I just got an ICX7250-24P, and as many have attested, it's too loud. I'm seeing ~65dB at start up that comes down to ~55dB steady state. At idle (only OOB MGMT interface active), I have the following stats:
Code:
SSH@icx7250-24p-router>show chassis
The stack unit 1 chassis info:
Power supply 1 (AC - PoE) present, status ok
Power supply 2 not present
Fan 1 ok, speed (auto): [[1]]<->2
Fan 2 ok, speed (auto): [[1]]<->2
Fan controlled temperature:
Rule 1/2 (MGMT THERMAL PLANE): 60.0 deg-C
Rule 2/2 (AIR OUTLET NEAR PSU): 33.5 deg-C
Fan speed switching temperature thresholds:
Rule 1/2 (MGMT THERMAL PLANE):
Speed 1: NM<-----> 93 deg-C
Speed 2: 82<----->105 deg-C (shutdown)
Rule 2/2 (AIR OUTLET NEAR PSU):
Speed 1: NM<-----> 58 deg-C
Speed 2: 49<----->105 deg-C (shutdown)
Fan 1 Air Flow Direction: Front to Back
Fan 2 Air Flow Direction: Front to Back
Slot 1 Current Temperature: 60.7 deg-C (Sensor 1), 33.0 deg-C (Sensor 2)
Slot 2 Current Temperature: NA
Warning level.......: 100.0 deg-C
Shutdown level......: 105.0 deg-C
I'm exploring ways to make this quiet, and note that what seems to be the accepted best approach is:
* Replace Fan 1 and 2 with Delta EFB0412VHD-F00's (10.1 CFM, 0.416 inH2O, 32.5dB): total CFM=20.2
* Add a splitter and mount Fan 3 on the ASIC (60mm x 60mm x 10mm seems to fit within the 1U height) - I see the Sunon MF60101V3-1000U-A99 recommended (12.2CFM, 0.08 inH2O, 22.5dBA).
From the perspective of the 40mm fans, this is a reduction in CFM from the stock (26.33 CFM, 2.04 inH2O, 53.8dBA), with 2 for total CFM = 52.66 CFM, so generates approx 40% of the airflow, which understandably makes Fan 3 on the ASIC necessary. With the ASIC, this comes to 32.2CFM if you just add it together. In reality it's probably not quite as good as that at exhausting the hot air given the orientation of Fan 3 on the ASIC, but maybe it's more effective at removing heat from the ASIC heatsink at the detriment of ambient temperatures within the chassis (which might be better?).
I'm a little hesitant to mount a fan on the ASIC as it seems like a destructive process (I see the recommendation to use self-tapping screws to screw into the heatsink). If anybody has a better recommendation on mounting methodology that would be good.
Has anybody tried the following, and if so what were your results?
1. A low noise adapter on the existing fans
2. Adding a third 40mm fan into the empty slot using a splitter (something like this?
Amazon.com: ThreeBulls PWM Fan Hub, PC CPU Cooling 4 PIN/3PIN Power Fan Extension Cable Hub Splitter Adapter Sleeved Case for 12V Desktop Computer Cooler Fans 1 to 5 Way (15.7 Inch) : Electronics)
It seems like the original fans actually have a pretty good noise/performance curve, so maybe just reducing the RPM with (1) would actually be a pretty good solution?
As for option (2), I found the following fans that seem like they may be good options for something slightly noisier than previously recommended, but closer to original performance:
Comments:
* (2) seems to only be available in batches of 100 on digikey, but purchasable in lower quantities from
bravo electro
* (8) is still not readily available outside of the US. US seem to have availability through
Bravo Electro
* I wasn't really sure how to quantify the effect of pressure difference across the configurations, except to note that all of the alternatives have a rather drastic drop. It would be good to understand what the static pressure of the system actually is so that the CFM results for each of these fans could be normalised (I think most of them have published P-Q curves so that should be possible). This is a TODO - I wasn't able to find any information quickly. It does say something that the OEM solution has such a high static pressure rating though; makes me think it's on the higher end - surely they wouldn't try to overspec it that much.
* The estimated total noise of the stock configuration in (1) for 2 fans doesn't match my observed results. My guess is that this is because (a) the calculation is idealised and the assumptions aren't perfectly aligned with reality and (b) I'm measuring the sound level of my entire rack, not just the switch. My rack has a baseline noise measurement of ~46dBA, which I wouldn't have thought would impact drastically. I suspect as others have already commented that the 53.8 dBA metric for the fan relates to the fan in free space, not counting the enclosure etc and any related harmonics that go along with that, which I know can contribute a fair bit. All of that to say; treat the estimated noise values with a grain of salt. They're probably indicative enough of trends but by no means accurate.
Resources:
* cumulative noise performance estimated using
Total dB level adding of incoherent or noncoherent uncorrelated sound sources signals combine sound levels two sources resultant level audio logarithmic decibel scale decibels or SPL sound pressure level add signal noise levels noncoherent incoherent random phase acoustic - sengpielaudio Sengpiel Berlin
* Interpretation of relative noise performance uses the equation 100 * 10 ^ (abs(NL1 - NL2) / 10) - decibels are a log scale, not linear.
Keen to hear if anybody else has experimented with these alternatives or if there are any obvious flaws in my thinking here; I haven't made a decision on way forward yet.
