
so I took a chance at replacing it with a Noctua NF-A9.
Of course, because of the "press" and the "internet" is full with positive reports of this brand and model. There exist at least two models, some are paired with additional modules.
It's easy to make fans creating lower noise. Typical methods are lower RPMs, lower air flow.
https://noctua.at/en/nf-a9-pwm/specification 4 pin, PWM speed regulation, nominal 2000 min^-1 at 12 V DC, with the low noise adapter (PWM signal) the RPM is lowered to 1550 min^-1, minimum 20% PWM duty cycle leads to (+/-20%) 400 RPM. Tempted to state this version is not for voltage based speed regulation.
https://noctua.at/en/nf-a9-flx/specification 3 pin, 1650, 1250 (Low-Noise Adaptor) and 1050 (Ultra-Low-Noise Adaptor) rpm. This is probably a voltage regulated unit, but far away from the RPM the device monitor does expect.
It usually works fine, the resported RPM in the dashboard is typically a bit lower than the stock fan, but the noise level is substantially quieter. The HDD temps increased slightly, but still typically <45C.
The RPM of these fans is very low - in fact ways below of what the hardware monitor does expect - it can happen easily that the speed is slightly reduced, by normal wear, by dust accommodated.
One rating you hardly ever see for fans is the "stall speed". That's the speed below which the fan cannot go, so it will stop. And it can be important when swapping out a fan. Others in the forum (not with a 214 I can recall) have had similar issues with Noctua fans, especially if you've used the "speed reducer" (which lowers the supply voltage). Typically, just slowly stepping up the voltage won't kick it on until it gets very high. So switching to "cool" when it's already stopped isn't going to fix it. But keeping it on "cool" permanently may (though that may defeat your efforts to keep it quiet).
Editing the .conf file may or may not help. I've found it to work with some parameters, but some seem to be overridden by the ReadyNAS fan control. Interestingly, there does not seem to be an RN214.conf in /etc/frontview/sensors, so you'll have to do as @StephenB did and see where /etc/sensors.d/system.conf points.
Thanks for the comments. I did review the specs from several manufavturers - including the OEM Delta Electronics fan, Noctua and Artic. I am aware of potential risks and prepared to make a different compromise, hence I am measuring temperatures under different situations. But, I accept I don't have full control of the fan, so my approach may indeed fail.
I'm not using the LNA. The max air flow of the original Delta fan is 57.9 CFM / 98.4 cubic m/hr @ 3200 RPM.
Noctua (currently istalled), max air flow is 46.4 CFM / 78.9 cubic m/hr @ 2000 RPM, obviously lower.
Artic F9 (was considering), max airflow is 43 CFM / 73.1 cubic m/hr @ 1800 RPM.
However, I was hoping that running the RN214 in "cool" setting, would still create sufficient airflow to maintain temperatures at smilar level to the OEM fan at a given setting. What I have found is that temperatures across the drives and CPU did not increase substantially between OEM and Noctua fan - but the audiable noise in my environment is night and day. Yes, if I had control over the fan, perhaps the OEM fan could be more tolerable, but I don't, so I'm trying an alternative approach.
Thanks @Sandshark, yes I suspect the 'stopped' fan error I am getting is to do with me trying out the different settings: "Quiet, Balanced, Cool". I have left it on "Cool" now, it probably 'works' at this setting.
The point is not the "cool" setting or the airflow in charge for the effective cooling airstream. Much more likely, the fan speed goes for whatever reason slightly below the limit defined in the system hardware/firmware, and the ~2000 RPM are at the lower limit in any case.
Leaving this alone, the OEM fan with it's two wire (correct?) or three wire (voltage and tach sensor) feels to me different from the NF-A9 PWM (pulse width speed regulation) with four wires (voltage plus PWM signal plus tach signal) but on the nominal voltage always. This could mean the fan does run on the nominal 12 V DC -if- the full cooling is required, and the voltage is set to the max.
G762 reads like the chip supporting PWM fan speed regulation, too. Now back to the question if the OEM fan had two or three wires (and a similar connector). Even on three wires, there can be PWM fan speed regulation, combined with the tach signal for fan rotation speed feedback.
Just because the fan is rotating ... many more things could be wrong.
Thanks @schumaku, the Noctua NF-A9 supports both PWN and variable voltage (if no PWM), and indeed I observe the reported fan speed varying - and I can hear it vary in speed too.
The OEM fan is 3-wire - it also has the tacho output, so it operates in a similar way - I believe from the available data, and from my observations.
The fan stalling, and not restarting, was a detectable problem. Other than that, I assume temperatures are the main area to monitor, I'm discarding power consumption based on the specs. I have run it for a month, temps seem under control, speed of fan does vary, should I be looking out for other potential failings that could cause bigger issues? (note, I do have several backups, just in case I fry the unit, even though I very much doubt that I will.)
Thanks @schumaku, the Noctua NF-A9 supports both PWN and variable voltage (if no PWM), and indeed I observe the reported fan speed varying - and I can hear it vary in speed too.
The OEM fan is 3-wire - it also has the tacho output, so it operates in a similar way - I believe from the available data, and from my observations.
Strange, can't find a plain NF-A9 model specs with three pins and PWM support, this includes the list of obsolete models, too. On your own, well, not my problem.
The point is not the "cool" setting or the airflow in charge for the effective cooling airstream. Much more likely, the fan speed goes for whatever reason slightly below the limit defined in the system hardware/firmware, and the ~2000 RPM are at the lower limit in any case.
The point is that the fan itself has a minimum speed, after which it stalls (stops running), which is what is being observed here. The OS sets the fan speed to below that speed, so it stops. Many fans require a sizable "kick" to get them going, which is why they come on full blast at power-on. So slowly ramping the speed back up either doesn't re-start it at all or takes a long time, during which the temperature rises quickly. This can result in fan "motorboating". Others have had similar issues with Nochua fans in a ReadyNAS, so that stall speed must be above that of the stock fan and below the point at which the OS is trying to set it. Netgear has never published the parameters for "cool", "balanced" and "quiet", but playing around with them, it appears that one of the things that change with that setting is the minimum speed. That supports why changing it in system.conf may have no effect -- the OS doesn't look at that and independently sets it. But permanently setting to "cool" may help -- it depends on what minimum speed it might set and the stall speed of the fan, both of which are unknowns.
@StephenBI tried adding 900 as the min speed line to the conf file, rebooted, but at 'Quiet" setting, the fan RPM went well below 600 (dashboard view) and then stalled again. It appears that setting is not observed on the RN214.
@SandsharkYes I read the Noctua starts at 5V (datasheet only says min RPM is 400), the OEM starts at <=4V (datasheet).
@schumakuThe A9 PWM can be used in non-PWM mode: FAQ
Can I run 4-pin PWM fans on 3-pin fan headers?
PWM fans come with a 4-pin connector for fully automatic speed control via your motherboard’s 4-pin PWM fan headers. Please note that 4-pin fans can also be connected to your motherboard’s 3-pin fan headers, though. When connected to 3-pin fan headers, the fan will run at full speed (unless the motherboard supports voltage based speed control).
I confirm the fan speed does indeed vary.
I'm still messing with this, will report any further progress. For now, 'Cool' setting appears to be working without issue.
FYI, the min_fan_speed_override file does work on my RN102 (running 6.10.9). The stock fan was running about 800 rpm without the override. I set it to 1500, and the speed went to that. It also changed the speed to 1000 when I set it there.
BTW, if you want to see the speed from ssh, you can install lm-sensors.
@StephenB min_fan_speed_override does indeed work on the RN214 also, so at least I can prevent the stall failure.
I decided to measure what the NAS was doing, it's quite odd in that it typically runs the stock fan at very low volts, ~2.8V I measured, which is not good for the Noctua NF-A9 at all. That fan starts at ~4.4V, but I contacted Noctua support and they provided me two different responses to the same ticket:
1st response "The NF-A9 PWMs' starting voltage is 4.4V at 435 RPM."
The next day a different person, 2nd response "NF-A9 PWM is 7V to 13.2V, its startup voltage is also 7V"
From my lab power supply, I confirmed my NF-A9 actually starts at 4.3V, but is unstable, at 4.4V it runs well.
I plotted the various output volts from the ReadyNAS 214 with the different system fan settings, having captured volts/fan RPM/temperatures, now I can make more sense of what's happening.
I found:
If RPM zero detected, the RN214 ramps voltage to 12V to attempt to get the fan started again.
RN214 drives the stock fan (Delta Electronics AFB0912HH ) at max 3250 (reported in dashboard) RPM, which is at ~8.5V. This is the max RPM in the datasheet, hence that makes sense, however it means the RN never drives the stock fan at max blow - which is at 12V. At 8.5V, my PC reported 2350 RPM, which seems correct as the rest of the V/RPM wer quite linear in the PC.
The RN214 reports the fan RPM incorrectly. After reading the voltages and reported RPM's, I hooked up both fans to a PC to capture the RPM, but controlled the voltage separately with my lab PSU. I found that the PC reported RPM's inline with the datasheets, the ReadyNAS was plain wrong at the RPM reporting on both stock and Noctua fans.
My NF-A9 fan stalls below 4.3V, and sometimes fails to restart - it stops and remains stopped, even when voltage increased. I observed this initially when controlled by my NAS, and then also when I controlled voltage with my lab PSU. Noctua support that the fan should restart, and I observe the RN indeed drives the volts to 12V, so my fan could in fact be faulty.
The RN drives the NF-A9 at 12V which achieves a real 2000 RPM, when it determines the need for max fan blow. The RPM is reported at 2850 - below the max RPM of the stock fan, which I assume allows it to reach the max 12V voltage, i.e. RN has two limits: max 12V, max 3250 RPM.
Perversely, the above behaviour means the RN drives both fans to similar-ish 'real' RPM, so that's probably why I observed very little differences in operating temperatures - more investigation needed/ongoing.
I am still playing with this, testing airflow, or better said, temperatures across the drives/CPU to determine if my replacement fan will actually work out. Fans do sometimes go wrong, so this might help others keep their NAS going a bit longer 😀. I'll share what I find.
Hi @aks-2 , any update to your investigation here?
I tried installing a Noctua NF-A9 FLX 3-pin fan, and had similar results as your A9-PWM version. The system would usually start, and run OK on any setting (as I recall), but then abruptly fail out of nowhere. If you found a setting that worked in SSH-land, please advise.
I also tried a 4-pin PWM fan, a Thermalright TL-9015W that runs at higher max RPM, thinking the issue might be RPM detection. In the garage, with the back panel removed but connected to the 3-pin header on the NAS, on power-up the fan spun up nicely. However, after I then reassembled the unit, brought it back inside, reconnected power and ETH, the unit won't spin up the fan at all. It's as if the fan settings got modified somehow, perplexing.
Anyway I'd like to run the Noctua with confidence. I could set it on Cool permanently but I fear the fan might still shut down in the middle of the night, during a backup or whatever. I don't want to cook my drives.
You'll see in another thread that I had a different disaster with my RN214, causing data corruption, that 'distracted' me for a while 😉!
Due to the issues running the Noctua, I reverted to the stock fan, whilst dealing with the corruption issue.
Let me check my notes for any more info I can offer. As far as I remember, the RN lowers the voltage too much, causing the fan to stall. Once stalled, the Noctua does not restart, even with 12V - although Noctua informed me that it should, so my fan must be a faulty example. They offered to send a replacement, but I did not take them up on that (yet).
I will check my notes and share anything else I can think of.
Why are you changing your fan, did the original fail?
The NAS has a setting to avoid frying your drives, so it will shut down before that if the temps get too high. Review your settings to enable: System / Settings / Alerts - tick "shut down system when the disk temperature exceeds safe levels" - but, I don't know what safe actually means here.
Thanks @aks-2 - in my case, no fan failures, just noisy.
I generated the minimum fan speed override command that @StephenB listed in your thread, except I set the min speed to 1500 instead of 1000.
So far, with about 24 hours run-time, it has worked. An hour ago, I swapped out the existing drives for 4 new HDD's, which caused the OS to go back to factory defaults. In this state, the fan failed (again). So I quickly added back in the override command, set the speed to Cool, rebooted, and it's working fine again. I also have alerts turned on - if the fan stops I'll get a notification so I can deal with it. Thanks again for all the data you shared, really helped to get closer to understanding the issue. The RPM reading is around 1580 (max speed for this fan is 1600). Per your notes, I can't be sure the fan is actually spinning at that rate. But I'm watching the CPU and drive temps, so far they're OK. Not as cool as the Delta fan on "Cool" setting though.
Also purchased a BeQuiet! Pure Wings 2 92mm fan, will give that one a try for grins.
